Question 3. Explain what it means to have a political ideology, and how ideology is distinct from party identification. Then, discuss the motives for adopting an ideology (not specific ideologies, but any ideology in general). In 1964, Philip Converse published a study of political attitudes, in which he claimed that very few people were ideological. Do you think things have changed since then? What are the reasons for changes or the lack of them? Is party identification or ideology a larger determinant of people’s political attitudes?
Question 4. Conscientiousness, openness, and authoritarianism are three personality traits that have been shown to impact political attitudes. First, describe each trait and how it affects a person’s attitudes and behavior in general. Then, explain how each is linked to one’s political attitudes, ideologies, and views on particular political issues
Question 5. Two potential problems with surveys are double-barreled questions and sensitive questions. Explain what a double-barreled question is and why it would be a problem. Then, give an original example (i.e., not the one from the lecture) and explain how to fix it. Next, explain the ways in which survey takers can see a question as being sensitive and what the risks are of including sensitive questions in surveys. Finally, explain one technique for improving responses to sensitive questions.
Question 6. Throughout the course, we have compared three theories of democracy: democratic elitism, pluralism, and participatory democracy. Describe each of these views and explain how they are different from one another. Then, discuss how these three viewpoints affect opinions on political socialization, the necessity of political knowledge, and the role of the mass media
Each question must be answer as esay format (3 paragraph or more) with a minimum of 300 words count
Please answer each question based on the book attach
YOU ONLY NEED TO ANSWER 3 OUT OF THE 4 QUESTIONS POST
PUBLIC OPINION
Fourth Edition
To my parents, Dale and Janice Clawson, for all their love and
laughter.
To my mother, Rachel Oxley, whose encouragement and optimism
never wavered, and to my entire family for supporting my endeavors.
Sara Miller McCune founded SAGE Publishing in 1965 to support
the dissemination of usable knowledge and educate a global
community. SAGE publishes more than 1000 journals and over 800
new books each year, spanning a wide range of subject areas. Our
growing selection of library products includes archives, data, case
studies and video. SAGE remains majority owned by our founder
and after her lifetime will become owned by a charitable trust that
secures the company’s continued independence.
Los Angeles | London | New Delhi | Singapore | Washington DC |
Melbourne
PUBLIC OPINION
Democratic Ideals, Democratic Practice
Fourth Edition
Rosalee A. Clawson
Purdue University
Zoe M. Oxley
Union College
Copyright © 2021 by CQ Press, an Imprint of SAGE Publications,
Inc. CQ Press is a registered trademark of Congressional Quarterly,
Inc.
All rights reserved. Except as permitted by U.S. copyright law, no
part of this work may be reproduced or distributed in any form or by
any means, or stored in a database or retrieval system, without
permission in writing from the publisher.
All third party trademarks referenced or depicted herein are included
solely for the purpose of illustration and are the property of their
respective owners. Reference to these trademarks in no way
indicates any relationship with, or endorsement by, the trademark
owner.
FOR INFORMATION:
SAGE Publications, Inc.
2455 Teller Road
Thousand Oaks, California 91320
E-mail: order@sagepub.com
SAGE Publications Ltd.
1 Oliver’s Yard
55 City Road
London, EC1Y 1SP
United Kingdom
SAGE Publications India Pvt. Ltd.
B 1/I 1 Mohan Cooperative Industrial Area
Mathura Road, New Delhi 110 044
India
SAGE Publications Asia-Pacific Pte. Ltd.
18 Cross Street #10-10/11/12
China Square Central
Singapore 048423
Library of Congress Cataloging-in-Publication Data
Names: Clawson, Rosalee A., author. | Oxley, Zoe M., author.
Title: Public opinion : democratic ideals, democratic practice / Rosalee A. Clawson,
Purdue University, Zoe M. Oxley, Union College.
Description: Fourth edition. | Thousand Oaks : Sage / CQ Press, [2021] | Includes
bibliographical references.
Identifiers: LCCN 2020017438 | ISBN 978-1-5443-9020-8 (paperback) | ISBN 978-
1-5443-9016-1 (epub) | ISBN 978-1-5443-9018-5 (epub) | ISBN 978-1-5443-9015-
4 (ebook)
Subjects: LCSH: Democracy. | Public opinion. | Political socialization. | Mass
media and public opinion.
Classification: LCC JC423 .C598 2021 | DDC 321.8—dc23
LC record available at https://lccn.loc.gov/2020017438
Printed in the United States of America
This book is printed on acid-free paper.
Acquisitions Editor: Scott Greenan
Editorial Assistant: Sam Rosenberg
Production Editor: Rebecca Lee
Copy Editor: Colleen Brennan
https://lccn.loc.gov/2020017438
Typesetter: Hurix Digital
Proofreader: Sally Jaskold
Indexer: Integra
Cover Designer: Janet Kiesel
Marketing Manager: Erica DeLuca
BRIEF CONTENTS
Tables, Figures, and Boxes
Preface
Acknowledgments
Part I What Should the Role of Citizens Be in a Democratic
Society?
Chapter 1 Public Opinion in a Democracy
Appendix to Chapter 1 Studying Public Opinion
Empirically
Part II Are Citizens Pliable?
Chapter 2 Political Socialization
Chapter 3 Mass Media
Chapter 4 Attitude Stability and Attitude Change
Part III Do Citizens Organize Their Political Thinking?
Chapter 5 Ideology, Partisanship, and Polarization
Chapter 6 Roots of Public Opinion: Personality, Self-
Interest, Values, and History
Chapter 7 Roots of Public Opinion: The Central Role of
Groups
Part IV Do Citizens Endorse and Demonstrate Democratic
Basics?
Chapter 8 Knowledge, Interest, and Attention to Politics
Chapter 9 Support for Civil Liberties
Chapter 10 Support for Civil Rights
Part V What Is the Relationship between Citizens and Their
Government?
Chapter 11 Trust in Government, Support for Institutions,
and Social Capital
Chapter 12 Impact of Public Opinion on Policy
Part VI What Do We Make of Public Opinion in a Democracy?
Chapter 13 Conclusion
Notes
Glossary
Index
About the Authors
DETAILED CONTENTS
Tables, Figures, and Boxes
Preface
Acknowledgments
Part I What Should the Role of Citizens Be in a Democratic
Society?
Chapter 1 Public Opinion in a Democracy
Theories of Democracy
What Is Public Opinion?
Defining Key Concepts
Empirical Assessments of Public Opinion
Themes of the Book
Appendix to Chapter 1 Studying Public Opinion
Empirically
Public Opinion Surveys
Experiments
Interviews
Focus Groups
Content Analysis
Conclusion
Part II Are Citizens Pliable?
Chapter 2 Political Socialization
Childhood Socialization
Parental Transmission of Political Attitudes
Generational and Period Effects
Genetic Inheritance of Political Attitudes
Conclusion
Chapter 3 Mass Media
What Should Citizens Expect from the Mass Media in a
Democracy?
What General Characteristics of the Mass Media
Shape News Coverage?
What Specific Characteristics of the Traditional News
Media Shape the Reporting of Political Events?
What About Fake News?
Are Citizens Affected by the Mass Media?
Media Effects in a Changing Technological
Environment
Conclusion
Chapter 4 Attitude Stability and Attitude Change
Are Americans’ Attitudes Stable?
Presidential Approval
Psychological Approaches to Attitudes
Conclusion
Part III Do Citizens Organize Their Political Thinking?
Chapter 5 Ideology, Partisanship, and Polarization
Converse’s Claim: Ideological Innocence
Ideological Identification
Party Identification
Polarization
Conclusion
Chapter 6 Roots of Public Opinion: Personality, Self-
Interest, Values, and History
Personality
Self-Interest
Values
Historical Events
Conclusion
Chapter 7 Roots of Public Opinion: The Central Role of
Groups
Race, Ethnicity, and Public Opinion
Rural Consciousness
Gender and Public Opinion
Conclusion
Part IV Do Citizens Endorse and Demonstrate Democratic
Basics?
Chapter 8 Knowledge, Interest, and Attention to Politics
How Knowledgeable, Interested, and Attentive Should
Citizens Be in a Democracy?
Are Citizens Knowledgeable about Politics?
Measuring Political Knowledge
Why Are Some Citizens More Knowledgeable Than
Others?
What Are the Consequences of Political Knowledge?
Are Citizens Interested in and Attentive to Politics?
Conclusion
Chapter 9 Support for Civil Liberties
Support for Democratic Principles
Are Americans Tolerant?
Sources of Tolerant Attitudes
Contextual Influences on Tolerance Judgments
Are Elites More Tolerant?
Civil Liberties Post-9/11
Conclusion
Chapter 10 Support for Civil Rights
Public Opinion and Presidential Candidates
Support for Civil Rights Policies
Conclusion
Part V What Is the Relationship between Citizens and Their
Government?
Chapter 11 Trust in Government, Support for Institutions,
and Social Capital
Trust in Government
Support for Institutions
Social Capital
Conclusion
Chapter 12 Impact of Public Opinion on Policy
Should Public Opinion Influence Policy?
Is Public Opinion Related to Policy?
Do Politicians Follow or Lead the Public?
Public Opinion and Foreign Policy
Conclusion
Part VI What Do We Make of Public Opinion in a Democracy?
Chapter 13 Conclusion
What Should the Role of Citizens Be in a Democratic
Society?
Are Citizens Pliable?
Do Citizens Organize Their Political Thinking?
Do Citizens Endorse and Demonstrate Democratic
Basics?
What Is the Relationship between Citizens and Their
Government?
What Do We Make of Public Opinion in a Democracy?
Notes
Glossary
Index
About the Authors
TABLES, FIGURES, AND BOXES
Tables
A.1 Question Wording and Response Options Matter 31
A.2 Support for the Death Penalty in a Survey-Based
Experiment 39
2.1 Children’s Descriptions of the President’s Duties 52
3.1 Making Sense of Subtle Effects 108
4.1 Stability of Individual Political Attitudes from 1958 to 1960
119
4.2 Aggregate Opinion Can Be Stable While Individual Attitudes
Change 124
5.1 Key Components of Black Political Ideologies 163
5.2 Measuring Political Ideology 164
6.1 Measuring Authoritarianism 189
6.2 Big Five Personality Traits 194
6.3 Ten-Item Personality Inventory 195
6.4 Measuring Egalitarianism and Individualism 201
6.5 Measuring Moral Traditionalism 203
7.1 Measuring White Identity 223
7.2 Latino and Asian American Party Identification among
Registered Voters, 2018227
8.1 Political Knowledge, June 2017 242
8.2 Political Knowledge, 1989–2007 251
8.3 Measuring Political Knowledge 253
8.4 Traditional and Gender-Relevant Political Knowledge 255
8.5 The Perils of Measuring Political Knowledge: Short-Answer
versus Multiple-Choice Questions 258
8.6 Demographic Differences in Political Knowledge 262
9.1 Assessing Public Tolerance of Atheists: Stouffer’s Survey
Questions 281
9.2 Least Liked Political Groups, 1978 and 2005 286
10.1 Religion and Likelihood to Vote for a Presidential
Candidate, 2016 311
10.2 Measuring Hostile Sexism 317
10.3 White and Black Support for Reparations, 2019 325
11.1 Assessing Public Trust: Survey Questions from the
American National Election Studies 347
11.2 Focus Group Discussions of Members of Congress 365
Figures
1.1 Party Identification, 1952–2016 23
A.1 Public Opinion toward the Death Penalty, 1991–2018 34
2.1 Children’s Evaluations of the President’s Job Performance
53
2.2 Parent-Child Correspondence of Party Identification by
Family Politicization and Parental Attitude Stability 62
2.3 Stability of Party Identification Over Time, Overall and by
Preadult Parent-Child Correspondence 64
2.4 Generational Differences in Attitudes, 2019 70
2.5 Genetic Versus Environmental Factors Influencing Political
Opinions 73
3.1 Where Do Citizens Obtain News Daily? 84
3.2 Sorting Out Causal Relationships 102
3.3 Political Tolerance by Framing Condition 105
4.1 Opinion Toward Government Spending, 1971–1989 122
4.2 Presidential Approval for Barack Obama and Donald Trump
126
4.3 Political Awareness in Zaller’s Attitude Change Model 133
4.4 Zaller’s Mainstream and Polarization Effects during Vietnam
War Era 136
5.1 Levels of Conceptualization among the American Public,
1956 and 2000 154
5.2 Relationships between Issue Opinions for the American
Public and Political Elites, 1958 157
5.3 Attitude Constraint and Attitude Stability among American
Public and Elites 160
5.4 Ideological Identification over Time, 1972–2016 165
5.5 Symbolic and Operational Ideology Classifications, 2008
167
5.6 Party Differences in Issue Opinions, 2016 170
5.7 Hypothetical Portraits of the American Public 173
5.8 Public Attitudes toward Government Provision of Services,
1984 and 2016 175
6.1 The Effect of Authoritarianism on Political Attitudes, 2016
191
7.1 Black–White Differences in Party Identification and Issue
Opinions, 2016 214
7.2 Racial Resentment among Whites, 1986 and 2016 217
7.3 Political Attitudes of Millennials by Race and Ethnicity 222
7.4 Gender Differences in Party Identification and Issue
Opinions, 2016 232
8.1 Misperceptions by News Source 248
8.2 Interest in Politics and Current Campaign, 1964–2016 267
9.1 Importance of Democratic Principles 277
9.2 Tolerance of Political Minorities, 1954 282
9.3 Tolerance of Speechmaking, 1954–2018 283
9.4 Tolerance of Least Liked Groups, Communists, and Atheists,
1978 and 2005 288
9.5 Public Opinion: Civil Liberties versus National Security 296
9.6 Public Support for Counterterrorism Policies, 2009 and 2010
300
9.7 Counterterrorism Policy Opinions Vary by Identity of Target
301
10.1 Support for Presidential Candidates, 1937–2019: Religion
310
10.2 Support for Presidential Candidates, 1937–2019: Gender,
Race, Ethnicity, and Sexual Orientation 313
10.3 Support for Same Schools, by Race 320
10.4 Support for School Busing, by Decade and Race 321
10.5 Support for Preferences in Hiring and Promotion for Blacks
323
10.6 Support for Gay Rights, 1977–2019 329
10.7 Support for Transgender Rights, 2017 and 2019 335
11.1 Public Trust in Government, 1964–2016 348
11.2 Trust in Government for Specific Demographic Groups,
2016 355
11.3 Racial Differences in Views Regarding the Police, 2016
356
11.4 Confidence in the Supreme Court, Executive Branch, and
Congress, 1973–2018 361
11.5 Approval of Institutions and Members of Institutions, 1992
363
11.6 Membership Declines for Civic Associations between Peak
Year of Membership and 1997 369
12.1 Public Opinion and Guns 379
12.2 Support for Gun Control Measures, by Partisanship 380
12.3 Consistency between Public Opinion and Public Policy 386
12.4 Citizen and Interest Group Influence on Public Policy 390
12.5 Foreign Policy Preferences of the American Public, 2014–
2019 404
Boxes
Box 1.1 Gendered Nouns and Pronouns 11
Public Opinion in Comparative Perspective
Box 2.1 Childhood Political Socialization in Europe 58
Box 3.1 Social Media in China 90
Box 4.1 Political Discussion in Social Networks 138
Box 5.1 Party Polarization across the Globe 179
Box 6.1 Authoritarianism across the World 193
Box 7.1 Attitudes toward Immigrants around the World 229
Box 8.1 Gender and Political Knowledge in Canada 256
Box 9.1 Support for Democracy around the Globe 278
Box 10.1 Support for Gay Rights across the World 331
Box 11.1 Levels of Public Trust in Other Nations 359
Box 12.1 Comparing Opinion-Policy Congruence across
Democracies 391
PREFACE
When we first tell people that we are political science professors, the
most common reactions are to launch into a discussion of politics or
to politely acknowledge our jobs and then change the subject. An
especially memorable encounter happened to Zoe in the early
2000s. Upon re-entering the United States after a trip to Montreal,
the U.S. border official asked about her job, which was expected. He
proceeded to ask a most unexpected question: What did she think
about then-president George W. Bush’s foreign policies? After
babbling for a few sentences, she changed the subject!
Ever since the 2016 U.S. presidential election, different reactions
have been common when people learn what we do for a living.
“These are interesting times to teach politics” or “You sure have a lot
to talk about these days” have become typical responses. Interesting
political times these certainly are, and we have attempted to capture
some notable political developments and trends in this fourth edition.
These include partisan disdain and polarization, fake news, white
racial identity, rural consciousness, and support for the norms of
democracy. We have also incorporated exciting and important new
public opinion data or scholarship on childhood socialization,
Millennials and Generation Z, the effects of media ownership, news
habits, the role of social media, ideological identification,
polarization, the lingering effects of slavery, racial and ethnic opinion
differences (including in the domain of athletics), immigration
attitudes, interest in politics, transgender rights, disability rights, trust
in the criminal justice system, public support for gun control, and
foreign policy opinions. All the while, we maintain a focus on
enduring questions in the study of public opinion.
Our pedagogical goals for this edition remain the same. We want
students to grasp how fascinating and important it is to study politics
generally and public opinion more specifically. What better way is
there to attain that goal, we think, than to discuss public opinion in
the context of democratic thought? After all, it is the particular
salience of public opinion within a democracy that makes its study so
vital and interesting. To that end, we situate the field’s empirical
research within a normative framework, specifically theories of
democracy, and focus on especially important and revealing studies
rather than tediously summarizing every available piece of research.
We organize the text into six parts, each of which poses a normative
question that is significant for democratic theory: What should the
role of citizens be in a democratic society? Are citizens’ opinions
pliable? Do citizens organize their political thinking? Do citizens
endorse and demonstrate democratic basics? What is the
relationship between citizens and their government? What do we
make of public opinion in a democracy? The chapters in each part
present evidence to help students answer the question at hand,
giving them both the content and context of public opinion. This
organization encourages students to understand and interpret the
empirical evidence in light of normative democratic theories, thus
enhancing their critical analysis skills.
We want students to appreciate the thrill of conducting research and
producing knowledge and to learn that conclusions about public
opinion emerge from original scholarship on the topic. Yet we also
want them to understand that no one piece of research is perfect and
that the ability to evaluate the strengths and weaknesses of a piece
of research is a vital skill. So we devote attention to explaining
specific studies in some depth throughout the text. Rather than
presenting only the conclusions that are drawn from a study, this
approach lets students see how those conclusions were reached,
exposes them in a fairly organic way to the range of research
methods used in the study of public opinion, and illustrates how the
choice of method influences the conclusions that researchers draw.
We thus use an “embedded” research methods approach throughout
the book rather than consigning methods to one stand-alone chapter.
In addition, we provide an Appendix to Chapter 1 that encapsulates
the basic information students need about key public opinion
methods.
This book includes other important pedagogical features. We focus
heavily on American public opinion, but Chapters 2 through 12
contain feature boxes called “Public Opinion in Comparative
Perspective” that highlight public opinion issues in a variety of
countries and serve to deepen students’ understandings of American
public opinion. A wealth of data is presented in more than eighty
tables and figures throughout the book to help students grasp
important research findings. Key concepts appear in bold in each
chapter and are listed at the end of each chapter. The key concepts
are also defined in the Glossary at the end of the book. Each chapter
also contains a list of suggested sources for further reading. Brief
explanatory annotations are provided with each suggested source to
guide students as they delve deeper into a topic.
ACKNOWLEDGMENTS
With each edition of this book, we find we require assistance from
fewer people as we revise. Having said this, the help we did receive
was extremely valuable. Rosie’s Human Basis of Politics students
and students in Zoe’s Public Opinion and U.S. Politics Seminar on
Partisanship courses provided useful feedback on the book, most
especially pointing out when material was not crystal clear or
organized well enough. For thoughtfully reviewing a chapter for this
edition, we thank Molly Scudder. Thanks goes to Walter Schostak as
well for helpful feedback along the way. Since the publication of the
first edition, we have been approached by many professors who
have used the textbook in their classes. We were heartened to hear
their (mostly!) positive comments, were happy to learn we were on
the right track with our approach, and welcomed their suggestions
for areas in need of improvement.
Everyone we worked with at CQ Press was supportive, professional,
and friendly, as had been the case with our previous books. We
especially thank Monica Eckman and Scott Greenan for their
enthusiasm for the project and guidance. Kate Scheinman and Sam
Rosenberg carefully and efficiently managed the submission of our
chapters as well as the preparation of material for publication. They
were supportive and cheerful throughout, even as we kept missing
our deadlines. For the excellent copyediting of this edition, we thank
Colleen Brennan. As our production editor, Rebecca Lee
shepherded the book through the final prepublication stages with
ease. We also thank Elaine Dunn for her diligent and speedy
copyediting on the first edition. Her initial feedback continues to
shape our work. We worked closely with others at CQ on earlier
editions of the book. For their never-ending encouragement and
wonderful advice, we thank Brenda Carter, James Headley, Elise
Frasier, and, most especially, Charisse Kiino.
We also thank the professors that CQ Press commissioned to review
the book as we were preparing to revise it for the fourth edition. Their
feedback was extremely helpful. They include Davida J. Alperin
(University of Wisconsin–River Falls), Ray Block Jr. (Penn State
University), Gar Culbert (California State University, Los Angeles),
Brian Frederick (Bridgewater State University), and David Kimball
(University of Missouri–St. Louis). We would be remiss if we did not
acknowledge the invaluable guidance provided by the faculty who
reviewed the manuscript for our earlier editions, including Scott
Basinger, Mark Brewer, John Bruce, Erin Cassese, Gar Culbert,
Johanna Dunaway, Howard Gold, Paul Goren, Richard Hofstetter,
Ted Jelen, Mary Fran T. Malone, James Monogan, Kimberley Nalder,
Tom Nelson, Shayla Nunnally, Kurt Pyle, Andrea M. Quenette, Adam
Schiffer, Robert Y. Shapiro, Mike Schmierbach, and Matt Wilson.
Along with these CQ-commissioned reviewers, many others
provided valuable and specific feedback on material for prior
editions. Their suggestions then continued to shape this edition of
the book. For that, we thank Ben Bauer, George Bizer, Richard Fox,
Cary Funk, Ewa Golebiowska, Mike Grady, Jennifer Jerit, Suzanne
Parker, Evan Reid, Walter Schostak, Keith Shimko, Bas van Doorn,
Ryan Whelpley, Jeremy Zilber, John Zumbrunnen, and, most
especially, Janice Clawson. For prior editions, we also received all
manner of help from many others, including Carol Cichy, Michelle
Conwell, David Hayes, Lisa Howell, Katsuo Nishikawa, Andrea
Olive, Bill Shaffer, and Helen Willis. We must also mention our many
Ohio State friends, who have supported and encouraged us
throughout. John Clark, Larry Baum, David Kimball, and Staci Rhine,
in particular, have shared their suggestions and wisdom along the
way. Finally, we will always owe a debt of gratitude to our graduate
school advisers—Paul Allen Beck, Thomas Nelson, and Katherine
Tate—and to our undergraduate mentors—Janet Martin and Bruce
Stinebrickner. Bruce Stinebrickner also gave us helpful feedback on
Chapter 1, which strengthened that chapter.
Throughout this book, we present results from many published
academic papers and books. We also incorporate public opinion
data from organizations that, at great care and expense, conduct
surveys of the American public. Fortunately for us and for all
students of public opinion, they make their results and, at times, their
raw data publicly available. Thus, we gratefully thank the American
National Election Studies, General Social Survey, Pew Research
Center, Gallup Organization, and Bright Line Watch, as well as many
news organizations, academic institutions, and commercial firms that
conduct opinion polls. Without the public opinion data these
organizations have gathered, our book would be much less rich.
Although it does not quite take a village to raise our children, we
have needed help from many to care for our sons. Knowing that they
were in the hands of loving and responsible caregivers and friends
enabled us to write without worry. For this, Zoe thanks Anna Ott,
Samantha Couture, Heather Hutchison, the talented teachers and
staff in the Schenectady City School District, and a long list of former
Union College students. Rosie thanks Pauline Wein, the wonderful
staff at the Patty Jischke Early Care and Education Center, the
dedicated teachers in the Lafayette School Corporation, numerous
coaches, and most especially her parents, Dale and Janice Clawson,
who are always willing to keep their grandson for days on end. Rosie
also thanks Lori Norris and Sharon Phillips for their assistance with
household matters.
Finally, we owe special thanks to our families. We were both raised
by parents who placed priority on education and who encouraged us
to pursue whatever channels most interested us. Their faith that we
would succeed in our chosen career paths provided us with the
confidence to try to do just that. Sadly, Zoe’s mother passed away
before this book was first completed. She was very pleased to learn
that we were writing a book, and we know that she would have been
proud to read it. When writing the first edition, Rosie’s husband was
deployed much of the time. He was around for most of the later
revisions, although Rosie is quite sure there were times he would
have preferred Iraq or Afghanistan to yet another conversation about
public opinion. When he’s not around (and even when he is), Rosie
depends heavily on her family—a big thanks to Dale and Janice
Clawson; Tammy, Mike, Troy, and Jared Harter; Jill, Scott, Liv, and
Sadie Castleman; and her Cleveland cousins. Our sons Alonzo and
Owen bring us tremendous joy. We are thrilled to say they enjoy
discussing with us their views of the world, political and otherwise.
We are fortunate to have very supportive husbands. Des and Dale
not only enjoy talking about politics and have useful computer skills
that we have put to good use, but they also do a disproportionate
share of the household and parenting duties when we are immersed
in writing. And they provide us with many needed distractions from
our work. We don’t know how we got to be so lucky.
PART I WHAT SHOULD THE
ROLE OF CITIZENS BE IN A
DEMOCRATIC SOCIETY?
CHAPTER 1 PUBLIC OPINION IN A DEMOCRACY
AMERICAN NATIONALISM. Populism. Nativism. Identity politics. Racism. Sexism. Antidemocratic
impulses. Support for authoritarianism. These are among the perspectives and attitudes of some
members of the American public that received significant attention from political commentators and
journalists during the lead-up to or since the election of Donald Trump as the U.S. president in 2016.
Over this same time period, others have commented on the rise of political interest and attention,
engagement in political protests, support for democratic socialism, and tolerance for diversity that have
characterized some segments of the American public. These dueling characterizations of the citizenry
also hint at other features of the contemporary political landscape: division and polarization.
Placing so much high-profile attention on the views of the public reminds us that in a democracy, such
as the United States, what the people think matters. Describing and analyzing citizens’ political
perspectives is a worthy endeavor. More broadly, in democratic nations we expect the public to have a
role in governmental decision making. Yet the precise role that citizens should play in a democracy has
been argued about for centuries. Whether the public actually can and really does live up to democratic
expectations is also a debatable topic. In the pages that follow, we explore the normative issues related
to how the public ought to function in a democracy. Throughout this book, we review empirical studies of
public opinion that describe how the public actually functions in America. We then link these studies
back to the normative theories of how citizens should behave in a democracy. Focusing on public
opinion from these two angles will, we hope, provide you with a broad understanding of this important
topic. We will also devote attention to most of the views of the public mentioned in the opening
paragraph, in particular describing whether these trends are unique to today’s political world or were
present before the 2016 presidential election.
THEORIES OF DEMOCRACY
A simple definition of democracy is “rule by the people.” What exactly, however, does rule by the people
mean? Answering this and related questions about democracy is neither easy nor straightforward. In
fact, many people across many centuries have devoted their lives to examining democracy and
delineating the proper characteristics of a democracy. Democratic theory is “the branch of scholarship
that specializes in elucidating, developing, and defining the meaning of democracy.”1 Among other
topics, democratic theorists deliberate over how the people should rule in a democracy (by voting
directly on all laws or by electing representatives for this task) as well as who should qualify as a
democratic citizen (all adults, only those who are educated, or some other group). Democratic theorists
also focus on citizens’ ruling capabilities and the role of the public in a democracy, as indicated by the
following overview of major democratic theories.
Classical Democratic Theory
The earliest Western democratic societies emerged in the city-states of ancient Greece. In Athens’s
direct democracy, for example, governing decisions were made by the citizens, defined as all nonslave
men of Athenian descent. All citizens were eligible to participate in the Assembly, which met at least
forty times per year. Assembly members debated all public issues, often at great length, before making
any final decisions. The Assembly tried to reach a consensus on all matters, and unanimous decisions
were preferred, under the belief that the common interest would only be realized when everyone
agreed.2 When unanimity was not possible, decisions were made via voting in the Assembly. The
implementation of the Assembly’s decisions was conducted by smaller groups of men, who had been
selected by lot or directly elected by the Assembly. These officials served for short periods of time and
were not allowed to serve multiple terms in a row. These procedures ensured that many different men
would serve in this executive capacity and that all citizens would have an equal chance of fulfilling these
roles.3
One of the few surviving descriptions of Athenian citizens and their democratic participation is contained
in Pericles’s oration at a funeral for fallen soldiers:
It is true that our government is called a democracy, because its administration is in the hands,
not of the few, but of the many; yet while as regards the law all men are on an equality for the
settlement of their private disputes, as regards the value set on them it is as each man is in any
way distinguished that he is preferred to public honors, not because he belongs to a particular
class, but because of personal merits; nor, again, on the ground of poverty is a man barred
from a public career by obscurity of rank if he but has it in him to do the state a service. … And
you will find united in the same persons an interest at once in private and in public affairs, and
in others of us who give attention chiefly to business, you will find no lack of insight into political
matters. For we alone regard the man who takes no part in public affairs, not as one who minds
his own business, but as good for nothing; and we Athenians decide public questions for
ourselves or at least endeavor to arrive at a sound understanding of them, in the belief that it is
not debate that is a hindrance to action, but rather not to be instructed by debates before the
time comes for actions. For in truth we have this point also of superiority over other men, to be
most daring in action and yet at the same time most given to reflection upon the ventures we
mean to undertake; with other men, on the contrary, boldness means ignorance and reflection
brings hesitation.4
As Pericles portrays, Athenian democracy was characterized by the active participation of public-spirited
men. In fact, he labeled “good for nothing” those men not taking part in public affairs. This passage also
alludes to other key characteristics of democratic citizenship that appear in classical models of
democracy, such as high levels of attention to and interest in political matters and the capability of
deciding matters in favor of the general interest rather than only to advance one’s own selfish interests.
Writing centuries later, Jean-Jacques Rousseau proposed a theory of democracy that has much in
common with the classical model. Rousseau strongly advocated popular sovereignty, the principle that
citizens hold the ultimate power in a democracy. He argued in The Social Contract that “sovereignty [is]
nothing other than the exercise of the general will” and “since the laws are nothing other than authentic
acts of the general will, the sovereign can act only when the people is assembled.”5 Rousseau also
distinguished the “general will” from the “will of all”: “the general will studies only the common interest
while the will of all studies private interest, and is indeed no more than the sum of individual desires.”6 In
other words, the general will is not determined by simply adding up every person’s individual opinions
but, rather, reflects what is in the best interest of the entire society. Procedurally, Rousseau favored a
direct democracy in which all citizens (restricted to property-owning free men) were to meet, discuss,
and decide on the content of the laws. As in the Athenian Assembly, Rousseau envisaged vigorous
legislative debate with a preference for unanimous decisions. Active political participation by the
citizenry served multiple purposes for Rousseau. It was the only method by which the general will could
be reached and enshrined in law. Active participation was also beneficial for the individual participants;
in other words, political participation had “intrinsic value … for the development of citizens as human
beings.”7
Rousseau’s theory did depart from classical democratic theory in two important ways. First, Rousseau
preferred that the citizens not be as involved in implementing the laws as they were in crafting
legislation. He placed less faith in the public’s ability to execute laws and proposed that a body of
administrators be selected for this duty.8 The administrators would be selected by the citizens and would
be expected to follow the general will but would be distinct from the citizen assembly. Second,
Rousseau’s vision of democracy relied on relative economic equality among citizens, as enshrined by all
free men having only a limited right to property. This does not mean that Rousseau favored strict
equality of property but, rather, that he opposed unlimited accumulation of wealth. Short of this, some
inequality was acceptable. Further, according to Rousseau, a citizen would not be able to make
decisions for the benefit of all if he were motivated by fear of losing his economic independence. The
right to enough property to make each citizen economically free from other citizens would prevent the
formation of groups motivated by economic self-interest. Rousseau feared that the existence of such
groups would undermine the creation of laws benefiting the common good.9 In short, economic
inequality could produce undemocratic effects.
Later democratic theorists and practitioners have criticized classical democratic theory as unworkable
for most societies. First, the city of Athens and Rousseau restricted citizenship rights to a degree that
has become unacceptable for many democracies. In both cases, only free men were citizens; women
and slaves were not given political rights. Further, the existence of a slave economy in Athens and the
reliance on women for unpaid domestic labor created much leisure time for the free men to participate in
government.10 The amount of time necessary to participate in the Assembly debates (forty times per
year!) is simply not feasible for most contemporary working adults. Second, most democratic polities are
larger than were the Greek city-states or the eighteenth-century towns of Rousseau’s Europe. In fact,
both the Greeks and Rousseau assumed that “[only] in a small state, where people could meet together
in the relative intimacy of a single assembly and where a similarity of culture and interests united them,
could individuals discuss and find the public good.”11 One of the primary reasons more modern
democratic theories, including those that follow, departed from the classical variants was to
accommodate popular rule in large, diverse, and populous nation-states. In fact, and as will become
clear as you proceed through the chapters of this book, some democratic theories have very much
evolved away from classical democracy in an attempt to speak to actual conditions in present-day
societies. In contrast, other theorists emphasize that classic democratic features are possible, even
needed, in modern-day complex societies. Finally, contemporary democratic theorists differ along other
criteria as well, such as their trust in the capabilities of the public.
Theories of Democratic Elitism and Pluralism
In contrast to classical democracy, theories of democratic elitism and pluralism do not allocate to
citizens direct involvement in governmental decision making. Rather, the citizenry exerts indirect control
by electing officials to represent their views and make decisions. This, of course, is the defining
characteristic of a representative democracy. Democratic elitists view frequent competitive elections
as the primary mechanism by which citizen preferences are expressed. Voters select their preferred
candidates, and the elected officials deliberate over and vote on the nation’s laws. These officials (or
political elites) are accountable to the public in that they must periodically run for reelection. Thus, the
elites have an incentive to represent the wishes of the public, and the will of the public will be reflected,
to some degree, in governmental decisions. Yet the daily decisions are made by the elites, who, by their
knowledge and expertise, are better able to make these decisions. Joseph Schumpeter outlines his
theory of democratic elitism as follows:
Suppose we reverse the roles of these two elements [the selection of representatives and the
decision-making power of the voters] and make the deciding of issues by the electorate
secondary to the election of the men who are to do the deciding. To put it differently, we now
take the view that the role of the people is to produce a government. … And we define: the
democratic method is that institutional arrangement for arriving at political decisions in which
individuals acquire the power to decide by means of a competitive struggle for the people’s
vote.12
Pluralists also view competitive elections as one important mechanism by which citizens hold elected
leaders accountable. Unlike democratic elitists, however, pluralists emphasize the essential role
performed by groups in representative democracies. Interest groups are collections of like-minded
individuals that attempt to influence elected officials and other governmental decision makers regarding
issues of concern to them. As intermediaries between the public and the elites, such groups are
especially important for transmitting the wishes of the citizenry to government officials in between
elections. According to pluralists, when many groups are actively engaged in debating public issues,
bargaining ensues among the groups and the public policies that result are compromises among the
various groups’ preferences.13 Because interest group leaders have the desire and knowledge to lobby
government officials, members of the public do not need to be actively involved to have their views
represented in lawmaking. For example, citizens who care about human rights do not need to write
letters to their elected officials but can, instead, have their concerns vocalized by an interest group such
as Amnesty International or Human Rights Watch. Leader responsiveness to public concerns should
result, argue pluralists.
Why have democratic elitists and pluralists proposed a more minor role for citizens in democratic
politics? Simply put, “the individual voter was not all that the theory of democracy requires of him.”14 In
practice, much evidence suggests that not all citizens are interested in or knowledgeable about politics,
that levels of citizen apathy run high, and that many do not participate in politics. This evidence,
collected by social scientists beginning in the 1940s, contributed to the development of democratic
elitism and pluralism.15 Indeed, it was the disconnect between dominant democratic theories and the
reality of life in existing democracies that focused theorists’ attention on actual democratic practices.16
Put another way, the theories of democratic elitism and pluralism were constructed by examining
contemporary democracies to determine what features they shared, particularly the levels of political
involvement and interest among the citizenry.17 Note that deriving a democratic theory based on
observations from existing democracies results in a very different theory than that which emerged from
ancient Athens. Having said that, other democratic theorists, such as the participatory democrats we
profile in the next section, interpreted the same social science evidence rather differently.
Contemporary democratic elitism and pluralism can trace their intellectual roots to earlier theorists of
representative democracy, such as the English philosophers Jeremy Bentham and James Mill and the
American James Madison.18 These earlier theorists, especially Madison, advocated that most people
are not capable of the democratic citizenship captured by Pericles in his funeral oration. In Federalist
No. 10, written in 1787, Madison argues that humans are self-interested and will pursue what benefits
themselves rather than the nation as a whole. In societies where the liberty of individuals to form their
own opinions and pursue their own goals is ensured, groups of similarly interested people will form. By
Madison’s definition, such groups, or factions, consist of citizens “who are united and actuated by some
common impulse of passion, or of interest, adverse to the rights of other citizens, or to the permanent
and aggregate interests of the community.”19 To overcome the negative effects of such factions, the
causes of which are “sown in the nature of man,” Madison proposes a republic in which a few citizens
are elected by the rest of the public to serve in the national government.20 In his own words,
The effect of [a representative democracy] is … to refine and enlarge the public views by
passing them through the medium of a chosen body of citizens, whose wisdom may best
discern the true interest of their country and whose patriotism and love of justice will be least
likely to sacrifice it to temporary or partial considerations. Under such a regulation it may well
happen that the public voice, pronounced by the representatives of the people, will be more
consonant to the public good than if pronounced by the people themselves, convened for the
purpose.21
Similar beliefs in the decision-making superiority of elite officials are reflected in the writings of
contemporary democratic elitists and pluralists. In an especially uncharitable view of the public, Joseph
Schumpeter states as fact “that the electoral mass is incapable of action other than a stampede.”22
More broadly, he argues that the public is capable of voting but little else and that therefore the elites
should be allowed to make decisions in between elections without public interference. Elite control over
decision making should also result in more stable governments, with fewer changes in policy due to
public impulses. Some theorists also emphasize that elites are more supportive of democratic norms
and values, especially the civil rights and liberties of marginalized and/or unpopular groups, than are
members of the public. In general, they suggest, this support for rights and liberties is beneficial to a
democracy where decision making is in the hands of the elite.23 The elites are not immune from public
pressures to restrict individual liberties but will typically sort out such issues among themselves, with a
preference toward maintaining such liberties.
Critiques of democratic elitism and pluralism have come from many quarters. As previously mentioned,
participatory democrats interpret the empirical evidence related to citizen participation vastly differently
than do democratic elitists and pluralists. Others have contradicted the pluralist assumptions that
interest groups will represent all points of view and that governmental officials are responsive to these
groups. Government officials can choose to ignore a group’s demands, especially when they believe the
group lacks widespread public support. For example, public outcry in favor of the principle of net
neutrality contributed to decision making at the Federal Communications Commission (FCC). In the
spring of 2014, the FCC proposed rule changes that would have permitted the content on some
websites to be transmitted more quickly than on other sites. Initially, a few interest groups and Internet
companies were active in opposing the proposed rules. Once word of these possible changes spread
more broadly, thanks in part to coverage on John Oliver’s HBO show Last Week Tonight, the FCC
received millions of public comments. Most people advocated for an open, neutral web whereby Internet
service providers cannot speed up or slow down the delivery of a website’s content. The FCC changed
course. In February 2015, they dropped their original proposal and instead voted in favor of new
regulations that promote net neutrality.24
Further, some groups possess more resources than others and thus have more influence over
policymaking. As well stated by political scientist E. E. Schattschneider decades ago, “The flaw in the
pluralist heaven is that the heavenly chorus sings with a strong upper-class accent.”25 This fact did not
go unnoticed by pluralists. Some accepted the inequality of political resources and argued that the
inequalities did not accumulate within certain types of people but, rather, were dispersed throughout
society. In other words, “individuals best off in their access to one kind of resource are often badly off
with respect to many other resources. … Virtually no one, and certainly no group of more than a few
individuals, is entirely lacking in some influence resources.”26 Pluralists, however, did not fully develop
the implications of group inequalities, an oversight that has been somewhat rectified by more recent
theorists in this area.27 Assumptions about noncumulative inequalities have also been challenged.
Business groups, these critics contend, occupy a privileged position in U.S. politics due to their wide
array of resources28 and indeed are more likely than the public or other types of interest groups to have
their wishes enshrined in public policy.29 For an example, let’s return to the topic of net neutrality. After
Donald Trump was inaugurated as president in 2017, he appointed a new chair of the FCC, one who
favors business deregulation. In pursuit of this goal and with the support of interest groups representing
the cable and telecommunications industries, yet despite millions of public comments that urged the
opposite, FCC members voted to overturn the net neutrality rules that had been adopted in 2015.30
Finally, Jack Walker’s assessment of democratic elitism takes quite a different form. He charges the
democratic elitists with changing “the principal orienting values of democracy.”31 Earlier democratic
theorists stressed the importance of citizen participation and the personal benefits that accrue to
individuals from this participation. In contrast, under democratic elitism, “emphasis has shifted to the
needs and functions of the system as a whole; there is no longer a direct concern with human
development. … [Elitists] have substituted stability and efficiency as the prime goals of democracy.”32
Participatory democracy, the final democratic theory we examine, represents a shift back toward the
developmental functions of democracy that Walker supports.
Participatory Democracy
As its name suggests, participatory democracy emphasizes the importance of political participation by
the public. Whereas participatory democrats recognize the need for representative democracy in nations
as large as the United States, they also see the possibility and benefits of more political involvement by
the public than is currently practiced.33 Because participation is linked to social class and wealth today,
participatory democrats advocate greater political involvement of all citizens as a means to redress
inequality. “This is not to say that a more participatory system would of itself remove all the inequities of
our society,” writes one theorist. “It is only to say that low participation and social inequity are so bound
up with each other that a more equitable and humane society requires a more participatory political
system.”34
Box 1.1 Gendered Nouns and Pronouns
“… factions, the causes of which are ‘sown in the nature of man …’”
James Madison, 1787
“… make the deciding of issues by the electorate secondary to the election of the men
who are to do the deciding.”
Joseph Schumpeter, 1976 (originally published in 1943)
“The individual voter was not all that the theory of democracy requires of him.”
Bernard Berelson, Paul Lazarsfeld, and William McPhee, 1954
When you read the preceding quotations, did they sound unusual to you? Did you stop and
wonder whether the original writers really meant their statements to refer only to men? Or are
women implicitly included as well? Would these statements have taken on a different meaning if
a female noun or pronoun had been used? What if Madison had stated that the causes of
factions are “sown in the nature of woman”? Would you have paused and wondered about that
statement? Today, writers often substitute “him or her” for “him” or even alternate using “him” and
“her” or “man” and “woman” when their statements apply equally to men or women. This was not
always the case, however, and certainly was not the norm in the 1700s or even as recently as
the 1950s. One way to determine whether the authors meant to refer only to men when they
wrote these sentences would be to read more writings by these authors to try to determine their
opinions regarding the political roles and rights of women. It is useful to bear in mind, however,
that women’s increasing involvement in politics has been accompanied by changes in language
use (not coincidentally). Early theorists might not have made their views toward women’s role in
politics known because this role was minimal, by law and by custom. Furthermore, when women
did engage in political activities, they were not viewed as political actors and could more easily
be overlooked. Thus, in some instances, it can be difficult to sort out whether these writers really
meant to refer to men only or whether by “man” they really meant “human.” We encourage you
not to just assume that using “man” implies women as well but, rather, to consider the time
period in which the author was writing and the nature of his or her conclusions regarding women
and men in politics. In other words, do stop and think when you encounter “him” or “man” rather
than merely breezing over these words.
Beginning with the first edition of this book (published in 2008), our approach was to alternate
using male and female pronouns when our statements were meant to apply to both women and
men. Unless otherwise specified, when we wrote “she” or “her” we could have also written “he” or
“him.” However, conventions regarding personal gender pronouns have changed during the
years that our book has been in print. This change reflects the reality that some people, including
but not limited to transgender individuals or people who identify as non–gender binary, recognize
that gendered pronouns such as “she” and “he” are not accurate for everyone. Gender-neutral
pronouns such as “ze,” “hir,” and “they” (used in the singular) thus have become more common
in the English language. With this new edition of our book, we also change our language, now
using the pronouns “she/her,” “he/him,” and “they/them” interchangeably.
This theory of democracy originated during the protest movements of the 1960s and also represented
dissatisfaction with the democratic elitist and pluralist models that were dominant at that time.35
Participatory democrats agreed with these other theorists that levels of disinterest and apathy ran quite
high among the American public, but they disagreed over the reason for these attitudes. Rather than
citizens being politically disinterested by nature or simply preferring to spend their time on other pursuits,
such as family, work, and leisure, participatory democrats argue that the political system, with its
relatively few opportunities for meaningful citizen influence, breeds apathy. To political scientist
Benjamin Barber, people “are apathetic because they are powerless, not powerless because they are
apathetic. There is no evidence to suggest that once empowered, a people will refuse to participate. The
historical evidence of New England towns, community school boards, neighborhood associations, and
other local bodies is that participation fosters more participation.”36 Citizen apathy is thus a problem to
be examined and solved rather than an accepted fact of political life in modern democracies.37
Participation in democratic decision making provides many personal benefits to those who engage in
this activity, according to participatory democrats. On this point, they agree with democratic theorists of
earlier eras, especially the nineteenth century’s John Stuart Mill.38 Citizens become more politically and
socially educated and can develop their intellect and character through political participation. By
communicating with and learning from other members of the public, individuals can look beyond their
own self-interest and come to know what is best for the community or nation as a whole. In short,
participation, in and of itself, can produce better democratic citizens.39 Peter Bachrach, in articulating
his vision of democracy as fostering individual self-development, states, “The majority of individuals
stand to gain in self-esteem and growth toward a fuller affirmation of their personalities by participating
more actively in meaningful community decisions.”40
According to some participatory democrats, a fully participatory society necessitates more citizen
involvement in decision making in governmental as well as nongovernmental institutions, such as the
workplace or school. As Bachrach asks, why should people be excluded from decision making by
private organizations when these decisions strongly affect their own lives and livelihoods?41 Further,
engaging in decision making at work and in other nongovernmental venues could increase
governmental participation. Engagement in workplace decision making fosters civic skills, provides
valuable experience, and, if effective, could create more confidence in an individual’s ability to influence
governmental decisions.42 The flip side of this argument is that the lack of involvement in decision
making in daily life might hinder political engagement: “After spending the day following orders without
question at the factory, a worker cannot be expected to return home in the evening to act like the civics
textbook’s inquiring, skeptical, self-actualizing citizen. Students who are taught primarily to obey
authority in school are not likely to grow into effective democratic citizens.”43
Skeptics of participatory democracy argue that the public does not respond to participatory opportunities
as the theorists contend they will. When barriers to political participation are eliminated or reduced,
citizens have not necessarily become more politically active. For example, the National Voter
Registration Act of 1993, more commonly known as the motor-voter bill, made voter registration easier
and, supporters alleged, would increase voter turnout once enacted. Even though registration rates did
increase in the wake of this reform, the levels of voter turnout did not substantially increase because of
motor-voter.44 More broadly, some scholars conclude that participatory democrats’ assumptions about
the public might be unrealistic.45 Rather than desiring to become more involved in politics, some citizens
actually dislike politics and wish to avoid the type of conflict that typically emerges during decision
making. Yet, other research reaches a different conclusion, demonstrating not only that people wish to
have more opportunities for meaningful deliberative democracy but that folks who are especially turned
off by the current political scene are particularly interested in this type of political participation.46 In short,
research has uncovered evidence that some citizens wish to engage with and learn from one another,
yet other citizens prefer to bypass any opportunity for deliberation, especially if the chance of
disagreement is high.
Democratic Theory and Public Opinion
As you can see, these theories of democracy are quite broad, addressing many features of democratic
governance. In our overview in this first section (Part I) of the book, we have highlighted aspects of the
theories that are most relevant for the study of public opinion. In particular, we have discussed how the
different theoretical perspectives answer this question: what should the role of citizens be in a
democratic society? This is a key issue that democratic theorists have long debated. In fact, we have
organized this textbook around fundamental questions that speak to democratic theory debates about
the public.
Part II of the book addresses an important question about the capabilities and competence of citizens:
are citizens pliable? Classical democratic theorists and participatory democratic theorists envision
citizens who hold informed, stable opinions based on reason and concern for the general will. At the
same time, these theorists believe democratic citizens should be open to persuasion from others but not
so open that their brains fall out. In other words, citizens should change their attitudes based on
information and evidence, not simply change their minds willy-nilly. As we have discussed, elite
democrats and pluralists have lower expectations for the public. They presume that many citizens’
opinions are ill-informed and that citizens are often influenced by political leaders, the media, and
reference groups in society. By examining the role of socialization in shaping political views, the effects
of the mass media on opinion, and the stability and instability of political attitudes, Part II of this book
addresses the pliability of the public.
Do citizens organize their political thinking? This critical question, addressed in Part III of the book,
speaks directly to the quality of public opinion. Classical democratic theorists and participatory
democratic theorists expect citizens to hold a wide range of political attitudes that are organized in a
meaningful fashion. For participatory democrats, it is crucial that citizens have a sophisticated
understanding of politics so they can voice their views and influence elected officials. If such
sophistication is lacking, identifying the reasons why is important for these theorists, so that corrective
action can be taken. Elite democrats, on the other hand, envision a citizenry that is much less proficient,
although they still want citizens to be competent enough to hold officials accountable at election time.
Part IV examines citizens’ appreciation for essential aspects of democratic citizenship and governance
by asking this question: do citizens endorse and demonstrate democratic basics? The democratic
basics we focus on are knowledge of, interest in, and attention to politics; support for civil liberties; and
support for civil rights. Whether the public is knowledgeable and interested enough for democratic
governance has long divided democratic theorists. Theorists also disagree on what level of citizen
support for civil liberties and civil rights is needed for a healthy democracy. Classical and participatory
democratic theorists, of course, want citizens to value these democratic basics. Elite democrats and
pluralists worry much less about such matters, primarily because they view elites as the key actors in a
democracy.
Part V of the book addresses a pivotal question about the nature of citizenship in a democratic society:
what is the relationship between citizens and their government? Classical and participatory theorists
want citizens to be actively involved in politics. Participatory democrats expect leaders to take public
opinion into consideration as they make decisions, which would lead citizens to trust their government.
Elite democratic theorists, in contrast, value trust in government for the stability it brings to the political
system, not because it is a function of citizens being pleased with the responsiveness of their
government. Further, elite democrats prefer that the public’s influence on government is largely limited
to voting in elections.
In the final section (Part VI) of the book, we pull it all together with this question: what do we make of
public opinion in a democracy? We review the theoretical debates and summarize the empirical
evidence, but ultimately we leave it to you to make sense of the role of citizens in a democratic society.
WHAT IS PUBLIC OPINION?
Public opinion is, on the one hand, a term that is familiar to most people and, on the other hand, rather
difficult to define. Popular conceptions of public opinion might include phrases such as “the voice of the
people.” For most of us, public opinion is probably best represented by the results from opinion polls,
such as those reported on the evening television news, in the newspaper, or on online news sites.
Among public opinion observers and scholars, many different definitions have been proposed. Although
researchers do not agree on one single definition of public opinion, some commonalities exist across
specific definitions. First, most emphasize that public opinion refers to opinions on governmental and
policy matters rather than on private matters (such as one’s favorite flavor of ice cream or favorite
movie). This characteristic is implied by a description of public opinion as “those opinions held by private
persons which governments find it prudent to heed.”47 Of course, what constitutes a private matter
might be in dispute. For centuries, the problems of domestic violence and rape within marriage were
considered to be private affairs best left to a married or intimate couple to resolve. Societal views on this
topic have changed, however, so that now people assume governments have to be involved in
addressing these serious problems.
Second, in recent decades a consensus definition of public opinion has emerged. As one example,
public opinion has been defined as “simply the sum or aggregation of private opinions on any particular
issue or set of issues.”48 In this view, public opinion refers to the preferences of individuals, tallied such
that each person’s opinion counts equally. Following the consensus, this is the definition that we use in
this book.
However, despite the consensus, some have raised important objections to defining public opinion as a
“one person, one vote” aggregation. One of the earliest critiques came from sociologist Herbert Blumer.
Society, according to Blumer, is organized hierarchically and “is not a mere aggregation of disparate
individuals.”49 Certain individuals have more influence over the formation and expression of people’s
opinions, and treating each person’s opinion as equal ignores this simple fact. For example, the leaders
of labor unions not only attempt to influence the opinions of their members but also present their
members’ views to government policymakers. Simply tallying up individuals’ opinions on a specific issue
also overlooks the dynamic opinion formation processes among groups and among people. In Blumer’s
words, public opinion “is a function of a structured society, differentiated into a network of different kinds
of groups and individuals having differential weight and influence and occupying different strategic
positions.”50 Blumer further attacks the “one person, one vote” accounting of opinions by arguing that
not all opinions are treated equally by government policymakers, in part because not all opinions of the
public actually reach these policymakers. Opinions that do not come to the attention of decision makers
will not influence their decisions.
Blumer directs his criticisms toward the public opinion polling industry, arguing that polls are incapable of
capturing public opinion as he understands the concept. By reporting the opinions from a random
selection of individuals, polls epitomize the “one person, one vote” aggregation of people’s preferences.
Not only are polls an unnatural forum for expressing one’s opinions, argues Blumer, but they also are
unable to capture the opinion formation process that he identifies. Opinion polls do not report, for
example, whether a poll respondent “belongs to a powerful group taking a vigorous stand on the issue
or whether he is a detached recluse with no membership in a functional group; whether he is bringing
his opinion to bear in some fashion at strategic points in the operation of society or whether it is isolated
and socially impotent.”51
Blumer wrote in 1948, at a time when public opinion polling was in its infancy. Opinion polls have grown
in use and influence since then, becoming the dominant method by which public opinion is assessed.
Further, as this one method has become dominant, there has been a narrowing in our understanding of
public opinion—a narrowing around the consensus definition previously described.52 Despite this,
Blumer’s insights are spot-on in the twenty-first century, argues Susan Herbst.53 Herbst, a public opinion
scholar, encourages us to reconsider what public opinion means in our digital age. Citizens engage in
political conversations through a variety of means these days: the Internet, cell phones (talking and
texting), and social media, to name a few. It is in these (often digital) exchanges where public opinion is
to be found and understood. Herbst labels these communication patterns “textured talk” and finds them
“so superior to the aggregation of anonymous individuals gathered in our artificial ‘publics’ produced by
polls.”54 In addition to providing more convenient outlets for political expression, new communication
technologies have also created audience segmentation. Should they choose to, citizens can rely on
digital sources that convey information on specific topics or that present information from only one
political viewpoint. For instance, political websites are ubiquitous on the Internet, and many of them are
very narrowly focused by issue or by ideology. As technology has led to a rise in public segmentation,
should public opinion continue to be defined as the aggregation of each individual’s opinion? Perhaps,
rather than an aggregate public, it is more accurate today to conceive of public opinion as arising from
multiple different publics.55
In her book The Politics of Resentment, Katherine Cramer argues forcefully that public opinion is much
more than what opinion polls (can) measure.56 As does Herbst, Cramer draws upon Blumer’s critiques.
Yet, Cramer finds meaningful public opinion in everyday, face-to-face conversations among people
rather than in the digital exchanges that Herbst explores. More specifically, by listening in on many
conversations among small groups of friends and colleagues, Cramer is able to examine how people
make sense of politics and public policies. She is particularly interested in what shapes people’s
understandings of social and political conditions, specifically individuals’ social identities as well as
messages communicated by the media and politicians. She further examines how these understandings
develop, change, or are reinforced via communication with other citizens. The nuanced portrait of public
opinion she presents is one that could not be captured by polling data.
Others have criticized the consensus approach by emphasizing the poor quality of public opinion as
assessed by polls, arguing that survey respondents often provide snap, top-of-the-head judgments.
Contrast this with public judgment, a state that exists when “people have engaged an issue,
considered it from all sides, understood the choices it leads to, and accepted the full consequences of
the choices they make.”57 Encouraging and cultivating thoughtful public judgment, according to this
view, is necessary if we want the public—and not only those people with specialized knowledge and
expertise—to govern in a democracy.
We mention these criticisms not because we find them superior to the consensus definition of public
opinion. Instead, we are sympathetic to these concerns because we find the “public opinion as
aggregation of individual views” definition too limiting. In addition to the concerns already outlined, we
are troubled that this consensus approach draws our attention only to one feature of public opinion: the
content of people’s political opinions. Although it is important to know how the public feels about an
issue, focusing only on the content of people’s opinions overlooks many equally important features of
public opinion. Understanding public opinion requires us to explore other topics, such as the sources of
those opinions, the processes by which opinions are formed and altered, the organization of an
individual’s opinions, and the impact of public opinion on public policy. In the chapters that follow, we
describe studies that illustrate a variety of definitions of public opinion. Along the way, therefore, we
touch on the many facets of public opinion. But, as will become evident, most scholars of public opinion
do rely on the consensus definition of public opinion, whether implicitly or explicitly.
DEFINING KEY CONCEPTS
Each of the chapters in this book addresses a specific aspect of public opinion in America. In these
chapters, you will repeatedly encounter a few of the same concepts and terms. We define those
concepts here, so that you will understand the later chapters more thoroughly.
Attitude and Opinion
Two terms that we use frequently in this book are attitude and opinion. These words are undoubtedly
familiar to you, and you will probably agree that they are similar to each other. They both have sparked
considerable attention to their meanings, however, and numerous definitions have been proposed for
each, especially for attitude. The term attitude is one of the most important concepts in psychology and
has been for many years. Over seventy-five years ago, a prominent social psychologist presented a
“representative selection” of sixteen definitions of attitude and then proposed his own comprehensive
definition.58 In the many decades since, still more scholars have discussed and debated the meaning of
attitude. Of the many possible definitions of attitude, we prefer this one: “Attitude is a psychological
tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor.”59 A
similar approach defines an attitude as “a general and enduring positive or negative feeling about some
person, object or issue.”60
These two definitions highlight some key features of an attitude. First, people hold attitudes toward
targets (“entity” or “person, object or issue”). In the realm of political attitudes, possible types of objects
for which we have attitudes are policy issues, political candidates or politicians, groups (such as the
National Rifle Association [NRA] or feminists), and institutions of government. Second, attitudes
represent an evaluation of an object, generally articulated as favorable or unfavorable, as liking or
disliking, or as positive or negative. So, in terms of specific political attitudes, your friend might favor
school prayer, dislike U.S. senator Elizabeth Warren, support the NRA, like conservatives, and
disapprove of Congress. It is also possible to have a neutral (neither favorable nor unfavorable) attitude
toward a target. Neutral attitudes might result from not being informed enough about an object to
evaluate it positively or negatively. Alternatively, you might assess certain features of an object positively
and other features negatively. If these cancel each other out and prevent you from an overall positive or
negative evaluation of the object, you might conclude that your attitude is neutral.
So, now, what is an opinion? Similar to an attitude, an opinion refers to a specific object and expresses
a preference, such as support or opposition, toward that object. As with attitudes, opinions vary in that
not everyone holds the same opinion toward an object. Though acknowledging these similarities, many
scholars distinguish between these two concepts by stating that an opinion is an expression of a latent
attitude. That is, whereas an attitude is not observable, an opinion is a verbal or written expression of
that attitude. Distinctions such as these are more common in the field of psychology than in political
science. In political science, you are not only likely to see the two terms used synonymously but are also
more likely to encounter the concept of opinion than attitude. We view these two terms as much more
similar than dissimilar and thus use them interchangeably in this book. This no doubt reflects our training
as political scientists, but it also reflects common use of the terms. In fact, in most thesauruses, opinion
and attitude are presented as synonyms of each other.
When thinking about a specific attitude or opinion, it is obviously important to consider its direction
(support vs. oppose, favorable vs. unfavorable, and so on). For the study of public opinion, we also
need to bear in mind two other characteristics of attitudes and opinions: extremity and importance. The
extremity of an opinion refers to whether support (or opposition) for the opinion object is slight or
strong. You might, for example, slightly favor U.S. intervention in foreign military conflicts but strongly
favor laws that prohibit testing cosmetics on animals. Attitude importance, in contrast, focuses on how
meaningful a specific attitude is to you or how passionately you care about the attitude. Although we
may have attitudes toward a wide range of political and social objects, not all of these attitudes will be of
equal importance, at least for most of us. The more important an attitude is, the less likely it is to change
over time and the more likely it will direct certain behaviors, such as thinking about the attitude object or
influencing our vote preferences for political candidates.61 Also, even though it is often the case that
more extreme attitudes are also more important, this does not necessarily have to occur.62 Take the two
examples presented here. You might have a more extreme opinion toward animal testing than U.S.
military intervention, but the latter opinion might be more important to you, especially when it comes to
evaluating national politics, such as the performance of political leaders.
Opinion Ingredients: Beliefs, Values, and Emotions
Specific political opinions do not stand alone in people’s minds. Instead, they are often related to, even
guided by, other mental constructs, most especially beliefs, values, and emotions. These three often
have evaluative content—content that can help to determine an individual’s specific opinion toward a
related entity. Beliefs are thoughts or information a person has regarding an attitude object, often
regarding what the person thinks to be true about the object. A person might, for example, believe that
the possibility of a very severe punishment, such as the death penalty, will not deter most people from
committing a serious crime. Someone possessing this belief would be more likely to oppose capital
punishment than would someone who believes in the deterrent power of death penalty laws. Beliefs
about the characteristics of members of social groups, such as Black people or Christian
fundamentalists, have a specific name, stereotypes. Stereotypes can be positive or negative, and
people can hold both positive and negative stereotypes toward the same group. Examples of positive
and negative stereotypes include viewing Black people as athletic or lazy and Christian fundamentalists
as charitable or intolerant. Believing certain stereotypes is often related to support for public policies that
affect the group in question. White Americans who believe most Black people are lazy, for instance, are
unlikely to support social welfare policies, especially compared with people who do not endorse this
stereotype.63
Values are specific types of beliefs. According to a prominent values researcher, “a value is an enduring
belief that a specific mode of conduct [instrumental value] or end-state of existence [terminal value] is
personally or socially preferable to an opposite or converse mode of conduct or end-state of
existence.”64 Examples of instrumental values include independence, responsibility, and self-control,
whereas examples of terminal values include a peaceful world, family security, and freedom. Unlike
other types of beliefs, values refer to ideals.
Values are also assumed to be quite stable over time for individuals, as highlighted by this definition: “By
values we mean general and enduring standards.”65 Whereas value change can and does occur,
stability is more common. Some have even argued that values are central to people’s political belief
systems, certainly more central than are attitudes.66 Further, much public opinion research
demonstrates that values are quite important in influencing people’s specific political attitudes. For
instance, opposition to social welfare spending is more likely among those who value responsibility, a
sense of accomplishment, and economic individualism and less likely among those who value
equality.67 Finally, certain values are more salient in American political culture than others in that they
guide political opinions more strongly. These include individualism, egalitarianism, and limited
government.68 Not all Americans value these three, to be sure, but whether a person values or does not
value each is related to opinions on many specific political matters.
Whereas beliefs are considered to be the cognitive components of attitudes, emotions make up the
affective component. Emotions are feelings that a person has toward the attitude object and are
oftentimes more consequential than beliefs in attitude formation.69 Emotions are especially common
when it comes to evaluating political individuals or groups. You might feel warmly toward a politician and
thus evaluate her (and even her job performance) highly. In contrast, fearing a politician would probably
lead to poor assessments of her but also might transfer into not supporting the issues that she supports.
Negative affect that is felt toward a specific group is commonly referred to as prejudice and can
influence attitudes toward politicians who are members of that group as well as policies designed to
benefit the group. Emotional reactions can also influence opinions toward political issues or public
policies.70 Anxiety that a foreign leader could detonate a nuclear weapon somewhere on U.S. soil could
lead a person to support a strong national defense and a preemptive foreign policy. Finally, people can
feel positively toward an attitude object but also hold negative beliefs about the object. For example,
someone could admire Latinos for their work ethic while at the same time hold negative stereotypes
about their intelligence or abilities.
Party Identification
Throughout this book, we present examples of many different political opinions. One opinion that we
refer to often, because it is a core opinion for many Americans and crucial to understanding the nature
of public opinion in the United States, is party identification. Party identification refers to a person’s
allegiance to a political party (typically the Democratic or Republican Party) or identification as
independent of a party. It is a self-classification rather than a description of the person’s behavior, as the
following excerpt from The American Voter, a classic study about voting first published in 1960,
highlights:
Only in the exceptional case does the sense of individual attachment to party reflect a formal
membership or an active connection with a party apparatus. Nor does it simply denote a voting
record, although the influence of party allegiance on electoral behavior is strong. Generally this
tie is a psychological identification, which can persist without legal recognition or evidence of
formal membership and even without a consistent record of party support. Most Americans
have this sense of attachment with one party or the other. And for the individual who does, the
strength and direction of party identification are facts of central importance in accounting for
attitude and behavior.71
In other words, a person could consider himself to be a Republican without ever formally registering as
such or without always voting for Republican candidates. Party identification is, instead, an attitude one
has about his attachment to a political party. Typically, then, to determine someone’s party identification,
a survey taker would not ask whom she voted for most recently but, rather, ask her whether she
identifies with a particular party, emphasizing the self-identification component of this attitude. To
illustrate, two examples of questions used by national survey organizations to assess the party
identification of the American public follow:
Generally speaking, do you usually think of yourself as a Republican, a Democrat, an
Independent, or what? [If Republican or Democrat:] Would you call yourself a strong
(Republican, Democrat) or a not very strong (Republican, Democrat)? [If Independent, Other or
No Preference:] Do you think of yourself as closer to the Republican or Democratic Party?72
No matter how you voted today, do you usually think of yourself as a Democrat, Republican,
Independent, [or] Something else?73
We highlight party identification here because it is important in American political culture for a number of
reasons. First, for an individual, party identification is quite stable over time, certainly more stable than
other political attitudes.74 When a change does occur, it is most likely to consist of people switching from
identification with one of the two major parties to considering themselves to be Independent or vice
versa. That is, switching from identifying with one of the parties to the other does not occur very often.
Second, party identification is a global attitude that is related to many specific political attitudes (such as
policy opinions or evaluations of political leaders). Third, people’s party identification can influence the
interpretation of newly encountered information. When learning of damaging information about a
Democratic president, for example, a Democrat is likely to interpret this information quite differently than
a Republican. Related to this, party identification can help a person to make sense of political issues
and topics, especially those that are unfamiliar. We elaborate on these and other aspects of party
identification throughout this book.
In Figure 1.1, we present the breakdown of Americans’ party identifications (as Democratic, Republican,
or Independent) in every presidential election year since 1952. Focus first on the partisans, those who
identify as Democratic or Republican (the solid lines). We see that American adults are much less likely
to identify with the Democratic Party now than they were in the 1950s and 1960s. Whereas one-half of
the population considered themselves Democrats in those decades, since 1988 less than 40 percent
have. This does not necessarily mean, however, that Americans are now identifying as Republicans in
much larger numbers. Rates of Republican self-identification have been very similar in recent elections
to what they were in the 1950s. Perhaps more noteworthy, the percentage of the public identifying as
Republican has not varied much (between 23 and 30 percent) over this time period.
The most significant change in party identification over the past fifty years is the switch from partisans to
partisan independence. In fact, Independents have been more common than either Democrats or
Republicans in five of the last eight presidential election years (refer to the top dotted line in Figure 1.1).
There was a substantial increase in Independents during the 1960s; only about 23 percent of the
population considered themselves to be Independent in 1960, but 35 percent did so in 1972. These
percentages, however, include people who lean toward supporting one of the major parties. That is,
when initially asked whether they consider themselves to be Democratic, Republican, or Independent,
they claim to be Independents. Yet, when then asked if they are closer to one of the parties, most of
these Independents do indicate closeness to one party. Removing these leaners from the analysis
presents a very different picture of Independents (see the dotted line at the bottom of the figure).
Although there are more pure Independents now than there were in the 1950s, the increase has not
been very large (from 6 percent in 1952 to 15 percent in 2016). Most of the increase in Independents,
thus, seems to have been among the leaners, people whose initial identity is as Independent but who
ultimately lean toward a party. Finally, in 2016 leaners were nearly equally likely to feel close to the
Democratic as to the Republican Party. Among the Independents that year, 30.3 percent indicated that
they felt closer to the Republican Party, whereas 29.7 percent leaned toward the Democrats.
Description
Figure 1.1 Party Identification, 1952–2016
Source: Analysis of American National Election Studies Cumulative (1948–2004), American National
Election Studies 2008, American National Election Studies 2012, and American National Election Studies
2016 data files.
EMPIRICAL ASSESSMENTS OF PUBLIC OPINION
As should have been clear from the democratic theory section earlier in this chapter, the main goal of
these theories is to present normative conclusions. That is, most theorists outline how a democratic
government and society ought to be structured, including what ought to be the role of the citizenry.
Democratic theories are not entirely normative, however. According to one view, the best models of
democracy “have been both explanatory and justificatory or advocatory. They are, in different
proportions, statements about what a political system or a political society is, how it does work or could
work, and statements of why it is a good thing, or why it would be a good thing to have it or to have
more of it.”75
To rephrase, theories of democracy can contain normative and empirical components. The empirical
features are statements about how a society actually is, based on observations of democratic societies.
These observations of reality can be important components of democratic theories and can complement
or contribute to a theory’s normative conclusions, as we have discussed with the theories of democratic
elitism and pluralism. In contrast to the normative focus of democratic theories, empirical analyses of
public opinion place primacy on accurately describing and explaining real-life phenomena. Any broader
conclusions, whether normative or otherwise, are of secondary importance for fields of empirical study.
Most public opinion scholars, and nearly all that we feature in this book, are empiricists. Examining
public opinion empirically requires, of course, that this phenomenon be measured in some way.
Measuring public opinion is not an easy or obvious task, however. What if, for example, you wanted to
know whether the public supports same-sex marriage. How would you determine public opinion on this
issue? You could stop people on the street, ask them if they agree gays and lesbians should be allowed
to marry their partners, and tally up the responses. You could read letters to the editor in the newspaper
to gauge public sentiment on same-sex marriage. You could look to elected officials’ statements about
the public to determine how citizens feel about this issue or assume that most people do support same-
sex marriage because the U.S. Supreme Court ruled in favor of such marriages. These are just a few
possibilities, and although each has its advantages (and disadvantages), none can be ruled out as
clearly inappropriate. In fact, all are examples of approaches that have been used, at one time or
another, to measure public opinion. The Appendix to this chapter discusses a variety of public opinion
research methods, methods that we will illuminate throughout the chapters of this book.
THEMES OF THE BOOK
By reading this book, you will learn a lot about public opinion in the United States. One way that we try
to fulfill this goal is by linking normative democratic theories with findings from empirical studies of public
opinion. Rather than only considering what researchers have concluded about public opinion,
throughout this book we discuss the democratic theory implications of a study or a body of research. We
hope this approach will encourage you to evaluate the public opinion research through the lenses of the
democratic theories outlined earlier and also to evaluate the democratic theories in light of the empirical
studies. This will deepen your assessments of these democratic theories and provide you with a broad
understanding of public opinion.
There are many, many empirical studies on public opinion. Summarizing all of these would be a
daunting task for us and would produce a book that would be tedious for you to read. Thus, we have not
attempted to discuss every relevant study on each topic. Instead, in each chapter, we focus on
prototypical and especially influential studies that bring the fundamental questions into sharp focus. Our
discussions of these prototypical studies are detailed and are meant to illustrate the strengths and
weaknesses of each study. To evaluate well the conclusions from a study, you need to know what is
good about the research and what its limits are. Using this approach will, we hope, encourage you to
consider how public opinion is studied as well as better understand how a choice of methodology can
influence the conclusions that researchers draw. We also anticipate that this feature of our text—what
we call “embedded methods”—will teach you about specific research methods better than if the topics
were presented in a separate chapter devoted only to methodology. If at any point you want to have
more information about a method you are reading about, however, you can always refer back to the
Chapter 1 Appendix for details about the specific method.
Finally, as we discussed earlier, the book is organized into parts. Each part poses a question that is
important for democratic theory, and the chapters in the part present evidence and arguments to help
you answer that question. We will not answer these questions, however. In fact, these questions do not
have “correct” answers. Instead, we present evidence and tools to help you think through the material
critically and challenge you to make your own judgment regarding the capacity of citizens to function
effectively in a democracy.
In Part II we address this question: are citizens pliable? We explore this question through specific
chapters devoted to the topics of political socialization, how the mass media shape public opinion, and
attitude stability. Chapters in Part III address this question: do citizens organize their political thinking?
We begin by examining whether individuals’ opinions are organized along a liberal-conservative
dimension or by partisanship. We also focus on other factors that might organize opinions, including
reference groups, personality, self-interest calculations, values, and historical events. The book then
moves on to this question in Part IV: do citizens endorse and demonstrate democratic basics? We focus
on how knowledgeable, interested, and attentive citizens are, and we investigate public support for civil
liberties and civil rights. Next, in Part V, we ask, what is the relationship between citizens and their
government? Specifically, we take up the topics of citizen trust in government and support for the
institutions of government. We also examine the relationship between public opinion and public policy. In
the final part (Part VI), we take a broad assessment of the role of public opinion in the United States,
asking, what do we make of public opinion in a democracy? The concluding chapter reviews normative
debates over the role of citizens in a democracy and summarizes the empirical evidence that speaks to
these debates.
KEY CONCEPTS
attitude 18
attitude importance 19
beliefs 20
citizen apathy 12
classical models of democracy 5
democratic elitists 7
democratic theory 4
direct democracy 5
emotions 21
empirical analyses 24
extremity 19
factions 8
interest groups 7
leaners 24
normative 24
opinion 19
participatory democracy 10
partisan independence 23
partisans 23
party identification 21
pluralists 7
popular sovereignty 5
prejudice 21
public judgment 17
public opinion 15
representative democracy 7
stereotypes 20
values 20
SUGGESTED SOURCES FOR FURTHER READING
Each chapter in this book ends with a list of readings and websites. If you wish to investigate
any of a chapter’s topics further, perusing these lists of suggested sources will be a good step
to begin your exploration.
Bachrach, Peter. The Theory of Democratic Elitism: A Critique. Washington, DC: University Press of
America, 1980.
Pateman, Carole. Participation and Democratic Theory. Cambridge, UK: Cambridge University Press,
1970.
Pateman, Carole. “Participatory Democracy Revisited.” Perspectives on Politics 10 (2012): 7–19.
Walker, Jack L. “A Critique of the Elitist Theory of Democracy.” American Political Science Review 60
(1966): 285–295.
These authors are proponents of participatory democracy. In these works, they present reasons for
supporting this theory and explain why democratic elitism is problematic. In her 2012 article, Pateman
contrasts participatory democracy with deliberative democracy.
Bryan, Frank M. Real Democracy: The New England Town Meeting and How It Works. Chicago:
University of Chicago Press, 2004.
Bryan presents an in-depth and illuminating analysis of the New England town meeting. Unlike
representative democracies, these meetings involve all eligible citizens coming together to debate and
vote on town matters. For this reason, Bryan labels the town meeting “real democracy.”
Cramer, Katherine J. The Politics of Resentment: Rural Consciousness in Wisconsin and the Rise of
Scott Walker. Chicago: University of Chicago Press, 2016.
Cramer’s research approach and conclusions illustrate her claim that “public opinion is not just what
polls measure” (p. 19). The characteristic of public opinion that she focuses on is rural consciousness,
demonstrating how such an identity can fuel a culture of political resentment.
Dahl, Robert A. Who Governs? Democracy and Power in an American City. New Haven, CT: Yale
University Press, 1961.
Schumpeter, Joseph A. Capitalism, Socialism and Democracy. 5th ed. London: Allen and Unwin, 1976.
The theories of democratic elitism and pluralism are discussed in these works.
Gilens, Martin, and Benjamin I. Page. “Testing Theories of American Politics: Elites, Interest Groups,
and Average Citizens.” Perspectives on Politics 12 (2014): 564–581.
This article is an example of scholars empirically testing normative theories of democracy, including
some of the theories described in this chapter.
Held, David. Models of Democracy. 2nd ed. Stanford, CA: Stanford University Press, 1996.
Macpherson, C. B. The Life and Times of Liberal Democracy. Oxford: Oxford University Press, 1977.
Macpherson and Held have organized the numerous variants of democratic theory into a manageable
number of models (four for Macpherson and nine for Held), some of which we have presented in this
chapter. For a description of similarities and differences among theories of democracy or an introduction
to specific theorists, these books are recommended.
Scudder, Mary F. Beyond Empathy and Inclusion: The Challenge of Listening in Democratic
Deliberation. Oxford: Oxford University Press, 2020.
Scudder examines how to make deliberation in large pluralistic societies more democratic. She argues
that listening is what makes deliberation meaningfully democratic. But listening to those with whom we
disagree is hard. She considers how to encourage democratic listening among citizens divided by deep
difference.
American Association for Public Opinion Research, www.aapor.org
AAPOR is the professional organization for public opinion researchers. This website contains
information about the organization, statements regarding the misuse of polls and poll results, material
concerning the ethics of public opinion research, guidelines for journalists who write about polls, and a
helpful (if somewhat dated) list titled “Fifty Books That Have Significantly Shaped Public Opinion
Research.”
http://www.aapor.org/
APPENDIX TO CHAPTER 1: STUDYING PUBLIC OPINION
EMPIRICALLY
IN THIS APPENDIX, we describe a variety of methods for empirically studying public opinion. As you
will see, each method has strengths and weaknesses. Further, some methods are better than others at
answering particular public opinion questions. As we proceed, we refer to studies examining death
penalty attitudes to illustrate how each method works in practice.
Public Opinion Surveys
Today, the most common method for assessing public opinion is via a survey or public opinion poll.
Most of us are familiar with polls or, at the very least, the results of polls. The survey results that we
frequently encounter (in the news media, on the Internet, and so on) are based on the responses
provided by a sample of people to the same list of questions. In scientific surveys, respondents are
selected randomly to represent a specific population (such as students at the University of Kansas,
residents of New Mexico, or citizens of the United States). Survey respondents answer a series of
questions, often by selecting one response from a list of options provided by the survey interviewer. For
example, to gauge public sentiment on the issue of capital punishment, a survey might include the
following question: “are you in favor of the death penalty for a person convicted of murder?” Those
being surveyed would respond by selecting “favor” or “oppose” or, in some cases, “no opinion” or “I don’t
know.” These types of questions, with a limited set of response options, are called closed-ended
questions.
Questions can be worded in a variety of ways, and the choice of which words to include can have
important, sometimes even dramatic, effects. To illustrate question wording effects, let’s examine two
ways the Gallup Organization has asked people about their death penalty attitudes. In October 2014,
Gallup polled a random sample of 1,017 adults living in the United States. Respondents were asked
whether they support the death penalty for convicted murderers (see Table A.1). Almost two-thirds of
respondents supported the death penalty. Just one month earlier, in September 2014, Gallup polled a
random sample of 1,252 U.S. adults but this time asked respondents to indicate which they favor more,
the death penalty or life in prison. Support for the death penalty dropped to 50 percent when
respondents had a choice of punishments for convicted murderers.1 This is a substantial difference, and
very different conclusions would be drawn about public support for the death penalty depending on
which survey question was used.
Table A.1 Question Wording and Response Options Matter
Question wording
Response
options
October 2014: “Are you in favor of the death penalty for a person convicted of murder?”
Favor 63%
Oppose 33%
No
opinion
4%
Response
options
(rotated)
September 2014: “If you could choose between the following two approaches, which do
you think is the better penalty for murder: the death penalty or life imprisonment with
absolutely no possibility of parole?”
Death
penalty
50%
Question wording
Life in
prison
45%
No
opinion
5%
Source: Data from Jeffrey M. Jones, Americans’ Support for Death Penalty Stable, Gallup, Washington, DC,
http://www.gallup.com/poll/178790/americans-support-death-penalty-stable.aspx?
utm_source=position4&utm_medium=related&utm_campaign=tiles
It is also important to pay attention to what response options are presented to respondents as well as
the order in which those options are provided. Take Gallup’s September 2014 question, for example.
When citizens were asked to choose between the two approaches, one-half of the respondents were
read the death penalty option first and the life in prison option second. The other half were read the
choices in the reverse order. The choices are rotated because of concerns about response order
effects.2 Quite simply, citizens’ opinions can be influenced by the order in which responses are
presented to them. In addition, note that no middle or undecided categories were provided to
respondents. As a result, citizens who were ambivalent or indifferent on the topic were unable to
express their views.
If you want to analyze public opinion over time, it is critical to ensure you compare survey data collected
using the same question wording and response options. For example, Gallup reported that 56 percent of
Americans support the death penalty in October 2018.3 To know whether that is an increase or
decrease in support since 2014, you need to know whether Gallup used the “are you in favor” or the “if
you could choose” question. The polling organization used the “are you in favor” version, so therefore 56
percent in favor is a 7 percentage point drop in support since the 2014 survey.
To illuminate another concern about question wording and response options, let’s discuss a question
used by the National Race and Crime Survey to assess opinions toward the death penalty. The wording
is, “Do you strongly oppose, somewhat oppose, somewhat favor, or strongly favor the death penalty for
persons convicted of murder?”4 Notice that response options from both points of view are provided in
the stem of the question, which is what survey researchers call a balanced question. In contrast,
Gallup’s October 2014 question refers only to the pro–death penalty position, which may encourage
respondents to answer in that fashion. As a result, balanced questions are considered superior.
So, when you come across poll results, it is important to know the question wording, the response
options, and the order in which those options were presented. Similarly, if you ever report the results of
an opinion poll, you also need to provide all that information. Otherwise, it is very easy to mislead,
whether intentionally or not, those who are reading your summary of the results.
In a perfect world it is also important to know the order in which survey questions are asked. For
instance, in the October 2018 Gallup survey, respondents were also asked whether they “believe the
death penalty is applied fairly or unfairly in this country today.” Forty-nine percent said it was applied
fairly, 45 percent said unfairly, and 6 percent had no opinion.5 This question was asked after
respondents provided their opinion on the death penalty. But what if this question had been asked
before? Respondents would have been primed to think about the fairness of the death penalty, which
could have influenced their support for the policy. Specifically, the 45 percent of respondents concerned
about the fairness of the death penalty might be less likely to say they favored the policy when that
http://www.gallup.com/poll/178790/americans-support-death-penalty-stable.aspx?utm_source=position4&utm_medium=related&utm_campaign=tiles
concern was fresh in their minds, and vice versa. Therefore, when you analyze a public opinion survey,
it is best to examine not only the question you are interested in but also the context in which that
question is situated. Unfortunately, researchers and especially journalists do not always provide the text
of the entire survey, so it is often difficult to evaluate whether question order effects are influencing the
results.
Public opinion polls have a number of advantages. Randomly sampling people from a specified
population allows us to draw conclusions about the opinions of the entire population. Why is that the
case? Because a random sample is one in which chance alone determines which elements of the
population make it into the sample. For example, let’s say you want to draw a sample of twenty-five
students from a class (or population) of one hundred students, and you want the opinions of the twenty-
five students to reflect the opinions of all one hundred students. How would you draw that sample? You
could have students write their names on slips of paper, collect the one hundred names in a hat, give it
a good shake to make sure the names are all mixed up, and then draw out twenty-five names.
Consequently, it would be chance alone that would determine which twenty-five students ended up in
your sample. When respondents are selected in this manner, and not on the basis of their specific
characteristics (such as race or political views), we can generalize the results from the sample to the
larger population from which the sample was drawn. The ability to draw such conclusions is known as
external validity. For polls that include only respondents who opt to participate, the results are
applicable only to those people who answered the survey questions. Because such poll results are
based on what is called a convenience sample, not a random sample, they cannot provide information
about a larger population.
Another advantage of surveys is that answering a closed-ended question is not very time-consuming, so
each respondent can answer many questions without being overly burdened. Also, many individuals can
be asked the same questions, again because the time commitment per person is not great. Providing
survey respondents the same questions with the same response options facilitates the tallying of results
(such as 56 percent of Americans support the death penalty) and also allows for a comparison of public
opinion over time, as we mentioned earlier. For instance, as shown in Figure A.1, public support for the
death penalty has fluctuated since 1991. In the early 1990s, three-quarters (or more) of the public
favored the death penalty, but by 2000 only two-thirds did. After 2001, perhaps because of the 9/11
terrorist attacks, support for the death penalty increased somewhat to 70 percent, but since the mid-
2000s, support has decreased over time with 56 percent of the public in favor of the policy in 2018.6
Surveys also have many uses. News media organizations use polls to measure the public’s political and
social opinions, and candidates conduct polls to determine which voters support them and why. Public
opinion scholars find surveys useful for assessing the content of the public’s opinions as well as
describing how people’s opinions differ. In particular, it is often interesting to examine whether different
groups have different attitudes on important issues of the day. A Pew Research Center national survey
asks respondents this question: “Do you strongly favor, favor, oppose, or strongly oppose the death
penalty for persons convicted of murder?”7 The Pew Research Center records respondents’ gender and
party identification in addition to their death penalty attitudes. The two favor options and the two oppose
options are combined to simplify the presentation of results. The Pew survey shows that men and
women hold different attitudes on the death penalty. In 2018, for example, 61 percent of men favored
the death penalty, whereas only 46 percent of women supported capital punishment for convicted
murderers. The gap in attitudes is even greater when you examine differences by party identification.
Seventy-seven percent of Republicans support the death penalty compared with 52 percent of
Independents and 35 percent of Democrats.8
Description
Figure A.1 Public Opinion toward the Death Penalty, 1991–2018
Source: Data from Death Penalty, Gallup, Washington, DC, https://news.gallup.com/poll/1606/death-
penalty.aspx
Note: Here is the question wording: “Are you in favor of the death penalty for a person convicted of murder?”
A specific type of survey, called a panel study or longitudinal survey, allows scholars to determine
whether people’s opinions have stayed the same over time. In a panel study, the same people are
asked their opinions on the same issues more than once. A study conducted by Robert Bohm and
Brenda Vogel illustrates the use of this type of survey to track people’s death penalty attitudes across
more than a decade.9 In the late 1980s, Bohm and Vogel surveyed college students at the beginning of
the semester during which they were taking a class on the death penalty. They resurveyed the students
at the end of the semester and then again a couple of years later. In 1999, Bohm and Vogel surveyed
these (now former) students for the fourth time. The researchers were interested in whether students’
attitudes changed after becoming more informed about the death penalty during the class and whether
that attitude change was lasting. They found that students were less supportive of the death penalty
immediately after taking the class, but over time the students reverted to their initial levels of support for
the policy. They concluded that information about the death penalty can influence citizens in the short
run but that views on the policy are largely driven by personality traits and values, which are more
influential than knowledge in the long run.
Panel studies are ideal for tracking changes in opinion across time, but it is important to note that
attrition is a potential weakness of such studies. Attrition refers to the drop-off in the number of
respondents over time. In the Bohm and Vogel study, for example, 120 college students were initially
administered the survey, but only 69 were still participating in the study by the fourth wave. If the
students who stopped participating were systematically different from the students who continued to
participate, we would need to be cautious about drawing conclusions from the study. Further, this
particular panel study did not use a random sample, which limits the external validity of the results;
however, many panel studies do rely on random samples, so their results can be generalized to the
broader population.
Public opinion data are collected using a particular survey mode, such as phone, Internet, mail, or face-
to-face. Each of these modes has strengths and weaknesses. Researchers should consider the topic,
population of interest, and survey context to determine which mode to use.10 For example, it is typically
better to ask sensitive questions over the Internet or via mail rather than face-to-face. Complex, long
questions are generally not appropriate for phone surveys. The Internet is useful to assess people’s
responses to video clips. In some cases the best approach is to combine modes, which is called a
mixed-mode approach. For example, a survey researcher might contact respondents via mail with a
postcard to inform them they have been chosen to participate in a study to be conducted by phone. The
2018 Gallup survey of death penalty attitudes mentioned earlier was conducted over the phone with 70
percent of the respondents contacted on cell phones and 30 percent on landlines.11 Obviously it is
https://news.gallup.com/poll/1606/death-penalty.aspx
critical to include cell phone numbers in the sample given their widespread use in society today. In the
United States 96 percent of adults own a cell phone.12 Furthermore, the number of homes without a
landline has increased dramatically over the past decade. Fifty-seven percent of homes in the United
States are cell phone only, and three-quarters of young adults rely completely on a cell phone.13
Do keep in mind that regardless of the particular survey mode, a random sample is necessary to
generalize the results of the sample to the population. Gallup, for example, uses a standard technique,
random digit dial, in which chance alone determines which phone numbers are called in their surveys.
This ensures that all phone numbers have the potential to be called rather than just those numbers on a
phone list. As a result, Gallup’s surveys are generalizable to the population.
In recent years the number of surveys conducted over the Internet has increased markedly due to their
ease and low cost. There is no random digit dial equivalent for Internet surveys, however, so
researchers take two different approaches to improve the representativeness of their sample. First, a
mixed-mode approach can be used to initially contact respondents by phone and then recruit them to
participate in an online survey. The researchers provide a computer and Internet access to those who
do not have both to ensure the sample is not biased. This approach can ensure a random sample, but it
takes time and money that can defeat the purpose of using an Internet survey.14
The second approach is for researchers to rely on opt-in Internet surveys. People who volunteer to
participate in a survey are different from a random sample because it is not chance alone that
determines their selection for participation in the survey. Therefore, when analyzing the survey
responses, researchers need to weight the respondents so that the sample mirrors the demographic
and political characteristics of a representative sample. The problem is there is no ironclad way to
determine exactly how well, or how poorly, the opt-in sample mirrors a representative one. Because
Internet surveys are here to stay, many public opinion scholars are working hard to better understand
the potential and limitations of opt-in surveys.15
Experiments
Another common method used by public opinion researchers is experimentation. Although there are
many types of public opinion experiments, in the most common form, the researcher manipulates a
feature of the study and then assesses individual or group responses. Imagine you wanted to know how
individuals respond to different types of news stories on the death penalty. You could assess this
experimentally by providing one type of news story to one group of participants in your study and
another type of news story to another group. After reading the news stories, these subjects, as
participants in an experiment are called, would be asked whether they support the death penalty.
Many news media studies use experimental designs just like that to see whether citizens’ opinions are
influenced by different news content. For example, Frank Dardis et al. created newspaper stories to
frame the death penalty in different ways.16 One story constructed the death penalty as an affront to
moral values (the morality frame), whereas another story emphasized that the policy was fundamentally
flawed because innocent people might be executed (the innocence frame).17 Some subjects read the
story with the morality frame, and others read the one with the innocence frame. Subjects then
completed a questionnaire that asked them to list the important factors they considered when
determining their opinion on the death penalty. Dardis et al. found that subjects exposed to the
innocence frame were more likely to mention innocence-related considerations as important factors in
determining their attitudes toward the death penalty than subjects presented with the morality frame.
Thus, the news frames shaped the ingredients of the subjects’ death penalty attitudes.
The two key features of experiments that distinguish them from other methods and that allow for
powerful causal conclusions to be drawn are manipulation and random assignment.18 Manipulation
involves the researcher varying access to information, events, or whatever is the focus of the research
among experimental participants. In the example we have been discussing, the researchers
manipulated exposure to news frames. Random assignment refers to the process by which people are
assigned to experimental groups. With random assignment, it is chance alone that determines which
subjects get in which condition. For instance, in the Dardis et al. experiment, subjects were randomly
assigned to read a story framed in terms of either innocence or morality. Individuals are randomly
assigned to groups, perhaps by flipping a coin to establish the person’s assignment, in the expectation
that individual characteristics that might be related to the study’s goals are equally likely to appear in all
groups. Because men are more supportive of capital punishment than women, for example, it is
important that not all men be assigned to the same group in an experiment designed to assess the
impact of news stories on how citizens think about capital punishment. Random assignment ensures
that chance, rather than a person’s characteristics, determines experimental group assignment.
With successful random assignment, a researcher can be very certain that any differences in opinions or
behaviors found across experimental groups are due to their exposure to the original stimulus (that is,
due to the experimenter’s manipulated feature). Experiments thus allow researchers to conclude that
one factor causes another—a feature of research designs called internal validity. The ability to draw
such causal conclusions is the primary advantage of experiments over other research methods. For
example, you could conduct a survey and ask people if they have read news articles framing the death
penalty in terms of innocence and whether they support the death penalty. If those who have read these
stories are less likely to favor capital punishment, it would be tempting to conclude that the innocence
frame influenced individuals’ opinions. But you could not rule out the possibility that those who opposed
capital punishment before exposure to the news stories were more likely to search out and read such
stories. So individuals’ political opinions might have influenced their news habits rather than the other
way around. If, however, you expose some people to the innocence news story and others to a story
framed in a different way, and you still find that those exposed to the innocence frame are less
supportive of the death penalty, you can be much more certain that the news frame influenced their
opinions.
Although experiments possess internal validity, they often have less external validity. That is, by using
convenience samples (such as college students enrolled in introductory mass communications courses,
as Dardis et al. did in their study) rather than random samples, experimenters cannot claim their sample
is representative of the broader population. One way to address this weakness is to include an
experimental design within a nationally representative survey. This method, called a survey-based
experiment, entails randomly assigning survey respondents to experimental conditions. This approach
“combine[s] the causal power of the randomized experiment with the representativeness of the general
population survey.”19
Mark Peffley and Jon Hurwitz used this approach in their national survey of race and death penalty
attitudes.20 They embedded an experiment in their survey by randomly assigning respondents to
receive one of three versions of a question about the death penalty (see Table A.2). In the baseline
condition, respondents received a death penalty question with no additional information. In the other two
conditions, respondents received information either about racial disparities or about innocence. Peffley
and Hurwitz’s results are fascinating. First, whites were substantially more supportive of the death
penalty than Black respondents across all three conditions. Second, whites and Blacks did not respond
in the same way to the different arguments. Support for the death penalty fell significantly among Blacks
when they were exposed to either the racial or the innocence argument. In contrast, whites were not
moved by the innocence argument, and they actually became more favorable toward the policy when
presented with the racial argument. Because respondents were randomly assigned to the conditions, we
can conclude with great confidence that the different arguments influence opinion on the death penalty.
Moreover, because the respondents were selected randomly from the U.S. population, the results of this
study apply to the American public in general.21 In other words, this study has both internal and external
validity.
An important aspect of scientific research, especially experimental research, is replication. Replication
occurs when scholars repeat a study to determine whether their results are consistent with prior
research findings. One study replicated the Peffley and Hurwitz research and did not find that whites
became more favorable toward the death penalty when exposed to the racial argument.22 Another
replication, by Peffley and Hurwitz themselves, found that only certain whites and Latinos became more
supportive of the death penalty when presented with the racial argument.23 Specifically, whites and
Latinos who blamed Black aggressiveness for crime became more favorable toward the death penalty
when presented with the racial argument. Where do these mixed results leave us? It is possible that
Peffley and Hurwitz’s original findings were timebound, meaning they held at that time but cannot be
generalized to a different time period. Or possibly the original findings were simply incorrect. Either way
it demonstrates the critical importance of replication to scientific research. Repeated studies are
essential to ensure that research findings are valid across different settings, people, and time periods.
Table A.2 Support for the Death Penalty in a Survey-Based Experiment
Baseline withno argument Racial argument Innocence argument
“Do you strongly
oppose,
somewhat
oppose,
somewhat favor,
or strongly favor
the death
penalty for
persons
convicted of
murder?”
“Some people say that the death
penalty is unfair because most of
the people who are executed are
African Americans. Do you
strongly oppose, somewhat
oppose, somewhat favor, or
strongly favor the death penalty
for persons convicted of
murder?”
“Some people say that the
death penalty is unfair because
too many innocent people are
being executed. Do you
strongly oppose, somewhat
oppose, somewhat favor, or
strongly favor the death penalty
for persons convicted of
murder?”
Baseline withno argument Racial argument Innocence argument
White
respondents
65% 77% 64%
Black
respondents
50 38 34
Source: Adapted from Table 5.1 of Mark Peffley and Jon Hurwitz, Justice in America: The Separate Realities of Blacks and Whites
(Cambridge, UK: Cambridge University Press, 2010), 158–159.
Note: Figures are the percentage of each group that somewhat or strongly favors the death penalty.
Interviews
Asking people about their political views is also accomplished by in-depth interviewing. Unlike surveys
in which hundreds (or thousands) of people are asked a series of closed-ended questions, interviewers
ask their respondents much broader, often open-ended questions. That is, interviewers typically do not
provide their respondents with a list of predetermined response options; rather, they allow the
interviewees to answer a question however they want. An interviewer interested in public opinion toward
the death penalty might ask the following question: “What do you think about the death penalty?” This
question encourages respondents to not just assess their overall opinion on the issue but consider the
roots of their opinion and perhaps even grapple with any contradictory thoughts they might have about
the policy. Topics such as racial disparities in the application of the death penalty, the deterrent effect of
the death penalty, or popular culture references to the death penalty might emerge in response to this
question. Note the question does not provide response options, thus allowing the respondent to answer
in multiple ways. The question prompts respondents to explain why they hold their opinions, and if
respondents do not volunteer such information, interviewers can follow up and ask them directly to
explain their perspectives. Such why questions, because they are open-ended, do not appear frequently
on opinion surveys, yet they can provide very useful information about public opinion.
Allowing respondents to decide what is most appropriate when answering questions results in
responses that are more likely to reflect their actual thinking (no matter how organized or how messy) on
the topic. By forcing respondents to select a preconceived option, surveys might not measure real
opinions on an issue. To take an obvious example, survey respondents confronted with the “favor” or
“oppose” option to a death penalty question will typically select one of these options even if their real
attitude is “I support the death penalty when I am certain that the person convicted of murder did,
without a doubt, commit the murder, but often one cannot be certain the person actually did commit the
murder and there are now many examples coming to light when incorrect decisions were made by
juries.” An in-depth interview is likely to capture the nuances of this person’s view, whereas a public
opinion survey with closed-ended questions simply cannot.
In-depth interviews can be especially useful when researchers are interested in understanding the views
of a particular group of people. For example, Sandra Jones conducted in-depth interviews with forty-
nine people active in the anti–death penalty movement to understand what mobilized them to get
involved.24 Jones found that many activists were motivated by moral outrage but that their outrage was
complex and nuanced. To illustrate, an African American male leader of the movement had this to say
about the death penalty:
Not only is it dehumanizing, but everything else that wraps around it is immoral. It’s immoral to
have another human being strapped down for the purpose of killing them. It is immoral to put
the warden in such a conflict. The one thing I’ve learned from doing this work is when I came to
it I had such a clear sense of who was good and who was evil. All that got blurred very quickly.
You can’t hate a guard who cries over an execution. You can’t hate a warden who is shaking
during an execution.25
It would be difficult, if not impossible, to capture the detailed richness of this person’s views about the
death penalty using a survey, but in-depth interviewing allowed the researcher to assess the fullness of
this activist’s perspective.
Because open-ended questions typically take longer to answer, the number of individuals participating in
an interview is usually much smaller than the number who respond to a survey. With a smaller number
of participants in a study, who have not been randomly selected to participate, it is inappropriate to draw
conclusions that can apply to a larger population. Thus, studies using interview respondents are often
criticized for not being representative of a larger population, a weakness that does not apply to surveys
of randomly selected individuals.
Focus Groups
Focus groups resemble interviews in a number of ways. Both are used by researchers to examine how
people think about political issues, and both use open-ended questions. The primary differences are that
focus group research is conducted on multiple people at once and consists of a group discussion that is
moderated and guided by a trained individual. Focus group researchers are often interested in learning
how individuals construct political issues in their mind, how people communicate about a particular
issue, and how an individual’s discussion of a topic responds to communication from others in a group.
In this way, focus groups are “a way to observe interaction among people that is important in
understanding political behavior that is not possible to observe using more traditional empirical
methodology.”26 To examine public opinion on capital punishment, for example, a focus group could be
used to assess how people discuss this issue, including which features of it are especially compelling or
relevant. Focus group participants could also be asked to read news articles or view movies about
capital punishment and then discuss their reactions to determine how a group constructs meaning from
such stories.
With the goal of understanding the complexity of citizens’ death penalty opinions, Diana Falco and Tina
Freiburger conducted six focus groups with twenty participants from Indiana County, Pennsylvania.27
The researchers asked the participants to brainstorm about their positive and negative beliefs about the
death penalty and to indicate their general opinion on the policy. Participants were also asked to read
various crime scenarios and evaluate whether they would support the death penalty in each situation.
Falco and Freiburger found that many citizens held both positive and negative views of the death
penalty and that almost all took characteristics of the offender or the victim into account as they
responded to the crime scenarios. The researchers concluded that the twenty citizens in these focus
groups have views on the death penalty that are much more complicated than suggested by “favor” or
“oppose” responses to a survey question. Because the focus group participants do not constitute a
random sample, the results cannot be generalized to the public as a whole. Nevertheless, these results
are still very important because they help scholars think more carefully about how to design survey
questions to more adequately measure the complexity of citizens’ subtle, and sometimes tangled, views
on the death penalty.
Content Analysis
The final method we profile here is content analysis. As its name indicates, content analysis is a
technique used to analyze the content of communication. More specifically, it has been defined as “a
research technique for the objective, systematic and quantitative description of the manifest content of
communication.”28 Content analysis can be applied to any type of communication, such as a news
media story, a speech by a politician, a popular television show, a tweet, or a novel. The primary object
of content analysis is to systematically summarize the content of the selected source or item. This is
done by selecting specific criteria of the communication to analyze and then carefully coding a selection
(such as stories or speeches) along these criteria. For example, a speech by Donald Trump could be
analyzed for the number of times the word “immigrant” is used; the number of times a topic, such as his
poll numbers, is mentioned; or whether he refers to himself in the third person.
In terms of public opinion research, many topics can be examined using content analysis. If a
researcher wishes to know how the news media present public opinion on an issue, such as capital
punishment, the content of news stories can be analyzed. Is public opinion represented as opinion
survey results or as quotations from individual people? Or are elected officials asked what they think the
public thinks about this issue?
Content analysis is also used in studies that seek to determine whether news media coverage is related
to public opinion. Recall the Dardis et al. experiment we discussed earlier. In that study, subjects were
exposed to news stories about the death penalty framed in terms of either morality or innocence. Dardis
et al. did not simply pull those media frames out of thin air; instead, they content-analyzed abstracts of
capital punishment news articles in the New York Times Index between 1960 and 2003 to identify
frames. By systematically analyzing what types of arguments were used in these abstracts, the
researchers were able to examine common frames used in the New York Times coverage. The morality
frame, for instance, included arguments about retribution, such as the “eye for an eye” rationale for the
death penalty. The innocence frame, in contrast, included arguments about the possibility that a person
on death row might be innocent due to a tainted or racist criminal justice system.29 Dardis et al. found
that the innocence frame received little attention prior to the 1980s but that it became a prominent frame
in the 2000s. The morality frame received significant attention in the 1970s but has been less prevalent
since then, although it continues to receive meaningful attention in the New York Times.
Historically, content analysis has been a labor-intensive method requiring humans to read and hand-
code textual materials. In more recent years researchers have developed computer-assisted techniques
that allow for the quick analysis of huge amounts of textual data. Researchers specify coding rules that
enable computers to analyze the massive amounts of information produced in society today. These “big
data,” such as data from tweets, Facebook posts, website content, and news articles, provide a trove of
information that can offer insights about public opinion.
Conclusion
These five methods—surveys, experiments, interviews, focus groups, and content analysis—are the
most common approaches used to assess public opinion. Surveys are by far the most frequently used
approach, whereas focus groups and content analysis are the least common. Each method has
advantages and disadvantages, and some methods are more appropriate than others for addressing
particular types of public opinion questions, as the chapters in this book further illustrate.
Last, most of these research methods require human participation. Conducting research on people
involves a host of ethical considerations. Chief among these concerns are that participants should
voluntarily agree to participate, they should offer their informed consent before the study begins, and
they should not suffer undue physical or psychological harm while participating in the study or afterward.
For a detailed discussion of these and other ethical matters involved when using people as research
participants, refer to The Belmont Report (listed in the Suggested Sources for Further Reading).
Key Concepts
attrition 35
balanced question 32
closed-ended questions 30
content analysis 42
convenience sample 33
experiments 36
external validity 33
focus groups 41
in-depth interviewing 39
internal validity 37
manipulation 37
mixed-mode approach 35
open-ended questions 40
opt-in Internet surveys 36
panel study or longitudinal survey 34
population 30
public opinion poll 30
question order effects 32
question wording effects 30
random assignment 37
random digit dial 35
random sample 33
replication 38
response order effects 31
sample 30
subjects 36
survey 30
survey-based experiment 38
survey mode 35
timebound 38
Suggested Sources for Further Reading
Aronson, Elliot, Phoebe C. Ellsworth, James Merrill Carlsmith, and Marti Hope Gonzales. Methods of
Research in Social Psychology. 2nd ed. Boston: McGraw-Hill, 1989.
Gilens, Martin. “An Anatomy of Survey-Based Experiments.” In Navigating Public Opinion: Polls, Policy,
and the Future of American Democracy, ed. Jeff Manza, Fay Lomax Cook, and Benjamin I. Page.
Oxford: Oxford University Press, 2002.
Krippendorff, Klaus. Content Analysis: An Introduction to Its Methodology. 2nd ed. Thousand Oaks, CA:
SAGE, 2004.
Krueger, Richard A., and Mary Anne Casey. Focus Groups: A Practical Guide for Applied Research. 5th
ed. Thousand Oaks, CA: SAGE, 2015.
Rubin, Herbert J., and Irene S. Rubin. Qualitative Interviewing: The Art of Hearing Data. 3rd ed.
Thousand Oaks, CA: SAGE, 2011.
Each of these sources provides a detailed overview of one specific research method: experiments,
survey-based experiments, content analysis, focus groups, and interviewing.
Asher, Herbert. Polling and the Public: What Every Citizen Should Know. 9th ed. Washington, DC: CQ
Press, 2016.
Dillman, Don A., Jolene D. Smyth, and Leah Melani Christian. Internet, Mail, and Mixed-Mode Surveys:
The Tailored Design Method. 4th ed. Hoboken, NJ: Wiley, 2014.
Asher’s book is an informative and readable introduction to all aspects of survey research. If you are
planning to conduct a survey yourself, the Dillman, Smyth, and Christian book is a must-read. It provides
practical advice and detailed examples of best practices.
“The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of
Research,” https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/index.html
“Standards and Ethics,” American Association for Public Opinion Research,
https://www.aapor.org/Standards-Ethics.aspx
The Belmont Report was produced by the National Commission for the Protection of Human Subjects of
Biomedical and Behavioral Research in 1979. This influential report established basic ethical principles
that should be followed by scholars conducting research on human participants. The American
Association for Public Opinion Research has developed standards, identified best practices, and
articulated a code of ethics for all public opinion researchers.
https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/index.html
https://www.aapor.org/Standards-Ethics.aspx
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists years and ranges from 1952 to 2016 in increments of 4. The vertical axis lists
percentage and ranges from 0 to 60 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year PercentageDemocratic
Percentage
Republican
Percentage Independent
with leaners
Percentage who are
purely independent
1952 49 29 23 6
1956 45 30 24 9
1960 46 30 23 10
1964 51 25 23 8
1968 46 24 30 11
1972 42 24 35 13
1976 41 24 36 15
1980 43 23 35 13
1984 38 28 35 11
1988 37 28 37 11
1992 37 26 38 12
1996 39 27 37 10
2000 36 23 40 12
2004 35 30 39 11
Year PercentageDemocratic
Percentage
Republican
Percentage Independent
with leaners
Percentage who are
purely independent
2008 36 28 39 12
2012 38 30 34 14
2016 38 31 34 15
Back to Figure
The horizontal axis lists years and ranges from 1991 to 2018. The vertical axis lists percentage and
ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated as follows:
Year Percentage in favor Percentage opposing Percentage with no opinion
1991 78 19 7
1995 79 13 9
2000 68 29 7
2001 69 27 8
2002 70 27 7
2004 65 31 7
2006 65 30 8
2008 65 31 7
2010 65 31 7
2012 65 33 7
2014 65 34 1
Year Percentage in favor Percentage opposing Percentage with no opinion
2015 65 37 3
2016 63 37 5
2017 60 43 5
2018 61 43 7
PART II ARE CITIZENS
PLIABLE?
D O U.S. CITIZENS hold stable political attitudes? Or are many
people pliable, with their attitudes frequently changing? These
questions highlight a core difference among the democratic theories
we profile in this book—beliefs regarding the political capabilities of
the public. Which view suggests a more capable public? Is it best for
people to hold opinions that stay the same over years or for people
to be open to new perspectives? Put another way, do citizens need
to be open to new ideas, fresh leaders, and policy innovations? Is it
possible to guarantee that without having citizens susceptible to a
politician’s every whim?
These are important questions for any student of the democratic
public to consider. Before reaching any firm answers, however, it is
important to know where people’s opinions originate, when and
under what conditions they are likely to change, for whom opinions
are likely to be stable versus changeable, and what role external
communication sources play in influencing opinions. These topics
are addressed in this part of the book in specific chapters on the
topics of political socialization, the mass media, and attitude stability
and attitude change.
CHAPTER 2 POLITICAL SOCIALIZATION
IN 1987, an eight-year-old girl named Betsy wrote a letter to her mayor soliciting some advice.
Journalists at National Public Radio learned of this letter, leading one of them (Noah Adams) to interview
her.
Noah: You wrote a letter to the mayor of New York, Mayor Koch.
Betsy: Right.
Noah: Tell me about that please.
Betsy: Well I wrote to him because my parents are getting divorced and I really don’t know who to turn
to. I just told him that my parents are getting divorced and my dad is with somebody else and I was just
getting used to something and now this and it’s really kind of hard on me and I’d like an opinion.
Noah: Why did you write to Mayor Koch?
Betsy: ’Cause he’s somebody who I thought he’s very good to us I guess because he’s the mayor and
he knows a lot of things and I thought he would know about this too.
Noah: Did you get an answer back?
Betsy: Yes.
Noah: What did he say?
Betsy: He … um … it’s very short. “Thank you for the letter. I was saddened to learn of the difficult times
you are experiencing now. It is important for you to share your feelings and thoughts with someone
during this time. I wish there was an easy solution to these problems but there is not. Please remember
that you are loved and that people care about you. All the best. Sincerely, Edward Koch.”
Noah: That’s nice. Was that reassuring to you in a way?
Betsy: No.1
Young Betsy had already developed views toward political leaders. Befitting her age, her image of
Mayor Koch was largely favorable, although subject to revision based on her encounter with him. In
other words, her political opinions were forming. The manner by which we all learn about politics and
develop political opinions is called political socialization. Put another way, “political socialization is the
process by which people acquire relatively enduring orientations toward politics in general and toward
their own particular political systems.”2 As we elaborate on in the next section, this process begins in
childhood. Also, as we demonstrate there, Betsy’s impressions of Mayor Koch were akin to those of
children from the late 1950s but diverge somewhat from childhood assessments of political leaders
today.
There are many sources of people’s political opinions. Important socialization agents include schools,
peers, and the news media. Primary among these, however, is the family. In fact, among early
socialization researchers, parents were thought to play the most influential socializing role.3 In the
pages that follow, we review the research supporting this conclusion, as well as discuss how the broader
political context influences developing political opinions. Recently, scholars have focused on an
alternative way that political attitudes are acquired—genetics. We review this hot, and somewhat
controversial, area of research near the end of the chapter.
Another way to think of political socialization is as the transmission of key political values and norms
from one generation to the next. This view of socialization focuses on how societies “inculcate
appropriate norms and practices in citizens, residents or members.”4 David Easton and Jack Dennis
were proponents of this approach, linking socialization to the maintenance of a democratic political
system.5 In particular, Easton and Dennis described the main goal of early socialization as fostering
confidence and trust in, as well as positive affect toward, the political system. They further argued that
the widespread holding of these attitudes is important for the persistence of a nation’s government.
Failure to transmit these norms to new generations of children could threaten a nation’s stability.
Consistent with Easton and Dennis’s view, successful socialization would result in citizens who support
the nation’s system of government and who respect political authority. Such outcomes would please
democratic elitists. First, socializing citizens in such a way could lead them to defer to political leaders
and the leaders’ expertise. This would preserve the dominance of elite decision making with lesser
involvement from the citizens, as democratic elitists prefer. Second, this type of socialization
emphasizes system support over individual development, a goal that democratic elitists support but one
that other democratic theorists, most especially participatory democrats, find worrisome.
In contrast, pluralists hope that socialization develops strong political identities and a clear sense of how
individuals’ interests are best represented in the political system. Especially with a clear sense of their
own similarities to the political parties, citizens can more easily pursue their interests and hold elected
officials accountable for representing them.6 Thus, pluralists would favor a socialization process that
results in strong partisan identification. To what degree does socialization accomplish the goals of these
democratic theorists? We return to this question at the end of the chapter.
CHILDHOOD SOCIALIZATION
Childhood socialization typically begins during the elementary school years, when children learn about
the political world and develop political orientations. From the fourth grade, “children move from near—
but not complete—ignorance of adult politics to awareness of most of the conspicuous features of the
adult political world” by the eighth grade.7 One of the earliest political attitudes formed is a highly
positive evaluation of the nation. Children believe that the United States is better than other nations and
at an early age develop a strong emotional attachment to the nation.8
Benevolent Leader Images
Another notable political orientation of elementary school children is their idealization of leaders,
especially the president. In one of the classic studies of childhood socialization, Fred Greenstein asked
fourth through eighth graders in New Haven, Connecticut, to rate specific political executives in 1958.9
Substantial majorities of children who knew these leaders rated them as “very good,” whereas barely
any children (less than 1 percent) rated the leaders as “bad.” For example, 71 percent of the children
evaluated the president’s job performance as very good, with a further 21 percent feeling that the
president was doing a “fairly good” job. These evaluations were higher than adult assessments of the
president. During the time of Greenstein’s study, 58 percent of the adult public approved of Dwight
Eisenhower’s performance. Similar positive assessments emerged in a study conducted by Robert Hess
and David Easton of second through eighth graders living in a Chicago suburb.10 These children were
asked to compare the president to “most men” on a number of characteristics. Large majorities of
children felt that the president is more honest, is more knowledgeable, and works harder. When asked
to evaluate the president as a person, 97 percent of students said the president is “the best person in
the world” or a “good person.”
The words children use to describe political leaders and their duties are quite interesting and further
demonstrate the positive attitudes children hold.11 Greenstein asked the children in his study, “What
kinds of things do you think the President does?”12 Some of their responses appear in the first column
of Table 2.1. These children generally described the leaders doing good deeds and providing for
people’s basic needs. Further, this benevolent leader imagery exists for most children in the absence
of factual information about the leaders. As the examples in the table demonstrate, some children do not
describe the leaders’ duties accurately, for example, assuming that the president gives people freedom.
Yet this does not prevent them from possessing positive attitudes about the leaders.
Certainly 1958 was a long time ago. Do we know if children’s views of the president are the same
today? Two replications of Greenstein’s work have been undertaken in the years since 1958. The first
was conducted in 2000 by Amy Carter and Ryan Teten among fourth- through eighth-grade children in
Nashville.13 The second, undertaken by Zoe Oxley, Mirya Holman, Jill Greenlee, Angela Bos, and
Celeste Lay, surveyed children in grades one through six at four different locations in the United States
(greater Boston, upstate New York, northeast Ohio, and New Orleans).14 Children in the latter study
were interviewed in 2017–2018 and were asked items from Greenstein’s as well as Hess and Easton’s
earlier work. The results from the more recent studies are similar to the earlier results in one important
respect: children continue to hold idealized and benevolent images of the office of the presidency. They
see the president as one of the most important adult roles, and the language children use when
describing what the president does is mostly favorable. However, as demonstrated in the right-hand
column of Table 2.1, negative images about the president’s activity have crept into children’s minds.
Table 2.1 Children’s Descriptions of the President’s Duties
“What kinds of things do you think the president does?”
1958 2017–2018
“Gives us freedom” (8th
grader)
“If someone’s being bad he tells them to stop” (1st grader)
“What kinds of things do you think the president does?”
1958 2017–2018
“[Does] good work” (6th
grader)
“Tries to protect our country, make laws and helps the country be
stable” (6th grader)
“Has the right to stop bad
things before they start” (5th
grader)
“Helps people in the country with difficult problems in their life” (3rd
grader)
“Is doing a very good job of
making people be safe” (4th
grader)
“Lies, [is] destroying our country” (4th grader)
“Deals with foreign countries
and takes care of the U.S.”
(8th grader)
“Goes to Florida, plays golf, talks with other political leaders, tries to
help our country, insults immigrants and people from other
countries” (6th grader)
Sources: Data for 1958 from Fred I. Greenstein, “The Benevolent Leader: Children’s Images of Political Authority,” American Political
Science Review 54 (1960): 939. Data for 2017–2018 from Angela L. Bos, Jill S. Greenlee, Mirya R. Holman, J. Celeste Lay, and Zoe M.
Oxley, Gender and Early Socialization Study.
In contrast, compared to earlier decades, children of today are much more likely to evaluate the
president himself negatively. Recall that nearly all of the children in Hess and Easton’s study considered
the president to be either a good person or the best person. By 2017–2018, a slight majority (51
percent) say the president is “not a good person,” whereas only 3 percent consider him to be the “best
person in the world.”15 When it comes to assessing job performance, children’s evaluations have
become much less favorable (see Figure 2.1). Not only did the percentage of children assessing the
president’s performance as “very good” drop from the highest in 1958 to the lowest in 2000 and 2017–
2018, but “bad” job responses changed from least to most common overtime. Notable also is the
similarity in responses for 2000 and 2017–2018.
Description
Figure 2.1 Children’s Evaluations of the President’s Job Performance
Sources: Data for 1958 from Fred I. Greenstein, “The Benevolent Leader: Children’s Images of Political
Authority,” American Political Science Review 54 (1960), 938. Data for 2000 from Amy Carter and Ryan L.
Teten, “Assessing Changing Views of the President: Revisiting Greenstein’s Children and Politics,”
Presidential Studies Quarterly 32 (2002), 457. Data for 2017–2018 from Zoe M. Oxley, Mirya R. Holman, Jill
S. Greenlee, Angela L. Bos, and J. Celeste Lay, “Children’s Views of the American Presidency,” Public
Opinion Quarterly 84 (2020): 141–157, DOI: 10.1093/poq/nfaa007.
Children come to have political attitudes from a number of different sources. In terms of their idealized
images of leaders, children transfer their generally positive feelings toward authority figures they
personally know (such as parents) to political leaders.16 That is, although children might not know
exactly what the president does, they understand that the president is a person of authority and
deserves respect. In addition, parents serve as agents of socialization by sharing information and
assessments of leaders with children. Although these adults may hold negative attitudes toward specific
leaders, they probably temper or sugarcoat their feelings when discussing politics with their children,
thus explaining why children’s attitudes toward leaders are generally more positive than adults’
attitudes.17 At least that was one main explanation for the political views held by children in the 1950s.
To account for the less positive assessments held by children in more recent decades, scholars have
pointed to increases in partisan animus among adults as well as in the negativity of news media
coverage of politicians.18 In other words, when the information environment surrounding children
contains more negative messages about political leaders, children’s views turn more negative.
Another important agent of childhood socialization is the school. Elementary school rituals, such as
reciting the pledge of allegiance and singing patriotic songs, foster patriotism and loyalty to the nation
among children. In school, children also learn to follow rules and obey authority figures, behaviors that
they pursue in nonschool settings as well.19 Elementary school curricula and teachers generally do not
directly inculcate children to hold specific political attitudes, such as support for a specific public policy.20
By high school, civics curricula have been shown to influence students’ levels of political knowledge and
trust in government, but curricular effects on political attitudes of elementary school children are
uncommon.21
Features of the political context, such as current events, also influence children’s attitudes. One study
assessed the opinions of Detroit-area children in grades four, six, and eight in 1966 and again in 1968.22
In 1968 children were less likely than in 1966 to believe the president is responsive to the people or that
the government is helpful to their families. Why? Children became more critical in part because of the
events that transpired during these two years, specifically riots in Detroit, the assassination of Martin
Luther King Jr., and the escalating war in Vietnam.
The Watergate scandal also had immediate and lingering effects on children’s images of the president.
In 1972, a burglary occurred in the Democratic National Committee headquarters in the Watergate
complex in Washington, DC. Amid allegations that he tried to cover up his involvement in this burglary,
President Richard Nixon resigned from office two years later. To examine the effects of this scandal
while events were still unfolding, Greenstein compared the attitudes of children in 1969–1970 with those
held in June 1973.23 Although these children viewed the president somewhat less positively in the
second time period, assessments of the president did not become significantly more negative during the
early 1970s. In a very specific domain, however, children’s attitudes toward the president did change.
Compared with four years earlier, in 1973 children were much more likely to believe that the president is
above the law (31 vs. 58 percent, respectively, expressed this view). Finally, the findings that children’s
evaluations of the president as a person were more negative in 2000 and 2017–2018 compared to the
1950s can probably be traced to Watergate. The Watergate era ushered in a sustained period of
increasingly negative views of government and politicians among the American public, including
children.
Age, Class, Ethnic, and Racial Differences
Although positive images of political leaders are fairly common among children, there are important
exceptions to this trend. Older children are substantially less likely to view leaders in an idealized
fashion.24 Further, in prior decades, children’s assessments of a president’s personal qualities (such as
honesty) became more negative as the children got older, but their evaluations of the president’s
governing-related characteristics (such as working hard and being knowledgeable) remained positive.25
More recently, however, children’s evaluations of all three of these traits become more negative as
children grow older.26
Significant class and racial differences also exist in children’s evaluations of political leaders. In 1967,
Dean Jaros, Herbert Hirsch, and Frederic Fleron surveyed children from Appalachia (specifically eastern
Kentucky).27 They selected this region because its higher-than-average levels of poverty and relative
isolation distinguish it from most middle- and upper-class regions of the United States. Jaros, Hirsch,
and Fleron’s results are strikingly different from those obtained by Greenstein or by Hess and Easton.
Appalachian children demonstrated much less positive attitudes toward leaders and the political system.
Whereas 77 percent of the fifth to eighth graders in the Chicago area, for example, believed that the
president works harder than “most men,”28 only 35 percent of the Kentucky children held this view. Also,
26 percent of the children in Appalachia believed that the president is “not a good person” compared
with only 8 percent of Chicago-area children. Rather than Greenstein’s benevolent leader, Jaros et al.
conclude that malevolent leader imagery is more common in eastern Kentucky. Their results are
important not only for what they demonstrate about political socialization in Appalachia, a region that is
not often studied, but also because they caution us against concluding that positive images of political
authority were universally held among American children in prior decades.
Idealized images of the president are less common among Black children than among white children. In
a 1969–1970 study of children’s attitudes, 32 percent of Black children possessed positive or idealized
assessments compared with 55 percent of white children.29 These racial differences generally exist at
all grade levels but are especially notable as children grow older. For example, whereas attitudes toward
the president and police officers were similar for Black and white second graders, by eighth grade Black
children held significantly more negative images than their white peers.30 More recently, the 2017–2018
childhood socialization study uncovered differences by race and ethnicity: assessments of the
president’s personal qualities and job performance were most positive among white school children and
most negative among Black children, with the views of Latino and Asian children falling in between.31
Racial differences also exist when we consider other political attitudes. White school children tend to
have considerably higher levels of political trust and efficacy compared with Black school children. Trust
assesses the degree to which individuals agree that political leaders are honest and act in the public’s
interest. Efficacy refers to the belief that one can influence the decisions of government officials and the
belief that these officials are responsive to public wishes. When levels of trust and efficacy by children’s
race were compared in the 1960s, Black children had consistently lower levels of efficacy than did their
white peers. Racial differences in trust, however, only emerged in research conducted after summer
1967, at which point levels of trust were lower among Blacks. Before then, white and Black children had
similar levels of trust. That year marked a time when the Black community as a whole became less
trusting of the government, in part because urban riots were occurring in the United States and the
policy gains achieved during the civil rights movement had seemingly ended.32
Ethnic and racial differences in these attitudes have persisted. In 2003 and 2004, Kim Fridkin, Patrick
Kenney, and Jack Crittenden surveyed white, African American, Latino, and Native American eighth
graders in and around Phoenix, Arizona.33 Compared with the minority students, white students
displayed more trust in government and higher levels of political efficacy. Native Americans had the
lowest levels of both trust and efficacy.
In one of the few recent examinations of younger children’s political attitudes, Christia Spears Brown,
Rashmita Mistry, and Rebecca Bigler conducted a small survey of African American children in the
aftermath of Hurricane Katrina.34 Katrina, a powerful hurricane, landed on the Gulf Coast in late August
2005, with devastating consequences. Massive flooding and property damage occurred, notably in New
Orleans. Governmental response to the hurricane was widely considered to be too slow and
inadequate. Meanwhile, Americans throughout the nation witnessed the crisis unfold on their television
sets. The New Orleans victims were disproportionately Black and poor, leading Brown and her
colleagues to explore Black children’s attitudes toward the government response and attributions of
responsibility for the inadequate relief efforts. Their sample was drawn from schools in one city—Los
Angeles—and contained only African Americans. Thus, their results cannot speak to racial differences in
relevant attitudes or to nationwide attitudes. Yet Brown, Mistry, and Bigler report a number of interesting
findings, particularly regarding age differences. Their youngest participants (second graders) were more
likely to evaluate President George W. Bush’s performance favorably than were the older respondents
(eighth graders). In terms of what was responsible for the delay in relief reaching the victims, the
younger children were most likely to credit logistical challenges (such as the difficulty in rescuing
thousands of people). The eighth graders, on the other hand, were more likely to believe racial
discrimination was a factor.
What might account for class, ethnic, and racial differences in children’s attitudes? According to one
approach, ethnic and racial minorities have less power than whites in the political system and less
reason to believe that political leaders will respond to their wishes. Furthermore, past ethnic and racial
discrimination at the hands of government (such as school segregation, police violence, and voter
disenfranchisement) has generated mistrust toward the government among affected group members.
Black, Latino, and Native American children are aware of these current and past realities, which
contributes to their having different attitudes from white children.35 In other words, this political reality
explanation posits that political attitudes respond to actual political events and phenomena.
Parental communication is also an important factor. Jaros, Hirsch, and Fleron attribute the Appalachian
children’s less favorable assessments of leaders and the political system to their parents’ views. Among
Appalachian adults, “there is a great deal of overt, anti-government sentiment. … Rejection of and
hostility toward political authority, especially federal authority, has long characterized the region.”36
Attempts to explain racial and ethnic differences in trust and efficacy also posit a role for parents. Fridkin
and her colleagues found that children who discussed politics with their parents had more positive
attitudes toward government but also that political discussions were more common in the homes of
middle-class white children than Black, Latino, Native American, or working-class white children.37
Fridkin, Kenney, and Crittenden also wonder about the nature of political discussions in the homes.
Negative views toward government (particularly the government’s past and present interactions with
minorities) might be shared between minority parents and children, they argue, more so than in white
households. Unfortunately, these researchers did not assess the content of family political discussions.
Their work, as well as that of Jaros, Hirsch, and Fleron, suggests that future studies of childhood
socialization should examine family conversations more fully.
Box 2.1 Childhood Political Socialization in Europe
Public Opinion in Comparative Perspective
Childhood socialization research has seen a recent resurgence in western Europe. Much like in
the United States, the scholarly community moved away from studying children, in part because
of an assumption that young children “lack basic competencies to deal with abstract and
complicated matters such as politics.”1 One question thus guiding newer scholarship is whether
children possess meaningful political perspectives. They do.
Children aged eight through twelve interviewed by Lena Haug display an interest in political
topics, with especially high interest in topics such as children’s rights, war, poverty, and the
environment.2 General knowledge of politics is evident among elementary school children, yet
gender differences also emerge by then, with girls displaying lower levels of knowledge.3
Children are also developing political orientations and an understanding of key political concepts.
Jan van Deth, Simone Abendschön, and Meike Wollmar explored children’s notions of good
citizenship, uncovering that as early as the first year of primary school, children differentiate
between characteristics such as helpfulness and obeying the law from being popular or wealthy.4
Conceptions of economic inequality and meritocracy among eight- to ten-year-old children were
the focus of Tanja Betz and Laura Kayser’s work.5 They conclude that children attribute adult’s
economic success to academic achievement earlier in life, a pattern that did not vary much by
the social class of the children studied.
These works provide important insights into the development of political views among children.
Furthermore, owing to the fact that children would have difficulty answering standard survey
questions posed to adults, these scholars use a variety of creative research approaches. One
way that Haug measured topic interest was by asking children to draw a picture about their
“World of Politics.” Van Deth and colleagues designed a survey that contained pictures
corresponding with question topics. Survey response options were presented as familiar objects
(such as smiley faces) or in familiar formats (multiple choices, as in a test). Betz and Kayser
opted not to survey children but instead conducted in-depth interviews, conversation-style.
1 Jan W. van Deth, Simone Abendschön, and Meike Wollmar, “Children and Politics: An
Empirical Reassessment of Early Political Socialization,” Political Psychology 32 (2011): 147.
2 Lena Haug, “‘Without Politics It Would Be Like a Robbery without Police’: Children’s Interest in
Politics,” American Behavioral Scientist 61 (2017): 254–272.
3 Alice Simon, “How Can We Explain the Gender Gap in Children’s Political Knowledge?”,
American Behavioral Scientist 61 (2017): 222–237.
4 Van Deth, Abendschön, and Wollmar, “Children and Politics.”
5 Tanja Betz and Laura B. Kayser, “Children and Society: Children’s Knowledge about
Inequalities, Meritocracy, and the Interdependency of Academic Achievement, Poverty, and
Wealth,” American Behavioral Scientist 61 (2017): 186–203.
PARENTAL TRANSMISSION OF POLITICAL ATTITUDES
As they move into adolescence, children begin to acquire specific political opinions to add to the more
general orientations toward government and political leaders gained during early childhood. Parents are
thought to be a key source of these political attitudes, perhaps even the most important source, as the
following quotation illustrates: “Whether the child is conscious or unaware of the impact, whether the
process is role-modelling or overt transmission, whether the values are political and directly usable or
‘nonpolitical’ but transferable, and whether what is passed on lies in the cognitive or affective realm, it
has been argued that the family is of paramount importance.”38
In 1965, Kent Jennings and Richard Niemi began a study to examine directly the similarity between
adolescents’ political attitudes and those of their parents.39 Their research—one of the most influential
political socialization studies conducted in the United States—improved on prior socialization studies in
important ways. Thus, we profile their study in this section. We begin with the socializing role played by
parents during their children’s adolescence and then explore whether attitudes acquired by the children
remain stable during their adult years. For both adolescent socialization and adult socialization, we
focus heavily on the acquisition of one important attitude—party identification.
Parental Transmission during Adolescence
Empirically, one could assess the influence of parents’ attitudes on their children by using a number of
approaches. One method involves surveying the children, asking them their political attitudes and also
asking them to report their parents’ attitudes. (Similarly, one could survey parents, querying them about
their and their children’s attitudes.) This approach is limited, however, because of the possibility that the
children either do not know their parents’ attitudes or assume that their parents’ attitudes are the same
as theirs. If the latter occurs, this projection could lead to the parents’ and children’s attitudes appearing
to be more similar than they really are. To avoid these problems, Jennings and Niemi surveyed children
and their parents separately, with members of each group completing their own questionnaires. In total,
1,669 high school seniors took part in the first (1965) wave of their study. For approximately one-third of
these students, their father was randomly selected to complete a questionnaire. The mother was
randomly selected for another third, and both parents were selected to be surveyed for the final third.
Another advantage of Jennings and Niemi’s study is that their research participants were selected to
represent the entire nation. Rather than studying parents and children from one city or one geographical
area, these researchers used a national sample. High schools across the nation were randomly
selected, with steps taken to ensure that this sample accurately represented the entire population of
high schools in the United States. Thus, the ninety-seven selected high schools included those from
cities, suburbs, and rural areas; those with varying numbers of students; those from every geographical
region of the nation; and both public and private schools. Within each selected school, fifteen to twenty-
one seniors (depending on the size of the school) were randomly selected to participate in the study.
This approach to selecting study participants, known as a national probability sample, allowed
Jennings and Niemi to make inferences from their participants to the entire nation of high school seniors
and their parents. With other methods, researchers must be more cautious in their conclusions.
Selecting participants from one’s local area, for example, does not allow a researcher to draw
conclusions about the entire nation. Further, if participants volunteer to participate instead of being
randomly selected, we cannot be certain that these self-selected participants’ attitudes mirror those of
the greater population. In fact, these people very likely may have more intense attitudes or be more
politically aware, factors that increase the likelihood that an individual will voluntarily participate in a
political survey.
To assess how thoroughly parents transmit their political attitudes to their offspring, Jennings and Niemi
compared a variety of political attitudes between parents and their children. One of their most significant
conclusions is that children are more likely to share their parents’ party identification than other
political attitudes. Nearly 60 percent of high school seniors had the same general partisanship as their
parents (for example, if the child was a strong Democrat, the parent was either a strong, weak, or
Independent-leaning Democrat). In only 7 percent of the parent-child pairs was one person a Democrat
and the other a Republican, or vice versa. This result led Jennings and Niemi to conclude that the
“transmission of party preferences from one generation to the next is carried out rather successfully in
the American context.”40 One significant difference in partisanship did emerge from their analysis: the
children were more likely to be politically independent than their parents (35.7 vs. 23.9 percent identified
as Independent, respectively). Unbeknownst to the researchers at the time, Jennings and Niemi’s data
capture a snapshot of a decades-long trend of Americans becoming more weakly attached to the
political parties, a topic we return to shortly.
As for political attitudes other than partisanship, such as opinions regarding school integration and
school prayer, and evaluations of political groups, these appear to be passed from parent to child less
often. Finally, there is little agreement between parents and offspring on their degree of cynicism toward
politicians and the political system. Overall, high school seniors are much less likely to be cynical than
their parents, a result that coincides with the childhood socialization research presented earlier.
Jennings and Niemi attribute this finding to the fact that schools serve as powerful socializing agents,
inculcating positive views of the nation (through rituals and curricula) while avoiding much critical
analysis of the U.S. government.
These results tell us something about adolescent socialization in the 1960s, but what about in more
recent decades? Fortunately, Jennings and Niemi were not done exploring political socialization in 1965.
The high school seniors they first interviewed in 1965 were reinterviewed three more times, the last time
in 1997. The children of the former high school seniors were also interviewed in 1997. This research
design permits an exploration of parent-child transmission for two different time periods, which is exactly
what Kent Jennings, Laura Stoker, and Jake Bowers did.41 In other words, they compared the
correspondence between the attitudes of the 1965 high school seniors and their parents (in 1965) with
the correspondence between these former high school seniors and their children (in 1997).
Jennings, Stoker, and Bowers found that, across a variety of political issues, the likelihood that a child in
his late teens will hold the same attitudes as his parents was largely the same in 1997 as it had been in
1965. Furthermore, parent-child correspondence was higher for party identification in 1997 than almost
all other issues. In a departure from 1965, however, parental transmission of two attitudes—toward gay
rights and abortion—was higher than for party affiliation. The authors attribute this to the high salience
and moral basis for both of these issues. In the end, they conclude that “the patterns of political
reproduction do not differ appreciably across the generations.”42 Jennings and Niemi’s initial results
were thus not timebound; that is, they were not a product of the political times of the 1960s.
Taking this work a step further, Jennings, Stoker, and Bowers explored when the transmission of party
identification from parents to their children would be enhanced. In particular, they identified family
characteristics that should result in parents providing frequent and clear cues regarding their political
views to their offspring. Two seem to be especially important: family politicization (the degree to which
parents are politically active and politics is discussed in the household) and parental attitude stability.
As demonstrated in Figure 2.2, the correspondence of party identification between parents and children
was indeed higher in more politically active and conversant families than low-politicization families. Also,
the more stable a parent’s party affiliation was, the more likely a child identified with the same party as
her parent. These two results held for both time periods, but the effect of both family characteristics was
stronger in 1997 than in 1965, as demonstrated by the larger gap between the light- and dark-colored
bars for 1997 than for the earlier time.
Description
Figure 2.2 Parent-Child Correspondence of Party Identification by Family
Politicization and Parental Attitude Stability
Source: Data from M. Kent Jennings, Laura Stoker, and Jake Bowers, “Politics across Generations: Family
Transmission Reexamined,” Journal of Politics 71 (2009): 789.
Note: Bars represent the correspondence of the parent’s and children’s party identification, derived from
multivariate regression analyses. Correspondence is measured on a scale of 0 to 1, with 1 indicating parents
and children shared the same party identification within every pair. A score of 0 would mean that parent and
child party identification is not the same for any of the pairs.
Another topic that political socialization scholars have examined is whether mothers or fathers are more
likely to pass on their party identification to their children. Early analyses, of the 1965 wave of the
Jennings-Niemi socialization data, demonstrate that when parents do not share the same party
affiliation, adolescents are a bit more likely to align with mom rather than dad.43 Zoe Oxley updated
these findings, exploring first the 1997 wave of Jennings and Niemi’s study, then a 2012 survey of high
school and college students.44 The 2012 survey is not ideal for exploring parent-child transmission
because the children were asked to report their parents’ partisanship, but it does permit a look at more
recent socialization patterns. The results demonstrate that mother-child congruence has continued to be
higher than that between fathers and children, although the parental gap is smaller in 2012 than it was
for 1997. Also, two of the 2012 results run counter to this overall finding. Children are more likely to
share their father’s partisanship when they discuss politics more with dad than mom or when dad does
the majority of the household tasks and child care. This pattern confirms what was expected (but not
directly tested) in earlier analyses: children’s party identification follows that of the parent that they
spend the most time with, whether discussing politics or not. Today, as in past decades, that parent is
typically the mother.45
Finally, examinations of parent-child transmission of party identification for all three time periods reveal
some evidence of same-sex transmission, with daughters being more likely to align with their mothers’
partisanship as sons align with fathers. Research conducted by Nicole Filler and Kent Jennings
demonstrates that same-sex parent-child congruence also exists for gender role attitudes, such as
preferences for women and men having equal roles in business and politics versus believing women’s
proper place is in the home as well as views toward the women’s movement.46 Indeed, for these
attitudes, same-sex transmission is even more prevalent than for party identification or for other political
opinions.
Do Preadult Attitudes Persist Into Adulthood?
Once children leave adolescence and enter adulthood, do their political attitudes remain the same? If
not, do they change in predictable ways? There are a number of methods to study adult socialization.
One of the most effective ways is to survey the same group of people when they are adolescents and
then again when they are adults. This method, called a panel or longitudinal study, is the approach
taken by Jennings and Niemi. As already mentioned, they interviewed their sample of high school
seniors four times: in 1965, 1973, 1982, and 1997. Although they were not able to reinterview all of the
1,669 seniors who had participated in the original study, they did reinterview nearly 1,000 of the
participants in all four waves of the study. This study design allows a comparison of the attitudes of
these individuals at various points in their life—as high school seniors, at twenty-six years old, at thirty-
five years old, and again at fifty years old—to directly assess whether their late-adolescent political
attitudes persisted into and throughout adulthood. This panel study has resulted in a rich array of
information and has produced a number of interesting insights into adult socialization.
In particular, we have learned a lot about the stability of party identification over time from analyses of
this panel study. Figure 2.3 demonstrates how stable the high school seniors’ party affiliation was from
the year of their high school graduation (1965) to their midtwenties (1973) to two other time periods:
1973–1982 (when the former seniors aged from twenty-six to thirty-five) and 1982–1997 (when they
aged from thirty-five to fifty). For now, focus your attention on the white bars for each time period. The
bar is shorter for the 1965–1973 time span than for the other two. This means that these respondents’
party identification was most likely to change between the ages of eighteen and twenty-six. After
passing through their midtwenties, their party affiliation remained more stable. In fact, the level of
stability was the same between the ages of twenty-six and thirty-five as between thirty-five and fifty.47
Finally, similar findings exist for opinions toward political issues such as racial policies, school prayer,
and political tolerance. Attitudes shifted around a fair bit when the members of the high school class of
1965 were in their early twenties but did not change as much throughout their later adulthoods.48
Description
Figure 2.3 Stability of Party Identification Over Time, Overall and by Preadult
Parent-Child Correspondence
Sources: Data for “All” bars from Laura Stoker and M. Kent Jennings, “Of Time and the Development of
Partisan Polarization,” American Journal of Political Science 52 (2008): 623. Data for other bars from M.
Kent Jennings, Laura Stoker, and Jake Bowers, “Politics across Generations: Family Transmission
Reexamined,” Journal of Politics 71 (2009): 794.
Note: Bars represent the stability of party identification across each time period, as measured by continuity
correlations.
This pattern supports an impressionable years model of attitude stability. The impressionable years,
typically late adolescence and early adulthood, can be a time of personal growth and development. As
an individual goes through her impressionable years, personal experiences (such as moving away from
the childhood home, beginning a career, or getting married) can have political ramifications.49 In
particular, her political views may fluctuate. According to this understanding, those in their early twenties
lack the experience to have consistent political opinions. As young adults grow older, they experience
fewer genuinely new events that influence their political opinions. Their political views then become
more firmly grounded in their past experiences and are thus more resistant to change.50 The
impressionable years model is a specific type of a broader class of socialization dynamics—life cycle
effects. The life cycle explanation presumes that people’s political attitudes are influenced by their age
(by their place in the life cycle). Theoretically, these effects could occur at many different stages in a
person’s life. Yet the most consistent effects of age on political attitudes are the ones we have already
mentioned: attitude volatility during the impressionable years followed by relative stability later in
adulthood.51
From the Jennings-Niemi panel data, we have learned that political opinions of many people undergo
changes as they enter adulthood. Now, let’s bring parents back into our discussion, particularly the
degree of parent-child attitude similarity at the end of adolescence. Return your attention to Figure 2.3,
focusing now on the colored bars. Recall that this figure presents the over-time stability of party
identification for Jennings and Niemi’s panel respondents for three different time periods. The gray bars
represent stability over time among those whose party affiliation did not correspond closely to that of
their parents’ party affiliation when they were eighteen years old. The black bars represent over-time
stability among preadults who had a high correspondence with their parents.
This analysis, which was conducted by Jennings, Stoker, and Bowers, uncovered two interesting
findings.52 First, from age eighteen to twenty-six, people who possessed a party affiliation similar to that
of their parents experienced less volatility in their partisanship compared to those whose partisanship
was dissimilar from their parents. As Jennings and his colleagues explain, “Those young adults entering
[adulthood] more securely attached to the political ‘apron strings’ of their parents were more likely to
withstand the novelties they were to encounter. Those less anchored in that way proved to be far more
vulnerable, and thus more apt to change.”53 Second, between the ages of twenty-six and thirty-five and
again between thirty-five and fifty, the stability of the party identification of these individuals over time
was much less dependent on whether they had shared their parents’ partisanship when they were
eighteen. The story here is much more about increases in stability over time as an individual leaves the
impressionable years (the gray and black bars for both of the later two periods are taller than for the first
period). In contrast, the gap between the gray and black bars is much smaller for either of the final two
time periods compared to the first.
In addition to parent-child attitude similarity, the degree of family politicization during adolescence also
influences attitude stability in adulthood. Yet, the findings from work by Elias Dinas might surprise you.54
Recall that in more politically active and conversant families, children are more likely to share their
parents’ partisanship than in less political families. As children enter adulthood and leave the family
home, however, it is actually the children of the more politicized families that are most likely to move
away from their parents’ party identification. How might this be? Young adults from more politically
engaged families carry this political interest with them as they embark on their new lives. They are more
likely to discuss politics and be aware of political happenings than their peers who were raised in less
politically tuned-in households. Exposure to new ideas and information results in some offspring from
politicized families revising their partisanship, most especially during their young-adult years. In contrast,
because children from less politicized households “end up less politically engaged as adults, their
partisanship is not as subject to the social and political forces that prompt partisanship changes.”55
Finally, the lingering effects of adolescent socialization on adult attitudes exist for political outlooks other
than partisanship. For instance, abortion opinions among adults are related to their adolescent religious
practices and the gender role attitudes expressed by their mothers while the children were growing up.
Julianna Pacheco and Rebecca Kreitzer reached these conclusions by analyzing longitudinal survey
data, data that were collected in an early wave from adolescents and their mothers and then in a later
wave after these adolescents had entered adulthood.56 As adults, support for abortion is lower among
those who had attended religious services more frequently as an adolescent compared to those who
were less religiously active in their youth. Furthermore, the more traditional one’s mother’s gender role
attitudes were during a child’s adolescence, the less supportive that person will be of abortion in
adulthood.
GENERATIONAL AND PERIOD EFFECTS
So far, we have primarily discussed the role that individuals and institutions play in shaping the political
attitudes of children and adolescents. The development of political opinions is also influenced by the
political context, such as specific political events and broader political trends.
When changes in the political context influence the political socialization of an entire age cohort, a
generational effect on political attitudes occurs. In other words, the opinions of an entire generation of
people can be influenced by the nature of the times. This is especially likely for those who are in their
impressionable years. Take, for instance, the formation of party affiliation among those who came of age
during the post–World War II period.57 They were socialized at a time when the political parties were
weakening on the national stage. Beginning in the 1950s, national politicians began to build personal
campaign organizations rather than tying themselves clearly to the national parties. Television coverage
of campaigns tended to focus on the candidates rather than the parties, and the candidates could
bypass the parties and their grassroots campaign organizations by using television to disseminate their
messages directly to the voters. These changes in the broader context influenced the partisanship of
individuals whose formative socialization occurred during this period. In particular, they were less likely
to identify strongly with either the Democratic or Republican Party than were cohorts who had been
socialized earlier, when parties had been more dominant.
As the political parties were changing nationally, regional developments were also afoot. Notably, white
voters in the southern United States began to shift their allegiance to the Republican Party in the 1960s.
During the decades immediately prior to this, Democratic support among whites in the South had been
strong, so strong that the region was known as the Solid South. Much to the chagrin of southern
Democratic whites, however, the national Democratic Party championed civil rights legislation in the
1960s. The partisan context in the region thus changed, with the Democratic Party becoming much less
popular over time. Following a generational effects model, this changed context produced effects on
partisanship in the region among those who were leaving adolescence. Southern whites who were in
their formative years during or after the 1960s were less likely than older whites in the region to adopt a
Democratic Party identification. Successive generations became less and less Democratic as they
entered adulthood.58 In fact, by 1984, southern whites who were just then becoming eligible to vote
were more likely to be Republicans than Democrats.
A period effect occurs when salient features of the political period influence the political attitudes of
many, regardless of age. Note that period effects differ from generational effects in one important way:
generational effects result when aspects of the political context shape the political attitudes only of
people who are similar in age (such as those who are in their impressionable years). A few examples
should serve to illustrate this difference. In fact, the studies we used to demonstrate generational effects
in the two previous paragraphs also uncovered period effects. Recall the first example. It demonstrated
that Americans were less likely to hold strong partisan attitudes after World War II because generations
of citizens were socialized in an era when the national political parties had weakened. At the same time,
across many age cohorts, the percentage of people identifying as Independents increased. This turn
away from the parties and toward partisan independence happened among younger citizens who were
just entering adulthood as well as among older adults.59 In other words, citizens of all ages were
somewhat influenced by the decline of parties on the national political stage, a period effect.
Similarly, a changed political context produced period effects in the South. Not only have new
generations of white Southerners displayed lower levels of identification with the Democratic Party than
prior generations, but also southern whites across many generations have been less likely to identify as
Democratic over time.60 For example, the average party identification of white Southerners who entered
adulthood in the 1960s was between Independent-leaning Democrat and weak Democrat. By 2004, the
average party affiliation of this cohort was between Independent and Independent-leaning Republican.
In other words, the partisanship of this group continued to change as they aged, even when they were
well past their formative years. The drift toward Republican Party identification occurred among other
generations of southern white adults as well. Key features of the 1960s help to explain why, most
especially the civil rights movement and the civil rights legislation that was passed by Congress during
that decade.
The Politics of Millennials and Generation Z
Generational differences abound in our society—in musical tastes, use of technology, clothing
preferences, and communication patterns, to name a few. Age cohorts can also differ from others in
terms of political outlooks and behaviors. More specifically, a political generation “refers to a group that
is [politically] distinctive in any number of respects by virtue of having experienced a specific set of
social, economic, technological, and/or political circumstances at a formative period in their lives.”61 In
other words, a generation of people—such as the Baby Boomers (born between 1946 and 1964) or
Generation Xers (1965–1980)—pass through their impressionable years around the same time and thus
are exposed to the same political and social context during this crucial time in their lives. If this exposure
results in a set of distinctive political views for the generation, a political generation exists.
What about today’s youngest generations? Do Millennials (1981–1996) and members of Generation Z
(born after 1996) differ politically from earlier generations?62 In terms of a variety of political views, yes,
although we can be more confident regarding the opinions of Millennials compared to Generation Zers.
The oldest members of Generation Z have only recently entered adulthood, and rigorous survey
research is typically conducted on Americans who are at least eighteen years old. Fortunately, though,
in 2019 the Pew Research Center conducted a detailed analysis of generational differences in political
views, for which they surveyed Generation Z members as young as thirteen.63 Their findings, some of
which appear in Figure 2.4, provide a snapshot of the developing views of this youngest generation.
Furthermore, among the first five topics displayed in the figure, Millennials’ and Generation Zers’
attitudes are clearly distinct from the other two generations (Generations Xers and Baby Boomers), in
the more liberal direction for all. Compared to older cohorts, members of the younger generations are
more supportive of racial equality, LGBT rights, and government involvement to address societal ills as
well as more likely to believe the Earth’s warming is due to human activity. One area in which the views
of Millennials and Generation Zers are somewhat different is gender identity. Members of Generation Z
are more supportive than Millennials of forms including gender options other than “man” and “woman”
and are more likely to know someone who prefers the use of gender-neutral pronouns. Finally,
differences across these four generations do not emerge regarding views toward the benefits of women
running for political office or for other gender role attitudes, such as the sharing of child-care
responsibilities by and the expectation of equal financial contributions from both parents in two-
heterosexual-parent households.64
Prior research by Pew, conducted in 2017 and 2018, documents other important differences in political
outlooks among Baby Boomers, Generation Xers, and Millennials.65 For instance, Millennials are more
likely to consider themselves political independents than are members of previous generations (44% vs.
39% for Generation X and 32% for Baby Boomers). Among those with partisan leanings, Millennials are
the generation most likely to be Democratic. Compared to older cohorts, Millennials are more supportive
of a path to citizenship for undocumented immigrants, government involvement to protect the
environment, legalizing recreational marijuana, the North American Free Trade Agreement, and
pursuing diplomacy rather than military might to ensure peace among nations. On the other hand, in
terms of trust in government as well as the issues of abortion and gun control, Millennials’ attitudes are
not much different than those of older generations.66 The lack of generational differences on gun control
opinions might surprise you, given the high profile gun control activism of Parkland, Florida, teens after
the February 2018 shooting at their high school. Furthermore, the youngest adults have come of age
during an era of more frequent mass shootings, at schools and elsewhere. Yet, many surveys have
found that Millennials’ views toward gun control tend not to be consistently more liberal than the views of
older Americans.67 Future research is needed to fully understand why and to determine whether
Generation Zers will be more supportive of gun control measures as they enter and advance through
adulthood.
Description
Figure 2.4 Generational Differences in Attitudes, 2019
Source: Data from “Generation Z Looks a Lot Like Millennials on Key Social and Political Issues,” Pew
Research Center, Washington, DC, January 17, 2019,
https://www.pewsocialtrends.org/2019/01/17/generation-z-looks-a-lot-like-millennials-on-key-social-and-
political-issues/
Note: The bars represent the percentage of respondents who agreed with each position.
Cataloging differences among generations is relatively straightforward, although explaining why such
differences exist is tricky.68 For one, at a given point in time, generations differ by when they went
through their impressionable years as well as their current age. Put another way, today’s Millennials’
attitudes might differ from Baby Boomers’ attitudes simply because they are younger rather than
because they were socialized in a different historical setting. Furthermore, generations differ in myriad
politically relevant ways, making it challenging to sort out which factors might contribute to generational
distinctiveness. Bearing this complexity in mind, a few possible explanations for the Millennial
generation’s distinctiveness have been proposed, although all would benefit from more rigorous
empirical testing. Coming of political age during a time of political polarization and government
stalemate has likely contributed to the Millennials’ weaker ties to the political party system. White
Millennials’ more liberal racial attitudes have been attributed to one key political event that happened as
many were entering adulthood: the election of Barack Obama as president.69 The fact that the Millennial
generation is more racially and ethnically diverse than older generations likely contributes to its
liberalism on issues such as equality, immigration, and government activism.70 That explanation likely
https://www.pewsocialtrends.org/2019/01/17/generation-z-looks-a-lot-like-millennials-on-key-social-and-political-issues/
applies to Generation Zers as well, given that this generation is even more diverse than are
Millennials.71 Other societal factors are probably also relevant for understanding the political
distinctiveness of Millennials and Generation Zers, such as the increase in nontraditional families and
the heightened visibility of gays, lesbians, and transgender individuals in popular culture and in
communities across the United States.
GENETIC INHERITANCE OF POLITICAL ATTITUDES
From our parents, we inherit, among other things, our hair color, eye color, and height. Could some of
our political attitudes also result from genetic inheritance? Yes, according to research conducted by
John Alford, Carolyn Funk, and John Hibbing.72 Not surprisingly, it is not at all easy to sort out what
portion of a person’s political attitudes might be genetically inherited, influenced by family during early
socialization, or influenced by other agents of socialization. After all, we cannot ask a newborn baby
whether she is a Republican or a Democrat at the moment of her birth, before her postnatal political
socialization begins. Alford, Funk, and Hibbing necessarily took a very different approach in examining
the relationship between genetics and politics by borrowing research approaches from scientists who
specialize in genetics.
Specifically, Alford et al. compared the similarity in political opinions of monozygotic twins (typically
known as identical twins) to the similarity of dizygotic twins (fraternal twins). The genetic material of
monozygotic twins is the same because the conception of these twins involved the fertilization of one
egg by one sperm. In contrast, dizygotic twins are conceived when two eggs are fertilized by two
separate sperm at the same time. Thus, these twins’ genetic makeup is only 50 percent identical (which
is the same for any pair of siblings). This fact regarding known differences in genetic makeup between
monozygotic and dizygotic twins is an important foundation of Alford et al.’s research. They also
assume, as do all who conduct research on twins, that the influence of environmental factors (parental
socialization, outside influences on attitudes, and so on) is roughly the same for each type of twin. This
does not mean that every twin’s attitudes are influenced precisely the same amount and in the same
way. Rather, the assumption is that the impact of environment on attitudes is the same, on average, for
monozygotic twins as it is for dizygotic twins.73
With this foundation in place, Alford, Funk, and Hibbing examined how similar the political opinions of
twins are separately for monozygotic versus dizygotic twin pairs. Unlike other studies of socialization
within families (such as Jennings and Niemi’s), Alford et al. do not need to compare parents’ attitudes
with their children’s; comparing twins’ attitudes across types of twins with known differences in genetic
inheritance is sufficient. Alford and colleagues parceled out the proportion of variation in attitudes that is
due to genetic heritability versus the twins’ shared environment (primarily parental and family
socialization but also shared school environments and exposure to the same media) versus their
unshared environment (any unique experiences, socialization that occurred after adolescence, and so
on). The statistical techniques they used are too complex to describe here, but a key premise of their
analysis is important to understand. If a political attitude is genetically inherited, then monozygotic twins
will be more likely (than dizygotic twins) to share the attitude with each other because monozygotic twins
share more genetic material. For attitudes that originate with environmental factors rather than genes,
the likelihood that a pair of twins will have the same attitude will not differ across monozygotic and
dizygotic twins.
So, what do Alford, Funk, and Hibbing conclude? They find that a portion of political attitudes is
genetically inherited, although the role of genetic inheritance differs across types of attitudes (see Figure
2.5). Of the three attitudes that they examined, the largest effect of genetic transmission occurs for
political ideology (whether one is conservative or liberal). Just over 40 percent of the variation in
whether these twins are liberal or conservative is due to genetic inheritance, compared with about one-
third for their unshared environment and about one-quarter for the shared environment. In contrast, the
unshared environment contributes a much larger share toward political opinionation (or whether
individuals have opinions on political matters), with genetic factors accounting for just over one-third,
and a negligible role for the shared environment. The attitude that is least likely to be genetically
inherited is party identification. The twins’ shared environment (which is assumed to be largely parental
socialization) is a much stronger influence on party affiliation than are genetics (.41 vs. .14,
respectively). This finding reinforces one that you learned earlier in this chapter: parental socialization is
an important source of individuals’ partisanship.
Description
Figure 2.5 Genetic Versus Environmental Factors Influencing Political
Opinions
Source: Data from John R. Alford, Carolyn L. Funk, and John R. Hibbing, “Are Political Orientations
Genetically Transmitted?” American Political Science Review 99 (2005): 160.
Note: Bars represent the proportion of estimated variability in political opinions due to each of the three
types of factors.
The publication of Alford, Funk, and Hibbing’s important study began a wave of other research exploring
genetics and politics. In addition to political ideology, it appears that other attitudes also have a
significant heritable component. These include the strength of a person’s party identification, trust in
government, interpersonal trust, core values such as egalitarianism, and agreement that voting is a civic
duty.74 Voter turnout and other forms of political participation have also been linked to genetics.75
Finally, Alford et al.’s finding that ideology is more closely related to genetics than party identification has
been replicated with a sample that includes twins as well as nontwin siblings of twins.76 Incorporating
nontwins into this research is an important development, primarily because twin-only designs are open
to generalizability critiques. In other words, we should be cautious in assuming that results gleaned from
studies of twins apply to all people, twins and nontwins alike.
How are genes linked to political attitudes? Answering this question is not easy. The consensus now is
that a complex chain is at work. Genes are known to influence many of our biological systems. These
systems can then condition our emotional reactions as well as the processing of incoming information,
both of which are probably related to our personality traits and core values, which in turn are related to
political ideology and issue attitudes. Throughout this sequence, environmental circumstances are also
at work. Conceptually, this is how genes are thought to be related to political attitudes.77
Research is now under way to explore some of the links in this chain, such as one study that explores
reactions to the September 11, 2001, attacks.78 An individual’s genetic makeup can influence how he
responds emotionally to a threatening situation, such as a terrorist attack. Some people felt angry
immediately after 9/11 and became more trusting of the government, whereas others were anxious and
experienced declines in trust. Anger can produce a desire for retaliation, in this case against the
responsible terrorists, and thus increased support for any person or body that can retaliate (such as the
federal government). In contrast, anxiety can be accompanied by beliefs that no person or organization
can control a situation. Furthermore, in the immediate aftermath of the 9/11 attacks, genes played a
more influential role on levels of governmental trust and environmental factors a less influential role
versus before that day. With the passage of time, during which no further attacks on U.S. soil occurred,
the influence of genes decreased back to pre-9/11 levels. In short, the shifting dynamics of public trust
resulted from the interaction between genetics and the environment, not genetic makeup or
environmental circumstances alone.79
The link between genetics and politics is an active area of public opinion scholarship. Nevertheless, not
everyone has climbed aboard the genetics train.80 In their 2005 article, Alford, Funk, and Hibbing
correctly predicted that some people would greet their finding of a genetic source for political attitudes
as “far-fetched, odd, even perverse.”81 Up until that time, much public opinion research, certainly
including political socialization, focused on environmental sources of attitudes. That is, factors external
to the individual were thought to play the largest role in developing political opinions. Peering internally,
all the way down to our genes, has the potential to upend conventional wisdom regarding opinion
formation.82 On the other hand, some socialization scholars have been open to this new research. Kent
Jennings, for instance, describes the genetics work as a “provocative addition to the political
socialization literature” and has called for integrating it with traditional approaches for studying
socialization.83
CONCLUSION
The acquisition of political attitudes begins fairly early in life and often with positive feelings toward the
nation and idealized views of political leaders. Would democratic elitists thus be satisfied with the
socialization of children in the United States today? Although it is true that levels of system support tend
to be high among children, as these theorists would hope, support is not uniformly high across all
children. For some children, such as those who are older, Black, Latino, Native American, or from
impoverished backgrounds, childhood attitudes are less benevolent. More worrisome for elite democrats
is that over recent decades children are more likely to have cynical attitudes toward those holding
political power. Elite democrats assume that citizens are neither interested in nor knowledgeable about
politics and rely on the expertise of elected officials to make governing decisions. As Americans,
including young Americans, have come to hold these officials in lower esteem, there is less deference to
political authority than elite democrats prefer. Even when children hold more uniformly positive attitudes
toward leaders, however, there is good evidence that these attitudes do not persist into adulthood,
further undermining elite democrats’ hope for socialization to produce system-supporting adults. For
instance, although the children of the 1950s trusted government and possessed idealized images of
leaders at that time, they “wound up rioting in the streets of Chicago or smoking dope in Vietnam or
working as carpenters under assumed names in Toronto” to avoid the Vietnam War draft.84
Adolescence brings with it the development of partisan leanings and opinions toward a variety of issues.
The family plays an important socializing role, particularly when it comes to party identification. Other
agents of socialization—schools, peers, and current political events—also shape political attitudes. Yet,
over years of socialization research, the family continues to emerge as an important shaper of children’s
attitudes, whether through genetic transmission or postnatal socialization. The imprint of our parents’
political attitudes is often still visible into adulthood, after a period of attitude instability that many of us
experience in our early twenties.
Pluralists would be pleased by the fact that the transmission of party identification continues to be more
complete than the transmission of most other political attitudes. Pluralists, after all, hope for socialization
to develop strong political identities before adulthood, chief among these being party identification. Yet
the development of preadult partisanship is not as complete as pluralists would prefer, partly because
the party affiliation of young adults is not entirely stable. This suggests that adolescents’ party
identification is not very crystallized, certainly not as crystallized as pluralists would want for citizens
about to reach voting age. Also of concern for pluralists is the movement away from strong party
affiliations that Americans have experienced over the past few decades. As generations of adolescents
have been socialized during a time of weaker national political parties, the effect has been to produce
more Independent voters who, in the pluralists’ eyes, are less able to have their interests represented
through the party system.
How would participatory democrats assess the state of political socialization? They would have a more
difficult task, given that much political science research on this topic has generally not focused on the
features of socialization that these theorists feel are most important. Having said this, we can point to
one conclusion from the research that certainly troubles participatory democrats: the fact that children
who are Black, Latino, Native American, or poor develop less trusting attitudes toward government at an
early age. The more trust in government citizens have, the more likely they will participate in politics
throughout their life. Participatory democrats worry that children who trust government less will grow up
to be adults who do not participate in politics, thus undermining the goal of political equality across
citizens that participatory democrats value so strongly.
More centrally for participatory democrats, teaching children to be active participants in democracy is
crucial. Most socialization research has not gauged the participatory skills and activities of children or
adolescents, certainly not to the degree that participatory democrats would like. Rather than
emphasizing which agents of socialization influence preadult attitudes most strongly, for example,
researchers could focus on whether parents and other agents pass on their civic engagement
perspectives to children.85 Philosophically, it is worth considering whether institutions in the United
States do or even should provide opportunities for civic skill development. Take universities. The
decades-long national trend has been “toward a higher-education system driven by the need to prepare
people for the workforce rather than the need to prepare people to participate in democracy.”86 Should
we worry that the civic mission of universities is slipping away? Participatory democrats do, just as they
would welcome more thorough assessment of courses and programs that develop in children,
adolescents, and young adults the civic skills that are relevant for political participation.
KEY CONCEPTS
adolescent socialization 59
adult socialization 59
attitude stability 62
benevolent leader imagery 52
childhood socialization 51
efficacy 56
family politicization 62
Generation Z 69
generational effect 67
genetic inheritance 71
ideology 72
impressionable years model 65
life cycle effects 65
malevolent leader imagery 56
Millennials 69
national probability sample 60
panel or longitudinal study 64
party identification 61
period effect 68
political generation 68
political opinionation 73
political reality explanation 57
political socialization 50
shared environment 72
socialization agents 50
trust 56
unshared environment 72
SUGGESTED SOURCES FOR FURTHER READING
Elder, Laurel, and Steven Greene. The Politics of Parenthood: Causes and Consequences of the
Politicization and Polarization of the American Family. Albany: State University of New York, 2012
Greenlee, Jill S. The Political Consequences of Motherhood. Ann Arbor: University of Michigan Press,
2014.
Sharrow, Elizabeth A., Jesse H. Rhodes, Tatishe M. Nteta, and Jill S. Greenlee. “The First-Daughter
Effect: The Impact of Fathering Daughters on Men’s Preferences for Gender-Equality Policies.” Public
Opinion Quarterly 82 (2018): 493–523.
Becoming a parent is one of the most transformative experiences of a person’s life, but can parenthood
shape one’s political views? Yes. In their book, Elder and Greene document a variety of issue opinions
for which the views of parents and nonparents, as well as mothers and fathers, diverge. Greenlee does
much the same in her thorough analyses of the opinions of mothers versus women who are not
mothers. Sharrow and her collaborators uncover evidence that fathers whose first child is a daughter
have more liberal views toward gender-equality policies. Having more daughters or a daughter who is
not a father’s first child does not alter these policy opinions.
Hibbing, John R. “Ten Misconceptions Concerning Neurobiology and Politics.” Perspectives on Politics
11 (2013): 475–489.
McDermott, Rose, and Peter K. Hatemi. “DNA Is Not Destiny.” In Oxford Handbook of Evolution,
Biology, and Society, ed. Rosemary L. Hopcroft. Oxford: Oxford University Press, 2018.
As the title of Hibbing’s article suggests, he rebuts ten common criticisms of biological approaches to
politics. He argues that political scientists need to move beyond false assumptions regarding biological
bases for political behavior and toward valid critiques of this approach. Not everyone agrees with
Hibbing. His essay is followed by eight that collectively present a variety of contrary views. McDermott
and Hatemi’s chapter presents research demonstrating that one’s genetic makeup does not solely
determine specific political behaviors or attitudes but rather interacts with environmental factors.
Lay, J. Celeste. A Midwestern Mosaic: Immigration and Political Socialization in Rural America.
Philadelphia: Temple University Press, 2012.
Hagerman, Margaret A. White Kids: Growing Up with Privilege in a Racially Divided America. New York:
New York University Press, 2018.
These two books explore adolescents’ views regarding race and diversity. Lay compares perspectives
regarding racial and ethnic diversity among teens in Iowa towns that were experiencing high levels of
immigration versus towns that were not. Hagerman focuses on the racial beliefs of affluent white teens.
McAdam, Doug. Freedom Summer. New York: Oxford University Press, 1988.
A fascinating study of college students who participated in the 1964 voter registration drive in Mississippi
that came to be known as Freedom Summer. Through interviews with many of the participants,
McAdam uncovers much about their motivations for volunteering on this project as well as the lifelong
effects this formative experience had on their political outlooks and behaviors.
Urbatsch, R. Families’ Values: How Parents, Siblings, and Children Affect Political Attitudes. Oxford:
Oxford University Press, 2014.
Families’ Values extends socialization research into new, or previously understudied, directions.
Urbatsch explores family socialization agents other than parents, such as brothers, sisters, spouses,
and children, finding that each can influence our political attitudes. He also studies a wide range of
political attitudes and behaviors, beyond the oft-studied topic of party identification.
“The Next America,” Pew Research Center, 2017, http://www.pewresearch.org/packages/the-next-
america
This section of the Pew Research Center’s website presents results from their analyses of generational
differences in the United States. Pew’s site contains loads of other information about Millennials and
Generation Z, with results from new surveys added often. Select “Millennials” or “Generation Z” under
their “Topics” link from Pew’s home page for the most recent additions.
GenForward Survey Project, University of Chicago, http://genforwardsurvey.com/
GenForward Surveys are conducted bimonthly, on a representative sample of eighteen- to thirty-four-
year-olds. As stated on their web page, their surveys are designed to pay “special attention to how race
and ethnicity shape how respondents experience and think about the world.” New reports, which
address a wide variety of political and social topics, are added throughout the year.
CIRCLE (The Center for Information and Research on Civic Learning and Engagement),
http://www.civicyouth.org
CIRCLE conducts and disseminates research related to the civic engagement of youth. Its website
contains a wealth of information, including research reports, tools for engaging in political and civic
activity, and links to other civic engagement sites.
http://www.pewresearch.org/packages/the-next-america
http://www.civicyouth.org/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists 1958, 2000, and 2017 to 2018. The vertical axis lists percentage and ranges
from 0 to 100 in increments of 10. Approximate data from the graph are tabulated as follows:
Rating 1958 2000 2017 to 2018
Very good 71 13 14
Fairly good 21 28 22
Not very good 4 20 19
Bad 1 28 36
Back to Figure
The horizontal axis lists characteristics. The vertical axis lists correspondence and ranges from 0 to 1.0
in increments of 0.1. Approximate data from the graph are tabulated as follows:
Characteristic Low correspondence value High correspondence value
Family politicization, 1965 0.44 0.63
Family politicization, 1997 0.2 0.72
Parental attitude stability, 1965 0.43 0.59
Parental attitude stability, 1997 0.2 0.42
Back to Figure
The horizontal axis lists periods and the corresponding age groups. The vertical axis lists correlations
and ranges from 0 to 1.0 in increments of 0.1. Approximate data from the graph are tabulated as
follows:
Period and the corresponding age
group All
Low preadult
correspondence
High preadult
correspondence
Period and the corresponding age
group All
Low preadult
correspondence
High preadult
correspondence
1965 to 1973, 18 to 26 years old 0.49 0.31 0.54
1973 to 1982, 26 to 35 years old 0.65 0.64 0.66
1982 to 1997, 35 to 50 years old 0.65 0.57 0.65
Back to Figure
The horizontal axis lists percentage of agreements and ranges from 0 to 100 in increments of 20. The
vertical axis lists various statements. Approximate data from the graph are tabulated as follows:
Statement
Percentage
of Generation
Zers in
agreement
Percentage
of
millennials
in
agreement
Percentage
of generation
Xers in
agreement
Percentage
of baby
boomers in
agreement
Racial inequality: Blacks are treated
less fairly than whites
63 60 54 49
Same-sex marriage: Gay and
lesbian couples being allowed to
marry is bad for society
15 15 25 30
Environment: Earth is getting
warmer mostly because of human
activity
55 58 48 45
Role of government: Government
should do more to solve problems
69 62 55 50
Gender categories: Forms asking
about a person’s gender should
include options other than man and
woman
49 50 40 38
Women in politics: More women
running for political office is good
for society
65 64 63 60
Back to Figure
The horizontal axis lists political opinions. The vertical axis lists proportions and ranges from 0 to 1 in
increments of 0.2. Approximate data from the graph are tabulated as follows:
Political
opinion
Proportion due to
genetic heritability
Proportion due to shared
environment
Proportion due to
unshared environment
Political
ideology
0.42 0.2 0.37
Political
opinionation
0.37 0.03 0.62
Party
identification
0.15 0.42 0.45
CHAPTER 3 MASS MEDIA
DRAMATIC TECHNOLOGICAL CHANGES in recent decades have had a major impact on how, when,
and whether citizens receive news. From the explosion of cable channels to the proliferation of
smartphones, citizens are able to get news when and how they want it, or they can choose to avoid
news altogether. Also, through social media citizens have unprecedented opportunities to shape and
make the news themselves; social media also provide malicious actors unprecedented opportunities to
spread fake news. Despite these extraordinary changes in the media environment, traditional media
outlets, such as network TV news and newspapers, remain influential, and powerful political elites and
institutions are still key drivers of news content.
We begin with a discussion of what we should expect from the news media in a democratic society.
Participatory democratic theorists have high expectations regarding the media’s role in providing citizens
with the information and tools they need to engage in the political process. Next, we examine general
characteristics of mass media ownership that shape news coverage. The media in the United States are
best described as corporate, concentrated, and conglomerate. We discuss the political implications of
those characteristics. Specific features of the traditional news media also influence how stories are
covered. We discuss the impact of news norms on the reporting of political events as well as whether
those norms have been undermined by technological changes that allow for niche news. We define fake
news and examine its role in politics today. We also grapple with the effects of media coverage on
citizens. We review how our understanding of media effects on public opinion has changed over time
and discuss the potential for citizens to use social media to influence the political process. Along the
way, we remind you that all media are not alike and illustrate how studying a variety of news media
provides added insight. We close by reviewing the evidence and considering whether the media play
their appropriate role in our democratic society.
WHAT SHOULD CITIZENS EXPECT FROM THE MASS MEDIA IN A
DEMOCRACY?
Before reading this section, stop and ask yourself: what do you expect from the mass media in a
democracy? How do you think the media should behave? Notice we are not asking how the media act in
reality. We will get to that in a moment. For now, we are interested in your ideal vision of how the media
should operate in a democratic society.
Probably the first thing that came to mind is that the media should be free from government control.1
Freedom of the press, of course, is a fundamental tenet in a democracy. The government should not
control the media by censoring stories or by forcing the publication of stories. Donald Trump, for
example, should not be able to stop the Washington Post from publishing a particular story, nor should
he be able to require the newspaper to cover certain events. In a democracy, a free press should also
be free from economic forces, such as market pressures and advertising dollars. Powerful economic
forces should not be able to prevent the media from covering important issues of the day and thus limit
debate.
A press free from governmental and economic control will have great benefits for citizens because such
a press should have several key characteristics. First, the media should act as an intermediary
between citizens and elites, providing both with the information essential for a well-functioning
democracy. Specifically, media organizations should provide citizens with the information and analysis
necessary to make smart decisions. Further, the media should cover how citizens think about issues so
that elites will be able to make educated decisions on behalf of their constituents. Second, a free press
should provide a forum for diverse views. Elite and citizen opinions from across the spectrum should
be presented so that debate can be wide ranging and critical or alternative voices will be heard. Finally,
the media should play a “watchdog” role.2 Because citizens cannot attend every city council meeting or
participate in every public hearing held by a federal agency, they rely on the media to scrutinize the
actions of public officials. The media should provide citizens with the information necessary to hold
government accountable and act as a check on the judicial, legislative, and congressional branches of
government. Implicit in this discussion is that the news media will convey accurate and relevant
information and will not let politicians’ and pundits’ misleading statements stand without challenge.3
This vision of the media is most consistent with the one held by participatory democratic theorists.
Participatory theorists, as you recall, want citizens to be actively engaged in the political process. By
providing accurate information, the media create knowledgeable citizens ready to participate in the give-
and-take of politics. Moreover, by presenting diverse viewpoints, the media ensure that both privileged
and marginalized voices will be heard. This facilitates participation on the part of all citizens (thus
redressing inequality in society, a core concern of participatory theorists). Participatory democratic
theorists believe that political participation makes people better citizens, and they see the potential for
the media to assist in that process.
Elite democrats, on the other hand, would be happy with a press relatively free from governmental
control. We say “relatively free” because these theorists might argue that there are circumstances under
which government officials should have influence over what the media publish. Take stories on national
security, for example. Elite democratic theorists might prefer censorship of such stories because they
think it is unnecessary for citizens to know the details about these policies. Elites are the decision
makers, so there is no need for the media to risk undermining national security by revealing information
to the public. Furthermore, elite democrats would be much less concerned with the effect of economic
forces on the media. From their perspective, media outlets should be responsive to market pressures
and advertising dollars so that citizens receive the kind of programming they want. As long as the media
provide enough basic information to citizens so they can vote, these theorists are not concerned with
the quality or diversity of news presented in the media.
This discussion of citizens’ expectations regarding the mass media largely presumes a separation
between citizens and the press. In today’s media environment, however, the line between citizens and
reporters, between news makers and news consumers, and between news and opinion is often blurred.
The Internet and social media provide “new, breathtakingly dynamic, and radically decentralized means
for people and organizations to communicate and cooperate with one another for political and civic
ends.”4 This gives some observers great hope that citizens will use these digital tools to enrich the
quality of democracy, whereas other observers caution that citizens are more likely to use these tools to
seek news from and communicate with like-minded individuals, leading to fragmentation and
polarization. Or perhaps even worse still, citizens will simply use the choices provided by new forms of
media to pursue entertainment and eschew news altogether.5 Another concern is that political elites and
citizens will use the Internet and social media to spread disinformation and propaganda. Facebook
announced in October 2019, for example, that it will not fact-check political advertisements on its social
network.6
WHAT GENERAL CHARACTERISTICS OF THE MASS MEDIA SHAPE
NEWS COVERAGE?
Let’s turn now to a discussion of the mass media in reality. Here we focus on general characteristics of
mass media ownership that influence news content. In the United States, the mass media can best be
characterized by the three Cs: corporate, concentrated, and conglomerate.7
The First C: Corporate
The first C is corporate because most media are owned by large companies. The primary goal of these
corporations is to make money for their owners and shareholders, not to serve the interests of citizens in
a democratic society. Because they are profit-driven companies, pleasing advertisers, not citizens, is
their main objective. And this concerns critics who argue that advertisers prefer stories that put people in
the mood to purchase their products rather than participate in politics. Moreover, in the world of
advertising, certain people are more valuable than others. Broadcast networks, for example, are most
interested in attracting viewers to their news programs who are between the ages of eighteen and forty-
nine, particularly women, who hold the purse strings in many households.8
Although most of you will not recall (but your grandparents will), citizens used to turn en masse to one of
three broadcast networks for their evening news: ABC, CBS, or NBC. That was it—the big three were
television news. That seems strange today with literally hundreds of available channels, several of which
are dedicated to news coverage. During that prior time period, these evening news broadcasts were
shielded from profit expectations. This changed in the mid-1980s as new owners took over ABC, CBS,
and NBC with an eye toward turning these evening news broadcasts into money-making enterprises.9
Around the same time, technological changes occurred that led to an explosion of cable channels,
resulting in a much more competitive environment. Not only were broadcast networks forced to compete
with cable news outlets (such as CNN, which debuted in 1980), but they also had to contend with
entertainment television (such as MTV, which debuted in 1981). Thus, just as the three broadcast
networks were asked to bring home the bacon, the competition became even stiffer.
Today the media environment is even more complex with traditional media, such as network television
news, newspapers, and radio, competing not only with cable news but also with social media, including
Facebook, YouTube, Twitter, and many other social networking sites. Critics raise concerns about how
social media companies use their access to user data to generate revenue from highly targeted
advertising. Privacy advocates shudder at the huge troves of data from its more than two billion users a
company like Facebook collects, analyzes, and shares (properly and improperly).10 Whether traditional
or social media, the goal of these companies is to make money.
The Pew Research Center conducted several surveys of citizens’ news habits between 1996 and 2012
that demonstrated the highly competitive and changing nature of the news business.11 To assess daily
news habits, Pew asked citizens whether they received news from particular sources “yesterday.” We
see that daily viewership of TV news (with no distinction made for local, nightly, or cable) did not decline
too much between 1996 and 2012 (see Figure 3.1), whereas radio news and especially print
newspapers saw significant declines in their daily usage.12 Americans increasingly obtained news from
the Internet and mobile devices. For instance, 39 percent of Americans in 2012 indicated they got news
yesterday from Internet or mobile sources, including cell phones, tablets, or other mobile handheld
devices.
In recent years the Pew Research Center has asked Americans how often they get their news from
various sources. In 2018, for the first time, more people said they got their news often from social media
(20 percent) than from print newspapers (16 percent). Thirty-three percent said they got their news often
from a news website or app, a 5 percentage point increase from 2016. Television remains the dominant
source for news with 49 percent of Americans getting their news often from local, national, or cable TV,
although this is a decline from 57 percent in 2016.13 These usage patterns demonstrate the competitive
pressures faced by news organizations.
A widely respected media scholar, Thomas Patterson, argues that traditional news organizations
“softened” their coverage in response to increased competition.14 In their rush to attract larger
audiences, the traditional media turned away from hard news coverage of important leaders and
policies in favor of soft news without a connection to public policy. To examine this supposition,
Patterson content analyzed a random sample of news stories from magazines, broadcast television
networks, national newspapers, and local papers. He found that soft news increased from roughly 35
percent of all stories during the early 1980s to almost 50 percent of all stories by 1999.15 Ironically,
Patterson argued that this increase in soft news coverage, intended to draw in news audiences, actually
drove people away from traditional platforms. For citizens who prefer hard news, the emphasis on soft
news is a turnoff; for those who favor soft news—presumably due to its entertainment value—there are
other, much more exciting options available that will draw their attention in the long run.
Description
Figure 3.1 Where Do Citizens Obtain News Daily?
Source: Data from “In Changing News Landscape, Even Television Is Vulnerable,” Pew Research Center,
Washington, DC, September 27, 2012, https://www.people-press.org/2012/09/27/in-changing-news-
landscape-even-television-is-vulnerable/
Although many scholars view the increase in soft news coverage as alarming, others argue that citizens
who turn to soft news media can actually learn about politics from those outlets.16 In particular,
politically inattentive citizens who have no desire to read the New York Times, for example, can still pick
up information as an “incidental by-product” of exposure to entertainment-oriented media, such as
daytime talk shows.17 Entertainment programs can also “provide deep and accessible insights into the
impacts of policies, the prevarications and real goals of public officials, and the distribution of wealth and
power.”18 In recent years Saturday Night Live’s skewering of Donald Trump comes to mind as an
example of such programming.
Jonathan Ladd argues that increased competition in the news industry resulting in tabloid-style news
and increased attacks on journalists by elites in both political parties has led to decreased confidence in
the press since the mid-twentieth century.19 For example, 23 percent of Americans said they had a
great deal of confidence in the press in 1973, whereas only 8 percent felt that way in 2016. Interestingly,
confidence bounced up to 13 percent in 2018.20 Perhaps the mainstream media’s award-winning
investigative coverage of the Trump administration stopped the decline in trust and restored at least a
few people’s faith in the institution.21 Overall, however, Ladd argues that this decline in trust is
problematic because the media are key to informing citizens about how well public policies are working
so they can hold political leaders accountable.
https://www.people-press.org/2012/09/27/in-changing-news-landscape-even-television-is-vulnerable/
The Second C: Concentrated
The second C is concentrated: a fairly small number of large companies own much of the media in the
United States. According to one study of thirteen media industries (newspapers, radio, television, search
engines, online news media, etc.), media ownership can best be characterized as “moderately
concentrated,” and concentration is increasing over time with levels varying across the different
industries. For example, there are high levels of concentration in online media and local radio, whereas
there are low levels in books and magazines.22 To illustrate concentration in local television news,
consider the CBS Corporation. CBS owns two television stations in each of these cities: Los Angeles,
New York, Philadelphia, Dallas-Ft. Worth, San Francisco, Boston, Detroit, Miami, Sacramento, and
Pittsburgh.23 Thus, what looks like a choice in news channels may actually be no choice at all.
Some media critics argue that media concentration affects news coverage and stifles discussion of
important issues in a democracy. Diverse news content is impossible if just a few companies control the
vast majority of the media, according to this perspective.24 To examine the effects of media ownership
on news content, Catie Bailard conducted an innovative study of the influence of media concentration
on coverage of the Supreme Court’s ruling in the Citizens United case.25 The Citizens United ruling had
significantly different implications for those media corporations that owned television stations than for
those that did not. Bailard exploited that fact to explore whether coverage of Citizens United differed
depending on television ownership.
The Citizens United decision established that independent political expenditures by corporations and
unions are constitutionally protected free speech. Any individual or group can donate as much money as
they please to groups that are independent of political parties and candidates, such as Super PACs (i.e.,
super political action committees). The groups then spend huge amounts of money on political
advertising, especially television advertising. Thus, media corporations that owned television stations
benefited immensely from the ruling, whereas other media corporations did not. As a result, Bailard
hypothesized that newspapers owned by media companies with television interests would cover the
Citizens United ruling more favorably than newspapers owned by companies that did not have
investments in television stations.
To test this hypothesis, Bailard identified newspapers owned by companies with no television stations, a
limited number of television stations, and a substantial number of television stations. Then she collected
439 articles that mentioned the Citizens United ruling published in those newspapers between March 1,
2009, and February 28, 2011. Next, she content analyzed whether the consequences of the ruling were
described in positive or negative terms.
Bailard found that newspaper coverage differed across the three types of media companies. For
example, 64 percent of articles in newspapers owned by companies with substantial television
ownership mentioned positive consequences of the Citizens United ruling, whereas only 47 percent of
newspapers with limited television stations and 45 percent of newspapers with no television stations did
so. She also found moderate differences in coverage of negative consequences of the ruling such that
newspapers with substantial television ownership were less negative than newspapers with parent
companies that did not stand to gain from increased television advertising.26
This pattern of news coverage exposed by Bailard raises serious questions about the effects of media
concentration on news content.27 The research has “important implications for the degree to which the
news produced by increasingly conglomerated and corporatized media companies may eschew neutral
or balanced coverage in favor of news frames that promote their own financial interests—a troubling
prospect for those who believe the news has a duty to provide citizens with a balanced and objective
perspective on the issues and policies facing the nation.”28
Gregory Martin and Joshua McCrain also examine the impact of media concentration on news
coverage.29 They take advantage of the fact that Sinclair Broadcasting, a company known for its
conservative political leanings, recently bought a number of local television stations. Martin and McCrain
compare news coverage on Sinclair-acquired stations with coverage on stations owned by other
companies within the same market area. They find that Sinclair-owned stations increased coverage of
national politics and provided more right-leaning content than prior to the acquisition relative to other
stations in the market. With this greater emphasis on national news, local television news stations are
less able to play a watchdog role, making it more difficult to hold local politicians accountable. In
addition, the greater conservative content suggests that Sinclair is trying to further its own political
agenda rather than report the news in a fair and objective fashion.
The Third C: Conglomerate
The third C is conglomerate. Not only do these large corporations own lots of different types of media,
they own lots of nonmedia companies as well. For example, the same company—the Walt Disney
Company—that airs ABC’s World News Tonight also owns eight television stations in major cities across
the country that “reach 22.5% of all U.S. television households” as well as ESPN, Hulu, theme parks
and resorts, movie studios, retail stores, and a variety of digital media.30 Many observers are concerned
with this conglomeration. The “willingness to criticize and scrutinize those with political and economic
power”31 is expected from a free press, yet do media conglomerates hold their own accountable? Media
critics from both ends of the ideological continuum worry about the effects of large, highly concentrated
media conglomerates. Liberals might be concerned about whether ABC news outlets will report on
environmental damage caused by Disney cruise ships, whereas conservatives might be concerned with
the impact of violence in Disney films on our culture. People across the ideological spectrum might
worry about whether ABC news outlets will report on the Walt Disney Company’s $3.5 million lobbying
efforts to influence the federal government on a range of issues, including radio and TV broadcasting,
trade, taxes, violence on television, and digital privacy.32
Exceptions to the Three Cs
Although corporate, concentrated, and conglomerate characterize much of the news media in the United
States, there are notable and important exceptions. National Public Radio (NPR) and the Public
Broadcasting System (PBS) are nonprofit media corporations whose goal is to provide high-quality,
noncommercial news and entertainment programming to American citizens. Both corporations are
membership organizations and receive significant financial support from their member stations (public
radio stations in the case of NPR and public television stations in the case of PBS). NPR and PBS are
also underwritten by corporations and private foundations. These underwriters do not air advertisements
in the traditional sense but do have their names mentioned during programming. A relatively small
portion of NPR’s and PBS’s annual budgets comes from federal funds.33 Conservative critics of NPR
and PBS regularly work to pass legislation to defund NPR and PBS. The Trump administration has
repeatedly called for ending funding to these news organizations.34 This is much to the dismay of
participatory democratic theorists because public radio and television stations play a crucial role in
providing citizens with information necessary to make educated judgments about the most important
issues of our day.35 Several studies demonstrate that citizens who obtain their news from PBS or NPR
are more informed about politics than citizens who get their news from commercial outlets (for more on
this topic, see Chapter 8).36 In a media environment dominated by conglomerates, noncommercial
outlets are an essential source of reliable and accurate hard news.
Another noteworthy exception to the dominance of media conglomerates is the minority press. Let’s
take the Black press as an example. Black newspapers tend to be independent or part of relatively small
family-owned chains.37 The Black press is affected by economic forces, but it is also driven by a mission
of advocacy. Its goal is to present the news from an angle that centers Black people, and it tries to
counter the often incomplete and erroneous coverage of minority communities that appears in the
mainstream media.38 As a result, the Black press provides a diversity of voices that are not heard in the
mainstream media.39 Black citizens find this particularly important, especially during periods of racial
strife.40 The circulation of Black papers also goes up during times of racial promise, as when Barack
Obama won the presidency.41 There are about 170 African American newspapers in the United
States.42 Just like mainstream newspapers, their circulation has decreased over time.43 Nevertheless,
by providing a forum for diverse points of view, Black newspapers act as an important, albeit small,
counterbalance to the nation’s major media outlets.
Other ethnic presses serve a similar function, and they are especially critical for citizens who prefer or
need news in languages other than English.44 The circulation for the top twenty Hispanic weekly and
semi-weekly newspapers has held steady since 2016.45 As Latinos become an increasingly important
electoral group, political scientists need to gain a better understanding of the content of Latino media as
well as its effect on public opinion. In general, much more research is needed on media outlets that
serve minority communities.
Finally, what about the Internet as an exception to the dominance of media conglomerates? Denis
McQuail, a noted media scholar, argues that the Internet “can unlock secrets and make much arcane
information readily available to ordinary people very rapidly, with an empowering effect.”46 The Internet
provides unprecedented opportunities for citizens to discuss and make the news. Blogs, for example,
turn readers into writers, and social media enable citizens to upload videos to YouTube, share political
content on Facebook, and tweet their own political commentary as well as retweet news stories and the
views of others.47 Furthermore, because few resources are required to set up a website, independent
media organizations can afford to establish a home on the Internet that would be impossible without the
technology.
The Internet is not a panacea, however.48 Citizens are much more likely to search the Internet for
entertainment news than for political news.49 Bloggers are not constrained by journalistic norms of
accuracy and objectivity (which we discuss later), and many have no qualms about reporting gossip as
fact or actively promoting falsehoods.50 Furthermore, many citizens get news from online sources that
share their point of view, undermining the potential of the Internet to provide a range of perspectives to
the public.51
Media conglomerates have also planted their flags in cyberspace. The same mega companies that
control traditional news outlets, as well as other companies, also garner significant traffic on their
Internet sites. These companies’ “brands” are so powerful that many citizens turn to them for
information; it just happens to be on the Internet rather than on television or in print. For example, a
study of tweets about immigration posted between January 20 and February 20, 2017, showed that links
to news articles from legacy media outlets—as the media organizations that existed before the Internet
are sometimes called—were the most common links shared.52 Nevertheless, the Internet provides
countless opportunities for average citizens, if they so desire, to seek out information from a range of
sources, including social media; reputable bloggers; independent media organizations, such as
ProPublica, the widely respected investigative journalism newsroom; and government, academic, and
nonprofit websites.
Box 3.1 Social Media in China
Public Opinion in Comparative Perspective
China is an authoritarian country. Its government censors information and abhors protest. One of
the authors of this book, Rosalee Clawson, traveled to China during the summer of 2019 on a
school-sponsored trip with her son, and they happened to tour Tiananmen Square on June 3rd,
the thirtieth anniversary of the government’s brutal crackdown on pro-democracy protests in the
square. Security was very tight with passport checks conducted by the military before they were
allowed to access the square. They received firm warnings from their tour guide not to unfurl any
school banners (or any banners at all) because the actions might be mistaken as starting a
protest. Surveillance cameras were ubiquitous on and around Tiananmen Square (as is the case
everywhere in China), and the government blocked even larger chunks of the Internet that day.
At the same time the Chinese government exerts this tight control over society, it also allows
over 1,300 social media companies and websites to operate in the country, and people post
millions of messages every day, including social media posts that criticize Chinese leaders and
the regime.1 Many observers believe the government pays people to post messages on social
media to “engage in ‘hand-to-hand’ verbal combat” rebutting those who are critical of the
government.2 Gary King, Jennifer Pan, and Margaret Roberts investigate this assumption by
systematically analyzing social media posts made by individuals paid by the Chinese
government. These posts were leaked from an Internet Propaganda Office. The leaked posts
provide valuable insights into the number, timing, and content of posts coming from people paid
by the regime.
King, Pan, and Roberts estimate that approximately 448 million posts a year come from
individuals paid by the Chinese government. That’s a lot of fabricated posts. The posts look like
they come from ordinary people, but they are actually made by government employees. In terms
of content, King, Pan, and Roberts find no evidence to support the conventional wisdom that paid
posts are challenging those who criticize the government, political leaders, or policies. Instead,
the vast majority (80 percent) are best characterized as “cheerleading.” These cheerleading
posts include such content as patriotic messages, inspirational quotes and slogans, thankfulness
for historical figures, and mentions of celebrations. Almost none of the paid posts engage in
arguments with others or comment on controversial issues.
These paid posts appear to be carefully coordinated by the government; there are bursts of
activity at certain critical times, such as around days that have historically been linked to protests
or when there are opportunities for protest, such as national holidays or events with large
crowds. Why does the Chinese government promote cheerleading messages during these
periods? King, Pan, and Roberts argue that the government uses these social media posts to
distract people from grievances and negative thinking. “Distraction is a clever and useful strategy
in information control” because it changes the subject and reduces anger.3 Distraction is more
effective at diffusing a situation than making arguments to directly challenge opposition views. As
a result, the cheerleading messages reduce the likelihood that people will participate in collective
action against the regime. This is why the paid posts are consciously timed to distract Chinese
citizens at the moments they are most susceptible to protest.
Overall then, the Chinese government doesn’t use social media to stifle criticism or pick fights
with ordinary people. Instead, the government posts celebratory messages intended to disarm
the Chinese people.
1 Gary King, Jennifer Pan, and Margaret E. Roberts, “How the Chinese Government Fabricates
Social Media Posts for Strategic Distraction, Not Engaged Argument,” American Political Science
Review 111 (2017): 484–501.
2 Ibid., 485.
3 Ibid., 496–497.
WHAT SPECIFIC CHARACTERISTICS OF THE TRADITIONAL NEWS
MEDIA SHAPE THE REPORTING OF POLITICAL EVENTS?
In this section, we discuss news norms that influence the reporting of political events by traditional
news organizations. Here we focus on norms that shape how journalists—not columnists or pundits—
decide what’s news, and we discuss which media organizations are constrained by these norms and
which aren’t. We also provide a critique of news norms.
News Norms
One of the most important norms that shapes news coverage is objectivity. In practice, journalists
define objectivity as providing both sides of an issue. To uphold this norm, journalists strive for balance
in their reporting. In the U.S. political context, this often means that a Republican viewpoint is balanced
with a Democratic one.53
A close cousin of objectivity is neutrality. According to this norm, journalists do not inject their personal
opinions into news coverage. Instead, they report on political events by presenting others’ viewpoints in
their stories, especially the viewpoints of official sources. Official sources include primarily government
officials but also other people who are powerful in society.54 By relying heavily on official sources,
journalists are able to do their jobs easily and efficiently. It also allows them to achieve the norm of
accuracy. Journalists work hard to ensure the information they report is correct. They perceive official
sources to be reliable, legitimate, and in the know; thus, journalists regularly turn to these sources in
their news coverage.55
Assigning journalists to newsbeats is another journalistic norm.56 Journalists are assigned to cover
specific institutions or topic areas. These are called beats. For example, major news organizations
assign journalists to the White House beat and the Pentagon beat. This allows journalists to gain
expertise on certain topics; develop relationships with key players, which is obviously important because
of the heavy reliance on official sources; and create familiar, reliable routines in a job where events are
constantly changing.57
Journalists are also influenced by norms related to newsworthiness. Conflict garners significant
attention because it is considered especially newsworthy. For example, media coverage of Congress
focuses heavily on partisan conflict within the institution and strife between the Congress and the
president.58 The emphasis on conflict is also obvious in news coverage of political campaigns. The
media tend to focus on the “horse race” aspect of campaigns: who’s ahead in the polls and who’s
behind, who has momentum and who doesn’t, and who’s leading in fundraising and who’s faltering.59
Even once politicians gain office, the news media still assess their every move in terms of competition
and gamesmanship rather than substance.60
These norms are a function of the traditional news media trying to reach as broad an audience as
possible as efficiently as possible. Many media corporations want to advertise to large audiences, not
just Republicans or Democrats. Thus, they provide objective news that will not tick off one side or the
other (or at least tick off both sides equally). This was not always the case; newspapers early in U.S.
history served the interests of powerful officials or political parties.61 Technological changes during the
mid-1800s, however, allowed publishers to print greater quantities of newspapers within a much shorter
time frame. This technological advance allowed for high circulation, which encouraged businesses to
own and advertise in newspapers. It became necessary for journalists to report the news in such a way
that newspapers would appeal to a wide audience in an efficient manner; thus, journalists adopted the
norms of objectivity, neutrality, accuracy, newsbeats, and newsworthiness.62
Do these news norms constrain the activities of all media organizations? No, they do not. Robert
Entman distinguishes media sources based on the extent to which they adhere to news norms.63 He
argues that traditional journalism (such as the New York Times or the CBS Evening News) has a
strong commitment to news norms, but advocacy journalism is committed to only some of these
norms. Advocacy journalism includes magazines with an ideological bent, such as The Nation on the left
and National Review on the right. These magazines strive for accuracy but have no interest in balancing
sources to follow the norm of objectivity. Tabloid journalism is much less committed, if at all, to news
norms. Tabloid journalism includes cable programs such as Tucker Carlson Tonight on Fox News, which
appeals to a conservative audience, and The Rachel Maddow Show on MSNBC, which speaks to a
liberal audience. Over the past three decades, the explosion of cable channels has allowed for profitable
niche programming. And because these programs do not need to appeal to a wide audience, the norms
of objectivity, neutrality, and accuracy have gone by the wayside. In fact, the draw of these cable news
programs tends to be the bombastic commentary of their hosts. The Internet, of course, has also
provided a platform on which tabloid journalism can thrive, and social media allow anyone to create
“news” without regard for journalistic norms of appropriate behavior.
Jeffrey Berry and Sarah Sobieraj refer to the venomous genre of political opinion media as the “outrage
industry.”64 They identify talk radio, cable television, and the blogosphere as key purveyors of outrage
and emphasize the economic incentives for programming that vilifies opponents, uses ideologically
extreme language, and engages in fear mongering. These media outlets produce outrage as their
product, which is highly profitable for both the corporate owners as well as the media personalities
involved, such as Rush Limbaugh (talk radio), Sean Hannity (cable), and Steve Bannon (blogosphere).
You’ll notice these three examples are from the political right. That is no accident. Berry and Sobieraj
argue that “outrage media are more abundant and mature” on the right, but they make clear the
phenomenon occurs on the left too, such as The Last Word with Lawrence O’Donnell.65
The outrage seeps into political discourse more generally as audience members rely on information
(and falsehoods) from these programs in their conversations with family, friends, and coworkers. This
makes it very difficult for constructive political conversations to occur among people who don’t share the
same worldview. As a result, the airing of the outrage media, unrestrained by traditional news norms,
results in negative outcomes for democratic deliberation.66
Even in traditional journalism, the commitment to news norms seems to be weakening. Research
suggests that the agenda of traditional news outlets is influenced by tabloid journalists and outraged
media personalities. For example, an analysis of the 2016 presidential election demonstrates that the
right-wing media were particularly adept at influencing the mainstream media to cover anti-Clinton
stories originating in outlets not committed to journalistic norms.67 In the summer of 2016, right-wing
websites were publicizing a book critical of Bill and Hillary Clinton. A review of the book with a headline
that insinuated Hillary Clinton used her position as secretary of state to steer donations to the Clinton
Foundation (“Cash Flowed to Clinton Foundation amid Russian Uranium Deal”) ran in the New York
Times. The article itself indicated there was no evidence to support this claim, but the story was shared
on social media and promoted by right-wing media as confirmation of wrongdoing. Overall, the
mainstream media coverage of both Clinton and Trump was negative during the 2016 presidential
campaign, but the media primarily focused on Clinton and scandal, whereas Trump’s coverage primarily
focused on his main campaign issue, immigration. This pattern of coverage is not what you would
expect from news organizations committed to norms of objectivity and accuracy.
Critiques of News Norms
Our discussion of news norms may have raised some concerns in your mind. On the one hand, these
news norms enable journalists to appeal to a wide audience, provide accurate information, and do their
work in an efficient manner. The norms also ensure that powerful political elites will be able to get their
messages out to the public, thus pleasing elite democratic theorists. On the other hand, some of these
news norms make it difficult, if not impossible, for the news media to live up to the ideal standards
proposed by participatory democratic theorists.
To begin, the norm of objectivity requires journalists to present two sides of an issue. But what if there
are more than two sides? Take abortion, for example. The debate is often characterized in the media as
pro-life versus pro-choice, with Republican elites supporting life and Democratic elites supporting
choice. Among the public, however, 41 percent of Americans do not fall neatly into either camp. Instead,
they believe that abortion should be available under certain circumstances, such as rape, incest, danger
to the life of the mother, or when some other clear need has been established. Furthermore, another 45
percent say that a woman should always be able to obtain an abortion as a matter of choice. Because
journalists rely so heavily on official, partisan sources, the abortion debate looks as if it is pro-life versus
pro-choice, yet only 14 percent of Americans believe abortion should never be permitted.68 If journalists
paid more attention to the opinions of average Americans, the news coverage would focus on the
circumstances under which abortion should be allowed rather than whether it should be allowed. For
many citizens, the issue is more complex and nuanced than simply for or against, which suggests there
might be room for compromise—something you would never know from media coverage of the issue.
We have discussed the limitations of objective reporting when there are more than two sides to an
issue, but what about when elites are in agreement and thus there is only one side of an issue? In those
circumstances, one of two things happens. Either important issues simply go unreported by the media
because there is no conflict to draw the attention of journalists, or, if the issue does get covered, it
appears as if there is no debate on the topic even though plenty of debate may be occurring among
those who are not powerful enough to be included as official sources. In matters of foreign policy and
national security especially, elites regularly stake out uniform positions.69 Thus, on some of the most
important issues of our day—war and terrorism—elites often present a united front, which leads the
press to act more as a tool of government than a watchdog. Indeed, in 2004, both the New York Times
and the Washington Post expressed regret for not publishing stories that challenged the George W.
Bush administration’s justifications for going to war in Iraq.70
The assignment of journalists to newsbeats results in journalists developing close relationships with
government officials on those beats, leading to the concern that journalists may become too cozy with
those officials. From this perspective, journalists act more like lapdogs than watchdogs. Newsbeats can
also lead to pack journalism as journalists assigned to the same beat end up covering the same set of
stories from the same perspective. Further, newsworthy events may be happening that do not get
reported because there is no journalist assigned to that newsbeat.71 For example, news organizations
regularly assign journalists to cover the Pentagon. Those journalists become familiar with weapons
systems, military buildups, and troop deployments; they attend press conferences; and they cultivate
key sources. Not surprisingly, stories emanating from the Pentagon are regularly featured in news
coverage. Reporters are not assigned to cover the Department of Veterans Affairs as a beat, however.
As a result, significant stories may be overlooked. An important story about the shoddy treatment of
wounded soldiers at Walter Reed Army Medical Center was missed for months, if not years.72 And, of
course, some stories may never be reported.
Finally, the norm of newsworthiness leads journalists to favor coverage of conflict and the strategic
aspects of political campaigns and governance over substantive issues. This is highly problematic for
those citizens who want journalists to cover the issues so they can evaluate whether political leaders are
addressing their problems.73 It is also troubling for participatory democratic theorists, who see a sharp
disconnect between what information journalists deem newsworthy and what information citizens need
to function effectively in a democratic society.
WHAT ABOUT FAKE NEWS?
Another issue of great concern to democratic theorists is fake news. For democracy to function, citizens
must be able to obtain factual information from the media. Information is fundamental to self-
governance. Concerns about fake news have increased exponentially since the 2016 presidential
election. A Russian disinformation campaign coupled with a winning presidential candidate whose
statements were judged by an independent fact checker to be false 70 percent of the time raises many
red flags in a democratic society.74
One study of fake news during the 2016 election focused on news coming from outlets that do not follow
traditional news norms and therefore have questionable editorial processes.75 The goal of the research
was to understand exposure to stories distributed by fake news outlets. This research connected Twitter
accounts and voter registration records to analyze the posts of 16,442 U.S. voters active on Twitter
during the campaign season. The study showed that a very small proportion of voters (1 percent) were
exposed to the vast majority (80 percent) of fake news stories and an even smaller proportion (0.1
percent) accounted for most (nearly 80 percent) of the sharing of information from fake news sources.
Conservatives and older individuals were more likely to be exposed to and spread fake news. The
researchers concluded that the circulation of fake news was “extremely concentrated” and that voters
were mostly exposed to news stories from legitimate news outlets.76
Another recent study compared the diffusion of true and false news stories on Twitter between 2006 and
2017.77 Here the stories were verified to be true or false by independent fact-checking organizations.
The research showed that false news spread on Twitter “significantly farther, faster, deeper, and more
broadly than truth,” and that humans rather than bots were doing the dissemination. The false news was
more novel than the true news, meaning that the topic tended to be different from the topics of other
tweets during the same time period, which may explain why the false news stories got people’s attention
and they retweeted them.
These two studies may seem contradictory, but it’s important to keep in mind that the first study
examined news coming from fake news outlets, whereas the second focused on false news stories as
identified by fact checkers. Together these studies suggest that relatively few people share or are
exposed to articles from fake news outlets, yet when those false articles are circulated, they can quickly
go viral.78 This research is especially troubling to participatory democratic theorists who worry that the
spread of false news stories interferes with citizens’ ability to be informed about political life and
undermines citizens’ efforts to have fact-based discussions with their fellow Americans.
Scholars, politicians, journalists, and citizens use the phrase “fake news” to refer to lots of different
things. To help get a handle on what fake news is, Jana Egelhofer and Sophie Lecheler make a
distinction between fake news as genre and fake news as label.79 Fake news as genre refers to
“deliberately created pseudo-journalistic disinformation,” whereas fake news as label refers to a “political
instrument to delegitimize news media.”80 Let’s discuss each of these in turn.
Egelhofer and Lecheler argue that political communication should only be deemed fake news if it has all
three of these characteristics: low facticity (meaning messages that are fully or partially untrue), created
with the intent to deceive, and presented in a journalistic format. For example, a website or newspaper
article that intentionally includes false information presented in a news format is fake news. In contrast,
a news story by a reporter who makes an honest error in reporting might be deemed poor journalism,
but it would not be considered fake news. A news parody site, such as The Onion, presents articles in a
journalistic fashion, but distorts facts for amusement, not to deceive, so it would not be classified as fake
news.81
Identifying whether information is presented in a journalistic fashion is fairly straightforward, and
journalists have developed reputable fact-checking procedures to determine facticity. It is more difficult
to assess the intention to deceive. Nevertheless, Egelhofer and Lecheler argue that the motivation
behind news is important because it helps distinguish between misinformation that is disseminated
unintentionally and disinformation that is purposely circulated.82 People exposed to fake news as genre
are more likely to have misperceptions about politics that then have an impact on their policy attitudes
and political beliefs. We explore this topic further in Chapter 8.
Turning to fake news as label, Egelhofer and Lecheler argue that some political actors have weaponized
the phrase to attack journalists and news outlets for reporting stories that challenge their messages.83
Political leaders, activists, and average citizens now call news they don’t like fake news. It’s a way to
summarily dismiss content they disagree with rather than explain why the news story is problematic.
This labeling can delegitimize the news media, undercut journalists’ ability to hold politicians
accountable, and in some cases, even pose a threat to journalists’ safety. Attacking the legitimacy of the
media reduces trust in the media, resulting in a “post-truth” era in which some citizens have a difficult
time sorting out who is lying and who isn’t. This raises fundamental and deeply troubling questions
about how citizens can operate effectively in a democratic society.84
ARE CITIZENS AFFECTED BY THE MASS MEDIA?
In this chapter, we have discussed what citizens should expect from the media in a democracy, and we
have addressed the empirical reality of the media in the United States. By now, you should have a good
feel for the general and specific characteristics of the media that shape news coverage. The question
remains, however, whether citizens are affected by the mass media. That is the topic we turn to in this
section.
The Hypodermic Model
Imagine you are going to the doctor to receive your annual flu shut. The doctor uses a hypodermic
needle to inject you with the vaccine. You leave her office with the medicine coursing through your veins
ready to fight off any flu bug that might come your way. Receiving a shot in a doctor’s office is an (often
unpleasant) experience to which we can all relate.
Now, let’s translate this phenomenon to the political arena. Take yourself back in time to the Great
Depression. Imagine you and your family sitting in your living room listening attentively to one of
President Franklin D. Roosevelt’s “fireside chats” on the radio. Or imagine you are in Munich, Germany,
during roughly the same period. Picture yourself reading newspapers over which Adolf Hitler has
exerted complete control. Are you being injected with messages from the mass media in the same way
a doctor injects you with medicine? Are the media messages so powerful and you so weak that
resistance is futile?
Bring yourself to the present and take a look around. Why are so many of your friends (and maybe even
you) wearing Under Armour apparel or carrying iPhones with the name brands prominently displayed?
Is it possible that the advertising campaigns of these companies have “injected” your friends with their
messages, thus getting them to buy overpriced products and provide free advertising for the company
all at the same time?
These examples illustrate what has been called the hypodermic model of media effects.85 The early to
mid-1900s saw a huge growth in advertising, numerous technological changes that allowed average
citizens access to the media, two world wars, the rise of dictators across Europe, and a powerful
president at home. All these factors led some observers to fear that the media could control citizens.
Underlying this fear were two assumptions: (1) that the media are extremely powerful and (2) that
citizens are not sophisticated enough to ward off media messages. The hypodermic model is certainly a
compelling metaphor, but is there evidence to support its view of media effects on citizens? It turns out
that systematic support for such wide-ranging, persuasive effects of the media never panned out.
Instead, scholars came to the conclusion that the media have relatively minimal effects on citizens’
political attitudes.
Minimal Effects Model
Whereas the hypodermic model viewed citizens as blank slates waiting to be written on by the mass
media, the minimal effects model of media influence has a more nuanced understanding of citizens.
From the minimal effects perspective, citizens’ slates are already marked up with a whole host of prior
attitudes and predispositions when they encounter media messages. Citizens rely on these existing
attitudes to help them sift through, evaluate, and often filter out media content. Thus, citizens are not
passively injected with messages from the media; instead, they are active receivers or rejecters of these
messages depending on their predispositions. As a result, the media have minimal effects on citizens’
political attitudes.
Evidence of the media’s minimal effects was provided by Paul Lazarsfeld, Bernard Berelson, and Hazel
Gaudet in their classic study of Erie County, Ohio, during the 1940 presidential campaign between
Franklin Roosevelt and Wendell Wilkie.86 Lazarsfeld et al. trained local interviewers to conduct in-home
interviews with a representative sample of Erie County residents between May and November 1940. To
be precise, this panel survey included six hundred people, each of whom was interviewed six times over
the course of the campaign. Thus, the design of the study allowed the researchers to track residents
over time to determine why people voted the way they did in November. In particular, Lazarsfeld and
colleagues were interested in the influence of campaign messages on citizens’ vote choices. By
studying one community in depth, the scholars were able to assess the campaign messages that were
circulating in the local media environment and examine what effect, if any, those messages had on
voters. In this way, Lazarsfeld et al.’s research provides evidence that allows us to assess the power of
the media to influence citizens’ political attitudes.
Lazarsfeld et al.’s research findings are striking. First, they discovered that a remarkable 50 percent of
citizens already knew in May for whom they were going to vote in November. Obviously, the campaigns’
media messages were not changing people’s choices because they had already made up their minds
before the campaign even got started. Nevertheless, Lazarsfeld et al. argued that political
communication still played an important role because it reinforced people’s existing decisions. Hence,
this was labeled the reinforcement effect.
Lazarsfeld et al. also identified an activation effect among those people who were initially undecided
about which candidate to support. The researchers demonstrated that campaign messages aroused
interest in citizens, which led them to pay more attention to the election; however, the fascinating thing
was that citizens did not pay attention to all aspects of the campaigns. Instead, citizens honed in on
particular magazine articles and newspaper stories that corresponded with their political predispositions.
In other words, citizens with Republican-leaning characteristics were more likely to seek out Republican-
leaning campaign news, whereas Democratic-inclined citizens sought out pro-Democratic media
content. This selective attention to the media activated citizens’ prior attitudes, which served to remind
citizens why they held those attitudes in the first place.87 Thus, citizens’ latent predispositions were
stimulated and strengthened by the news stories. Rarely were those predispositions challenged, and
when they were, citizens were anchored by their predispositions and therefore resistant to change. By
November, citizens’ preexisting attitudes became crystallized, encouraging them to vote for the
presidential candidate who was consistent with their values and predispositions all along.
Finally, Lazarsfeld and colleagues found little evidence of a conversion effect. In other words, very few
citizens actually changed from one candidate to another during the course of the campaign. We might
expect that citizens who had few existing attitudes would be susceptible to campaign messages and
thus to conversion; however, those same citizens who did not have strong predispositions also did not
expose themselves to campaign news. In other words, those most likely to be persuaded were the least
likely to come across the persuasive messages. Conversion, then, was a rare phenomenon.
Lazarsfeld et al. also argued that the media’s influence was limited because many citizens relied on
conversations with politically engaged friends and family, rather than the mass media, to obtain
information about the presidential campaign. The researchers described the process as a “two-step
flow of communication.”88 First, highly interested citizens would gather campaign information from
newspapers and the radio. These people were called “opinion leaders.”89 Second, the opinion leaders
would talk about the election with their friends and family, passing on information about candidates and
issues to those who were much less caught up in the campaign. Therefore, Lazarsfeld et al. did not
dismiss the influence of the media entirely because clearly the opinion leaders were gathering
information from news organizations, but they did emphasize that personal contacts were more
influential for most everyday, average citizens.
Subtle Effects Model
When scholars did not find evidence that the media had widespread persuasion effects, many lost
interest in studying the influence (or lack thereof) of the media. Research in this area was dormant for
quite a while. Maxwell McCombs and Donald Shaw reversed that trend, however, with their research on
the agenda-setting role of the media in the 1968 presidential election.90 Their study marks the beginning
of the subtle effects model era of media research.
Agenda Setting
McCombs and Shaw acknowledged that the media cannot change people’s minds on the issues of the
day, but they argued the media can set the agenda by influencing which issues the public perceives as
important. Their study was based on a comparison between what a random sample of Chapel Hill, North
Carolina, voters said were the key issues in the 1968 presidential election and the actual campaign
coverage in the news media relied on by voters in that community. McCombs and Shaw found a strong
relationship between the issues emphasized by the media and those issues deemed important by the
voters. For example, the mass media devoted a significant amount of coverage to foreign policy and
law-and-order issues, and Chapel Hill voters indicated those topics were major campaign issues. Thus,
the media set the agenda by establishing which campaign issues are considered important in the minds
of voters. The media “may not be successful much of the time in telling people what to think, but it is
stunningly successful in telling its readers what to think about.”91
McCombs and Shaw’s research breathed new life into the study of media effects and spurred a new
generation of scholars to further investigate agenda-setting effects.92 In 1982, Shanto Iyengar, Mark
Peters, and Donald Kinder, for example, tackled a significant question left unanswered by McCombs
and Shaw’s research. Because McCombs and Shaw’s conclusions were based on comparing
aggregated cross-sectional survey data to media content, they were not able to demonstrate that media
coverage caused voters to consider certain issues more important than others. To explain further,
McCombs and Shaw had surveyed Chapel Hill voters at one time (therefore collecting what scientists
call cross-sectional data) and then lumped them all together (meaning they aggregated the voters) to
compare what voters as a group indicated were their campaign priorities with what issues were covered
by the media. Thus, their research was not able to establish that the news coverage caused individual
voters to consider particular issues important. Why not? Because perhaps it was the case that the
media simply reflected the priorities of the voters. The media might have anticipated the interests of
voters and therefore covered those issues they thought would draw the largest audience. Figure 3.2
illustrates this conundrum. Thus, the question remains: do the media cause voters to think certain issues
are important, or do voters think certain issues are important and the media cover those issues as a
result?
To untangle this causal relationship, Iyengar, Peters, and Kinder conducted an agenda-setting
experiment using citizens of New Haven, Connecticut, as subjects.93 By paying subjects $20 to
participate in their experiment, they were able to recruit a group of participants who mirrored the
characteristics of New Haven citizens. Six days in a row during November 1980, subjects reported to an
office at Yale University that had been transformed into a casual setting for television viewing. The
researchers “encouraged participants to watch the news just as they did at home.”94 In this way, they
tried to make the context as natural as possible to ensure their results could be generalized beyond the
experimental setting. Therefore, by recruiting a mix of people to participate in their study and creating a
comfortable setting for watching the news, Iyengar et al. took steps to ensure their experiment was high
in external validity.
Figure 3.2 Sorting Out Causal Relationships
When the subjects arrived on the first day, they were asked to complete a questionnaire on political
topics. Embedded in this survey were questions that asked subjects to rate the importance of several
national problems. Over the next four days, subjects watched videotapes of the prior evening’s network
newscast, or so they thought. On the last day, subjects completed another questionnaire that repeated
the problem-importance questions.
Now, there are three crucial details here. First, the newscasts were not truly from the night before.
Instead, the experimenters created newscasts based partially on what had been shown the night before
but with specific types of stories either added or deleted. Second, the experimenters created two
different versions of the newscast. In one version, stories describing problems with U.S. defense
capabilities were inserted in the middle of the broadcast, whereas no such stories were included in the
other version. Thus, the researchers had complete control over the characteristics of the experimental
treatment (the newscasts). And third, subjects were randomly assigned to view either the newscasts that
emphasized weaknesses in U.S. military preparedness or the newscasts that did not mention the issue.
In other words, it was chance alone that determined whether subjects saw the defense-related news
stories or whether they saw newscasts without those stories. As a result of this random assignment, the
subjects in the two conditions were essentially the same. Overall, then, this process of random
assignment of subjects to conditions and experimenter control over the treatment ensured that the only
difference between the two groups was that one viewed newscasts with the defense stories and the
other did not. Thus, if the subjects in the two conditions expressed different opinions on the final
questionnaire, we know that it is due to the experimental treatment because all other factors were held
constant.
And, indeed, this was just the case. Subjects who viewed the newscasts emphasizing the problems with
U.S. military preparedness changed their opinions and rated defense issues as much more important in
the postexperiment questionnaire than in the initial questionnaire. Before viewing the newscasts, the
subjects ranked defense as the sixth most important out of eight problems. After watching the
newscasts, defense jumped to the second most important problem. Furthermore, their attitudes on the
importance of other issues did not change, and subjects in the control condition did not change their
ranking of the importance of defense as a national problem. In sum, the media coverage affected what
people saw as important, thereby setting the agenda.
Priming
In addition to studying agenda setting, Iyengar and colleagues examined media priming in their
experiments.95 The researchers hypothesized a priming effect whereby the issues emphasized by the
media would become the same issues citizens used to evaluate political leaders. Thus, media coverage
sets the standards by which politicians are judged. For example, if the media cover defense topics, then
a president’s performance on that issue becomes a salient factor shaping opinion toward the president
in general. This is exactly what they found. After viewing stories on the inadequacies in the defense
system, subjects’ views of President Jimmy Carter on that issue were a stronger predictor of their overall
evaluation of Carter than in the condition in which subjects did not see stories on defense. In other
words, the defense stories primed citizens to evaluate the president along those lines.
In addition to priming issues, the media can also highlight particular traits, such as experience or
competence, on which citizens will evaluate political leaders. Recall our discussion of the media’s
emphasis on scandal-related coverage of Hillary Clinton during the 2016 presidential election. The
media coverage raised questions about her honesty and integrity, and at least some of the stories
implied she was corrupt. Therefore, honesty and integrity were made particularly salient to citizens,
which likely led them to weigh those traits more heavily in their evaluation of her than if they had been
exposed to media coverage focused on positive traits or on political issues. By calling attention to
scandals, the media primed citizens to judge Clinton accordingly.
Framing
In addition to agenda setting and priming, scholars have also identified media framing effects. Framing
is defined as “the process by which a communication source, such as a news organization, defines and
constructs a political issue or public controversy.”96 Media frames identify which aspects of a problem
are relevant and important, and they imply which characteristics of a problem are not significant. They
also influence which aspects of a story are remembered.97 A framing effect occurs when media frames
influence public opinion on the issue being framed. To illustrate frames and framing effects, we first turn
to another classic study conducted by Shanto Iyengar.98
Iyengar examined television coverage of poverty, concluding that news stories were framed in either
episodic or thematic terms. Episodic frames focused on individual poor people, whereas thematic
frames emphasized poverty as a societal problem. For example, an episodic story on poverty might
focus on a young single mother who is trying to make ends meet after losing her job. In contrast, a
thematic story might discuss the nation’s poverty rate. Obviously the topic of both stories is poverty, but
one focuses your attention on the characteristics of the poor person, whereas the other leads you to
think about poverty as a problem faced by the country as a whole. Iyengar found that the episodic frame
dominated news coverage during the early to mid-1980s—two-thirds of the stories on poverty were
framed in terms of particular victims of poverty.
Do these media frames influence public opinion? Iyengar answered this question by conducting an
experiment to test whether the different frames influenced how people assign responsibility for poverty.
Subjects were randomly assigned to view either a thematic or episodic story on poverty. Those exposed
to the episodic frame were significantly more likely to hold individuals responsible for causing and
solving their own poverty and less likely to point to societal factors. In contrast, when the coverage
emphasized the general phenomena of poverty, citizens were more likely to point to societal causes and
solutions and less likely to hold individuals responsible for their poverty. Ironically, media coverage that
highlights individual people and their plight leads citizens to point the finger of blame at the poor
themselves. Overall, then, the dominance of the episodic frame in media coverage of poverty has clear
implications for how citizens think about the issue.99
Thomas Nelson, Rosalee Clawson, and Zoe Oxley also examined media framing effects, but they took
the research a step further by specifying the psychological mechanism that leads to such effects.100
These scholars studied media coverage of a Ku Klux Klan (KKK) rally held in Chillicothe, Ohio. They
identified two frames used by local television news stations to cover the event: free speech and public
order. A newscast using the free speech frame emphasized the right of the KKK to speak and included
images of KKK leaders speaking before a microphone. Several Klan supporters were interviewed and
said they wanted to hear the KKK’s message. One man said, “I came down here to hear what they have
to say and I think I should be able to listen if I want to.”101 In contrast, a newscast with the public order
frame focused on the possibility that violence would erupt at the rally between protestors and the KKK.
The news story included images of police officers standing between the Klan members and the
protesters. A bystander who was interviewed said, “Here you have a potential for some real sparks in
the crowd.”102
To examine what impact these frames had on tolerance for the KKK, Nelson et al. conducted an
experiment using actual news coverage of the rally. They recruited college students enrolled in
introductory political science courses to participate in the experiment. These subjects were randomly
assigned to view either the free speech frame or the public order frame and then were asked to
complete a survey that included a variety of questions, including two measuring tolerance for the KKK.
The first question asked, “Do you support or oppose allowing members of the Ku Klux Klan to hold
public rallies in our city?” The second asked, “Do you support or oppose allowing members of the Ku
Klux Klan to make a speech in our city?”103 Subjects responded on 7-point scales ranging from strongly
oppose (1) to strongly support (7). Those exposed to the free speech frame were significantly more
likely to support the KKK’s right to rally and speak than those in the public order condition (see Figure
3.3). The free speech frame increased political tolerance for the KKK by more than one-half of a point
on a 7-point scale. This is both a statistically and substantively significant increase in support for the
KKK’s right to participate in the public arena.
Description
Figure 3.3 Political Tolerance by Framing Condition
Source: Data from Thomas E. Nelson, Rosalee A. Clawson, and Zoe M. Oxley, “Media Framing of a Civil
Liberties Conflict and Its Effect on Tolerance,” American Political Science Review 91 (1997): 572.
Note: Higher numbers on 7-point scales indicate greater tolerance.
Nelson et al. also collected data in their experiment to understand the psychological mechanism leading
to these framing effects. Previous scholars had hypothesized an accessibility model to explain why
priming and framing effects occur. This perspective emphasizes that citizens are limited information
processors operating in a complex political world. Because there is no way people can deal with all the
information in their environment, they make judgments based on the most readily available
considerations. The political context, such as news frames, makes certain concepts more accessible
than others. In turn, these accessible concepts influence how citizens evaluate the issue that is being
framed. For example, the free speech frame makes concepts such as freedom and liberty accessible.
Thus, when citizens are asked whether the KKK should be allowed to rally after exposure to the free
speech frame, freedom and liberty are uppermost in their minds. These accessible concepts encourage
citizens to support the KKK’s rights. At least that is the mechanism according to the proponents of the
accessibility model.
Nelson, Clawson, and Oxley, however, suggest an importance model instead. They argue that not all
equally accessible concepts have an equal effect on political evaluations. In other words, just because
freedom and liberty are accessible does not mean they will automatically influence citizen judgment.
Nelson et al. propose a more thoughtful model of information processing, which says that citizens will
judge some accessible concepts more important than others. And those important or relevant concepts
will be the ones that influence opinion.
To test these competing hypotheses, Nelson et al. randomly assigned subjects to either an accessibility
or importance condition. In the accessibility condition, subjects were asked to respond to series of letter
strings flashed on their computer screens. Subjects had to indicate whether each letter string was a
word or a nonword. The task included words made accessible by the free speech frame (such as
freedom, liberty, independence, and rights) and by the public order frame (such as violence, disorder,
danger, and disturbance). How quickly subjects responded to the words indicates the accessibility of the
words. This reaction time task is a standard method that psychologists use to measure accessibility.104
Nelson et al. found that, regardless of the framing condition, the public order and free speech concepts
were equally accessible. Thus, differences in accessibility could not explain why subjects were more
tolerant in the free speech framing condition than in the public order framing condition.
Nelson and colleagues provide evidence, however, that the importance model explains how framing
effects occur. In the importance condition, subjects were asked to evaluate the importance of certain
values related to free speech and public order. For example, subjects were asked to indicate “how
IMPORTANT each of these ideas is to you when you think about the question of whether or not the Ku
Klux Klan should be allowed to make speeches and hold demonstrations in public”: “Freedom of speech
for all citizens is a fundamental American right.” and “There is always a risk of violence and danger at
Ku Klux Klan rallies.”105 The researchers found that public order values were deemed significantly more
important after exposure to the public order frame and that free speech values were viewed as slightly
more important in the free speech framing condition. As a result, these important values were weighted
more heavily when determining support for the KKK’s right to participate. In sum, frames influence which
values citizens view as most important to the matter at hand, which leads to changes in public opinion
regarding the issue.106
This research offers a more redeeming view of citizens. Rather than being buffeted around willy-nilly by
whichever considerations are made most salient by a media frame, as suggested by the accessibility
model, citizens engage in a more thoughtful process of weighing the importance of certain values as
they form their opinions.107
Much work on framing effects has been done using experimental methods. Experiments, of course, are
wonderful tools for testing causal hypotheses. There is also a lot of research on the content of media
frames using content analysis techniques. Two recent studies focused on visual images and media
frames. For example, one study examined political cartoons in the era of Black Lives Matter (BLM).108
Anish Vanaik, Dwaine Jengelley, and Rolfe Peterson analyzed over five hundred cartoons by award-
winning cartoonists published between August 2012 and January 2016. They compared the pre-BLM
movement cartoons about racism to the cartoons published after the movement began. They mark the
start of the BLM movement as August 2014 when Michael Brown, an 18-year-old African American man,
was shot and killed by a police officer in Ferguson, Missouri. Vanaik et al. demonstrated that cartooning
on racism increased after the killing of Brown. Furthermore, they showed that a “police frame” increased
from 6 percent to 33 percent of all cartoons on racism in the BLM era. This police frame included
imagery portraying “the police as quick to use excessive force in their encounters with Black citizens,
increasingly militarized, and overwhelmingly white, and lacking diversity.”109 This police frame reflected
the emphasis of the BLM activists.
Another study analyzed the visual images associated with stories on immigration in news magazines.
Emily Farris and Heather Silber Mohamed examined immigration pictures in Time, Newsweek, and U.S.
News and World Report between 2000 and 2010.110 They found that news magazines frame
immigration as a threat by disproportionately portraying undocumented immigrants and focusing
significant attention on the border and border enforcement. Photos, for example, showed men in
handcuffs and border patrol agents detaining men. These studies provide important insights into how
critical issues of our day are framed in the mass media.111
In sum, agenda setting, priming, and framing constitute what are known as subtle media effects.
Researchers in this tradition have not found the widespread persuasion effects suggested by the
hypodermic model, nor is their evidence consistent with the minimal effects model. Instead, researchers
have shown how the media can influence public opinion by (1) affecting what the public thinks about, (2)
affecting which issues and traits shape evaluations of leaders, and (3) affecting which considerations
are viewed as most important when assessing a political issue. It can be challenging to keep these
different subtle effects straight, so Table 3.1 provides a brief description and example of each.
Table 3.1 Making Sense of Subtle Effects
Subtle
Effect Description Example
Agenda-
setting
effect
The media emphasize certain issues, and as
a result, people perceive those same issues
to be most important.
The media give a lot of attention to
immigration so people believe that issue is
important.
Priming
effect
The media emphasize certain issues or
traits, and as a result, those issues or traits
are most salient as people evaluate political
leaders.
The media give a lot of attention to
immigration so people evaluate Donald
Trump based on his handling of that issue.
Framing
effect
The media construct issues by identifying
which aspects of a problem are most
relevant and important, and as a result,
people weigh those aspects more heavily as
they form their opinions on an issue.
The media construct immigration as a
threat by focusing on immigrants as
criminals and thugs so people weigh those
considerations heavily in their attitudes
toward the issue of immigration.
Limits on Subtle Effects
Are there limits on subtle media effects? Can the media set the agenda to such an extent that we would
consider them to be controlling the agenda? Can the media prime issues so much they overwhelmingly
determine how candidates and politicians will be judged? Can the media frame political issues and
therefore manipulate public opinion? These are important questions that emerge out of the research on
subtle effects.
James Druckman moves us toward answering these questions by examining whether there are limits on
framing effects.112 Specifically, he asks, who can successfully frame an issue? He argues that citizens
look to trusted, credible elites for guidance when determining their issue positions. He demonstrates that
credible communication sources effectively frame public opinion, whereas less credible sources are not
able to do so. As with Nelson, Clawson, and Oxley’s research on framing effects, Druckman provides a
more redeeming view of citizens. His study shows that citizens are not simply the victims of
manipulation on the part of elites; instead, citizens react to cues that make sense—whether the elite is a
credible source. Druckman and his colleagues also demonstrate that citizens respond to partisan
cues.113 They find that when citizens believe the issue at hand is one polarized along party lines,
citizens will more likely follow a frame endorsed by their party even when that frame is a “weaker” one.
Additional work by Dennis Chong and Druckman reminds scholars that the real world of politics may put
limits on framing not apparent in most research on framing effects.114 In actual political debates, frames
rarely go uncontested. Competing frames are part and parcel of political discourse. In many framing
studies, however, participants are exposed to one-sided messages and then almost immediately asked
their opinions on the issue at hand. Perhaps not surprisingly, researchers find that the messages shape
public opinion. But what happens when citizens are faced with competing frames over the course of a
policy debate or political campaign that might last weeks or months? Although the findings on this topic
are quite complex, the simple answer is that when citizens are exposed to competing frames at one
time, the messages tend to cancel each other out. When citizens receive messages at different times,
however, they weigh the most recent message more heavily in their political judgments. Chong and
Druckman conclude that, if you are an advocate of a particular frame, you would be well served to
promote your message early to influence initial attitudes, often to combat other competing messages,
and late to make sure your message is the last one heard by the public.115
MEDIA EFFECTS IN A CHANGING TECHNOLOGICAL ENVIRONMENT
As the media environment has evolved over the past few decades, technology has been a catalyst
driving many changes. Partisan news outlets, for example, are more common than they used to be, in
part because the explosion of cable channels has made it possible for companies to turn a profit even
with small audiences. Whereas media organizations of the past had to attract large audiences to
increase revenues, technological developments have made it cheaper to produce and disseminate
news these days. Therefore, catering to a narrow partisan audience is a viable option. Another
significant new technology is, of course, social media. Social media platforms are ubiquitous in
American society, but do they affect public opinion? We address social media effects on the public
shortly. As for a discussion of the impact of partisan news on citizens’ political attitudes, that appears in
the Partisan News and Polarization section of Chapter 5. Finally, television is by no means a new
technology, but television production techniques and values have certainly changed over time. We turn
now to research that explores how these developments affect the public.
“In-Your-Face” Politics
Several recent studies have examined the effects of incivility on political life.116 For example, Diana
Mutz’s path-breaking research examines the political impact of incivility on television.117 Mutz starts with
the premise that powerful social norms govern face-to-face interactions. These norms dictate the
appropriate physical distance between people in a social interaction and guide us to distance ourselves
from people with whom we disagree. Television violates these social norms in two ways. First,
technological changes now enable cameras to shoot tight angles, bringing political figures into closer
spatial proximity to television viewers than is normal for a face-to-face interaction. It is now
commonplace for cameras to be “zoomed in so close that even portions of the speaker’s head are
routinely cut off and out of frame.”118 Second, norms of face-to-face social interaction require polite
conversation. Much of the political discourse on television, however, is anything but polite. It is standard
fare for politicians and pundits to disagree vehemently while shouting over one another. Discourteous
interactions, not expressions of civility, are the norm. As a result of these two characteristics of political
discourse on television, Mutz argues that citizens overwhelmingly experience politics as “in-your-face.”
What is the impact of in-your-face politics? Mutz designed an experiment to investigate the effect of in-
your-face politics on citizens. Specifically, she created her own television program modeled after
standard political talk shows. She recruited two actors (equally attractive white males between the ages
of thirty-five and forty-five) to play the part of congressional candidates competing for an open seat in
Indiana. The two candidates and a moderator followed a script in which there was disagreement on
several issues. In the “civil” condition, the candidates followed norms of politeness. For example, one
candidate said, “I don’t disagree with all of your points, Bob, but …” and then articulated his position. In
the “uncivil” condition, the substantive content was the same, but it was no longer delivered in a polite
way. Instead, the candidate said, “What Bob is completely overlooking is . …”119 The experiment also
included a manipulation of camera angles. In the “close-up” condition, the candidates’ faces filled the
screen, whereas tight close-ups were not included in the other condition, and the camera was at a
“medium” distance. This experimental design resulted in four conditions. Subjects were randomly
assigned to view either a civil, close-up program; a civil, medium program; an uncivil, close-up program;
or an uncivil, medium program. Mutz found that subjects were more physically aroused by the close-ups
and the uncivil discourse. These in-your-face conditions also caused subjects to list more of the
candidates’ arguments when asked to recall points made during the discussion. These findings suggest
there are benefits to watching in-your-face programming. Arousal can lead to greater attention to
politics, and recalling content from a political talk show can help citizens function effectively in a
democracy.
Mutz was also concerned, however, about whether in-your-face politics would affect citizens’
perceptions of the legitimacy of the opposition.120 Here the effects are not so positive. In a second
experiment using the same talk show stimulus, Mutz asked subjects to recall arguments made by the
candidates, evaluate both candidates on a feeling thermometer, and indicate for which candidate they
would vote if they had the opportunity to do so. Again in this experiment, Mutz found that subjects were
able to recall more in the uncivil and close-up conditions, including better recall of arguments made by
candidates they did not prefer. Importantly, when subjects were exposed to both the close-up shots and
the uncivil exchange, their attitudes toward the candidate they didn’t like were more negative and the
arguments made by the disliked candidates were perceived as less legitimate. To summarize these
findings, Mutz says subjects are “remembering more, but respecting it less.”121 In another experiment,
Mutz shows that uncivil debates lower subjects’ trust in politicians, political institutions, and the political
system. Overall, Mutz’s research suggests that televised in-your-face politics are problematic for
constructive debate in a democratic society.
Social Media Effects
Social media are a mainstay of American politics. Both Republican and Democratic political leaders
make active use of social media to connect with their supporters. During the 2016 presidential primary,
Democratic candidate Bernie Sanders was dubbed “Facebook royalty” in a New York Times
headline.122 Members of Congress on both sides of the aisle live tweeted the 2019 impeachment
hearings. And of course some might argue that Donald Trump governs by tweet. In July 2017, Trump
tweeted, “My use of social media is not Presidential—it’s MODERN DAY PRESIDENTIAL.”123
It is clear that political elites embrace social media as a technology to communicate with their
supporters, but how many citizens use social media for political purposes? A study by the Pew
Research Center shows that 39 percent of U.S. adults on Twitter tweeted at least once about national
politics during a one-year period from June 2018 to June 2019. But, the overwhelming majority of tweets
during this time frame, 97 percent, came from just 10 percent of Twitter users.124
Jason Gainous and Kevin Wagner argue that social media are particularly important to study despite the
relatively low percentage of citizens who use social media for political purposes because (a) these users
are highly politically engaged, (b) even small numbers of citizens can influence elections in competitive
districts, and (c) the number of users is increasing rapidly.125 Gainous and Wagner make several strong
claims about the impact of social media on politics. As a two-way form of communication, social media
allow users to choose what information to access and what to share with others. Importantly, citizens are
able to embed themselves in a social network of like-minded individuals free from traditional journalistic
gatekeeping. Gainous and Wagner demonstrate that many citizens prefer news that is consistent with
their point of view, and those citizens are most likely to use Twitter and Facebook to gather political
information. Social media provide citizens with the ideal tool to selectively expose themselves to
information with which they agree. Political candidates, interest groups, and elected officials are all
happy to accommodate this desire for agreeable content via social media because it also allows them to
circumvent media gatekeepers. Why would Donald Trump want to be asked questions at a news
conference when he can simply tweet his message to his followers?
Jaime Settle agrees that social media can be politically consequential even though many people do not
post content that is explicitly political, although her focus is quite different than Gainous and Wagner’s.
In her book Frenemies, she makes a compelling argument that people’s Facebook feeds contain much
information that is politically informative.126 By this, she means that a variety of different types of posts
can send signals about individuals’ political preferences. Sharing an article from a news source that is
well understood to be liberal, for instance, suggests the person who shared the story is liberal, even if
the article is not about politics. Lifestyle habits that correlate with political issues or supporters of one of
the political parties (such as hunting or attending church) can carry political meaning (Republican, in
these cases). Even status updates about shopping could tip off one’s political identity. Consider for a
moment your images of people who shop at Whole Foods and purchase hybrid cars versus those who
prefer Walmart and pick-up trucks.
What impact on Facebook users does such politically informative content have? As Settle put it, “People
draw political inferences based on even the non-political content that their social contacts post, and the
recognition of that [political] identity begins a cascade of processes that result in strengthened
attachment to one’s own political party and heightened negative judgment of one’s opponents.”127 She
uncovers many notable outcomes, including that individuals perceive Facebook posters to be more
ideologically extreme if the poster appears to support the opposing political party. Posters that support
one’s in-party are perceived as less extreme. Out-party supporters are also judged to be less
knowledgeable about politics and less likely to engage in critical thinking than are in-party supporters.
Finally, Facebook users are more likely than nonusers to believe that the U.S. political parties are
demographically and socially distinct from each other. Settle attributes this conclusion to the inferences
users make of the politics of the people who appear in their Facebook feeds.
CONCLUSION
Do the mass media live up to democratic ideals? Overall, participatory democratic theorists would say
no. In a perfect world, the media should be free from government and economic control, and they
should inform and educate the public, provide a forum for diverse views, and hold government officials
accountable. The reality is much different. The media in the United States are best characterized by the
three Cs: corporate, concentrated, and conglomerate. Soft news is on the rise, tabloid journalism is
increasingly influential, fake news is a concern, and in-your-face politics is problematic for democratic
debate. Further, adherence to news norms often results in news that is biased toward the perspectives
of powerful officeholders. Participatory democratic theorists would argue, however, that all is not lost
because nonprofit media, the minority press, and the Internet offer alternatives to the dominance of
conglomerates. Social media offer unprecedented opportunities for citizens to consume news and
create and share their own political content, yet the reality is that many citizens use social media to
simply seek out information that confirms their preexisting views and, along the way, make (not
necessarily accurate) inferences about the politics of others. Exposure to and sharing of fake news is
concentrated among a small group of people, yet the potential exists for falsehoods to go viral. Many
observers worry about both fake news as genre and fake news as label, and the implications of each for
democratic deliberation.
Elite democratic theorists have a much different view about whether the media live up to democratic
ideals. Compared with participatory democratic theorists, they have much lower expectations for citizens
in a democracy, and as a result, they also have much lower expectations for the media. Because the
media are relatively free from government control and provide citizens with enough information to go to
the polls and cast a ballot, elite democrats are pleased. The influence of economic forces, the emphasis
on official sources, and the increase in soft news, tabloid journalism, and in-your-face politics simply do
not raise the same concerns for elite democrats.
Are citizens influenced by the mass media? The answer to that question has changed over time.
Scholars originally proposed a hypodermic model of media effects, which said that the media were
extremely powerful and would persuade unsophisticated citizens with their messages. This model went
by the wayside, however, when little evidence was found to support it. Next, the minimal effects model
emerged. This model argued that citizens would filter media messages through their preexisting
attitudes. Instead of converting citizens to a new point of view, media messages were more likely to
reinforce and activate current predispositions.
Most recently, scholars have found substantial evidence to support a subtle effects model of media
influence. This tradition argues that the media influence citizens through agenda setting, priming, and
framing; the media influence what citizens think about, which issues or traits citizens bring to bear when
evaluating political leaders, and which considerations shape their thinking on political issues.
Overall, both participatory and elite democratic theorists can find things to like about the subtle effects
model. On the one hand, elite democrats would find it natural for citizens to take cues from the media.
Citizens are not expected to follow politics day in and day out; thus, it makes sense that the media
would provide guidance for what issues are important and how politicians and issues should be
evaluated. On the other hand, participatory democratic theorists would be pleased that citizens take in
media messages in a thoughtful way and do not simply fall prey to elite manipulation.
KEY CONCEPTS
accessibility model 106
accuracy 92
activation effect 99
advocacy journalism 93
agenda-setting effects 101
concentrated 85
conglomerate 87
conversion effect 100
corporate 83
fake news as genre 97
fake news as label 97
forum for diverse views 81
framing effects 103
hard news 84
horse race 92
hypodermic model 98
importance model 106
intermediary 81
legacy media outlets 89
minimal effects model 99
minority press 88
neutrality 92
news norms 91
newsbeats 92
newsworthiness 92
nonprofit media corporations 88
objectivity 91
official sources 92
outrage industry 93
pack journalism 95
priming 103
reinforcement effect 99
social media 111
soft news 85
soft news media 85
subtle effects model 100
tabloid journalism 93
traditional journalism 93
two-step flow of communication 100
watchdog 81
SUGGESTED SOURCES FOR FURTHER READING
Clawson, Rosalee A., and Dwaine H. A. Jengelley, “Media and the Politics of Groups and Identities in
Tumultuous Times,” Politics, Groups, and Identities 6 (2018): 785–787.
Jengelley, Dwaine H. A., and Rosalee A. Clawson, “Media Coverage and Its Impact on the Politics of
Groups and Identities,” Politics, Groups, and Identities 7 (2019): 148–150.
These introductory pieces provide a brief overview of two sets of articles on media and the politics of
groups and identities. The articles cover a range of topics, including framing police and protesters in
post-Ferguson news coverage; how digital-native news influences the legacy press’s coverage of
transgender issues; and the gender composition of images of members of Congress in Twitter feeds and
online.
Iyengar, Shanto, and Donald R. Kinder. News That Matters. Chicago: University of Chicago Press,
1987.
In a series of classic experiments, these authors provide evidence of agenda setting, priming, and
framing effects.
Matsaganis, Matthew D., Vikki S. Katz, and Sandra J. Ball-Rokeach. Understanding Ethnic Media:
Producers, Consumers, and Societies. Thousand Oaks, CA: SAGE, 2011.
This compelling text provides insight into the creation, nature, role, and consumption of ethnic media.
Shapiro, Robert Y., and Lawrence R. Jacobs, eds. The Oxford Handbook of American Public Opinion
and the Media. Oxford: Oxford University Press, 2011.
This edited volume pays particular attention to changes in technology and the mass media.
The Great Hack. Directed by Karim Amer and Jehane Noujaim, Netflix, 2019.
This documentary tells the story of Cambridge Analytica’s use of Facebook user data to try to influence
elections.
Accuracy in Media, www.aim.org
Media Research Center, www.mrc.org
These two groups monitor the media from a conservative perspective.
FAIR (Fairness & Accuracy in Reporting), www.fair.org
Media Matters for America, www.mediamatters.org
These two groups monitor the media from a progressive perspective.
Pew Research Center, State of the News Media, https://www.pewresearch.org/topics/state-of-the-news-
media/
The Pew Research Center conducts empirical research on the news media and publishes regular fact
sheets that provide all sorts of useful information about changes in American journalism, including
ethnic media and digital news.
Pew Research Center, “Explore Local News Habits in Your City with Our New Interactive,”
https://www.pewresearch.org/fact-tank/2019/03/26/explore-local-news-habits-in-your-city-with-our-new-
http://www.mrc.org/
http://www.fair.org/
http://www.mediamatters.org/
https://www.pewresearch.org/topics/state-of-the-news-media/
https://www.pewresearch.org/fact-tank/2019/03/26/explore-local-news-habits-in-your-city-with-our-new-interactive/
interactive/
This interactive site provides detailed profiles of local news attitudes and behaviors in cities across the
United States.
https://www.pewresearch.org/fact-tank/2019/03/26/explore-local-news-habits-in-your-city-with-our-new-interactive/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists years and ranges from 1996 to 2012 in increments of 4. The vertical axis lists
percentage and ranges from 0 to 100 in increments of 10. Approximate data from the graph are
tabulated as follows:
Year Watched TVnews yesterday
Read daily
newspaper
yesterday
Listened to news on
radio yesterday
Got online or mobile
news yesterday
1996 59 50 43 No data
2000 57 48 42 No data
2004 60 42 40 24
2008 58 35 35 28
2012 57 29 34 40
Back to Figure
The horizontal axis lists rallies and speeches. The vertical axis lists tolerance and ranges from 2 to 6.
Approximate data from the graph are tabulated as follows:
Activities Tolerance in free speech frame Tolerance in public order frame
Rallies 4 3.4
Speeches 4.1 3.6
CHAPTER 4 ATTITUDE STABILITY AND ATTITUDE
CHANGE
THROUGHOUT THIS BOOK, we present many examples of the public’s political attitudes, at times
noting whether or not these attitudes have changed over time. Chapter 1, for example, demonstrates
that Democratic Party identification is less common and partisan independence more common now than
they once were. In later chapters, we learn that the percentage of the public holding a liberal ideology
has not changed very much over the decades, all the while the percentage claiming no ideological
identification has declined. Americans’ interest in politics is a bit lower than in the past. Citizens have
warmed to the idea of some nontraditional presidential candidates (e.g., women, Jews, Blacks). Trust in
government has dropped precipitously since the 1960s. Public support for a variety of gun control
measures has been fairly steady over the past decade.
Sometimes, then, public views stay the same across time. Sometimes they do not. Which is more likely?
Put another way, do you think attitude change is common, or is attitude stability the norm? In Chapter
2, we learned how children and adolescents develop their political attitudes. After this early socialization,
do political attitudes remain the same throughout an individual’s adulthood? If attitude change occurs,
what causes this change? And, from a normative standpoint, what does it say about the public if attitude
instability is more common than stability? We address these questions in this chapter, first by examining
some evidence for attitude change and some for attitude stability. Next, we discuss factors that explain
fluctuations in presidential approval, one of the most important political attitudes held by citizens. In the
remainder of the chapter, we draw heavily from the discipline of psychology as we overview theories of
attitude formation and change that have greatly influenced how political scientists study public opinion.
ARE AMERICANS’ ATTITUDES STABLE?
One way to assess whether people’s attitudes remain the same over time is to survey people at one
time about their political opinions and then ask them about the same opinions later. Recall from Chapter
2 that this approach, known as a panel study, was the one used by Kent Jennings and Richard Niemi
to assess attitudes among high school seniors and their parents at various ages for both groups. Across
the years, a few panel studies have been conducted as part of the American National Election Study
(ANES) series, the most recent in 2008–2009. ANES surveys have been carried out at least every two
years, coinciding with presidential and congressional elections, since 1948.1 Although in most years the
ANES is a cross-sectional study—a survey using a new representative sample of adults—panel
studies have been conducted periodically, whereby previous respondents are reinterviewed.
Individual Attitude Change
One of the very first empirical analyses of political attitude stability was conducted by Philip Converse
using data from the ANES 1956, 1958, and 1960 panel study.2 This study questioned respondents
about salient political issues of the day in each of the three years. These issues included domestic and
foreign policy issues such as school desegregation, federal aid to education, and military aid to fight
communism. Respondents were also asked their party identification. Comparisons of respondents’
political attitudes in 1958 with their attitudes in 1960 are presented in Table 4.1. To measure the degree
to which individuals’ attitudes were stable, Converse calculated tau-b correlation coefficients, which are
presented in the first column of Table 4.1. When tau-b equals 1.0, everyone’s attitude was the same in
1960 as it had been in 1958. The smaller the value of tau-b, the more people’s attitudes fluctuated over
these two years. Another way to compare attitude stability over time is to determine how many people
kept the same opinion at two times. We present our results from such an analysis in the second column
of the table. The figures here are the percentage of respondents whose opinions stayed on the same
side of an issue or whose party allegiance was to the same party (or who remained politically
independent) from 1958 to 1960. Political attitudes are considered stable by this measure even if
someone changed from strong to weak agreement or from strong to weak partisanship over time. The
final column of the table presents the percentage of respondents who expressed no opinion toward the
policy issues in either 1958 or 1960, or in both years.
As we see from the results in Table 4.1, party identification was the most stable political attitude over
these two years, considerably more than any of the other attitudes. For most people, their party
affiliation remained the same from 1958 to 1960. Converse attributes this to the fact that attitudes and
affect toward groups, such as political parties, help to organize the political opinions of many people.
Because party identification was the only attitude surveyed “that touches on pure affect toward a visible
population grouping,” it came as no surprise to Converse that these attitudes remained quite stable over
two years.3
Table 4.1 Stability of Individual Political Attitudes from 1958 to 1960
Political
attitude
Correlation between
attitudes in 1958 and
1960 (tau-b)
Percentage holding the
same attitude in 1958 and
1960
Percentage having no
opinion in at least one
year
Party
identification
.73 85.7 N/A
School
desegregation
.43 57.5 15.8
Employment
discrimination
.41 60.0 17.9
Guaranteed
employment
.41 56.5 15.2
Isolationism .39 59.6 16.3
Political
attitude
Correlation between
attitudes in 1958 and
1960 (tau-b)
Percentage holding the
same attitude in 1958 and
1960
Percentage having no
opinion in at least one
year
Federal aid to
education
.38 57.2 13.8
Foreign
economic aid
.34 48.0 21.4
Foreign
military aid
.32 56.7 24.9
Federal
housing
.29 40.7 33.9
Sources: Data from Philip E. Converse, “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. David E. Apter
(New York: Free Press, 1964), 240; Analysis of American National Election Studies 1956–58–60 Panel Data File.
In contrast, attitude instability was much more common for the policy issues, with only between 40 and
60 percent of the public holding the same opinion over time, versus nearly 86 percent for party
allegiance. Among the policy issues queried, citizens’ attitudes toward racial employment discrimination,
isolationist foreign policy, and school desegregation were the least likely to change. Opinions about
foreign aid and whether the federal government should be involved in providing housing and electric
power were especially likely to fluctuate. To explain why there were these differences in attitude stability
across issues, Converse points to the fact that some issues, even domestic ones, were less significant
to citizens in the late 1950s.4 Stable attitudes are more likely toward objects that tend to influence
people’s everyday lives (such as jobs or schools) or toward salient groups (such as Blacks). When
issues are more remote, attitude instability is more common.
People’s political attitudes could change for a number of reasons. A recent study of individual attitude
change regarding same-sex marriage highlights a few of these factors. Using data from three waves of a
panel survey (2008, 2014, and 2016), Hye-Yon Lee and Diana Mutz explore support for same-sex
marriage.5 The percentage of people who changed their views between 2008 and 2014 was 29 whereas
16 percent changed between 2014 and 2016. Nearly all of this change was in the direction of higher
support. As for why people changed their minds, Lee and Mutz uncover three explanations. Citizens
who became more supportive of same-sex marriage over these years (compared to those who did not)
had an increase in the number of gay and lesbian friends and relatives in their lives, became less
religious, and obtained a higher level of education. Of these, increases in interpersonal contact with
gays and lesbians had the largest influence on attitudes.
Citizens’ opinions are also susceptible to change in the face of societal events. Current events can
provide new information to citizens as well as perhaps new understandings of issues, both of which can
produce changes in attitudes. Converse considered this possibility but ruled it out as an explanation for
the attitude instability he observed. In the late 1950s, high-profile news events did occur that could have
influenced attitudes toward some issues that Converse studied, but not all. The public witnessed
standoffs and violence in school desegregation cases in the southern United States, such as the
integration of Central High School in Little Rock, Arkansas, in 1957 when army soldiers were required to
escort Black students into the formerly all-white high school. Yet for one of the issues where the public’s
attitudes were the least stable, the role of the federal government in providing housing, there were no
changes in federal policy or other relevant newsworthy items during these years that might have
produced public attitude change. Converse also witnessed that the average correlations between
people’s attitudes were the same and were similarly low between 1956 and 1958, between 1958 and
1960, and between 1956 and 1960. We might expect that attitudes would fluctuate more the longer the
interval between measuring the attitudes. This was not the case, leading Converse to turn to
explanations for instability other than responses to changing events.
Ultimately, Converse concludes that many citizens’ policy opinions are meaningless and might be better
characterized as nonattitudes than attitudes.6 Converse argues that “large portions of an electorate do
not have meaningful beliefs, even on issues that have formed the basis for intense political controversy
among elites for substantial periods of time.”7 If public attitudes were well developed, carefully
considered, and based on detailed information, according to Converse, they probably would not change
as much as he had observed. Converse’s statements might seem extreme and unnecessarily harsh,
especially when you consider that, according to Table 4.1, a majority of the public did continue to hold
opinions on the same side of most issues between 1958 and 1960. Examining temporal attitude stability
was only one portion of Converse’s research, however, and his conclusions were based on a variety of
other analyses. In particular, Converse demonstrated that most people’s opinions are not well grounded
in broader beliefs (such as liberalism or conservatism) and are not well organized.8 We discuss these
other conclusions more thoroughly in Chapter 5.
Yet, even thinking only about Converse’s evidence of attitude change, it is fair to ask how many
Americans should have stable attitudes for us to conclude that the public’s policy opinions are
meaningful. Is our standard 100 percent? If so, then any deviation from this might lead to conclusions
similar to Converse’s. In fact, it was empirical evidence such as this that fostered the development of the
theories of democratic elitism and pluralism. Recall from Chapter 1 that these theorists compared
empirical research about the public’s capabilities with classical democratic theorists’ expectations about
the public and concluded that the public was not living up to the classical model. By using a different
standard, however, perhaps one that recognizes some policy issues are complex and are not central to
most people’s daily lives, we may conclude that levels of attitude stability apparent in the late 1950s
seem reasonable and not too low.
Collective Attitude Stability
We turn now to an examination of the stability of collective public opinion, that is, the aggregate
political opinions of the public. In the most detailed analysis of collective opinion over time, Benjamin
Page and Robert Shapiro analyzed results from public opinion surveys conducted over fifty years. Page
and Shapiro’s conclusions paint a very different picture than that emerging from studies of individuals. In
their own words, “The American public, as a collectivity, holds a number of real, stable, and sensible
opinions about public policy and … these opinions develop and change in a reasonable fashion,
responding to changing circumstances and to new information.”9
Page and Shapiro’s evidence for this conclusion consists primarily of comparisons of responses to
survey questions that were asked at least two different times between 1935 and 1990. For 58 percent of
these questions, aggregate public opinion did not change significantly over time. For instance, in 1942,
approximately 75 percent of the American public believed that the United States should have an active
rather than isolationist foreign policy. Fourteen years later, the percentage of Americans holding this
attitude was also about 75 percent. Opinions on this topic did fluctuate a bit in the intervening years, but
within a narrow range of 72 to 80 percent supporting activism over isolationism.
Turning to domestic matters, Page and Shapiro observed that opinions about government spending on
a variety of programs did not fluctuate very much during the 1970s and 1980s. Consistently high
percentages of the public (68–77 percent) felt that the government was spending too little to fight crime,
whereas consistently low percentages (12–25 percent) felt that government was spending too little on
welfare (see Figure 4.1). Attitudes toward government spending on Social Security were especially
stable during the late 1980s (the years for which opinion data are available). The item displayed in
Figure 4.1 that shows the most variation was beliefs about education spending. Whereas about 49
percent of the public felt the government spent too little on education in the early 1970s, this percentage
gradually increased during the time period, reaching 68 percent in 1989.
Description
Figure 4.1 Opinion Toward Government Spending, 1971–1989
Source: Adapted from Figures 2.2, 4.1, and 4.3 of Benjamin I. Page and Robert Y. Shapiro, The Rational
Public: Fifty Years of Trends in Americans’ Policy Preferences (Chicago: University of Chicago Press, 1992),
49, 120, 126.
While Page and Shapiro found many examples of collective attitude stability, they also uncovered some
instances of attitude change. Unlike what we might expect given Converse’s results, however, Page and
Shapiro observed that collective attitude change was often modest in size and that large attitude swings
over short periods of time were not very common. Opinion change was more likely to be gradual (slow,
steady shifts in Americans’ beliefs, often in response to changes in the broader social or political
climate) than to fluctuate (increases in support followed by decreases, or vice versa, over a short period
of time). We just observed such a gradual increase with education spending. Americans’ opinions have
also changed slowly over time for other domestic policy issues, such as racial integration of schools and
capital punishment. As for capital punishment, support for the death penalty steadily increased from 47
percent in 1966 to 78 percent in 1985. More recent research on death penalty attitudes shows that
support for the policy continued to increase into the mid-1990s but then declined in the 2000s.10 In
2018, the Gallup organization reported that 56 percent of Americans supported the death penalty for
people convicted of murder.11 Consistent with Page and Shapiro’s argument that aggregate opinion
responds to the political environment, the decline in support for the death penalty occurred as the media
paid more attention to the “innocence frame,”12 that is, the possibility that innocent people might be
executed.
Page and Shapiro also explored the prevalence of abrupt opinion changes among the public.
Abruptness, or a change of at least 10 percentage points within a year, was especially likely for foreign
policy issues. On closer examination of these cases, Page and Shapiro conclude that these sudden
opinion changes often happened in the wake of real-world events: “wars, confrontations, or crises, in
which major changes in the actions of the United States or other nations have understandably affected
calculations about the costs and benefits of alternative policies.”13
Collective Public versus Individual Citizens
Throughout their book, Page and Shapiro contrast their image of aggregate public opinion as
reasonable, meaningful, and largely stable with the image of the public presented in the empirical work
of Converse and in the writings of America’s founders, most especially James Madison and Alexander
Hamilton, who asserted that the public was subject to fits of passion and that public opinion was
expected to be volatile. How might we explain these discrepancies? Which view is correct?
To answer these questions, we should first consider the differences in examining individual opinions
versus the aggregate opinion of the public. Aggregate opinion can appear quite stable even if many
individuals change their opinions, provided that the individual changes cancel each other out. Imagine,
for example, that we surveyed ten people in 2018 and six supported increasing government spending
for fighting terrorism (see Table 4.2). Two years later, we surveyed the same ten people. Four
respondents, Alonzo, Owen, Lei’ana, and Sadie, held the same attitudes, whereas the others had
changed their minds. Madison, Cameron, and Kyree initially favored increasing spending but later
supported spending decreases. Khalil, Jared, and Olivia wanted less spending in 2018 but favored
spending more in 2020. Even with all these changes, our aggregate result is the same: six of ten
respondents supported spending more money to fight terrorism in 2018 and 2020. The same
phenomenon can occur when surveying representative samples of the American public. Many changes
at the individual level over time can still result in temporal stability (or near stability) at the aggregate
level. Thus, both Converse and Page and Shapiro are correct, in that their conclusions are appropriate
for their level of analysis (individual or collective).
Table 4.2 Aggregate Opinion Can Be Stable While Individual Attitudes Change
Ten survey
respondents Interviewed in 2018
Reinterviewed in
2020
Substantial individual
change
Alonzo Increase Increase Stable
Owen Increase Increase Stable
Lei’ana Increase Increase Stable
Madison Increase Decrease Change
Cameron Increase Decrease Change
Kyree Increase Decrease Change
Khalil Decrease Increase Change
Jared Decrease Increase Change
Ten survey
respondents Interviewed in 2018
Reinterviewed in
2020
Substantial individual
change
Olivia Decrease Increase Change
Sadie Decrease Decrease Stable
Aggregate stability 6 Increase, 4
Decrease
6 Increase, 4
Decrease
In describing the aggregation process, however, these researchers present different assumptions about
what the aggregate results are really measuring. Page and Shapiro argue that individual changes,
though perhaps frequent, are relatively minor deviations from true, enduring beliefs and values. Thus, “a
poll or survey can accurately measure collective preferences as defined in terms of the true or long-term
preferences of many individual citizens.”14 Converse, in contrast, views the aggregation process as
covering up the shallow, fleeting opinions of individual citizens. “It is quite possible,” he writes, “thanks to
the hidden power of aggregation, to arrive at a highly rational system performance on the backs of
voters, most of whom are remarkably ill-informed much of the time.”15
Normatively, we might wonder whether it is better to think of the public as individual citizens or as a
collection of individuals. Each approach is reasonable and well grounded in democratic theory. Many
democratic theories highlight the characteristics and capabilities of individual citizens, although theories
reach different conclusions as to whether citizens are capable of ruling in a democracy. Yet, in positing a
role for citizens to influence public policy, as most democratic theories do, these theories must also
focus on aggregate public opinion. Discussions of whether and to what extent the wishes of the public
should influence the decisions of elected officials naturally revolve around the wishes of the aggregate
public rather than the wishes of an individual citizen. Aggregating individuals’ opinions is necessary for
representation to occur, whether at the national, state, or congressional district level. In other words, if
we want to know whether the national government is responsive to the wishes of the public when it
comes to counterterrorism policies, for instance, we turn to aggregate measures of public opinion toward
these policies. So whether we adopt Converse’s conclusions that individual attitudes are unstable or
Page and Shapiro’s arguments that the aggregate public holds largely stable opinions depends in part
on whether we are concerned about the democratic public as individual citizens or as a collective whose
views might influence governmental decisions.
PRESIDENTIAL APPROVAL
One of the most important political attitudes studied by political scientists is presidential approval.
Politicians, journalists, and political junkies are also quite interested in this concept. Presidential
approval refers to the public’s level of approval or disapproval of the president’s job performance. For
decades, the Gallup organization has conducted daily telephone surveys with national, representative
samples of U.S. citizens to track approval of the president. Gallup asks, “Do you approve or disapprove
of the way [the incumbent] is handling his job as president?” Gallup releases its latest approval numbers
for the president based on a three-day rolling average of these tracking polls.16 Some political
observers are so obsessed with these numbers that they glue themselves to their computers or
smartphones waiting anxiously for the data to be released.
Why are these approval ratings so important? Because they speak volumes about a president’s political
power. Strong approval ratings are good news for a president’s reelection chances, and they boost the
electoral fortunes of members of Congress in his party who support his agenda.17 Higher approval
ratings also enable the president to be more successful in his interactions with Congress.18 In general,
high approval ratings can be thought of as a form of political capital that allows the president to pursue
his goals.
Presidential approval ratings for Barack Obama and Donald Trump are presented in Figure 4.2. (Take a
look at the Presidential Job Approval Center on the Gallup website for job approval ratings from Truman
to Trump.19) One thing to note about these ratings is that presidential approval often declines over time.
Barack Obama took office with approval ratings in the mid-60s, but those declined over time. His
approval rating bounced back somewhat during his last months in office, although it only reached the
high 50s then. This initial popularity is often referred to as the honeymoon period. Presidents take
office with substantial goodwill from the public and, quite important, from other political elites. Some
presidents who are reelected also see an uptick in their approval around the time of this reelection, as
Obama did in the fall of 2012. As time goes by however, the president is faced with many difficult and
controversial problems, and inevitably he makes decisions that tick off one group of people or another.20
Although the average citizen may not track a president’s every move, political elites are paying close
attention, and some begin to complain about the president’s decisions. The media transmit, and may
even exaggerate, those criticisms, so that citizens are increasingly exposed to negative coverage of the
president. As a result, public approval for the president falls over time.21 Additionally, people tend to
expect new presidents to be able to accomplish key policy goals and to have favorable personal
characteristics, such as sound judgment and high ethical standards. Over time, as presidents fail to live
up to these expectations, which can be unrealistically high, a public expectations gap emerges. The
larger the gap between what the public expects of a president and what the president actually
accomplishes, the lower the president’s approval.22
Description
Figure 4.2 Presidential Approval for Barack Obama and Donald Trump
Source: Data from “Presidential Job Approval Center,” Gallup, Washington, DC, 2019,
https://news.gallup.com/interactives/185273/presidential-job-approval-center.aspx
If you have already looked closely at Figure 4.2, you will have noticed that Donald Trump did not
experience a honeymoon period in approval ratings. As one Gallup analyst put it three months into his
presidency, “Already a trendsetter by earning the lowest initial job approval rating [45%] of any president
and falling below 40% approval in record time, Trump’s recent 35% and 36% approval ratings are the
lowest of any president in his first year.”23 Also notable is how relatively stable Trump’s approval ratings
have been over time.24 As of October 31, 2019, only 11 percentage points separated the highest (46)
and lowest (35) percentages of the public approving of his performance, a much smaller range than for
other presidents.
Another important feature of the presidential approval data is the way the public reacts to foreign policy
events. Specifically, a rally round the flag effect occurs when presidential popularity surges in the wake
of a foreign policy event involving the United States, especially an unexpected crisis.25 In the most
dramatic example of this effect, after the 9/11 attacks, President George W. Bush’s ratings soared to 90
percent approval as citizens’ patriotism and feelings of unity were activated by the tragedy of that day.26
As expected, over time Bush’s ratings declined from this extraordinarily high level. Barack Obama’s
approval also received a boost, of more than 10 percentage points, in response to the May 2011 killing
of Osama bin Laden, the leader of al-Qaeda. Foreign policy events can clearly benefit a president’s
approval ratings, but war can take a toll on approval. For instance, a study of presidential approval
ratings from 1948 through 2008 demonstrates that war casualties and the financial costs of war can
dampen a president’s popularity.27
Citizens’ views of presidential job performance are influenced by another crucial factor—the economy.
Whether scholars use objective measures of economic well-being, such as the unemployment rate,
or citizens’ perceptions of the state of the economy or both to explain presidential approval, it is clear
that economic variables have a significant impact on citizens’ evaluations.28 As Obama was well aware,
a poor economy wreaks havoc on a president’s popularity. The bump in approval he experienced as a
result of bin Laden’s death quickly dissipated as economic conditions weighed heavily on many U.S.
citizens during summer 2011. In contrast, the public usually evaluates the president more favorably
when the national economy is strong. That has not been the case for Trump, however. During his first
two years in office, economic growth and the unemployment rate were similar to what they were during
Bill Clinton’s last year in office. Yet, Clinton’s approval rating was around 60 percent compared to
Trump’s average rating of 39 percent.29
Finally, and perhaps not surprisingly, there are also partisan differences in presidential popularity.
Indeed, partisan divergence is becoming the defining characteristic of citizens’ approval of the president.
Partisan polarization was at a record high during Obama’s final year in the White House. For those 12
months, on average, 89 percent of Democrats approved of his performance compared to 12 percent of
Republicans, a 77 point gap. That record was eclipsed during Trump’s second year in office, when the
gap was 79 points (87 percent Republican versus 8 percent Democratic approval).30 Partisan dynamics
are also present when conditions that typically affect approval ratings change. In the past, citizens
identifying with the party other than the president’s displayed larger approval gains when the situation
warranted (a rally effect or strong economy) and larger declines when conditions became less favorable,
compared to members of the president’s party.31 During Obama’s presidency, however, a new pattern
emerged, according to recent research conducted by Kathleen Donovan and her colleagues.32
Presidential approval did not budge much for Democrats or Republicans as national economic
conditions improved, although Independents did change their ratings in line with the state of the
economy. In contrast, Democratic citizens had “gone all-in on the president already, leaving little room
for their approval to grow” whereas Republicans were loath to reward Obama for a changed economy.33
https://news.gallup.com/interactives/185273/presidential-job-approval-center.aspx
The strong influence of partisanship on approval ratings also explains why Trump did not have a
honeymoon period and why his ratings have been so stable despite an improved economy.34
Overall, this aggregate-level research on presidential approval shows that it is not as stable as collective
public opinion on issues tends to be. Yet at the same time, changes in presidential approval are not
random but are reactions to political events and economic conditions, suggesting at least a somewhat
sophisticated and capable citizenry. There is a caveat, however. Among partisans, changes in approval
are becoming rarer as the public relies more on partisanship than real-world conditions when evaluating
the president. This reality might not surprise elite democrats, but it is worrying for participatory
democrats, who would prefer that evaluations of the president’s job performance be updated as
conditions warrant.
PSYCHOLOGICAL APPROACHES TO ATTITUDES
Psychologists have studied attitudes more thoroughly than scholars in any other academic discipline
and have provided important theories and approaches for political scientists who examine political
attitudes. In trying to understand attitude formation and change, psychologists place primary emphasis
on individuals. Some psychological theories, for instance, focus on the influence of individual
characteristics, such as motivations and core values, on opinions. Other approaches examine how
people learn and process new information, especially persuasive communication meant to change
people’s attitudes. Psychologists have also postulated many different routes to attitude change.
Although we do not have space here to thoroughly review all the work on attitudes in psychology, we
highlight approaches that are especially relevant to political attitude formation and change.
Functions of Attitudes
One approach to understanding attitude formation is to consider the functions attitudes serve for
individuals. Functional theories of attitudes highlight the motivations people have for holding the
attitudes that they do.35 For some people, attitudes serve a utilitarian function. People hold positive
feelings toward objects that provide benefits, such as support for a tax policy that will increase their
income. Negative attitudes are held toward objects that prevent people from satisfying their needs or
that inflict punishment. A second type of function is ego defense. Attitudes fulfill this purpose by
protecting an individual’s ego or self-image from threats. These threats could be internal, such as
denying certain negative characteristics about oneself. In particular, prejudice can result from ego-
defense mechanisms: “When we cannot admit to ourselves that we have deep feelings of inferiority, we
may project those feelings onto some convenient minority group and bolster our egos by attitudes of
superiority toward this underprivileged group.”36
Some individuals have a desire to understand the world around them, and attitudes can help to fulfill this
goal. Attitudes can, in other words, serve a knowledge function. In particular, our existing attitudes can
help us sort through newly encountered information. Imagine, for instance, that someone holds a
positive attitude toward immigrants and learns that a new policy is meant to help ease the process of
immigrating to the United States. They can use the knowledge that their attitudes toward immigrants
provides to arrive at support for the immigration policy quickly without (or before) learning all the details
of the new policy.
Attitudes that allow individuals to express their core beliefs and values perform a value-expressive
function. People who are deeply religious, for example, will feel that their identity is being actualized if
they hold political views that reflect their religiosity, such as support for prayer in school. Somewhat
related to a value-expressive motivation is social adjustment.37 If social adjustment is an important
goal for individuals, they will express attitudes that conform to the views and values of peers in their
social network.38
According to functional theories of attitudes, the reasons an attitude remains stable or the processes
that result in attitude change depend on which function the attitude serves.39 Attitudes that serve a
value-expressive function, for instance, are not very likely to change because they are grounded in an
individual’s concept of self. Attitude change will only come about if a person’s self-concept changes.
Imagine a pacifist who, in the face of actual instances of unprovoked aggression toward a defenseless
group, grows dissatisfied with holding pacifist values. As another example, attitudes that help a person
make sense of the world will probably change when they cease to provide such knowledge, such as
when the person’s views toward a group are found to be based on incorrect information. If you believed
that women are not as decisive as men, you would probably evaluate female politicians less positively
than male politicians regardless of whether the specific politicians are decisive or not. Repeated
encounters with decisive women might cause you to update your stereotype about women and then
alter your attitudes toward female politicians.
Functional approaches to attitudes also suggest when persuasive appeals will succeed. Direct attempts
to change someone’s attitude will be more successful if the arguments are geared toward the function
that the attitude is serving. For instance, in one series of studies, college students were exposed to
different arguments about the institutionalization of the mentally ill.40 Messages that highlighted
cherished values (particularly loving and responsibility) persuaded people who held value-expressive
attitudes, whereas messages about the opinions of peer groups (such as being told 70 percent of
college students favored institutionalization) persuaded people who expressed social adjustive attitudes.
Learning Theory
Learning theory provides another approach to attitudes. Some of our likes and dislikes are acquired
through experience. When we are exposed to some stimuli, we receive positive reinforcement, such as
feelings of happiness or rewards. With other stimuli, we receive negative reinforcement, such as a
punishment or feelings of fear. Children, for instance, could learn to like the Republican Party because
they hear their parents frequently saying positive things about the party. Likewise, children who hear
their parents using derogatory words to describe members of a minority group can develop prejudiced
attitudes toward that group.
A basic premise of the learning theory of attitudes, then, is that, with repeated exposure to attitude
objects paired with positive or negative adjectives, we can learn to like or to dislike those objects. More
broadly, this theoretical approach assumes that attitudes are obtained much like habits are. In our daily
lives, we learn about the attributes of attitude objects and the feelings associated with them; from these
cognitions and affect, our attitudes develop. Attitudes are, in this view, a by-product of the learning
process and individuals are rather passive actors in the acquisition of attitudes. Once attitudes are
acquired in this way, they will tend to remain stable until new learning occurs.
According to learning theory, attitudes can also be influenced when the affect that is associated with one
attitude object is transferred to a related object, a phenomenon called a transfer of affect. Individuals
hold positive feelings toward objects that they like and negative feelings toward disliked objects. These
feelings can be passed on to associated objects, resulting in either a new attitude or attitude change,
depending on whether an attitude already exists for the associated object. In politics, this principle helps
us understand why politicians try to connect their opponents to stimuli that induce negative emotions
(such as fear and disgust) in most Americans. It also explains why politicians surround themselves with
objects that evoke positive feelings for most people, such as the American flag. Politicians hope that the
positive feelings citizens have for the flag will transfer to themselves, thus resulting in positive attitudes
toward the politicians.
During the 2016 presidential primary season, a Democratic candidate ran a campaign advertisement to
encourage the transfer of positive affect toward America to himself. The ad, which was simply titled
“America,” contained a few shots of the candidate who aired it, Bernie Sanders. Mostly, though, the ad
was “a montage of American images—farms, cities, windmills, offices, coffee shops and kitchen tables
—set to the well-known Simon & Garfunkel song of the same name [‘America’].”41 Compared to other
ads that were aired during the 2016 presidential election, this Sanders ad made citizens feel happier,
more hopeful, and better about American democracy. It also boosted Sanders’s favorable ratings, just
as predicted by affect transfer theory.42
Communication-Induced Attitude Change
Much attention in psychology has been placed on understanding how and when communication can
influence attitudes. Some of this communication is directly designed to change people’s attitudes. In
contrast to such explicitly persuasive communication, other messages can influence attitudes without
necessarily attempting to do so. Examples of the latter include news media stories. By providing new
information and understandings of issues, media coverage of current events can influence attitudes
even if the journalists’ intent is not to do so. Of the many models of communication and attitude change
that exist, we highlight one that is especially relevant to understanding the formation and change of
attitudes in response to political communication: John Zaller’s receive-accept-sample model.43
Zaller was interested in how the flow of political information can influence citizens’ opinions. In particular,
he examines political information that appears in the mass media. “This coverage,” he writes, “may
consist of ostensibly objective news reports, partisan argumentation, televised news conferences, or
even paid advertisements, as in election campaigns.”44 Collectively, he refers to this communication as
elite discourse, although he does not always specify whether the elites are politicians, candidates,
news reporters, or experts quoted in news stories. In fact, he often means to convey that any and all of
these types of elites can provide political information to the citizenry—information that may come to
influence the public’s opinions.
To understand the dynamics of public opinion formation and change, Zaller draws heavily on
psychologist William McGuire’s research on persuasion.45 The first process they both explore is
reception. For people to be persuaded by a political message, they must first receive it. That is, they
must be exposed to the message, and they must comprehend it. If this message reception does not
occur, it would be impossible for the message to change citizens’ minds.
In addition, Zaller argues that the effect of elite discourse on attitudes will not be the same for all people.
In other words, key individual characteristics mediate the effects of communication on attitudes. Chief
among these is political awareness, or the degree to which citizens follow political matters closely.
Citizens who are politically aware are much more likely to receive political messages from elites than
those who do not follow politics closely.
Once citizens have received a message, Zaller argues that persuasion can occur only if they also agree
with the content of the message. Notably, at this acceptance stage of persuasion, political awareness
works differently than at the reception stage. Here awareness makes it less likely that individuals will
accept the premise of an argument that they have received. Simply stated, people who pay more
attention to politics are “better able to evaluate and critically scrutinize the new information they
encounter.”46
Given these two relationships between political awareness and reception, on the one hand, and
acceptance, on the other, those who possess either low or high levels of awareness are not likely to
have their attitudes influenced by elite discourse. Those with low levels of awareness are unlikely to
receive political messages, whereas those highly aware are likely to rebut any political arguments they
do receive. This means that those with the highest probability of having their attitudes changed by
political discourse are people who possess moderate levels of awareness, a pattern displayed visually
in Figure 4.3. Moderately aware citizens do receive some political messages but do not possess a large
enough store of relevant knowledge with which to resist communication-induced attitude change.
As for other individual attributes that influence whether political messages will result in attitude change,
Zaller focuses special attention on people’s existing political orientations. These orientations, which he
calls political predispositions, encompass core values, enduring beliefs, and experiences.
Predispositions are the ingredients for specific policy opinions (those who value equality will have
different political opinions from those who do not value equality, for example), and they also mediate
whether people accept the premises in elite discourse that they receive. Individuals are not likely to
accept a message that is inconsistent with their predispositions. A conservative citizen who encounters
a stream of liberal messages, such as arguments in favor of increasing government spending on social
welfare policies, is unlikely to adopt the position of these messages because they conflict with
preexisting political views.
Description
Figure 4.3 Political Awareness in Zaller’s Attitude Change Model
Source: Based on arguments contained in John R. Zaller, The Nature and Origins of Mass Opinion
(Cambridge, UK: Cambridge University Press, 1992).
In Zaller’s model, then, people are not necessarily passive vessels into which political opinions are
poured. Instead, some citizens consider incoming arguments and decide whether to accept them. As
Zaller correctly argues, however, the ability of people to know which messages coincide with and which
contradict their predispositions increases with political awareness. In other words, those who are more
politically aware are better able to see the connections between elite discourse and their own
predispositions, whereas “politically inattentive persons will often be unaware of the implications of the
persuasive communications they encounter, and so often end up ‘mistakenly’ accepting them.”47 In
addition, sometimes cues exist in the elite discourse to facilitate this connection. Such cues provide
details regarding the implications of a political argument for a person’s predispositions. Often these cues
are in the form of the partisan or ideological leanings of the source of the message. For example, when
a Republican politician criticizes a Democratic politician’s foreign policy, Republican citizens are more
likely than Democrats to accept this argument because they can assume, based on the source of the
message, that the position advocated therein coincides with their existing political dispositions.
From this foundation, Zaller argues that citizens’ opinions are not necessarily altered once they have
received and accepted a political message. Instead, Zaller argues that individuals accept messages but
only update their opinions when they are called on to do so, such as when discussing politics with
friends or when asked by a polling organization to answer a survey question. At this point, people
sample among the considerations that are stored in their memory related to the political issue at hand.
Considerations, as Zaller defines them, are “any reason[s] that might induce an individual to decide a
political issue one way or the other.”48 If, at the time respondents must state their opinions, there are
more considerations at the top of the head that favor the issue, they will indicate support for it. If they
gather more opposing considerations from memory, however, an opposing opinion will result. This
sampling is the final stage of Zaller’s receive-accept-sample model.
What accounts for the balance of pro- versus anti- considerations that are accessible in people’s minds?
The content of elite discourse at the time. Zaller considers environments in which there is elite
consensus on an issue so most messages articulated by elites favor the same side of the issue. At other
times and for other issues, political elites will disagree over an issue and the flow of political information
will contain messages both favoring and opposing the issue. When one message dominates elite
discourse, there will be few details to cue citizens as to whether the arguments are consistent with their
predispositions. This is because politicians from both dominant political parties and from varying
ideological backgrounds will express support for an issue during times of elite consensus.
Therefore, in a one-message environment, more politically aware people will be more likely to receive
and, due to the absence of key cueing details, accept the dominant message. Zaller calls this the
mainstream effect. For example, during the early years of American involvement in Vietnam, most
political elites supported America’s role. Consequently, in 1964, citizen support for the Vietnam War
increased with political awareness, regardless of whether citizens were liberal or conservative (see
Figure 4.4). In contrast, elite disagreement over a policy issue, especially when the disagreement falls
along party or ideological lines, results in the polarization effect. As the Vietnam War carried on, elite
agreement disappeared. More and more Democratic and liberal elites expressed opposition to the war.
As a result, attentive liberal citizens began to oppose the war, whereas attentive conservative citizens
continued to express support for the war. As the polarization effect predicts, the opinions of those with
lower levels of political awareness did not differ by political predispositions. Either these individuals did
not receive the countervailing messages about the war or they accepted all received messages
(because they were less able to discern the connection between the messages’ arguments and their
own predispositions).
The Vietnam War happened a long time ago, so you might be wondering whether there are more recent
examples of these communication effects. Recent evidence of the mainstream effect is hard to come by,
perhaps because elites from across the political aisle are not very likely to agree on policy issues these
days. In contrast, for an essay published on the twentieth anniversary of the book that introduced his
model, Zaller presented figures demonstrating the polarization effect for a range of issues.49 These
include public opinion regarding tax increases, abortion, gun control, immigration, and global warming.
For all these issues, the opinions of the public diverge by partisanship among those who are most
aware of politics, whereas partisan gaps are absent among the less attentive public. Another issue
where public opinion patterns coincide with the polarization effect is the Affordable Care Act (a.k.a.
Obamacare).50 Democratic members of the public were more supportive than Republicans of this law
when it was passed in 2010 and have remained so since then. The partisan divide is larger among
those who are more (versus less) knowledgeable about specific features of the law. Furthermore, this
gap has actually widened over time. When health care reform measures were instituted in the past (e.g.,
Medicare), as citizens gained experience with the policies, partisan divisions over them fell away. Not so
with Obamacare, due largely to the partisan-based elite messaging surrounding this health care policy.
Resistance to Persuasion
Psychologists have also examined when people are especially likely to resist persuasive
communication. The result of resistance is that attitude change will not occur in the face of persuasive
messages. Scholars who explicitly focus on resistance, however, push beyond this outcome-focused
definition. Instead, they discuss resistance in terms of individual motivations to withstand change or the
processes by which people ensure the persuasive communication will fail to change attitudes.51
Resistance is more likely (1) for people who consider themselves not easily susceptible to persuasion,
(2) for attitudes that are linked to one’s core values, and (3) when people have been forewarned that a
persuasive message is coming.52 In contrast, buttering people up before attempting to persuade them
can reduce resistance. Why? Because one reason for resisting a persuasive message is a motivation to
preserve a positive concept of oneself. If you have just been told that you possess a positive trait, this
motivation is lessened and thus resistance to persuasion is less likely.53
Description
Figure 4.4 Zaller’s Mainstream and Polarization Effects during Vietnam War
Era
Source: Adapted from Figure 6.1 of John R. Zaller, The Nature and Origins of Mass Opinion (Cambridge,
UK: Cambridge University Press, 1992), 103.
How we think of our attitudes is also related to whether those attitudes are resistant to persuasion.
Imagine for a minute that the 2020 Democratic presidential primary is occurring in your state and that
the only two candidates on the ballot are Elizabeth Warren and Bernie Sanders. If you prefer Warren,
this preference might exist because you support her or because you oppose Sanders. As it turns out, if
you think about your opinion in terms of opposition, it will be more resistant to change. A series of
experiments demonstrated this effect.54 Participants were given information about fictitious candidates
and were encouraged to think of their candidate preference in terms of support or opposition. Note that
they were not encouraged to prefer a different candidate but, rather, to think about their attitude in terms
of supporting their preferred candidate or opposing the preferred candidate’s opponent. They were then
exposed to a counterattitudinal message (such as learning that their preferred candidate had been
involved in a scandal) and then asked their candidate preference again. Participants who thought of
their initial attitude in terms of opposition showed less attitude change after receiving the
counterattitudinal message. The researchers argue that this result occurs because negative information
(such as opposition) is a more significant contributor to attitudes than are positive details, thus making
negatively framed attitudes more resistant to change. In fact, the powerful effect of negative information
on evaluations of politicians has been well documented in public opinion research, perhaps helping us
to understand why negative campaign advertisements are effective.55
Motivated Reasoning
When forming political opinions, most people are motivated in pursuit of one or more goals.
Psychologists and political scientists have identified two types of goals that are particularly likely to
motivate political reasoning: accuracy goals and directional goals. The former motivate people “to seek
out and carefully consider relevant evidence so as to reach a correct or otherwise best conclusion,”
whereas directional goals direct a person to “apply their reasoning powers in defense of a prior, specific
conclusion.”56 In the American context, where so much of politics is organized around political parties,
this “specific conclusion” is often one that is consistent with one’s partisanship. When partisan
motivated reasoning occurs, citizens will search out information that coincides with their partisanship
and quickly accept such supporting information without giving it much thought. At the same time,
reasoners motivated by directional goals “strive to maintain their existing evaluations [by] discounting,
counterarguing, and otherwise dismissing information running counter to their preferences.”57 In these
ways, party identification can act as a screen through which incoming information is filtered.
Box 4.1 Political Discussion in Social Networks
Public Opinion in Comparative Perspective
Many forms of communication can influence people’s political attitudes. Persuasive messages
transmitted through the mass media are just one example, albeit an important one. Some people
also are exposed to political topics and arguments through conversations that occur in their
social network. Social networks consist of “people with whom a person communicates on a
direct, one-to-one basis.”1 Mutz focuses her attention on in-person communication rather than
communication via online networks, such as social media platforms that have become more
popular since her research was conducted.
Conversations with family, friends, and others in one’s network cover a wide range of topics, of
course. How likely is it that politics is discussed? As it turns out, there is substantial variation
across nations in how often people discuss politics with other citizens. Political conversations are
common in Israel, Norway, the Netherlands, and Greece but much less likely to occur in Great
Britain, Singapore, and Canada. The United States falls in between these nations. In comparison
with eighty other nations, U.S. citizens engage in political discussions slightly more than
average.2
Distinct from the frequency of political conversations is the composition of one’s social network,
particularly the degree to which networks include people whose political views differ from their
own. Examining survey data from twelve nations, Diana Mutz compared people’s political beliefs
with those of their conversation partners (excluding spouses). The similarity of views is much
higher across discussion partners in the United States compared with other nations. In other
words, on a day-to-day basis, Americans are less likely to be exposed to political beliefs and
arguments that differ from their own. Mutz speculates that Americans’ residential choices
contribute to their less diverse social networks. Most people do not select their neighborhood
based on the political views of their neighbors, of course. Yet this decision can be based on core
values that are related to political beliefs. As Mutz puts it, “People are likely to choose
environments because they are populated by ‘people like me’ in the sense of shared lifestyles,
values, or even market position.”3
Does technology facilitate or impede social network diversity? Among people living in emerging
economies, such as India, Kenya, and Lebanon, smartphone users have more diverse
networks.4 This is particularly the case for those who spend a lot of time on Facebook or other
social media platforms. Social media users in Colombia, for example, are 17 percentage points
more likely to interact with people of a different religion than are nonusers. Colombians who use
social media are also 19 and 22 points more likely to engage with, respectively, people of a
different race or ethnicity or who support a different political party. Similarly large differences in
network diversity exist between social media users and nonusers in Tunisia and Mexico. Note
that the Pew survey on which this report is based asked about interactions that occur in
respondents’ daily lives, but did not specify whether these interactions take place in-person or
electronically. It is reasonable to conclude respondents had both types in mind.
1 Diana C. Mutz, Hearing the Other Side: Deliberative versus Participatory Democracy
(Cambridge, UK: Cambridge University Press, 2006), 10.
2 Ibid., Chap. 2.
3 Ibid., 47.
4 Laura Silver and Christine Huang, “In Emerging Economies, Smartphone and Social Media
Users Have Broader Social Networks,” Pew Research Center, Washington, DC, August 22,
2019, https://www.pewinternet.org/2019/08/22/in-emerging-economies-smartphone-and-social-
media-users-have-broader-social-networks/
We could provide many examples of partisan motivated reasoning among the American public.58
Partisans are more likely to support a novel policy proposal when it is endorsed by a politician of their
own party compared to a member of the other party.59 In other words, the partisanship of an endorsing
politician can strongly influence public attitudes toward a new policy initiative, even before many details
of the initiative are known. The relationship between presidential approval and the state of the economy
is another example. Recall that earlier in this chapter we reviewed research concluding that approval of
presidents is no longer very responsive to changes in economic conditions. Citizens of the president’s
party tend to support him and out-party citizens tend to oppose him, no matter the state of the economy.
To explain their finding, the authors of this work turn to partisan motivated reasoning. Instead of
grounding approval in real-world conditions or the president’s actual performance in specific policy
domains, citizens turn to their partisanship to determine whether they approve of the president.60
Partisan motivated reasoning can also influence the interpretation of factual information, as
demonstrated by Brian Gaines, James Kuklinski, Paul Quirk, Buddy Peyton, and Jay Verkuilen’s study
of public opinion about the Iraq War.61 Gaines et al. conducted panel surveys with University of Illinois
students between October 2003 (roughly six months after the U.S. invasion of Iraq) and December 2004
(soon after the United States turned authority over to the Iraqi Provisional Government). These scholars
examined whether partisanship influenced (1) citizens’ factual beliefs about the situation in Iraq, (2) the
interpretation of those facts, (3) opinions about George W. Bush’s handling of the war in Iraq, and (4) the
relationship among facts, interpretations, and opinions about Bush.
Gaines et al. analyzed whether some partisan groups were more accurate than others when asked how
many U.S. troops had been killed in Iraq. They found that citizens of all partisan stripes were fairly
accurate in assessing the number of casualties. Moreover, as the number of troops killed in action
increased over time, citizens updated their estimates accordingly, regardless of their party identification.
However, despite the accuracy of their beliefs about the number of troop casualties, citizens’ opinions of
Bush’s handling of the war diverged a great deal based on partisanship. Strong Republicans were the
most approving of Bush, with weak Republicans the next most supportive group. At the other end of the
continuum, strong Democrats were the most disapproving of the president, with weak Democrats not too
far behind.
Now, all groups basically agreed on the number of casualties, so why did their opinions of Bush differ so
much? To answer that question, Gaines et al. examined how the partisan groups interpreted the factual
information about casualties. Specifically, they asked, “When you think about [the number of] U.S.
troops being killed in the military action in Iraq since the May announcement that major combat
operations had ended, do you think of that number as very large, large, moderate, small, or very
small?”62 An overwhelming proportion of strong Democrats indicated that the number of casualties was
either very large or large. In sharp contrast, approximately 80 percent of strong Republicans saw the
casualties as moderate, small, or very small. Even as casualties rose across time, strong Republicans
did not change their interpretation of the number of casualties. Overall, the results clearly suggest that
citizens’ party affiliations shaped how they interpreted the number of casualties.
Thus far, the results we have discussed compare the beliefs, interpretations, and attitudes of different
partisan groups. Gaines et al. also conducted an individual-level analysis to sort out whether it was the
factual beliefs or the interpretation of those beliefs that influenced citizens’ attitudes toward Bush’s
handling of the war. They found that factual beliefs about the number of casualties had no effect on
citizens’ evaluations of the president, whereas people’s interpretation of how large the casualties were
had a substantial influence on their views of the president. “In other words, the meanings that people
gave to their factual beliefs about troop losses, not the beliefs themselves, drove their opinions toward
https://www.pewinternet.org/2019/08/22/in-emerging-economies-smartphone-and-social-media-users-have-broader-social-networks/
the war.”63 Notably, these meanings were heavily influenced by party identification and, thus, can
explain why Republicans viewed Bush’s handling of the war much more favorably than did Democrats.
Gaines et al. demonstrate that even when Democrats and Republicans agree on the facts in a situation,
they might have very different understandings of what those facts mean. Research conducted by Brian
Schaffner and Cameron Roche concludes that partisan motivated reasoning can also produce partisan
disagreement on the facts themselves.64 The 2012 presidential election was waged in large part over
who was best prepared to bring the U.S. economy fully out of recession: the sitting president, Barack
Obama, or his challenger Mitt Romney. The early October release of the monthly jobs report was thus
greatly anticipated by many. This report is prepared by a government agency, the Bureau of Labor
Statistics, that is generally perceived to be credible and nonpartisan. In October 2012, they announced
that the unemployment rate in September had dipped to 7.8 percent, the lowest during Obama’s
presidency to date. Media attention to this news was widespread. Furthermore, many outlets noted that
the report would likely benefit Obama’s reelection efforts, such as this example: “The improvement [in
the unemployment rate] lent ballast to Mr. Obama’s case that the economy is on the mend and
threatened the central argument to Mitt Romney’s candidacy, that Mr. Obama’s failed stewardship is
reason enough to replace him.”65
How did the public respond to this information? In particular, was Democrats’ and Republicans’
knowledge of the actual unemployment rate accurate? A nationally representative survey was in the
field during the month of October, with some people completing the survey before the monthly jobs
report was released and some completing it after. Schaffner and Roche used this naturally occurring
experiment to examine responses to a factual question about the unemployment rate before and after
the report’s release, separately by partisanship at both times. They found that Democrats’ guesses
about the unemployment rate were on average lower after the report was released than before,
accurately reflecting the fact that the rate had dropped. On the other hand, Republicans’ estimates of
the jobless rate were, on average, higher after the release than before. Both patterns can be attributed
to partisan motivated reasoning. In this situation, partisan-directed goals coincide with providing an
accurate response for Democrats but not for Republicans. Partisan motivations took over for many
Republicans, either in the form of counterarguing the unemployment news (“I know I saw that the rate
had dropped to 7.8 percent, but that can’t be correct because it runs counter to what I think of Barack
Obama’s effectiveness as president, so it must be higher”) or partisan cheerleading (“I know the rate is
7.8 percent, but that is a good number for Democrats, so I’m going to say it is actually higher to express
my opposition to Barack Obama”).66
Research on partisan motivated reasoning demonstrates, over and over again, that party identification
functions as a critical filter for citizens. Partisans, especially strong ones, are motivated to interpret the
world in a way that is consistent with their predispositions. We were revising this textbook during the
early weeks of the U.S. House of Representatives’ presidential impeachment inquiry of Donald Trump,
an event that seems likely to invoke partisan motivated reasoning. The inquiry was launched in
September 2019 following allegations that Trump threatened to withhold military aid to Ukraine unless
officials in that country investigated one of Trump’s Democratic opponents in the 2020 presidential
contest. Not surprisingly, Democratic citizens were strongly in favor of the House initiating an
impeachment inquiry, whereas Republicans were not.67 Partisans also differed in their assessments of
Trump’s alleged actions regarding Ukraine. Polling conducted in October 2019, for instance, indicated
that 59 percent of Republican citizens considered Trump’s request that Ukraine investigate a
Democratic presidential candidate to be appropriate and that 72 percent of Republicans believed Trump
was acting in the national interest in his interactions with Ukraine. In contrast, 84 percent of Democrats
thought Trump’s request was inappropriate and a whopping 94 percent believed he was acting in his
own personal interest.68
Attitudes Follow Behavior
Many people prefer to maintain consistency across their beliefs, attitudes, and attitude-relevant
behaviors. Someone who attends a rally in support of the Republican Party, for example, probably has a
positive attitude toward the group and views the attributes of the party positively (such as believing that
Republicans support policies that are good for the country). If inconsistencies exist among these
elements, individuals will act to restore order. If someone else has negative attitudes toward the
Republican Party but still attends a rally, perhaps at the request of a friend, they might change their
attitudes so that they are positive. These examples illustrate key postulates of cognitive dissonance
theory, presented by Leon Festinger in the 1950s.69 Festinger argued that when an individual holds
cognitive elements (beliefs, attitudes, or knowledge of one’s behaviors, for example) that are in
opposition, cognitive dissonance exists. This dissonance produces negative arousal and the motivation
to either eliminate the dissonance by changing the dissonant cognitive element or reduce the
dissonance by adding more consonant elements. The person who attends a Republican Party rally but
has a negative attitude toward the party might be unable to change this attitude but could instead
reduce dissonance by thinking about positive characteristics of the party. They might, for instance, think
that the Republican Party is better than the alternatives and that some Republican politicians are
competent and pursue policies that are good for the country.
Festinger and colleagues conducted a number of studies in which they demonstrated that if people are
forced to engage in behavior that contradicts their attitudes, their attitudes can change as a result.70
Once the behavior is performed, after all, it cannot be unperformed, so attitude change is one way to
eliminate the discrepancy between the behavior and the attitude. Although this theory was not
developed specifically for understanding political attitudes, political opinions could respond to behavior,
as recent work in political science suggests.71 For instance, citizens could first vote or otherwise
express their support for a candidate, then form or even change their attitudes toward specific policies
so that their attitudes coincide with the candidate’s. Ethnic or racial prejudice could also develop from
behavior. In particular, engaging in racially hostile actions (whether physical or verbal) could produce
prejudicial attitudes in order to reduce or eliminate cognitive dissonance. This framework has been
applied to partially account for racism in the American South in the long aftermath of slavery and while
racial segregation was prevalent in that region as well as anti-Semitism in pre–World War II Germany.72
In a challenge to cognitive dissonance theory, Daryl Bem proposed a different explanation to account for
the findings from cognitive dissonance studies. Bem’s approach, called self-perception theory, argued
that attitude change such as that observed in Festinger’s research was not the result of correcting an
unpleasant state of dissonance but, rather, occurred because people infer their attitudes from their
behaviors. Bem describes his theory as follows:
When we want to know how a person feels, we look to see how he acts. Accordingly, it seemed
possible that when an individual himself wants to know how he feels, he may look to see how
he acts, a possibility suggested anecdotally by such statements as, “I guess I’m hungrier than I
first thought.” … Individuals come to “know” their own attitudes, emotions and other internal
states partially by inferring them from observations of their own overt behavior and/or the
circumstances in which this behavior occurs.73
Bem’s self-perception theory does not presume that people use their behaviors as a guide to determine
all their attitudes. Instead, the theory assumes that such self-perception is most likely when attitudes are
weak or ambiguous. Because of this, some have argued that self-perception is better at accounting for
attitude formation than attitude change.74 If a person is uncertain about their attitude, using their recent
behavior as a guide can help shape a new attitude. If, in contrast, the person’s attitude is well developed
and strong, it is unlikely to change in the face of a contradictory behavior. Perhaps more important, the
stronger individuals’ attitudes are and the more certain they are about these attitudes, the less likely
they are to even engage in behavior that contradicts the attitudes. In short, whether through cognitive
dissonance reduction or self-perception mechanisms, it seems as if attitudes can follow from behavior.
CONCLUSION
In summarizing Bem’s self-perception theory, psychologists Alice Eagly and Shelly Chaiken write:
In general self-perception theory maintains that people function as relatively superficial
information processors who merely generalize their attitudes from currently available external
cues. By this account, people look to see what their recent behavior has been and assume that
their attitudes are congruent with this behavior.75
This view of attitudes—as shallow and changeable—is also present in some of the other approaches
that we have reviewed in this chapter. Zaller, for one, assumes that most people, especially those who
are not highly politically aware, do not possess stable attitudes but, rather, hold a mix of attitude-relevant
considerations in their heads. People’s opinions at any point in time will reflect which of these
considerations they bring to mind. This model thus provides support for elite democratic and pluralistic
theorists’ assumptions regarding the citizenry, especially their views that the public’s political opinions
are ill considered and subject to frequent changes. Yet Zaller’s model indicates that these conclusions
are not appropriate for all people and for all situations. Reality, in other words, is more complex than
some elite democrats and pluralists imply.
Eagly and Chaiken go on to describe another way of understanding attitudes: “A view of attitudes as
relatively enduring tendencies suggests instead that people generalize their attitudes from internal data;
they have stored their attitudes in memory and retrieve them from this internal source when called upon
to make an attitudinal judgment.”76 From this description, we envision attitudes as enduring and stable,
as judgments that are well linked to core values and beliefs. Enduring attitudes are likely if these
attitudes serve important functions for individuals. If people hold attitudes because they serve an ego-
defensive or a value-expressive function, the attitudes will be fairly resistant to change. Learning
theorists remind us that attitudes formed from repeated exposure to attitude objects have become
habitual and are also unlikely to change. Even though these attitudes might be stable, the functional and
learning approaches do not assume that all stable attitudes have been carefully considered. In contrast,
learning theorists assume that people rather passively acquire attitudes without thinking through why
they might like or dislike an attitude object. Ego-defensive attitudes, such as racial prejudice, provide
similar examples of superficial thought. When feeling threatened, some individuals respond by
expressing discriminatory attitudes toward members of other racial groups rather than considering the
actual source of the threat, which could in fact be their own feelings of inferiority.
Some democratic theorists have assumed and some political scientists have demonstrated empirically
that citizen attitudes are quite pliable. One of our goals in this chapter has been to present these
arguments and this evidence. We also have presented alternative evidence to encourage you to
consider whether attitude instability should be viewed negatively, that is, as evidence that the citizenry is
not living up to the model presented by classical democratic theorists. Page and Shapiro, for example,
demonstrate that when collective opinion changes it is often in rational response to external cues, such
as world events or a changed social climate. Likewise, presidential approval can be driven by events,
including international crises and economic conditions. Further, we turned to psychology to demonstrate
the variety of approaches available to understand the nature of attitudes and the mechanisms by which
they can change. Models of communication-induced attitude change and motivated reasoning are
especially relevant for political attitudes. Much political activity is designed to try to influence the
opinions and judgments of political actors, whether they are citizens or politicians. Understanding
whether and how political communication influences citizens’ opinions has been the focus of much
public opinion research.
At this point, you should have some thoughts about the following questions: are citizen attitudes
fleeting? If they are, do we care? In particular, does the presence of unstable citizen attitudes undermine
the public’s ability to evaluate political issues and communicate its preferences to political leaders? Is
democratic governance, in other words, threatened by attitude instability?
KEY CONCEPTS
acceptance 132
attitude change 117
attitude instability 119
attitude stability 117
cognitive dissonance theory 143
collective public opinion 121
considerations 134
cross-sectional study 118
ego defense 129
elite discourse 132
expectations gap 127
financial costs of war 127
functional theories 129
honeymoon period 126
knowledge function 129
learning theory 131
mainstream effect 135
nonattitudes 120
objective measures of economic well-being 127
panel study 118
partisan motivated reasoning 139
perceptions of the state of the economy 127
polarization effect 135
political awareness 132
political predispositions 133
presidential approval 125
rally round the flag 127
receive-accept-sample model 134
reception 132
resistance 136
self-perception theory 143
social adjustment 130
social networks 138
tracking polls 125
transfer of affect 131
utilitarian function 129
value-expressive function 130
war casualties 127
SUGGESTED SOURCES FOR FURTHER READING
Albarracin, Dolores, Man-pui Sally Chan, and Duo Jiang, “Attitudes and Attitude Change: Social and
Personality Considerations about Specific and General Patterns of Behavior.” In The Oxford Handbook
of Personality and Social Psychology, 2nd ed., ed. Kay Deaux and Mark Snyder. Oxford: Oxford
University Press, 2018.
Banaji, Mahzarin R., and Larisa Heiphetz. “Attitudes.” In The Handbook of Social Psychology, 5th ed.,
Vol. 1, ed. Susan T. Fiske, Daniel T. Gilbert, and Gardner Lindzey. Hoboken, NJ: John Wiley & Sons,
2010; updated version published November 2015, DOI: 10.1002/9780470561119.socpsy001010.
These chapters provide detailed discussions of recent research on attitudes in the field of psychology.
Arceneaux, Kevin, and Ryan Vander Wielen. Taming Intuition: How Reflection Minimizes Partisan
Reasoning and Promotes Democratic Accountability. New York: Cambridge University Press, 2017.
The authors of this book demonstrate that partisan motivated reasoning is less likely to occur when
people process new information in a careful, reflective way.
Gronke, Paul, and Brian Newman. “FDR to Clinton, Mueller to? A Field Essay on Presidential Approval.”
Political Research Quarterly 56 (2003): 501–512.
Donovan, Kathleen, Paul M. Kellstedt, Ellen M. Key, and Matthew J. Lebo, “Motivated Reasoning, Public
Opinion, and Presidential Approval,” Political Behavior, prepublished February 23, 2019, DOI:
10.1007/s11109-019-09539-8.
Gronke and Newman’s article provides an excellent review of the early research on presidential
approval. Donovan and her coauthors’ more recent work links approval to theories of motivated
reasoning.
Harrison, Brian F., and Melissa R. Michelson. Listen, We Need to Talk: How to Change Attitudes about
LGBT Rights. New York: Oxford University Press, 2017.
Through a series of innovative experiments, Harrison and Michelson conclude that people can be
persuaded to change their attitudes toward gay and lesbian rights when exposed to an unexpected
message from someone with whom they share an identity. Professional football fans, for instance,
became more accepting of LGBT rights when a supportive message was communicated by the
Commissioner of the National Football League.
Zaller, John R. The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press,
1992.
The Nature and Origins of Mass Opinion after 20 Years. Special issue of Critical Review, 24 (2012).
Zaller’s book introduced his receive-accept-sample model. The special issue of Critical Review provides
a more recent assessment of the model.
Gallup Presidential Job Approval Center, https://news.gallup.com/interactives/185273/presidential-job-
approval-center.aspx
“How Popular Is Donald Trump?,” FiveThirtyEight, https://projects.fivethirtyeight.com/trump-approval-
ratings/?ex_cid=rrpromo
These two websites provide extensive presidential approval data. The Gallup site allows you to analyze
levels of presidential approval among Democrats, Republicans, and Independents as well as by various
demographic characteristics.
https://news.gallup.com/interactives/185273/presidential-job-approval-center.aspx
https://projects.fivethirtyeight.com/trump-approval-ratings/?ex_cid=rrpromo
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists years and ranges from 1971 to 1989 in increments of 2. The vertical axis lists
percentage support and ranges from 0 to 100 in increments of 10. Approximate data from the graph are
tabulated as follows:
Year
Percentage who
think too little is
spent on fighting
crime
Percentage who
think too little is
spent on education
Percentage who think
too little is spent on
Social Security
Percentage who
think too little is
spent on welfare
1971 73 49 No data 20
1973 70 51 No data 21
1975 71 52 No data 21
1977 70 58 No data 11
1979 74 56 No data 11
1981 78 56 No data 20
1983 70 60 No data 20
1985 71 62 51 24
1987 71.5 62 56 20
1989 75 67 56 22
Back to Figure
The horizontal axis lists years and ranges from 2009 to 2019. Barack Obama’s presidency was between
2009 and approximately mid-2016. Donald Trump’s presidency is after mid-2016 to present. The vertical
axis lists the approving percentage and ranges from 0 to 100 in increments of 10. Approximate data
from the graph are tabulated as follows:
Year Percent approving
Year Percent approving
2009 68
2010 50
2011 45
Mid-2011 40
2012 45
Late 2012 52
2013 50
2014 41
2015 45
2016 50
Mid-2016 45
2017 35
2018 37
2019 40
Back to Figure
The horizontal axis is labeled political awareness and lists low, moderate, and high. The vertical axis is
labeled probability of attitude change and ranges from 0 to 1 in increments of 0.1. Approximate data
from the graph are tabulated as follows:
Level of political awareness Reception Acceptance Attitude change
Level of political awareness Reception Acceptance Attitude change
Low 0.1 0.9 0.1
Moderate 0.5 0.5 0.4
High 0.9 0.1 0.9
Back to Figure
In both the graphs, the horizontal axis is labeled political awareness, listing low, moderate, and high,
and the vertical axis is labeled percentage support Vietnam War, ranging from 0 to 100 in increments of
20.
Graph A is titled, The mainstream effect, 1964. Approximate data from the graph are tabulated as
follows:
Level of political
awareness
Percentage of conservatives
supporting Vietnam War
Percentage of liberals
supporting Vietnam War
Low 40 40
Low to moderate 60 52
Moderate 68 55
Moderate to high 63 55
High 77 75
The second graph is titled, B. The polarization effect, 1970. Approximate data from the graph are
tabulated as follows:
Level of political
awareness
Percentage of conservatives
supporting Vietnam War
Percentage of liberals
supporting Vietnam War
Low 43 48
Low to moderate 57 62
Level of political
awareness
Percentage of conservatives
supporting Vietnam War
Percentage of liberals
supporting Vietnam War
Moderate 61 60
Moderate to high 75 48
High 75 37
PART III DO CITIZENS
ORGANIZE THEIR POLITICAL
THINKING?
ARE PEOPLE’S POLITICAL OPINIONS on different issues related
to one another? That is, is there some consistency across views, or
does knowing citizens’ views on one issue not help predict their
views on other issues? Assessing consistency can be tricky, but one
common yardstick is political ideology. With this approach, a person
with all conservative views would be considered to have more
consistent attitudes than someone with a mixture of liberal and
conservative views. But is ideology the best yardstick? And, if it is,
what is the best way to measure the degree of ideological thinking
and ideological organization of people’s political opinions? The first
chapter of Part III opens with these topics. That chapter, as well as
the other two chapters in this part, then move beyond ideology to
consider a range of factors that might shape people’s attitudes.
Partisanship, personality, self-interest, core values, historical events,
and social groups are the topics explored.
Why should we care if the public organizes its political thinking?
Answering this question brings us back to the normative topic of
citizen competence. Can citizens function effectively in a democracy
if their political views are not well organized? The chapters in this
section speak to questions such as these; we hope you will ponder
them as you proceed.
CHAPTER 5 IDEOLOGY, PARTISANSHIP, AND
POLARIZATION
“With just over two weeks to go until Election Day, the country looks and sounds more like the
Divided States of America.”1
“We disagree, fiercely, about almost everything. We disagree about terror and security, social
justice, religion in politics, who is fit to be a judge, and what democracy is.”2
“Americans on both the left and the right now view their political opponents not as fellow
Americans with differing views, but as enemies to be vanquished.”3
PROBABLY NOTHING in the preceding quotations surprises you. If we know anything about American
politics today, after all, it is that the public is politically divided. Deeply divided. Bitterly divided. Right?
Actually not. Conventional wisdom notwithstanding, Americans do not disagree so strongly about many
political issues. Sure, public divisions exist, but they do not divide the public so far apart as is commonly
assumed. There is something to the final quotation above, however: people’s views of their (real or
perceived) political foes have turned markedly negative. We pick up these topics in our discussion of
political polarization at the end of the chapter.
First, though, the chapter focuses on political ideology and partisanship. News coverage of politics and
government often conveys the ideological or partisan leanings of political figures, such as members of
Congress, Supreme Court justices, interest groups, and commentators. Are ideology and party also key
identifications for citizens? That is, do members of the public think about politics in ideological terms or
does ideology help to structure people’s opinions toward specific issues? Not so much, as we discuss
next. We then turn to the topic of partisanship, demonstrating that party affiliation is a core political
identity for many Americans. Throughout, as always, we will tie specific topics to normative democratic
theory.
CONVERSE’S CLAIM: IDEOLOGICAL INNOCENCE
Liberalism and conservatism are the two dominant ideologies in U.S. politics. A political ideology is “an
interrelated set of attitudes and values about the proper goals of society and how they should be
achieved.”4 Put another way, “If an attitude is a strand of feeling, then an ideology is a rope of
intertwined attitudes and related fibers.”5 Two aspects of these definitions are worth emphasizing. First,
an ideology consists of attitudes that are coherent and related to one another.6 Second, an ideology
does not refer to just any set of related attitudes but, rather, to beliefs about society and especially the
proper role of government. In the American context, conservatives emphasize order, tradition,
individual responsibility, and minimal government intervention, particularly in economic matters.
Liberals, in contrast, believe that government intervention in the economy is sometimes necessary to
combat features of the free market (such as discrimination and low wages). Liberals also value equality,
openness to dissenting views, and civil rights.
According to many democratic theorists, citizens and politicians need to communicate effectively with
one another so that, among other reasons, citizens can evaluate the performance of elected officials
and these officials can know the political preferences of the citizens. Communication between citizens
and leaders is enhanced if the two groups talk about politics using the same terms. This does not occur
in the United States, however, at least according to Philip Converse’s classic work on political ideology.
Conducting his research in the late 1950s and early 1960s, Converse concluded that political elites are
much more likely than citizens to organize the political world ideologically along a liberal-conservative
continuum. Not only are citizens less likely to think about politics ideologically, the terms liberal and
conservative carry little meaning for many people. Public understanding of political debates is
threatened by such a lack of understanding. As Converse put it, “The more impoverished [a citizen’s]
understanding of the term [conservative or liberal], the less information [the term] conveys. In the limiting
case—if he does not know at all what the term means—it conveys no information at all.”7 Further, low
knowledge of ideology and infrequent ideological reasoning is, at least to some, evidence that the public
is not capable of democratic citizenship. Converse’s work ignited a firestorm of research, with many
scholars trying to resurrect a more respectable view of citizens’ capabilities. After detailing Converse’s
original argument and the methodology on which his study relies, we discuss some of the research that
challenges Converse’s arguments.
The overarching goal of Converse’s research was to examine the belief systems of citizens and elites.
He defines a belief system as “a configuration of ideas and attitudes in which the elements are bound
together by some form of constraint or functional interdependence.”8 Although Converse prefers the
term belief system rather than ideology, he does admit that the two are closely related. Further, as we
will soon see, to determine whether the public’s beliefs are joined in coherent systems, he uses the
liberal-conservative ideological dimension as one of his gauges. As for attitude constraint, Converse
refers to the degree to which we could predict a specific attitude of someone knowing her attitude
toward a different political object. When a belief system is present, “if a person is opposed to the
expansion of social security,” we can judge that “he is probably a conservative and is probably opposed
as well to any nationalization of private industries, federal aid to education, sharply progressive income
taxation, and so forth.”9 Constraint, for Converse, means that people’s political attitudes are related to
each other because they derive from an overarching worldview (such as a political ideology).
Do People Demonstrate Ideological Thinking?
Converse’s overall conclusion, as we have already mentioned, was that elites were much more likely to
possess belief systems compared with the general public. This conclusion is based primarily on his
analyses of the 1956, 1958, and 1960 American National Election Studies (ANES) panel study. To
understand and evaluate Converse’s conclusions, you need to know what evidence he used to make his
arguments. Thus, we summarize Converse’s study in depth. For his first analysis, Converse examined
the degree to which respondents in 1956 used ideological language in response to questions about the
political parties and the two major-party candidates for president. This series of questions began: “Is
there anything in particular that you like about the Democratic Party? Is there anything in particular that
you don’t like about the Democratic Party?” Respondents were then asked for their likes and dislikes of
the Republican Party and the candidates (Democrat Adlai Stevenson and Republican president Dwight
Eisenhower).
Such open-ended questions allow respondents to discuss politics using their own terms and language,
thus providing important insights into how people conceive of the political world. From his analysis of the
responses, Converse categorized the public into five groups based on the degree to which people used
an abstract benchmark, such as the liberal-conservative ideological continuum, to evaluate the parties
and politicians.10 Those individuals who did use this continuum, such as by differentiating the parties
based on ideology and correctly linking specific policy positions of the parties to this ideology, were
labeled Ideologues.11 The second group—Near Ideologues—included people who used ideological
labels such as liberal or conservative but perhaps did not fully understand the meaning of these terms or
did not use ideology as their primary tool for evaluating politics. One example here is a man who liked
both the “liberalness” of the Democrats and the “conservative element in the Republican Party.”12 All
told, only about 11.5 percent of the public was classified as either an Ideologue (2.5 percent) or a Near
Ideologue (9 percent; see the lighter bars in Figure 5.1).
Description
Figure 5.1 Levels of Conceptualization among the American Public, 1956
and 2000
Sources: Adapted from Table I of Philip E. Converse, “The Nature of Belief Systems in Mass Publics,” in
Ideology and Discontent, ed. David E. Apter (New York: Free Press, 1964), 218, and Table 10.1 of Michael
S. Lewis-Beck, Helmut Norpoth, William G. Jacoby, and Herbert F. Weisberg, The American Voter Revisited
(Ann Arbor: University of Michigan Press, 2008), 279.
Far more common were people who conceptualized politics in terms other than ideology. The largest
category was Group Interest citizens, who made up 42 percent of the respondents. These individuals
tended to discuss the parties and candidates in terms of whether they favor the interests of specific
groups, such as the man who disliked Republicans because “they are more for big business” or the
woman who liked that the Democrats “have always helped the farmers.”13 Nature of the Times was the
label Converse applied to his fourth group. People in this group (24 percent of the public) linked the
parties or candidates with the current state of the nation. More specifically, parties in charge during
times of peace or prosperity were evaluated more favorably than were those who ruled during war or
economic downturns. An example from the 2000 ANES nicely illustrates this type of reasoning. One
respondent explained his dislike of Republican candidate George W. Bush as follows: “He is a
Republican. I have been in construction for the last 33 years and every time there has been a
Republican in office, I’ve been in the unemployment line.”14
The final group in Converse’s classification evaluated the parties and candidates on grounds other than
issues. No Issue Content citizens included those who used personal characteristics to evaluate
candidates, were not sure what either political party stood for (even when they identified with one of the
parties), or did not follow politics closely enough to discuss parties or candidates. One person in 2000,
for instance, liked Bush because of “his sincerity. He surrounds himself with good people and he is well
connected.”15 This final group made up 22.5 percent of the citizenry, nearly twice that of the Ideologues
and Near Ideologues combined. To Converse, these results clearly demonstrated that most members of
the public do not think about political parties and candidates ideologically.
Replication of this part of Converse’s study in recent decades has been rare. Most recently, Michael
Lewis-Beck, William Jacoby, Helmut Norpoth, and Herbert Weisberg have done so, allowing us to see
whether Converse’s findings have stood the test of time.16 Relying on the likes and dislikes questions in
the 2000 ANES (regarding George W. Bush, Al Gore, and the Republican and Democratic parties),
these researchers categorized respondents into the five groups first created by Converse. As
demonstrated by the darker bars in Figure 5.1, the number of Ideologues among the public was higher
in 2000 than in the 1950s, whereas the Group Interest category was smaller. Despite these changes,
the two largest categories continued to be Group Interest and Nature of the Times, each containing 28
percent of the public. Following closely behind was No Issue Content with 24 percent. As was the case
in the 1950s, in 2000 these three categories were all significantly larger than the Ideologues (10.5
percent) or Near Ideologues (9 percent).
Do People Recognize Ideological Terms?
Moving on and mostly moving away from open-ended questions, Converse next assessed the degree to
which people could recognize the terms liberal and conservative. Even if ideological reasoning was
uncommon among citizens, public understanding of these terms could be more common. To address
this possibility, in 1960 Converse asked respondents, “Would you say that either one of the parties is
more conservative or more liberal than the other?” Those answering yes were then asked which party is
more conservative and then why they characterized that party as more conservative. Nearly 40 percent
of the respondents either did not recognize these terms or were unable to attach any meaning to the
terms.
Among those who did identify the ideological leanings of each party and did attempt to discuss the
meaning of conservatism, there was variation in the correct use of the terms and in the breadth of
ideological thinking apparent in the answers. Converse concluded that about 17 percent of the
respondents did not correctly apply the terms or did not provide a correct meaning for conservatism,
whereas 29 percent provided correct meaning but demonstrated only a narrow understanding of the
ideologies. Typically, these respondents discussed ideology only in terms of which party spends more
money and which saves more. Republicans are more conservative, one person explained, because
“they vote against the wild spending spree the Democrats get on.”17 The remaining respondents, about
17 percent, recognized the ideological terms, identified the Democrats as liberal and the Republicans as
conservative, and displayed a more thorough understanding of liberalism and conservatism. Although
this number is larger than the 11.5 percent that displayed ideological thinking in response to the open-
ended questions about parties and candidates, it is still a small percentage of the public.
Are Individuals’ Attitudes Constrained and Stable?
So far, Converse’s analyses suggest that most people do not use the liberal-conservative ideological
spectrum to organize their political thinking. This, however, does not mean that the political views of
most people are unorganized. Perhaps beliefs are organized along other criteria. Attitude constraint, in
other words, may exist among the public even though the political worldview that is constraining the
attitudes is not liberal or conservative ideology. Converse tested this assumption two ways. First, using
the 1958 ANES data, he looked at the relationship between a number of issue opinions to see whether
opinions toward an issue (such as federal education aid) correlated with opinions on another issue
(such as public housing). Because liberals tend to support federal government spending on education
and government provision of public housing, and conservatives tend to oppose both, if most of the
public organize their issue opinions along this ideological continuum, we would expect these opinions to
be highly correlated among the public. Yet what if people who support federal education aid also tend to
oppose public housing? This would suggest a different organizing framework. If this is the case, we
would still see high correlations between the attitudes, albeit in the opposite direction, with support for
one issue correlated with opposition to the other.
Examining the relationships between four domestic issues and three foreign affairs issues, Converse in
fact found very low correlations among the public, leading him to dismiss the possibility that the public’s
beliefs are constrained along any dimension. Further, he compared the correlations of the public with
those of political elites (in this case, congressional candidates) and found that belief constraint is much
higher among the elites. Figure 5.2 presents the average correlations separately for domestic issues
and foreign issues and then for a comparison between all domestic and all foreign issues. As you can
see, correlations (and thus belief constraint) were higher among elites than the public for all three
comparisons.
Description
Figure 5.2 Relationships between Issue Opinions for the American Public
and Political Elites, 1958
Source: Adapted from Table VIII of Philip E. Converse, “The Nature of Belief Systems in Mass Publics,” in
Ideology and Discontent, ed. David E. Apter (New York: Free Press, 1964), 229.
Second, Converse compared people’s issue attitudes in 1956 with their opinions on the same issues in
1958 and again in 1960. This is the analysis we presented at the beginning of Chapter 4. As you recall,
the levels of attitude stability were quite low. The correlation (expressed in this case with tau-b
coefficients) between opinions on school desegregation in 1958 and 1960 was .43, whereas the
correlation over time on the issue of federal housing assistance was .29. In contrast, respondents’ party
identification was much more stable across these two years (.73), demonstrating that party affiliation
does not change as much as do issue opinions. Further, Converse found that people’s issue opinions
fluctuated as much between 1956 and 1958, and between 1958 and 1960, as they did between 1956
and 1960. Given the longer time frame of the last period, we might expect less stability than over a two-
year period. But this pattern was not apparent in Converse’s data, leading him to conclude that the
public “contains significant proportions of people who, for lack of information about a particular
dimension or controversy, offer meaningless opinions that vary randomly in direction during repeated
trials over time.”18
Groups as a Source of Belief Constraint
Although Converse argued that most Americans did not possess an ideologically constrained belief
system, he did find one source of belief constraint among the public—attitudes toward social groups.
Converse’s respondents were asked two policy questions that referenced African Americans (or
Negroes, the common label in the 1950s). One queried public support for the federal government to
ensure public schools are desegregated, and the other assessed whether the government should
ensure that African Americans are not discriminated against in employment and housing. The
correlation between opinions on these two items was .57, much higher than the average relationship
among public opinion toward the entire range of domestic policy issues that Converse examined (see
Figure 5.2). Further, the correlation between these two items among elites was actually lower than for
the public (.31).
In sum, Converse’s conclusions were that (1) the public does not think about political parties and
candidates ideologically, (2) recognition and correct use of the terms liberal and conservative are quite
rare, (3) constraint across a variety of issue positions is low, and (4) attitude consistency over time is
low. Citizens do appear, however, to organize their political opinions around views of prominent groups.
Elites, in contrast, use ideology to organize their political thinking, as is evident by their higher levels of
attitude constraint. If these results seem to confirm elite democrats’ assumptions that the public is not
well equipped for democratic governance, they should. Empirical findings such as Converse’s led to the
development and refinement of the theory of democratic elitism. Elite democrats assume that the public
is neither engaged in nor well informed about politics, which should contribute to their low levels of
ideological understanding and use of ideology to organize their thinking. Other theorists, particularly
those with a more optimistic view of the public’s capabilities, found Converse’s work limiting and looked
to other explanations to account for his findings.
Critiques of Converse
Converse’s research received much attention at the time of publication, has spurred countless
commentaries and studies (some supporting and some opposing his conclusions), and is still influencing
public opinion scholars today. His work was referred to as “celebrated” and “influential”; but it was also
called “notorious” by one scholar,19 and another described it as an “enduring milestone” and a
“millstone,” the latter because of the “misleading criteria Converse used to assess political competence
and electoral responsibility.”20 As these quotations suggest, Conserve’s work was not well received by
all. Over the years, critiques have come from many quarters. We next briefly summarize and evaluate
key counterarguments.
Were Converse’s results due to the nature of the times? Several people have argued that the 1950s
was an especially nonideological time in the nation’s politics, thus producing the low levels of ideological
thinking measured by Converse. This was a decade of (relatively speaking) consensual politics.
Disagreements between the political parties were minor, the political environment was not dominated by
the discussion of conflictual issues, and the public was not very tuned in to politics. Politics during the
1960s and 1970s was much more ideologically contentious. Battles raged over civil rights, the United
States was involved in what became a controversial war in Vietnam, the economy took a downturn, riots
broke out in many cities, and President Richard Nixon was forced to resign as a result of the Watergate
burglary. These salient issues increased public attention to politics and divided the Democrats and
Republicans quite publicly as the two parties openly debated their differences over these matters.
Did these more ideologically contentious times result in citizen views that were more ideologically
grounded during this time period? Probably. Attitude constraint did tick up a bit in the 1960s and
1970s.21 Some of the increase in 1972, for example, was due to the ideological nature of that year’s
presidential election.22 The two candidates—Republican president Richard Nixon and Democrat George
McGovern—proposed ideologically distinct platforms during their election campaigns. McGovern, in
particular, was frequently described as representing the very liberal wing of the Democratic Party. The
events of and debate over the Vietnam War also seemed to produce public attitudes that were more
stable. In Converse’s own words, “The crescendo of political turmoil associated with the later stages of
the Vietnam war was producing somewhat firmer opinions on key subjects.”23 More generally, attitude
constraint and stability are higher for certain types of issues. Issues that are grounded in religion,
morality, or civil rights show less fluctuation over time and are more tied to ideological principles than are
other types of issues (such as economic or foreign policy).24 The former issues are more likely to be on
the public agenda now than they were in the 1950s. As the issue context changes, then, public belief
systems can become more constrained.
Even so, we have not witnessed very high levels of attitude constraint among the public in the decades
since Converse’s initial work. Figure 5.3 presents average correlation coefficients (Pearson’s r) across
pairs of issues for both 1980 and 2004. For each year, relationships between key issues of the day were
examined.25 In 1980, the average correlation between the issue pairs for the public was only .12,
whereas it was .22 in 2004.26 A similar, albeit broader, analysis of attitude constraint reaches a similar
conclusion.27 Two scholars examined the correlation between all pairs of policy issues appearing on the
ANES between 1972 and 2012. They demonstrate that the average correlation between issue opinions
is .16 and that this correlation has only increased slightly over time (by .01 per decade).
Next we turn to levels of attitude stability in the decades since Converse’s work was published. You can
see in Figure 5.3 that stability is higher than constraint (Converse also found this). However, levels of
attitude stability are not terribly high. These results come from analyses of the 1972–1976 and the
2000–2004 ANES panel studies. For the first period, respondents were asked about three issues
(abortion, school busing, and defense spending) using the same question wording in both 1972 and
1976. The average correlation between people’s responses in these two years was .5.28 More recently,
respondents were asked about several policy issues using the same question wording in both 2000 and
2004. These included preferred levels of government spending on a variety of policies (such as
environmental protections, welfare, child care, and aid to the Black community) as well as abortion and
legal protections against discrimination for gays and lesbians.29 Stability over time across these four
years was slightly lower than in the 1970s, with an average correlation of .44.
Description
Figure 5.3 Attitude Constraint and Attitude Stability among American Public
and Elites
Sources: Attitude constraint for public (1980) and elites (1980) and attitude stability for public (1972–1976)
and elites (1980–1984): Adapted from Figures 1 and 5 of M. Kent Jennings, “Ideological Thinking among
Mass Publics and Political Elites,” Public Opinion Quarterly 56 (1992): 426, 432. Attitude constraint for public
(2004): Analysis of 2004 American National Election Study Data File. Attitude stability for public (2000–
2004): Analysis of American National Election Studies 2000–02–04 Panel Data File.
Recall that Converse compared attitude constraint for the public with that of the elites. Replications of
this portion of his analysis have been rare, but one study did examine both elite and public attitude
constraint and stability. The results from the public are the ones that we have been looking at in Figure
5.3; the elite results are also presented there. These elite data come from surveys of individuals
attending the Democratic and Republican Party conventions in 1980 and 1984. Unfortunately, more
recent analyses of elite opinion have not been conducted, so we must rely on this analysis. A quick
glance back at this figure demonstrates that elites had higher levels of both constraint and stability than
did the public during this time.30 This result confirms one of Converse’s key conclusions: elites possess
more ideologically constrained belief systems than do citizens. This conclusion does not appear to be
timebound, again suggesting that Converse’s overall findings were not restricted to the 1950s.
Converse’s method—survey research—has also come under attack. Some argue that asking citizens
specific, focused questions does not allow them to reveal the complexity of their political thinking or the
connections that they make across policy issues. Furthermore, questions that are designed to assess
liberal or conservative ideologies are not likely to pick up the presence of other ideologies among the
public. Libertarians, for example, prefer minimal government involvement in the economy but also
believe that individual civil liberties need to be protected from government intrusion. These views are
thus a mixture of liberalism and conservatism. If we used Converse’s method for measuring attitude
constraint (and his assumption that ideologies are arrayed along a liberal-conservatism continuum),
libertarians would show low levels of constraint even though their attitudes are derived from an
overarching ideology.31
To better measure other ideologies as well as to examine the process by which people reason about
political matters, some researchers prefer in-depth interviewing over surveys.32 During an interview
session, a researcher asks someone very broad questions. The questions are designed to allow
interviewees to discuss what is important to them rather than pointing them toward specific topics as
surveys generally do. Furthermore, interviewing allows an examination of the reasoning process in
addition to policy opinions, whereas survey research primarily assesses the latter. Put another way,
political ideology might best be gauged through interviewing because this method provides for a
contextual analysis of one’s thoughts. Others have argued that listening to people discuss politics using
their own terms and at their own pace is better than survey questions or even interviews for uncovering
whether and how people organize politics ideologically. Thus, ethnographic research is preferred by
some, whereby scholars place themselves in a setting or community, observe individuals, ask questions,
and more generally interact with people.33
Finally, some argue that Converse overlooked ideological strains among minority communities. When he
examined how people answered open-ended questions about political parties and candidates, the most
common category of responses referred to groups (refer back to Figure 5.1). These Group Interest
citizens did not display ideological thinking, at least according to Converse’s standard. Others disagree,
arguing that social groups play such a prominent role in American society and politics that group-linked
political thinking should be considered ideological. In American society, this is particularly the case with
race. As one scholar explains,
If society is organized around race, and racial conflict is part of everyday life, and if our stories
of the world are also organized around race, race is profoundly political and profoundly
ideological. … To use Converse’s language, there are a number of linking mechanisms
between blacks’ social locations, their racial identities, and various (generally unsatisfactory)
aspects of their social, economic, cultural and political worlds.34
Works examining Black political ideology have identified a range of political ideologies among African
Americans.35 In contrast to how liberalism and conservatism were defined earlier in this chapter, Black
political ideologies encompass views toward the status of Black people in society and the proper
strategies for improving this status, whites and other races, and Blacks’ interactions with the state (see
Table 5.1). Group-based perspectives, in other words, play a prominent role in these ideologies.
Delineating these ideologies is difficult using national surveys such as the ANES because such surveys
are often designed with the political views of majority whites in mind. Thus, researchers have designed
and fielded surveys of Blacks with questions that were created to tap into attitudes and concerns
relevant to this community.
Summary
Some critics of Converse have chipped away at his conclusions. The levels of attitude constraint and
stability that he documented were probably low because of the political environment of the 1950s. More
ideologically polarizing times reveal more coherent belief systems among the public. The survey method
also makes it difficult to uncover idiosyncratic ideologies or ideologies other than the dominant ones of
liberalism or conservatism. On the whole, however, many of Converse’s key findings have not been
substantially undermined. Constraint and stability have increased some over time, but not tremendously
so, and public levels fall far short of elite levels for both. There is also little evidence that the public
routinely thinks about politics ideologically. Donald Kinder and Nathan Kalmoe reach much the same
conclusion in their recent book, a book we profile more fully in the next section. They go a step further,
though, and note that this stability in ideological innocence came about despite changes to American
politics and society that could have had the opposite effect. “Educational transformation, party
polarization, revolutionary change in information dissemination, fundamental alteration in gender and
race relations: impressive as these changes have been, equally impressive is how little effect they have
had on how the American electorate understands politics.”36
Table 5.1 Key Components of Black Political Ideologies
Ideology Central beliefs
Radical
egalitarianism
(Dawson)/liberal
integrationism
(Harris-
Lacewell)
Believes that racism is widespread in America and that it is spread primarily
through racist institutions but that the nation’s ideal of equality for all points toward
a colorblind society. Black justice will only be achieved by demanding equality
from the state and working across races.
Black
nationalism
Views racial categorization and racial oppression as dominant features of society.
Whites will oppose attempts by Blacks to gain full equality. Proposes a Black
nation, generally not a separate political unit but instead a community within the
broader society with separate cultural traditions and needs. Emphasizes strong
unity among Black people.
Black feminism Opposes the racism of whites and the sexism of white and Black men. Race and
gender intersect to oppress Black women in multiple ways. Will form political
alliances with non-Black people, unless racism is present.
Black
conservatism
Believes in individual responsibility rather than government to improve the status
of Black people. Has faith that free markets do not discriminate. Black equality
should be achieved through economic progress rather than through programs that
provide Blacks with special consideration, such as affirmative action.
Disillusioned
liberalism
(Dawson only)
Views racism among whites as ingrained and unlikely to disappear, so political
alliances with whites are discouraged. Racial equality is to be pursued instead by
increasing Black political and economic power.
Sources: Michael C. Dawson, Black Visions: The Roots of Contemporary African-American Political Ideologies (Chicago: University of
Chicago Press, 2001), Chaps. 1, 3–6; Melissa Harris-Lacewell, Barbershops, Bibles, and BET: Everyday Talk and Black Political
Thought (Princeton, NJ: Princeton University Press, 2004), Chap. 1.
IDEOLOGICAL IDENTIFICATION
Another way to assess ideology among the American public is to examine levels of self-identification. Do
Americans tend to identify more as liberals or conservatives or neither? The ANES has tracked
ideological identification for the past few decades. The exact wording of this question appears in Table
5.2. Note that there are three categories each for liberalism and conservatism, depending on the degree
to which someone identifies with one of these ideologies. People whose ideological orientation is neither
liberal nor conservative could select the midpoint on the scale (labeled “Moderate; middle of the road”).
Finally, the end of the survey question explicitly encourages those who “haven’t … thought much about”
where they fall on the scale to say so.
Ideological identification from 1972 to 2016 appears in Figure 5.4.37 The solid lines in this graph present
trends over time in percentages of the public identifying as conservative, moderate, or liberal. In 1972
and 1976, the percentages of moderates and conservatives were nearly equal, at just over a quarter of
the public each. Since then, conservatives have outnumbered both moderates and liberals. Nearly one-
third of the public identified as conservative in 2016. Liberals had been the smallest of the three
ideological groups before 2008. Liberals and moderates were essentially the same portion of the public
in 2008 and fairly close in 2012, whereas liberals slightly outnumber moderates (26 to 24 percent) in
2016.
Table 5.2 Measuring Political Ideology
“We hear a lot of talk these days about liberals and conservatives. Here is a seven-point scale on
which the political views that people might hold are arranged from extremely liberal to extremely
conservative.”
“Where would you place yourself on this scale, or haven’t you thought much about this?”
1. Extremely liberal
2. Liberal
3. Slightly liberal
4. Moderate; middle of the road
5. Slightly conservative
6. Conservative
7. Extremely conservative
Source: American National Election Studies Cumulative (1948–2016) data file codebook, https://electionstudies.org/wp-
content/uploads/2018/12/anes_timeseries_cdf_codebook_var
https://electionstudies.org/wp-content/uploads/2018/12/anes_timeseries_cdf_codebook_var
Description
Figure 5.4 Ideological Identification over Time, 1972–2016
Source: Analysis of American National Election Studies Cumulative (1948–2012) and American National
Election Studies 2016 data files.
Focus your attention now on the dashed line, which demonstrates how many people did not know their
political ideology or had not thought enough about it to classify themselves. Between 1972 and 1984
more people chose not to self-identify than did those who claimed any specific ideological identification.
In other words, there were more “Haven’t thought much about it” responses than either conservatives,
liberals, or political moderates during these years. The number not classifying reached a high of 35.8
percent in 1980. Beginning in 1988, however, conservatives began to outnumber those without an
ideology. The number of nonidentifiers has decreased gradually ever since. At 18 percent, 2016 saw the
lowest proportion of the electorate indicating that they were not able to identify their ideology since the
ANES began asking this question. Having said that, there is still nearly one-fifth of the public that has
not thought enough about their ideology to self-identify. To us, this is further evidence in support of
Converse’s conclusions that a significant portion of the public does not think ideologically.
Symbolic versus Operational Ideology
With more self-identified conservatives than liberals or moderates among the public today, Americans
must support conservative positions on specific policy issues. Not so, say Christopher Ellis and James
Stimson.38 In their work, Ellis and Stimson draw a clear distinction between symbolic and operational
ideology. Symbolic ideology is what we have just been discussing, an individual’s ideological
identification (as conservative, moderate, or liberal). In contrast, operational ideology “is grounded
more explicitly in concrete decisions, what citizens think the government should or should not be doing
with respect to important matters of public policy.”39 In other words, operational conservatives favor
conservative issue positions, such as increased spending on defense and crime, tax cuts, and the death
penalty. In contrast, support for government social welfare policies, environmental protection, abortion,
and gay rights are indicators of operational liberalism.
Ellis and Stimson’s research highlights a paradox regarding ideology in the United States. Symbolically,
Americans are most likely to be conservative (as we had demonstrated in Figure 5.4). Operationally,
however, Americans tend toward liberalism.40 To reach this latter conclusion, Ellis and Stimson
examined responses to over 7,500 survey questions conducted between 1952 and 2010. These items
queried public views toward a wide range of domestic issues. Across all of these questions, the public
supported the liberal policy position 48 percent of the time compared to 34 percent for the conservative
position. Furthermore, liberal positions were favored over conservative ones in every year except 1952.
In addition to examining ideology for the entire nation, Ellis and Stimson also examined ideology at the
individual level. Specifically, they grouped individuals based on their symbolic and operational
ideologies.41 They excluded moderates as well as those who did not self-identify an ideological
orientation, thus leaving four possible combinations. People can have symbolic and operational
ideologies that are both conservative or both liberal. These categories are labeled, respectively,
consistent conservatives and consistent liberals. Two groups of ideological mismatches are also
possible: symbolic conservatives who are operationally liberal and symbolic liberals who are
operationally conservative. Figure 5.5 presents the classification of Americans into these four groups for
2008, the most recent year provided by Ellis and Stimson. Among those people whose symbolic and
operational ideologies match, consistent liberals substantially outnumber consistent conservatives.
Turning to those people whose symbolic and operational ideologies differ, the largest group is
composed of symbolic conservatives who are operationally liberal. Indeed, this group is almost as
numerous as consistent liberals. Finally, practically nonexistent are individuals with a liberal symbolic
ideology paired with operational conservatism.
How might we explain these national and individual-level patterns of symbolic and operational ideology?
Ellis and Stimson point to a number of factors, including that symbolic ideology is a personal identity. Put
another way, it reflects how people think of themselves. As it turns out, more Americans are comfortable
thinking of themselves as conservative rather than as liberal. This is due, in part, to the popularity of a
conservative identity outside of politics, most notably in reference to religious worldviews and lifestyle
choices. “Along with church on Sunday,” write Ellis and Stimson, “imagine living by conventions—
marriage, family, children, and work—and you have a lifestyle often called conservative. ‘Conservative’
in this context means conventional behavior and appearance, playing by the established rules, and
fitting into established social patterns.”42 Some individuals who think of themselves as conservative in
these nonpolitical domains transfer this identity to politics, thus self-identifying as politically conservative
even if they do not hold conservative policy views.
Description
Figure 5.5 Symbolic and Operational Ideology Classifications, 2008
Source: Data from Figure 5.5 of Christopher Ellis and James A. Stimson, Ideology in America (Cambridge,
UK: Cambridge University Press, 2012), 98.
What Is the Meaning of Ideological Identification?
Given that ideological innocence appears widespread among the American public, “what do we make of
the fact that many Americans seem quite happy to describe themselves in ideological terms?”43
Answering this question motivated Kinder and Kalmoe to write their book Neither Liberal nor
Conservative. The result is a comprehensive exploration of ideology, addressing both the ideological
innocence research that began with Converse’s study and work examining ideological self-identification.
Among their notable conclusions, Kinder and Kalmoe make the case that many self-identified
moderates are actually nonideological. Put another way, moderates do not place themselves in the
middle of the liberal-conservative scale because their views regarding government are in between
liberalism and conservatism. Instead, moderates share a number of political characteristics, such as low
levels of engagement with politics and low stability of issue opinions over time, with those who refuse to
select an ideological category when asked (the “Haven’t thought much about it” respondents). “All things
considered,” Kinder and Kalmoe conclude, “the ‘moderate’ category seems less an ideological
destination than a refuge for the innocent and confused.”44
If ideological identification is meaningful for individuals, it should influence a number of political attitudes
and behaviors. By and large, it does not, argue Kinder and Kalmoe.45 Other than among the most
politically informed members of the public, ideology generally does not predict which candidate people
vote for in presidential elections. Ideological identification is weakly related, if at all, to specific issue
opinions, a conclusion that is in line with Ellis and Stimson’s findings.46 Finally, liberals and
conservatives tend not to differ in their evaluations of the state of the economy. In the end, Kinder and
Kalmoe suggest that ideological identification reflects, rather than drives, political opinions and
experiences. For instance, voting in a presidential contest where candidates present differing ideological
platforms can alter people’s ideological identification. Young adults were more likely to move their
ideology in the conservative direction if they voted for conservative candidate Ronald Reagan instead of
Jimmy Carter in 1980. In short, ideological identification is not a driver of political opinions, decisions,
and perceptions. What is? Party identification, say Kinder and Kalmoe, and substantially so.47
PARTY IDENTIFICATION
All the while many citizens neither think ideologically nor organize their political attitudes along a single
conservative-liberal dimension, party identification is one of the most meaningful political attitudes that
individuals hold. As we discussed in Chapter 1, party identification refers to a person’s psychological
attachment to a political party or identification as independent of a party. Not surprisingly, party
identification is the best predictor of which party a person will vote for in an election. But party
identification has effects that extend far beyond that. Recall the discussion of partisan motivated
reasoning in Chapter 4, demonstrating that party affiliation can influence our perceptions and
interpretations of new policy information. Party identification can color how citizens view political
candidates. For example, research shows that when citizens know a candidate’s party, they rely on that
knowledge to evaluate the candidate rather than relying on information about the candidate’s issue
positions, even when those positions are inconsistent with the party label.48 Partisanship is also related
to citizens’ issue attitudes, as we demonstrate shortly. In this section, we also discuss research
exploring the topics of negative partisanship and partisan independence. Along the way, we will illustrate
that party identification is a core identity for many, albeit not all, Americans.
Party Identification and Issue Opinions
Party identification has a significant impact on a wide variety of policy opinions, as you can see from the
2016 ANES data presented in Figure 5.6.49 For these analyses, we have included leaners in with the
partisan categories of Republican and Democrat. As you probably remember from Chapter 1, leaners
are those individuals who first indicate they are politically independent but when then asked whether
they lean toward one of the two parties, they identify one toward which they lean. To begin, let’s
consider the long-standing and fundamental difference between Democrats and Republicans on social
welfare policies. Since the Great Depression, Democrats have championed more expansive social
welfare programs than Republicans have. In 2016, for example, nearly 60 percent of Democrats
supported more government services even if it meant an increase in spending, whereas only 13 percent
of Republicans favored that position. Note, however, that Republicans are not opposed to government
spending across the board; 69 percent of Republicans supported increasing government spending on
defense, whereas just 34 percent of Democrats approved of increased spending in this area.
Since the 1960s, racial issues have been a key dividing point between the parties, with Democrats
taking positions that are more pro–civil rights. In 2016, there was a large gap between partisans, 39
percentage points, in their support for efforts to improve the social and economic position of Blacks. Half
of Democrats are interested in addressing the concerns of Black Americans compared to only 11
percent of Republicans. In recent decades, cultural issues have divided Democrats and Republicans.
Take abortion, for example. Sixty percent of Democrats endorsed the most pro-choice stance when
asked their views on abortion policy, whereas 28 percent of Republicans favored that position.
We would be remiss if we did not mention the policy attitudes of the 15 percent of Americans who
identified as pure Independents in 2016. As the data in Figure 5.6 demonstrate, pure Independents are
independent for a reason. They do not consistently side with one party or the other. On opinions
regarding more government services and military spending, their views are more similar to Democrats,
whereas they look somewhat more like Republicans on the issue of aid to Blacks. When it comes to
abortion, their opinions fall in between Democrats and Republicans.
Description
Figure 5.6 Party Differences in Issue Opinions, 2016
Source: Analysis of American National Election Studies 2016 data file.
Notes: Respondents placed themselves on 7-point scales for the government services, military spending,
and aid to Blacks questions. The bars represent the percentage of respondents who placed themselves on
the more-government-services, more-spending, and more-government-effort side of the scales. For the
abortion question, the bars represent the percentage of respondents who chose the most pro-choice stance
from a list of four options.
Positive or Negative Partisanship?
Are Democrats Democratic because they have a positive attachment to the Democratic Party or
because they have a negative reaction to the Republican Party? Early writing on party identification, and
even more recent scholarship, suggested that partisanship largely resulted from a positive attachment to
one’s party.50 Increasingly, though, scholars have explored negative partisanship, or the phenomenon
of identifying as a partisan primarily through negative association with the other party. Evidence of such
negative partisanship was clearly on display during the 2012 and 2016 presidential elections, including
that partisans’ negative views of the other party were linked to a greater tendency to vote and, among
voters, to vote only for candidates of one’s party.51
This does not mean that Americans are no longer positively affiliated with their political party. To the
contrary, negative assessments of the other political party and its candidates coexist alongside positive
views toward one’s own party and candidates. For instance, Republicans can feel warmly toward their
party and coolly toward the Democratic Party and ascribe positive personality traits to Republican
candidates while believing Democratic candidates possess negative traits.52 Furthermore, when given a
choice between acting in ways that help one’s party versus hinder the other party, the former seems to
be the default reaction. In a series of studies conducted by Karyn Amira, Jennifer Cole Wright, and
Daniela Goya-Tocchetto, participants were asked which of two articles a political website should display:
one that presents their party favorably or one portraying the other party negatively.53 Two-thirds of
individuals selected the favorable own party story. However, what happens when people are first
confronted with a message that threatens their partisanship, such as a story detailing voter fraud
committed by party members? They are more likely to want the website to publish the unfavorable
article about the opposing party. In other words, preferences switch.
This brief overview of recent research suggests that positive and negative views of the political parties
interact in complex ways. At the end of the day, whether Americans’ individual partisan affiliations are
grounded more strongly in positive or negative attachments with the political parties is an open question.
More research is needed to examine this topic.54 What is clear, however, is that negative partisanship is
fueling a type of political polarization, a topic we take up in more detail later in this chapter.
Partisan Independence
As noted in this and other chapters, when it comes to issue opinions and vote preferences, partisan
leaners have a similar profile to identifiers of the party toward which they lean. Despite this similarity,
when initially asked whether they think of themselves as Democratic, Republican, or Independent,
leaners choose Independent. Understanding what motivates this expression of partisan independence,
among both leaners and pure Independents, was the focus of research conducted by Samara Klar and
Yanna Krupnikov for their book Independent Politics.55 In a political environment awash with partisan
conflict and gridlock, negative images of parties and partisans abound. In contrast, political
independence is valued and often portrayed positively. Identifying as partisan can thus be embarrassing,
yet can be avoided by declaring oneself an Independent. Indeed, Klar and Krupnikov empirically
demonstrate that when the political context supplies negative information regarding partisans, the
tendency to identify as an Independent increases, primarily among people who are keen on making a
good impression. Importantly, people’s issue opinions do not shift depending upon whether they
encounter negative images of partisans nor is it likely that their actual partisan leanings change. It is the
expression of their (non)partisanship that is affected.
Klar and Krupnikov also document a variety of consequences that flow from this desire to avoid publicly
identifying as partisan.56 Across a set of interesting experiments, they find that after reading a news
article or even a statement about partisan disagreement in American politics, people were less likely
than those who read about partisan unity to engage in the following behaviors: reveal their partisanship
or candidate preference on a social network, take a partisan sticker as a thank-you for participating in a
research study, display the partisan sticker if they took one, and be willing to try to persuade a friend to
support their preferred party at the polls. Note that these results emerge only among people who like to
make a good impression on others. Finally, nationally representative survey results point to a similar
conclusion: those who are more sensitive about the impression they make on others engage less in
explicitly partisan activities. Sure, these citizens might cast a private ballot for candidates of their
preferred party but, write Klar and Krupnikov, that “should be cold comfort for parties. Rather, parties
should be much more concerned about what these people do not do … Candidates and parties need
ordinary people to influence their friends and spread the party’s message to their networks. Candidates
and parties succeed to the extent that ordinary voters are vocal, clear, and public in their support.
Candidates and parties lose all of this when people try to avoid partisanship.”57
POLARIZATION
Political polarization refers to a gap in political opinions whereby Americans are divided in their
preferences. Furthermore, as commonly understood, polarization implies not a narrow gap but instead a
chasm with the public divided into camps on opposing sides. Such polarization could occur for opinions
toward specific policy issues, for ideological orientations, or for partisanship. To illustrate, we drew a
hypothetical graph of polarization (see panel C of Figure 5.7), in this case exploring the relationship
between partisanship and ideology. The lighter bars present ideological categories among self-identified
Democrats, whereas the darker bars do the same for Republicans. Note that in this illustration all of the
Democrats are liberal and all of the Republicans are conservative. This is what we usually mean by
polarization: people are divided into extreme camps with no one in the middle.
Description
Figure 5.7 Hypothetical Portraits of the American Public
Source: Figure created by authors.
Contrast this image of polarization with the other two panels of Figure 5.7. These are also hypothetical
graphs, presenting possible distributions of ideology separately for Democrats and Republicans. The top
panel shows ideology and partisanship to be rather intermingled. Democrats are arrayed along the
entire ideological spectrum, as are Republicans. The most common ideological identification for both is
moderate, yet there are liberal Democrats and liberal Republicans as well as conservative Democrats
and conservative Republicans. Focus your attention now on the middle panel. Here we still see rather
wide ideological distributions for both partisan groups. However, some shifting has occurred: democrats
have moved to the left as Republicans have shifted right. The distributions are no longer centered on
moderate. Instead, the Democrats are centered on the slightly liberal category. The Republican
distribution peaks at conservative. Finally, there are more liberals than conservatives among Democrats,
whereas conservatives outnumber liberals among Republicans.
Which of these possible options most accurately portrays the American public today? The middle one.
Indeed, the stylized graph in panel B is fairly close to the actual relationship between ideology and
partisanship that existed in the 2010s. Panel A, in contrast, is not that far off from the American public of
1972. Put another way, over the past few decades, partisans have sorted themselves somewhat more
into their proper ideological camps, yet this sorting falls short of the polarization displayed in panel C. A
number of public opinion scholars have identified this trend, including Matthew Levendusky in his book
The Partisan Sort.58 Levendusky further demonstrates that partisan sorting is driven by people
changing their ideological identification rather than their partisanship.59 Over time, for instance,
Republicans have been more likely to self-identify as conservative rather than moderate or liberal. The
phenomenon of conservatives opting to change their partisanship from Democratic to Republican has
been much rarer. This result should not be surprising, given our earlier discussion that partisanship is
more likely than ideology to be a central political identity for most Americans.
Another way to explore political polarization among the public is to examine opinions toward specific
issues. Is the public divided over key political issues of the day with two groups clustering on opposite
sides of these issues? Not really. As it turns out, for many issues, more Americans take centrist than
extreme views, a pattern that has not changed very much over time. Take attitudes toward whether the
government should provide services to the public, even if doing so would require extra government
spending. In 1984, one of the first years that the ANES probed public views on this topic, the most
common response was smack dab in the middle between preferring fewer versus more government
services (see Figure 5.8). Very few people selected either of the options at the extreme ends of this
response scale. Public attitudes were little changed in 2016. Sure, fewer people selected the middle
option than they had thirty-two years earlier, but this was not because more people were holding
extreme opinions. Instead, more people still opted for the middle position rather than any other one. This
preference for centrism rather than extremism exists across many political issues. Even for issues over
which Americans disagree with one another, the gap between the two sides tends not to be very large.60
Opinions toward abortion are illustrative of this point. If the public were polarized on this issue, we would
see a group of pro-choice Americans who support a woman’s right to obtain an abortion in all
circumstances and a pro-life group who do not support abortion in any circumstances. Instead, some
people are pro-choice (45 percent in 2016), some are pro-life (14 percent), and others support abortion
in some circumstances but not others (41 percent).61 Even on a hot-button issue such as this,
disagreement does not often take the form of deep division. This is further evidence that Americans are
not polarized.
Description
Figure 5.8 Public Attitudes toward Government Provision of Services, 1984
and 2016
Source: Analysis of American National Election Studies Cumulative (1948–2004) and American National
Election Studies 2016 data files.
The public is thus not polarized, yet many people assume that polarization exists in America.62 Why? A
primary reason is that U.S. politicians have become more polarized in recent decades. A few decades
ago, the ideology of members of Congress were fairly well, but not perfectly, sorted by partisanship.
Most Republicans were conservative, although a few liberal Republicans served in Congress. Similarly,
whereas there were some conservative Democrats in the House and the Senate, most were best
characterized as liberal. Members of Congress, in other words, looked like the middle panel of Figure
5.7. Today, panel C of that figure is a more accurate representation of political elites.63 Democrats in
Congress are liberal; Republicans are conservative. That is, partisans in Congress are divided into
polarized ideological camps. This elite polarization has not gone unnoticed by the news media or the
public. Polarization among national political elites is a much more common topic among news reporters
than was the case a few decades ago.64 Coupling this with the tendency of the news media to highlight
conflict, and a portrait of a society deeply divided with strong rhetoric coming from both sides emerges in
daily newscasts.65 Perhaps it is not surprising then that the public is more likely now than in past
decades to believe the parties are different from one another.66 Fully 80 percent of citizens saw
important differences in the stances of the parties in 2012 compared to only 58 percent in 1998. The
perceived ideological distance between the parties has also grown over those same years. The public
sees the Republicans as more conservative than they used to, all the while classifying Democrats as
more liberal today than in the past. Finally, this elite polarization has contributed to the partisan sorting
of the public. In the words of Levendusky, “As elites pull apart to the ideological poles, they clarify what it
means to be a Democrat or a Republican. Ordinary voters use these clearer cues to align their own
partisanship and ideology.”67
Affective Polarization
In short, elites are polarized but the public is not, at least when it comes to holding extreme political
opinions. As for feelings toward parties and politicians, however, we do see evidence of public
polarization. Affective polarization, the combination of dislike of one’s political opponents and
favorable feelings toward political allies, is on the rise. One common method for measuring feelings
toward political groups and politicians is via a feeling thermometer. This survey question asks
respondents whether they feel warm, cold, or neither toward specific groups. The warmest rating on this
thermometer is 100 degrees; the coldest is 0 degrees. Over time, partisans in the public have given
colder ratings to the opposite party.68 During the 1980s, for instance, the average Republican rating of
Democrats was 45 degrees, whereas the average Democratic rating of Republicans was 44 degrees. In
2016, these ratings were, respectively, 25 and 27 degrees. In 1980, the most common 10-degree
interval selected by Democrats to rate Republicans was 41–50 degrees. In 2012 and 2016, the most
common interval was 0–10 degrees, the lowest possible interval. The same pattern existed for
Republican assessments of Democrats. Across these decades, in contrast, partisans’ feelings toward
their own party remained relatively stable. That tells us that Democrats and Republicans do not feel
colder toward both parties, but only toward the opposing party. Furthermore, negative feelings regarding
the nonidentified party are directed much more toward party leaders and candidates rather than
partisans among the citizenry.69 Finally, affective polarization is especially apparent among individuals
who perceive that the two political parties diverge substantially in their issue positions.70 Specifically, the
more individuals perceive that the opposing party holds issue positions far away from their own, the
more they view that opposing party negatively and their own party positively.
Affective polarization is also apparent in ways other than feeling thermometer ratings of the parties.
Partisan divergence in the personality trait assessments of presidential candidates has increased.71
Democrats in the electorate today are more likely to ascribe positive traits (intelligent, knowledgeable,
etc.) to Democratic than Republican candidates; the opposite pattern emerges for Republican citizens.
In contrast, in the past it was not uncommon for people to think that both parties’ presidential candidates
possessed some positive traits. In addition, stereotyping of party supporters is more common today than
it used to be, in ways that favor one’s own party.72 In particular, American partisans are more likely in
recent times to believe that typical supporters of their party possess positive personal traits. Supporters
of the other party are, you guessed it, stereotyped more negatively. In an “us versus them” mentality,
“they” are presumed to be closed-minded, hypocritical, and selfish, whereas “we” are honest, intelligent,
and open-minded. There is even evidence that some partisans engage in the extreme act of
dehumanizing the other side. When presented with the familiar Ascent of Man image, which pictorially
depicts millions of years of human evolution, people rate members of the party they oppose as less fully
human than members of their own party. This pattern emerged for both Democrats and Republicans,
across multiple studies conducted in 2016 and 2018.73
Outside of the political realm, discomfort with interparty marriage is on the rise.74 In 1960, only 5 percent
of partisans expressed displeasure at the thought of a child marrying someone affiliated with the other
party. Fifty years later, one-third of Democratic parents indicated they would be somewhat or very
unhappy if one of their children were to marry a Republican. Meanwhile, nearly one-half of Republicans
would be unhappy to see their child marry a Democrat. Unhappiness with interparty marriage is
especially likely if the hypothetical in-law is expected to discuss politics frequently.75 Feelings toward
partisans may even crop up when reviewing résumés or selling tickets to a college football game. One
experimental study asked participants to review the credentials of two (hypothetical) finalists for a
college scholarship. Both of the finalists were high school seniors. One was the president of the Young
Democrats, the other the president of the Young Republicans. Nearly 80 percent of the Democrats and
Republicans in this study selected the finalist with the same partisanship as them, even when the finalist
from their own party had a lower grade point average than the out-party finalist.76 As for college football
tickets, participants in an experiment demonstrated a preference to (hypothetically) sell tickets to a
game to a supporter of their political party rather than a partisan opponent. They were also willing to
receive a lower price for the ticket if it was purchased by someone of their party versus the opposing
party. Note that all of the study participants were fans of rival teams (e.g., Alabama and Auburn), the
tickets were purportedly for the game between these rivals, and the hypothetical buyer was described
as a fan of the other team.77
We thus see significant evidence of affective polarization among the American public, for both political
and nonpolitical judgments. A variety of explanations for the increase over time in this type of
polarization has been suggested by scholars.78 One of the most compelling of these is Lilliana Mason’s
argument that social sorting is a main contributor.79 Social sorting occurs when multiple social
identities, such as racial, religious, and ideological identities, coincide with partisanship. Tracking party
coalitions over time, Mason demonstrates that the Democratic and Republican parties are each more
socially homogeneous than they used to be.80 Democratic Party identifiers are more likely to be liberal,
Black, and/or secular, for example, whereas conservatives, whites, religiously active Christians, and the
wealthy are more likely to be Republican. This was not always the case. Back in the day, such identities
were more intermingled. Some Democrats were liberal, some were conservative (recall the top panel of
Figure 5.7). Wealthy Northerners were Republican, but wealthy Southerners were Democratic. And so
on. The degree to which identities overlap or cross-cut is important politically, as Mason describes: “A
single group identity can have powerful effects, but multiple identities all playing for the same team can
lead to a very deep social and even cultural divide. Those with cross-cutting partisan, religious, and
racial identities are more likely than socially sorted citizens to welcome the opposing team into their lives
and to consider them as fellow citizens.”81
At the individual level, social sorting is related to affective polarization.82 Specifically, Mason reveals that
the gap between warm in-party versus cold out-party feelings is larger for people whose social identities
line up with their partisanship. Those who are not so strongly socially sorted do display a warm feeling
bias toward their own party, but the feelings gap is not as large as for their sorted counterparts. Social
sorting also contributes to affective polarization in the nonpolitical domain. The more people’s identities
are aligned with partisanship, the less likely they wish to connect socially with members of the opposite
party, whether that be via marriage, friendships, or living side-by-side as neighbors. Mason also
explores whether issue disagreement is related to affective polarization, finding that, for the most part, it
is not. Put another way, individuals do not view their political opponents negatively because these
opponents hold different issue positions. Rather negativity can arise between partisans despite
agreement on political issues. This insight is reflected in the title of Mason’s book, Uncivil Agreement,
and has implications for governing: “the social homogenization of the parties has made it difficult for
partisans to learn to like, or even humanize, their partisan opponents. They are stereotyped, vilified, and
rejected out of hand. The unfortunate truth of this, however, is that these deep social divisions are
allowing opportunities for policy compromise to go unnoticed.”83
Box 5.1 Party Polarization across the Globe
Public Opinion in Comparative Perspective
The United States is not the only democracy where party elites are polarized. In many other
nations, political parties within a country take very different policy stances from one another. Yet,
in some countries, party convergence exists, whereby party platforms are rather more similar to
each other. Such natural variation in the degree of party polarization across nations allows
scholars to examine whether elite polarization influences public opinion. In the most
comprehensive such analysis to date, Noam Lupu examines one effect of polarization: the
likelihood that citizens will identify with a political party.1 In his study, Lupu analyzes thirty-seven
nations from across the globe. All are democracies, although there is a mix of newer and older,
more established democracies.
Across these nations, Lupu concludes that party polarization is indeed related to partisanship
among the public. In nations where the parties are more polarized, more citizens affiliate with a
political party than in nations with less polarization. Furthermore, individual citizens who perceive
that their country’s parties are polarized are more likely to be partisans than are people who
believe the parties of their nation are similar to each other. Why might polarization affect
partisanship holding? The “further apart the political parties, the easier it may be for citizens to
distinguish among their electoral options. And if citizens can more clearly distinguish parties, they
may find it easier to form a party attachment.”2 Put another way, if national parties hold similar
positions, it can be hard for voters to figure out which of those parties would best represent their
issues and interests.
Ultimately, Lupu argues that polarization can be positive for democracy. When parties take
clearly divergent views in a nation, the populace knows where the parties stand. This not only
makes it easier for citizens to predict what a party will do if elected, but also makes it easier to
then hold these elected officials accountable when the next election rolls around.
1 Noam Lupu, “Party Polarization and Mass Partisanship: A Comparative Perspective,” Political
Behavior 37 (2015): 331–356.
2 Ibid., 334.
Partisan News and Polarization
Turn on cable television news stations Fox or MSNBC any weeknight and you will see talk show hosts
and guests launching criticisms at their political opponents. The same is true if you visit many political
websites, where the tone of commentary is often nastier and more vitriolic than the cable shows. Surely,
then, the media have played a role in polarizing the attitudes of the public. In particular, scholars and
others have presumed that partisan news, or media outlets that cover and present the news from a
specific political point of view, fuel polarization. As it turns out, reality is more complex. Let’s not forget
that, as we just demonstrated, citizens are not actually polarized when it comes to issue opinions.
Perhaps, though, partisan news has the potential to polarize the public, and thus if our political news
becomes even more partisan, issue polarization could emerge in the future. We discuss research that
speaks to that topic. First things first, though: how many people actually consume partisan news?
Compared with prior decades, partisan or ideologically oriented news outlets are definitely more
common today. Whether it be cable television stations, talk radio, websites, or podcasts, citizens now
have many choices for strictly liberal or strictly conservative news.84 Not only do these liberal and
conservative media sources exist, but people’s choice of news sources is related, at least in part, to their
own political predispositions. This process is known as selective exposure, whereby people choose to
consume news that coincides with their existing political views.85 Some people search out like-minded
news, in other words. For example, the audiences of the Fox News cable channel and Rush Limbaugh’s
radio show contain more conservatives than liberals. Liberals are more prevalent among viewers of
shows on the MSNBC cable channel or the HuffPost online site.86 We do need to be careful not to dwell
on these patterns of selective news exposure, however, particularly because the audiences for partisan
outlets are still smaller than for mainstream news.87 The print and digital daily circulation of legacy
newspapers (those that originated in print version) was nearly 30 million in 2018, and the average daily
viewership of the evening news on ABC, NBC, or CBS was 5.3 million that year. Fewer people (1.25
million) tuned into cable television news shows, some of which lean to one partisan side or the other, on
a typical evening in the year. The audiences for most political websites are smaller still. In late 2018,
only a handful of political digital-only news sites received more than 10 million unique visitors over an
entire month. The average length of time for each visit to these sites was two minutes.
Determining whether consumption of partisan news leads to issue polarization is tricky. Many viewers of
shows on the Fox News cable channel do not become conservative because they watch the shows,
after all, but rather watch the shows because they are already conservative. Comparing the political
views of liberal versus conservative news consumers is thus not a good approach for determining
whether these news outlets polarize the public. Instead, as we discussed in Chapter 3, experiments are
better methods for examining whether news content influences people’s attitudes. By randomly
assigning people to watch a conservative or a liberal news show, we can conclude whether the content
of these shows changed attitudes and, specifically, whether these attitudes became more extreme.
Using this approach, scholars have demonstrated that partisan news can lead to issue polarization. In
particular, when liberals watch liberally oriented news, their views become even more liberal.
Conservatives hold even more conservative views after consuming conservative news.88
These results are based on having study participants watch partisan news regardless of what their
normal viewing habits are. That is, think of these results as demonstrating what can happen to attitudes
when people are forced to watch partisan news. But what happens when you take into account the fact
that some people prefer to watch partisan news, whereas others much prefer entertainment over any
type of news? One set of experiments demonstrates that partisan news viewing leads to opinion
polarization only among those who prefer entertainment, most likely because they are not familiar with
the political rhetoric and arguments present in the shows.89 They are more open to persuasion and end
up holding more extreme attitudes after watching partisan news. For those preferring news, in contrast,
“ideologically congenial shows appear to have relatively small effects among those who choose to watch
them … these shows end up preaching to the choir.”90 Taking these experimental results out of the lab,
it appears as if partisan news could, in real life, polarize the opinions of those people who generally
prefer to watch entertainment shows. Yet these are precisely the citizens who are unlikely to consume
partisan news because they much prefer entertainment over news. If they do not watch political shows
on Fox or MSNBC, their opinions cannot be directly influenced by these shows. In short, then, partisan
news shows do not seem to produce issue polarization in the real world.91 On the other hand, the nasty
tone and fiery rhetoric hurled at political opponents on these shows has likely contributed to affective
polarization, as we discussed in Chapter 3.92
CONCLUSION
Few, if any, works have influenced public opinion scholarship as much as Converse’s “The Nature of
Belief Systems in Mass Publics.” The piece has been called the “foundation stone of political-behavior
research,”93 and a veritable cottage industry of research on public ideology arose after its 1964
publication, not all of it favorable. However, many of Converse’s core conclusions have stood the test of
time. Chief among these is that members of the public tend not to see the political world in ideological
terms. One of Converse’s primary concerns, of course, was the degree to which citizens and elites think
about and discuss politics using the same terms. Converse found that elites are much more likely to
possess ideologically constrained belief systems. Elite constraint has rarely been studied since
Converse’s work was published, but research examining this topic has confirmed Converse’s finding. As
Kinder and Kalmoe put it in their book, “The ideological battles under way among American political
elites show up as scattered skirmishes in the general public, if they show up at all.”94 This empirical
evidence, as we noted earlier in the chapter, is in line with one of the key tenets of elite democratic
theory. In contrast, if citizens do not understand the nature of elite policy decisions, it might become
more difficult for them to evaluate and constrain, if necessary, elite behavior. This undermines
democratic governance, certainly the type of governance assumed by participatory democrats, who
hope that citizens will fairly routinely monitor the actions of leaders.
Converse also concluded that group-oriented thinking is common among the public. Nowhere is this
reality more obvious than in the case of partisan affiliation. Political parties are the main groups active in
American politics, and much political activity is organized by or around the parties, from elections to
legislative bodies in state capitals and Washington, DC. As one public opinion scholar put it, “Elite
competition is partisan competition: Democrats and Republicans define the issues of the day and set
the terms of the debate. … For ordinary citizens, party is how they understand the political world.”95 Just
as pluralists expect, political parties help to orient the public toward political debates and structure public
thinking about issues. If disdain of parties pushes more and more people to publicly identify as
independent, however, pluralists will be concerned.
What do democratic theorists make of partisan sorting and increased affective polarization among the
public? To the degree that partisan sorting has enabled citizens to line up their ideology and issue
preferences with their partisanship, participatory democrats and pluralists would be pleased, especially if
such sorting allows the public to better recognize their issue preferences within the dominant elite
discussions and then act on these preferences in the voting booth. On the other hand, negative
partisanship and affective polarization worry participatory democrats. Animosity toward one’s opponent
is not typically a precursor to meaningful political involvement. In fact, high levels of negativity and
distrust on all sides might just turn citizens away from politics. Furthermore, if political attitudes and
activity are motivated by negative partisanship, democratic accountability is threatened. Rather than
hold elected officials accountable for their actions and policy decisions while in office, partisan animus
can drive citizen behavior regardless of what these officials actually do.
Finally, if the public tends not to think ideologically and if many members of the public do not organize
their beliefs along an ideological continuum, is it fair to conclude that public attitudes are fleeting and not
well reasoned? As our discussion of party identification might have suggested, we are not yet prepared
to draw that conclusion, and we hope you will wait as well. In the next two chapters, we describe other
sources of people’s political beliefs. Chapter 6 examines personality, values, self-interest, and historical
events. We return to the topic of groups in Chapter 7, in this case membership in and attitudes toward
demographic groups. Whether these alternative sources are as politically meaningful as or more
meaningful than ideology and whether they revive a view of the public as more competent for
democratic politics than Converse concluded are topics that we encourage you to consider as you read
the next two chapters.
KEY CONCEPTS
affective polarization 176
attitude constraint 153
attitude stability 157
belief system 152
Black political ideology 162
consensual politics 159
conservatives 152
ethnographic research 161
ideologically contentious 159
ideology 151
in-depth interviewing 161
liberals 152
libertarians 161
negative partisanship 171
open-ended questions 153
operational ideology 166
partisan news 180
partisan sorting 174
party identification 168
political polarization 172
selective exposure 180
social sorting 178
symbolic ideology 166
SUGGESTED SOURCES FOR FURTHER READING
Arceneaux, Kevin, and Martin Johnson. Changing Minds or Changing Channels? Partisan News in an
Age of Choice. Chicago: University of Chicago Press, 2013.
Levendusky, Matthew. How Partisan Media Polarize America. Chicago: University of Chicago Press,
2013.
Stroud, Natalie Jomini. Niche News: The Politics of News Choice. Oxford: Oxford University Press,
2011.
If you want to learn more about partisan news and polarization, these books are good places to start.
Stroud examines partisan selective exposure and demonstrates that news choices have consequences
for citizens’ attitudes and behaviors. The other two books ask whether partisan news polarizes the
public. Yes, concludes Levendusky. No, say Arceneaux and Johnson.
Hajnal, Zoltan L., and Taeku Lee. Why Americans Don’t Join the Party: Race, Immigration, and the
Failure (of Political Parties) to Engage the Electorate. Princeton, NJ: Princeton University Press, 2011.
Lavine, Howard G., Christopher D. Johnston, and Marco R. Steenbergen. The Ambivalent Partisan:
How Critical Loyalty Promotes Democracy. Oxford: Oxford University Press, 2012.
These two books devote attention to people with weak, nonexistent, or shifting partisan ties. Hajnal and
Lee identify the pathways to partisanship and nonpartisanship separately for African Americans,
Latinos, Asian Americans, and whites. The Ambivalent Partisan examines individuals who experience a
mismatch between their party identification and their evaluations of the Democratic and Republican
parties.
Is Democratic Competence Possible? Special issue of Critical Review 18 (2006).
This special issue of the journal Critical Review contains a republication of Converse’s 1964 “Nature of
Belief Systems” paper (which is currently out of print) along with twelve essays that comment on
research conducted on this topic since 1964. The issue ends with an interesting response essay by
Converse in which he addresses some of the arguments of his critics.
McCarthy, Nolan. Polarization: What Everyone Needs to Know. New York: Oxford University Press,
2019.
The title of this book says it all: it provides an overview of many facets of political polarization.
Barber, Michael, and Jeremy C. Pope. “Conservatism in the Era of Trump.” Perspectives on Politics 17
(2019): 719–736.
Barber and Pope identify multiple strands of conservatism in the contemporary Republican Party.
Noel, Hans. Political Ideologies and Political Parties in America. Cambridge, UK: Cambridge University
Press, 2013.
Noel traces the origins of American conservatism and liberalism, most notably the forging and
development of these ideologies among activists, intellectuals, and political commentators. Partisanship.
Symposium in Advances in Political Psychology 39 (2018).
Party Polarization. Special issue of Public Opinion Quarterly 80 (2016).
Scholars have been very actively researching partisanship and polarization in recent years. These two
special journal issues cover a range of interesting topics and present many new findings.
“Political Typology Quiz,” Pew Research Center, 2017, https://www.people-press.org/quiz/political-
typology/
Ripley, Amanda, Rekha Tanjarla, and Angela Y. He, “The Geography of Partisan Prejudice,” The
Atlantic, March 4, 2019, https://www.theatlantic.com/politics/archive/2019/03/us-counties-vary-their-
degree-partisan-prejudice/583072/
For a fun exercise, take the “Political Typology Quiz” and find out your political type. Pew’s typologies
include some ideologically based ones (Steadfast Conservatives, Business Conservatives, and Solid
Liberals, for example). Partisan animosity is not similarly strong across the United States. Check out the
partisan prejudice map on The Atlantic site if you want to see where your county, or other counties,
stand.
https://www.people-press.org/quiz/political-typology/
https://www.theatlantic.com/politics/archive/2019/03/us-counties-vary-their-degree-partisan-prejudice/583072/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists five groups. The vertical axis lists percentage and ranges from 0 to 50 in
increments of 10. Approximate data from the table are tabulated as follows:
Group Percentage in 1956 Percentage in 2000
Ideologues 2.5 10
Near ideologues 9 9
Group interest 42 28
Nature of the times 24 28
No issue content 22.5 24
Back to Figure
The horizontal axis lists the issues. The vertical axis lists correlations in gamma and ranges from 0 to 1
in increments of 0.2. Approximate data from the graph are tabulated as follows:
Issues Correlations of public Correlations of elite
Domestic issues 0.23 0.52
Foreign issues 0.23 0.38
Between domestic and foreign issues 0.1 0.23
Back to Figure
The horizontal axis lists attitude constraint and attitude stability. The vertical axis lists correlations in r
and ranges from 0 to 1 in increments of 0.2. Approximate data from the graph are tabulated as follows:
Characteristic Public correlations, 1970sto 1980
Public correlations, 2000
to 2004
Public elites, 1980 to
1984
Characteristic Public correlations, 1970sto 1980
Public correlations, 2000
to 2004
Public elites, 1980 to
1984
Attitude
constraint
0.12 0.21 0.46
Attitude
stability
0.5 0.42 0.72
Back to Figure
The horizontal axis lists years and ranges from 1972 to 2016 in increments of 4. The vertical axis lists
percentage and ranges from 0 to 50 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year Percentage identifyingas conservative
Percentage
identifying as
moderate
Percentage
identifying as
liberal
Percentage not
identifying
1972 26 28 18 27
1976 26 25 17 33
1980 28 20 17 35
1984 28 23 18 30
1988 31 22 17 29
1992 30 22.5 20 27
1996 32 23 18 26
2000 30 22.5 19 28
2004 31 25 18 22.5
2008 31.5 21 21 22
Year Percentage identifyingas conservative
Percentage
identifying as
moderate
Percentage
identifying as
liberal
Percentage not
identifying
2012 31.5 21 21 21.5
2016 31 23 26 17
Back to Figure
The horizontal axis lists the classifications. The vertical axis lists percentage and ranges from 0 to 50 in
increments of 10. Approximate data from the graph are tabulated as follows:
Classification Percentage
Consistent conservatives 7.5
Consistent liberals 24
Symbolically conservative, operationally liberal 22
Symbolically liberal, operationally conservative 1.5
Back to Figure
The horizontal axis lists percentage of agreements and ranges from 0 to 100 in increments of 20. The
vertical axis lists opinions. Approximate data from the graph are tabulated as follows:
Opinion
Percentage of
Republicans,
including leaners, in
agreement
Percentage of
Independents
in agreement
Percentage of
Democrats,
including leaners,
in agreement
Government services: Provide more
government services even it means
an increase in spending
15 41 59
Military spending: Defense spending
should be increased
68 45 35
Opinion
Percentage of
Republicans,
including leaners, in
agreement
Percentage of
Independents
in agreement
Percentage of
Democrats,
including leaners,
in agreement
Aid to blacks: Government should
make every effort to improve the
social and economic position of
blacks
12 23 50
Abortion: By law, a woman should
always be able to obtain an abortion
as a matter of personal choice
28 43 60
Back to Figure
In all the graphs, the horizontal axis lists characteristics, and the vertical axis lists percentage, ranging
from 0 to 50 in increments of 10.
The first graph is titled, A. Ideology and partisanship intermingled. Approximate data from the graph are
tabulated as follows:
Characteristic Percentage of Democrats Percentage of Conservatives
Extremely liberal 3.5 1
Liberal 15 5
Slightly liberal 15 10
Moderate 40 35
Slightly conservative 15 25
Conservative 10 20
Extremely conservative 2 4
The second graph is titled, B. Partisan sorting. Approximate data from the graph are tabulated as
follows:
Characteristic Percentage of Democrats Percentage of ConservativesCharacteristic Percentage of Democrats Percentage of Conservatives
Extremely liberal 8 0
Liberal 20 1
Slightly liberal 33 3
Moderate 27 15
Slightly conservative 7 25
Conservative 5 45
Extremely conservative 0 10
The third graph is titled, C. Polarized. Approximate data from the graph are tabulated as follows:
Characteristic Political party Percentage
Extremely liberal Democrat 30
Liberal Democrat 45
Slightly liberal Democrat 25
Moderate Neither 0
Slightly conservative Republican 25
Conservative Republican 45
Extremely conservative Republican 30
Back to Figure
The horizontal axis lists fewer services on the left end and more services on the right end. The vertical
axis lists percentage of agreements and ranges from 0 to 60 in increments of 10. Approximate data from
the graph are as follows:
The graph shows seven cluster bars, each cluster consisting of data from 1984 and 2016. The first
cluster shows data for fewer services, and the seventh cluster shows data for more services.
Seven percent of the population were in agreement for fewer services in 1984, and 11 percent were
in agreement in 2016.
The heights of most bars increase toward the fourth cluster and then decrease.
Nine percent of the population were in agreement for more services in 1984, and 9.5 percent were
in agreement in 2016.
CHAPTER 6 ROOTS OF PUBLIC OPINION:
PERSONALITY, SELF-INTEREST, VALUES, AND
HISTORY
AFTER THE PUBLICATION of Philip Converse’s path-breaking research in 1964, which we discussed
at length in the previous chapter, much of the debate among public opinion scholars was driven by his
findings regarding the public’s lack of ideological sophistication. In fact, some observers argued that too
much attention was paid to this debate over ideology, distracting scholars from how citizens really do
think about politics.1 In this chapter, we move beyond Converse and his critics to address this question:
if ideology doesn’t organize most citizens’ opinions, what does? Donald Kinder has offered an answer to
that question. He encourages scholars to consider the pluralistic roots of public opinion.2 Specifically,
he suggests five factors that might influence citizens’ attitudes: personality, self-interest, values,
historical events, and group attitudes. Note that by pluralistic roots Kinder means many different roots.
He is not referring to the pluralist theory of democracy.
The research on the public’s lack of ideological sophistication leaves us with a view of citizens as not
being competent to play an active role in governance, which bolsters the elite democratic theorists’
argument that citizens should be removed from the policymaking process. In contrast, the research on
the pluralistic roots of public opinion resurrects a more positive view of the average citizen. From this
perspective, citizens are capable of holding reasoned, complex opinions derived from meaningful
political factors, such as values and group identity. Pluralists are especially pleased by the reliance on
group-based thinking. Participatory democratic theorists are also encouraged that citizens’ opinions
have some logic underlying them, yet we will see they are not always thrilled by the particular logic that
drives public opinion.
In this chapter we discuss four of the five factors—personality, self-interest, values, and historical events
—that shape public opinion. Each section focuses primarily on one of the factors, but you will quickly
notice that these forces are not mutually exclusive. In many instances, more than one of the factors
influences public opinion on a particular issue. In the next chapter, we focus on the fifth factor by
discussing the central role of groups in shaping public opinion.
PERSONALITY
You might recall hearing about Cambridge Analytica. The British political consulting firm garnered
worldwide attention in 2018, when news broke that they had acquired information from nearly 90 million
Facebook users. During the 2016 U.S. presidential election, Cambridge Analytica used personal data to
develop psychological profiles of users and then target political advertising to them based on these
profiles.3 Although some have questioned whether it is possible to develop complex profiles based on
people’s social media activity,4 the Cambridge Analytica operation was premised on a relationship that
social scientists identified decades ago: personality influences political views.
Back in 1950, Theodor Adorno and his colleagues introduced the concept of an “authoritarian
personality.”5 They defined authoritarianism as a set of personality traits, including submissiveness to
authority, a desire for a strong leader, general hostility and cynicism toward people, strict adherence to
convention, and a belief that people should be roundly punished if they defy those conventions.6 These
traits appear most often in people exposed to strict and rigid child-rearing practices. In other words,
children whose parents dole out a “relatively harsh and threatening type of home discipline” are more
likely to have authoritarian personalities as adults.7
This harsh upbringing leads to frustration among children, which ultimately gets redirected toward
outgroups; anger and resentment toward parents becomes displaced onto people who are considered
weak and inferior. As a result, authoritarians demonstrate high levels of intolerance for outgroups. This
theory was developed shortly after World War II as a way to make sense of anti-Semitism. Over the
years, Adorno et al.’s work on the authoritarian personality has been challenged on a number of
theoretical and methodological fronts.8 Nevertheless, the concept has continued to be a compelling one
for social scientists.
More than fifty years after the publication of Adorno’s research, Karen Stenner’s work reinvigorated the
study of authoritarianism.9 According to Stenner, “Authoritarianism is an individual predisposition
concerned with the appropriate balance between group authority and uniformity, on the one hand, and
individual autonomy and diversity, on the other.”10 She labels people who value sameness and
conformity to group norms as authoritarians and people who value diversity and individual freedom as
libertarians.
Stenner takes pains to point out that the differences between authoritarians and libertarians are not
simply a matter of political ideology.11 In earlier research, some scholars had conflated conservatism
and authoritarianism. But there is an important distinction. For example, people whom Stenner labels as
“status quo conservatives” do not mind diverse viewpoints, as long as that diversity is stable over time.
In contrast, authoritarians are bothered by the diversity in and of itself. Another way to think of the
distinction is that status quo conservatives do not like change, whereas authoritarians are pleased with
change as long as it moves people toward greater “oneness and sameness.”12 (Also note that we briefly
discussed libertarians in Chapter 5. There we discussed libertarianism as a political ideology, whereas
here Stenner is using the term to refer to a personality predisposition.)
Stenner argues that authoritarianism and libertarianism constitute very broad, normative worldviews
about the way society should function and that these worldviews influence citizens’ opinions across a
wide range of political, racial, and moral issues. Because authoritarians value conformity and obedience
to authority, they favor policies that stifle diversity and enforce sameness across citizens. On the other
end of the continuum, libertarians value freedom and difference and thus oppose coercive government
policies or policies that discourage individuality.
Stenner provides evidence of the impact of authoritarianism on public opinion by analyzing national
survey data from the General Social Survey (GSS). But before we dive into Stenner’s research findings,
we must first discuss how she measures the concept of authoritarianism. Stenner argues that the best
way to determine whether someone is an authoritarian is to examine his or her beliefs about child-
rearing practices. Child-rearing values “can effectively and unobtrusively reflect one’s fundamental
orientations toward authority/uniformity versus autonomy/difference.”13 The GSS asks respondents to
rank child-rearing values by indicating which qualities are “most desirable” for a child to have and which
ones are “least important” (see Table 6.1). According to Stenner, authoritarians place a high value on
obedient, neat, and well-mannered children, whereas libertarians are partial to inquisitive and
responsible children with good judgment.14
Stenner examined data from surveys conducted by the GSS between 1990 and 2000; she focused her
attention solely on white respondents.15 Her analysis showed that authoritarians are significantly more
likely than libertarians to be racially, morally, and politically intolerant. In terms of race, authoritarian
citizens tend to oppose interracial marriage, support housing segregation, and say they would not vote
for a Black presidential candidate. When it comes to moral issues, authoritarians oppose homosexuality,
believe in compulsory school prayer, and think that pornography should be banned. Citizens who are
authoritarian are also opposed to civil liberties for groups from both the left and right sides of the
ideological spectrum. For example, authoritarians do not think that homosexuals or racists should be
allowed to teach in a college or university. In addition, authoritarians are more punitive than libertarians.
Authoritarians are more likely to support the death penalty, believe that courts are too soft on criminals,
support wiretapping, and own a gun. Taken as a whole, Stenner provides strong evidence that white
authoritarian citizens abhor difference and diversity and believe people should be harshly punished for
disobedience.
Table 6.1 Measuring Authoritarianism
Authoritarian child-rearing values Libertarian child-rearing values
“That a child obeys his or her parents
well”
“That a child is interested in how and why things
happen”
“That a child is neat and clean” “That a child has good sense and sound judgment”
“That a child has good manners” “That a child is responsible”
Sources: Karen Stenner, The Authoritarian Dynamic (New York: Cambridge University Press, 2005), 164–165; “General Social Surveys,
1972–2018: Cumulative Codebook,” National Opinion Research Center, March 2019,
http://gss.norc.org/documents/codebook/gss_codebook , 416–422.
Note: The General Social Survey asks respondents to rank order thirteen child-rearing values using a series of questions: (a) “Which
three qualities listed on this card would you say are the most desirable for a child to have?” (b) “Which one of these three is the most
desirable of all?” (c) “All of the qualities listed on this card may be desirable, but could you tell me which three you consider least
important?” (d) “And which one of these three is least important of all?” Stenner constructed her authoritarianism scale based on the
ranking of the six values presented in this table.
Stenner’s research also examined the effect of normative threats on citizens with authoritarian
predispositions. To authoritarians, a normative threat is a situation in which oneness and sameness are
called into question. It is worth quoting Stenner at some length on her definition of normative threats:
In diverse and complex modern societies, the things that make us one and the same are
common authority and shared values. The conditions most threatening to oneness and
sameness, then, are questioned or questionable authorities and values: that is, disrespect for
leaders or leaders unworthy of respect and lack of conformity to or consensus in group values,
norms, and beliefs.16
http://gss.norc.org/documents/codebook/gss_codebook
When citizens are faced with normative threats, authoritarianism becomes activated and thus has a
stronger effect on intolerance of difference. In other words, those citizens with an authoritarian bent
become even more intolerant when they are in a situation in which “diversity and freedom ‘run amok.’”17
In a series of survey-based experiments, Stenner provides evidence to support this interaction between
authoritarianism and normative threats. In one condition, interviewers read subjects a (fictitious) news
story about how public opinion is increasingly divided in the United States on important political issues.
This news story represents a normative threat to oneness and sameness because it focuses on division
among the public. After exposure to the story, authoritarians became more intolerant and libertarians
became more tolerant when asked their opinion about whether whites have a right to keep Black people
out of their neighborhoods. In another condition, interviewers read subjects a (fictitious) news story
about how strong and trustworthy U.S. presidents have been and how citizens can expect high-quality
presidential candidates in the future. This story is a normatively reassuring one to authoritarians
because it emphasizes authority. After exposure to this normative reassurance, authoritarianism and
libertarianism were in effect deactivated, and the two groups became more similar in terms of their
levels of racial tolerance as a result. As Stenner puts it, “They each let down their defenses in the wake
of such reassurance, rendering authoritarians calmed and libertarians complacent, the latter fairly
characterized as ‘asleep at the wheel’” in this condition.18
Stenner also examined the interaction of authoritarianism with normative threats in other domains
(political and moral) and found the same pattern of results: normative threats activated authoritarianism
and thus led people with authoritarian leanings to be more intolerant and people with libertarian
predispositions to be more tolerant. These research findings are problematic for participatory democratic
theorists. Recall that participatory democratic theorists believe that citizens can come together to
discuss political issues in a productive way. Citizens can learn from one another, which will help them
look beyond their self-interest and gain an understanding of what is best for the nation as a whole.
Stenner’s research, however, pokes a big hole in this argument. Her findings suggest that if
authoritarians are exposed to diverse viewpoints during conversations with their fellow citizens, they
would view that as normatively threatening and thus become even more intolerant; however,
participatory democratic theorists can take some solace from Stenner’s results showing that libertarians
exposed to diverse viewpoints become more tolerant.
Marc Hetherington and Jonathan Weiler also examine authoritarianism and its impact on policy attitudes
by analyzing data from national surveys conducted in the 1990s and 2000s. These scholars argue that
authoritarianism is best thought of as a worldview about what is right and what is wrong.19 Like Stenner,
they measure authoritarianism by assessing citizens’ child-rearing values. Those who “score high in
authoritarianism have (1) a greater need for order and, conversely, less tolerance for confusion or
ambiguity, and (2) a propensity to rely on established authorities to provide that order.”20 Hetherington
and Weiler demonstrate that authoritarians and nonauthoritarians (that is, those whom Stenner would
call libertarians) have very different attitudes toward many of the most prominent issues of our day. As
demonstrated in Figure 6.1, these patterns were still present in 2016. Only 45 percent of the citizens
who score the highest on the authoritarian scale support same-sex marriage, whereas 88 percent of
those who score lowest support marriage equality. Fifty-seven percent of the least authoritarian
individuals believe that the use of wiretaps by the government since the 9/11 attacks has gone too far,
compared to 30 percent of authoritarians. Support for building a wall along the U.S. border with Mexico
was 32 percentage points higher for authoritarians than nonauthoritarians (42 vs. 10 percent).
Authoritarianism is also related to views regarding race in U.S. society. For instance, two-thirds of
authoritarians but only one-quarter of nonauthoritarians believe that Blacks should overcome prejudice
and work their way up without special favors. Taken as a whole, these are striking differences in opinion
based on citizens’ level of authoritarianism. Furthermore, differences in issue opinions between
authoritarians and nonauthoritarians have existed for decades.21
Description
Figure 6.1 The Effect of Authoritarianism on Political Attitudes, 2016
Source: Analysis of American National Election Studies 2016 data file.
One final item of note: the 2016 results that we have discussed so far incorporate Americans of all races
and ethnicities, whereas most studies of authoritarianism, including Stenner’s as well as Hetherington
and Weiler’s, examine the views of whites only. For some issues, the relationship between opinion and
authoritarianism looks essentially the same whether the analysis includes people of different races and
ethnicities or only whites. For others, however, the gap between authoritarians and nonauthoritarians
becomes larger if only whites are examined. This is the case for the third and fourth issues in Figure 6.1.
When only whites are included in the analysis (see the bottom of the figure), authoritarians’ support for a
border wall increases as does the belief that Blacks do not deserve special favors. This pattern
demonstrates that authoritarianism can be a larger driver of political views for whites than for people of
color, a topic we return to later in this section.
In addition to providing further evidence that authoritarianism affects attitudes toward a range of issues,
Hetherington and Weiler make two new contributions to the study of authoritarianism. First, the scholars
argue that studies showing nonauthoritarians become more tolerant in the face of threats are flawed
because nonauthoritarians do not actually perceive the information they are exposed to as
threatening.22 To illustrate this point, consider Stenner’s experiment again. The normative threat she
exposed her subjects to was disagreement among the public, which is a threat to the oneness and
sameness valued by authoritarians. Divided public opinion, however, is not a threat to libertarians; in
fact, they probably appreciate diversity in citizens’ views. According to Hetherington and Weiler, when
nonauthoritarians are confronted with truly threatening situations, they become more authoritarian, not
more tolerant. Indeed, the scholars demonstrate that nonauthoritarians who are very worried about the
war on terror (and therefore presumably feel threatened by it) are much more likely to support
warrantless wiretapping than nonauthoritarians who are not worried at all about the war on terror. In fact,
worried nonauthoritarians are as supportive of wiretapping as authoritarians, who are pro-wiretapping
regardless of their level of concern about the war on terror. This finding helps explain why the
percentage of citizens who were willing to forgo civil liberties increased after the terrorist attacks of 9/11.
Authoritarians were already willing to sacrifice civil liberties for security prior to 9/11, but
nonauthoritarians weren’t willing to do so until they felt threatened by the tragic events of that day. They
changed their minds and moved in a more authoritarian direction, which led to an aggregate increase in
support for security over civil liberties. (See Chapter 9 for further discussion of attitudes toward civil
liberties post-9/11.)
Hetherington and Weiler’s second contribution connects authoritarianism and political parties. They
argue that the most prominent issues of our day are ones on which authoritarianism matters a great
deal, such as gay rights, race, immigration, and the war on terror. Political parties have taken distinct
positions on these issues, with the Republican Party taking the authoritarian stance and the Democratic
Party taking the nonauthoritarian perspective. Even though citizens tend to be fairly moderate on these
issues, over the past two decades they have “sorted” themselves into one party or another so that their
authoritarian leanings correspond with their party identification. In other words, authoritarians are
increasingly aligning themselves with the Republican Party and nonauthoritarians are increasingly
identifying with the Democratic Party.23 More recent work by Hetherington and Weiler concludes that
this trend has persisted.24 As a result, because authoritarianism is a fundamental worldview, not simply
different policy preferences, the gap between the two parties seems quite large. Among citizens, the
average Republican and average Democrat may not be all that different, but their differences stem from
deep-seated understandings of right and wrong, making American politics seem quite polarized.25
Box 6.1 Authoritarianism across the World
Public Opinion in Comparative Perspective
In her book The Authoritarian Dynamic, Karen Stenner argues that the concept of
authoritarianism helps us understand intolerance of difference not just in the United States but
across the world “from Switzerland to China to Nigeria to Azerbaijan.”1 She analyzed the World
Values Survey, which includes survey data collected from over 110,000 people in fifty-nine
nations between 1990 and 1998. She measured authoritarianism based on people’s evaluation
of child-rearing practices. Across all these nations, Stenner demonstrated that authoritarians are
more likely to be intolerant across a range of racial, political, and moral issues. For example,
authoritarian people are more opposed to homosexuality, abortion, divorce, and racial integration
than libertarians. Further, authoritarians are more likely to believe that jobs should be reserved
for the native-born, that maintaining order is more important than free speech, and that fighting
crime is more important than progressing toward a humane society. More recently,
authoritarianism has been linked to anti-migrant attitudes and support for right-wing populist
political parties throughout Europe.2
What is so fascinating about Stenner’s work is that she demonstrates the linkage between
authoritarianism and intolerance in a wide variety of countries across the globe. This is
noteworthy because conflict within countries is often attributed to factors specific to that country.
So, for example, when observers try to explain the breakup of Yugoslavia during the 1990s, they
often point to long-standing ethnic hatreds that were unleashed after the 1980 death of Tito, the
country’s powerful and charismatic communist leader. In France, conflict surrounding the
assimilation of North African immigrants is most often explained by cultural and religious factors
unique to that country and those particular immigrants. But Stenner argues that regardless of the
particular groups involved or whether difference is based on culture, religion, race, ethnicity, or
tribes, the primary factor that explains intolerance is authoritarianism. Why? Because
authoritarianism is a universal personality predisposition. Stenner concludes, “Authoritarianism
rather consistently produces a predictable cluster of sociopolitical stances varying in target and
form, but never in function: the animating spirit throughout is to limit difference in people, beliefs,
and behaviors.”3
1 Karen Stenner, The Authoritarian Dynamic (New York: Cambridge University Press, 2005),
129.
2 Marc Hetherington and Jonathan Weiler, Prius or Pickup? How the Answers to Four Simple
Questions Explain America’s Great Divide (Boston: Houghton Mifflin Harcourt, 2018), Chap. 7.
3 Stenner, The Authoritarian Dynamic, 116.
Authoritarianism is not the only personality predisposition examined by political scientists. In recent
years, scholars have also investigated the relationship between specific personality traits and political
attitudes. There are five traits, the Big Five, that personality researchers have identified as fundamental
to understanding individuals’ attitudes and behaviors: openness to experience, conscientiousness,
agreeableness, extraversion, and emotional stability (see Table 6.2). These traits stem from genetic
differences, and they form early in childhood, remaining largely stable throughout a person’s lifetime.
Scholars measure these personality characteristics by asking survey respondents to indicate whether
particular traits apply to themselves (see Table 6.3).26 For example, a respondent is conscientious if she
thinks of herself as dependable and self-disciplined and does not consider herself disorganized and
careless. Someone open to new experiences would also think of himself as complex but neither
conventional nor uncreative.
Table 6.2 Big Five Personality Traits
Personality trait Description
Openness to
experience
Individuals with this trait respond positively to a wide range of experiences;
they appreciate complexity and novelty.
Conscientiousness Individuals with this trait engage in socially prescribed impulse control; they
appreciate rules and norms.
Agreeableness Individuals with this trait are kind and communal; they are not antagonistic
toward others.
Extraversion Individuals with this trait are active and energetic; they are sociable and
outgoing.
Emotional stability Individuals with this trait are steady and resilient; they are not nervous and
anxious.
Source: Based on Alan S. Gerber, Gregory A. Huber, David Doherty, Conor M. Dowling, and Shang E. Ha, “Personality and Political
Attitudes: Relationships across Issue Domains and Political Contexts,” American Political Science Review 104 (2010): 111–133.
Look around your dorm room or apartment (or, if you are not in your living space now, imagine its
contents). Do you see calendars, an iron, and a laundry basket? What about books and music—are
there many types of books and CDs, or are your book and music collections rather homogeneous? Do
you have any art supplies in your room or evidence of travel (such as an airline ticket, travel books, or
souvenirs)? Are your living quarters neat or messy? Well lit or quite dark? According to research in
psychology, your answers to these questions are likely to be related to whether you are conservative or
liberal.27 A study of college students and new college graduates demonstrates that conservatives tend
to have neater, better lit, and better organized (hence the presence of calendars and laundry supplies)
residences. Liberals, on the other hand, live in messier and darker rooms that contain a wider variety of
books, music, and artistic supplies as well as travel literature and souvenirs.
Table 6.3 Ten-Item Personality Inventory
“Here are a number of personality traits that may or may not apply to you. Please write a number
next to each statement to indicate the extent to which you agree or disagree with that statement.
You should rate the extent to which the pair of traits applies to you, even if one characteristic applies
more strongly than the other.”
Disagree
strongly 1
Disagree
moderately 2
Disagree
a little 3
Neither agree nor
disagree 4
Agree a
little 5
Agree
moderately
6
Agree
strongly 7
I see myself as:
1. _____ Extraverted, enthusiastic.
2. _____ Critical, quarrelsome.
3. _____ Dependable, self-disciplined.
4. _____ Anxious, easily upset.
5. _____ Open to new experiences, complex.
6. _____ Reserved, quiet.
7. _____ Sympathetic, warm.
8. _____ Disorganized, careless.
9. _____ Calm, emotionally stable.
10. _____ Conventional, uncreative.
Source: Samuel D. Gosling, “Ten Item Personality Measure (TIPI),” http://gosling.psy.utexas.edu/scales-weve-developed/ten-item-
personality-measure-tipi
Note: Each of the Big Five personality dimensions is measured by answers to two of the trait pairs: Openness to experiences: agreeing
with 5, disagreeing with 10; Conscientiousness: agreeing with 3, disagreeing with 8; Agreeableness: agreeing with 7, disagreeing with 2;
Extraversion: agreeing with 1, disagreeing with 6; and Emotional stability: agreeing with 9, disagreeing with 4.
This intriguing work demonstrates that personality traits can be linked to both ideological leanings and
behavior. Being open to new experiences is related not only to having a messier room and owning more
books but also to holding a liberal ideology. Conscientious individuals are neater, more organized, and
more likely to be conservative. The openness-liberalism and conscientiousness-conservatism
connections are quite robust, having been found in a number of studies.28 To explain these
relationships, one scholar pointed to features of these two ideologies: “Liberalism is conceived of
embracing change and proactive policies, whereas conservatism is likened to personal responsibility,
caution, and maintaining order.”29
The Big Five traits are also associated with party identification. Given that the Democratic Party is the
more liberal party in the United States, it should come as no surprise to learn that the trait of openness
is related to Democratic Party identification. Republican identifiers are quite high in conscientiousness,
again not surprisingly.30 Beyond which party a person identifies with, Big Five traits are related to the
strength of one’s party identification and whether someone even identifies with a party.31 People high in
extraversion or high in agreeableness are more likely to have a partisan identity and have stronger party
identifications than do those who are low in either trait. Extraverts are sociable whereas agreeable
individuals are communal, thus likely leading them to be “drawn to the affective, social benefits of party
identification … [or] to the communal and cooperative components of joining a political ‘team.’”32 In
contrast, those high in openness to new experiences are less likely to affiliate with a political party.
Partisan ties help to structure the political world, providing cues and context with which to interpret
political debates and events. Those who are high in openness do not desire structure to the degree that
those low in openness do, perhaps making partisanship less appealing to the former.
According to research conducted by Alan Gerber and his colleagues, the Big Five also have a
substantial impact on citizens’ issue opinions, particularly toward economic and social policies.33 The
scholars measured economic policy attitudes by assessing citizens’ views on the proper role of
government in the provision of health care and attitudes toward taxes on the wealthy. Social policy
attitudes were measured by asking citizens about their abortion attitudes and whether they supported
civil unions for gay and lesbian couples. They find that citizens who are open to experiences are more
liberal on economic and social issues, probably because they respond more positively to new programs
and initiatives. Conscientious individuals, however, are more conservative on economic and social
issues; they are rule followers and value self-discipline, which makes them more amenable to
conservative policies. People who are agreeable have more liberal economic attitudes, most likely
because they view those policies as helpful to others, yet they are more conservative on social issues,
perhaps because they see social policies such as same-sex marriage as disruptive to the community. In
contrast, emotionally stable citizens are more conservative on economic policy. Because they are less
anxious, they are less likely to see the need for government to step in and regulate the economy.
Gerber et al. also consider whether the relationship between personality traits and political attitudes
differs for whites and Blacks. Why might the relationship differ for the two groups? Because whites and
Blacks operate in substantially different political environments and the effects of personality on policy
attitudes is context-specific. That is, political context influences how individuals interpret government
policies, which then affects whether a particular personality trait leads to more or less support for those
policies. For example, given historical and current discrimination against them, Black Americans
perceive liberal economic policies as helping those who have been systematically denied opportunities
to succeed in the marketplace. Thus, Gerber et al. conclude that Black people tend to support liberal
economic policies regardless of their level of conscientiousness. Christopher Johnston, Howard Levine,
and Christopher Federico find the same, and they also conclude that Blacks’ economic policy views are
unrelated to traits such as openness to new experiences or authoritarianism. On the other hand, traits
and economic policy opinions are related for whites and, to a lesser extent, Latinos.34 They explain
these findings largely by pointing to the minimal variation in Blacks’ economic policy opinions compared
to much larger variation among whites, with Latinos falling in between. When a group’s members’
attitudes do not vary very much, there is little room for personality traits (or other characteristics, for that
matter) to influence attitudes. Although this conclusion does seem reasonable, Johnston and his
coauthors admit that it is tentative, largely because most national surveys do not contain enough
respondents of color for thorough exploration of racial and ethnic differences. More work is needed to
fully explore these dynamics.
SELF-INTEREST
It seems incredibly intuitive that self-interest would have an important effect on our policy attitudes.
When considering human nature, it certainly seems as if people are looking out for number one. Indeed,
James Madison argued that a representative form of government is the best form of government
because citizens are too focused on their narrow self-interest, whereas representatives have the
wisdom to “best discern the true interest of their country.”35 Elite democratic theorists have used this
argument to justify why elites (rather than citizens) should have central decision-making roles in politics.
Despite the intuitive—even compelling—nature of the claim that citizens follow their self-interest, there is
actually quite limited evidence to support the proposition. On policy opinions ranging from government
spending to government health insurance to race and gender issues to foreign policy, scholars have
found only weak or nonexistent effects of self-interest.36 For example, several studies showed that white
nonparents were as likely to oppose school busing as a means to achieve racial integration as white
parents with school-age children.37 Instead of self-interest, racial prejudice was a key factor influencing
attitudes on school busing: prejudiced citizens were more opposed to busing, whereas nonprejudiced
citizens were more supportive of the policy (regardless of whether or not the citizens had kids). Racial
prejudice is also a much stronger predictor of people’s attitudes regarding government spending on
domestic programs than are characteristics that are linked to whether one would benefit personally from
such spending (such as by being unemployed).38 Other research indicates that citizens’ evaluations of
the nation’s economy are more important than their own personal economic circumstances when
assessing the political party in power.39 In other words, general concerns about society—what political
scientists call sociotropic concerns—trump pocketbook issues when citizens evaluate their
government.
There are a few instances, however, when self-interest does influence citizens’ policy attitudes.
Homeowners are more likely to favor property tax cuts than nonhomeowners.40 Smokers are more
opposed to cigarette taxes and bans on smoking than nonsmokers.41 And gun owners are less
supportive of gun restrictions than people who do not own guns.42 These examples illustrate the
conditions under which self-interest can be influential: when the effects of a policy are visible, tangible,
large, and certain.43 The innovative studies we profile next provide further evidence of such conditions
triggering self-interest.
Robert Erikson and Laura Stoker demonstrate that self-interest can have powerful and long-lasting
effects when citizens are faced with circumstances in which their lives might be severely disrupted and
even put in jeopardy.44 Specifically, Erikson and Stoker examine what happens to the political attitudes
of young men when they are faced with the prospect of being drafted for military service. In 1969 in the
midst of the Vietnam War, Republican president Richard Nixon instituted a new policy that assigned
numbers (1 through 366) to draft-eligible men based on their birth dates. The men assigned low
numbers were called up first for duty, whereas the men assigned high numbers were virtually assured
they would not be drafted. In effect the policy randomly assigned some men to be vulnerable to being
sent to war in Vietnam and others not to be. This was a perfect case in which self-interest should have
shaped public opinion because those with low draft numbers faced “a (relatively) high likelihood of being
forced to abandon all personal plans and undertakings and to take part in a potentially life-threatening
war. As one’s lottery number increased, one’s vulnerability decreased.”45
As luck would have it, a representative sample of young men affected by this draft policy were
interviewed in 1965 and reinterviewed several times later as part of an ongoing panel study to examine
political attitudes and socialization. (This is the Jennings and Niemi panel study we discussed at some
length in Chapter 2.) The data collected included the respondents’ birth dates, which allowed Erikson
and Stoker to determine the draft number assigned to each male respondent. As a result, the
researchers were able to use these data to investigate whether vulnerability to the draft changed young
men’s political attitudes. Indeed, they found striking evidence that being assigned a lower draft number
influenced attitudes in several ways. When reinterviewed in 1973, men with lower numbers were more
likely to think the war in Vietnam was a mistake than those with higher numbers. In addition, compared
with the men who held high draft numbers, the men vulnerable to the draft were less likely to have voted
for Nixon for reelection and were more likely to express a liberal ideology and liberal issue positions.
Remarkably, when interviewed twenty-four years later in 1997, the men who had been assigned low
numbers continued to be more likely to report that the war was a mistake than those assigned high
numbers. Erikson and Stoker also present evidence that the vulnerable men reconsidered their
partisanship, which led them to become more Democratic and largely stay that way into later adulthood.
Finally, it was not only these men who were politically affected by their draft numbers. Their parents
were as well.46 Parents of sons who had low draft numbers were more likely to vote in the 1972
presidential election than were parents of sons with high numbers, especially if the families with low-
draft-number sons lived in a town where at least one soldier had been killed in the Vietnam War.
Self-interest also comes into play for a current public policy challenge: how to address the opioid
overdose crisis. In particular, Justin de Benedictis-Kessner and Michael Hankinson demonstrate that
both financial and spatial self-interest influence public opinion regarding this issue.47 They conducted a
nationally representative survey experiment in 2018. To examine financial self-interest, they assessed
support for a proposal that would allocate state funding to treat opioid addiction. Half of the respondents
were told that to pay for this new funding, taxes would increase by $55 for people earning more than the
state’s median income, whereas those earning less than the median would see their taxes increase by
$5. In contrast to this income-based proposal, the other half of the respondents were asked about an
overdose rate-based option. For that proposal, localities with higher rates of opioid overdoses would pay
more than localities with lower rates (again, a $55 vs. $5 increase in taxes).
Did financial self-interest play a role in people’s thinking about these proposals? Yes, it did. For both
options, those who would see a larger tax increase were less supportive of the policy proposal, by
nearly 10 percentage points, than those whose taxes would increase by only $5. Other personal
characteristics were also related to policy support. Overall, for example, partisanship was more strongly
predictive of policy opinions than financial self-interest, with Republican citizens displaying less support
than Democrats for both options. Yet, when self-interest and partisanship were both taken into
consideration, partisan differences reduced for the income-based proposal. This was because lower-
income Republican citizens, who would bear less of a financial sacrifice than their wealthier peers, were
more favorable toward the proposal than were other Republicans.
Spatial self-interest concerns came to the fore for another policy proposal examined by de Benedictis-
Kessner and Hankinson. This proposal asked respondents whether they would be in favor of a new
clinic to treat opioid addiction. Half were told that the new clinic would be built two miles from their
house. For the other half, the clinic was to be situated only one-quarter of a mile away. Support was
substantially higher when the proposed clinic was to be two versus one-quarter of a mile away (53 and
39 percent, respectively). For this proposal, however, the role of self-interest did not differ for
Republicans and Democrats. Both groups were less willing to support the construction of a clinic if it
were to be located near their homes. Ultimately, the authors imply that government spending might be
more palatable to the public than infrastructure projects when it comes to opioid treatment policy.
“Policymakers may leverage self-interest specifically among lower-income conservative voters to build
broad coalitions of policy support,” they write, whereas “spatial self-interest consistently provokes
NIMBY [not in my backyard] opposition toward new treatment clinics.”48 In terms of why self-interested
concerns are relevant for this particular policy domain, de Benedictis-Kessner and Hankinson point to
the high salience of the opioid overdose crisis as well as the financially or spatially concentrated
burdens that may fall on some people to address the crisis.
VALUES
Values are “general and enduring standards.”49 They are abstract beliefs about how the world should
work. As such, values constitute citizens’ core principles, guiding their understanding of right and wrong.
Thus, it makes sense that citizens’ values would influence their specific policy positions.50
Scholars have identified two fundamental values that influence public opinion: egalitarianism and
individualism.51 Egalitarianism is the belief that citizens should be equal regardless of their personal
characteristics.52 In the U.S. context, egalitarianism emphasizes equality of opportunity, not necessarily
equality of results. In other words, egalitarianism is the belief that all citizens should have the chance to
achieve rather than the belief that all citizens should be guaranteed equal outcomes. Individualism is
the belief that citizens should get ahead by virtue of their own hard work; people should “pull themselves
up by their own bootstraps” and rely on their own ingenuity.
Egalitarianism and individualism are abstract concepts and therefore difficult to measure. Nevertheless,
political scientists have devised a set of survey questions to assess these concepts. In particular,
Stanley Feldman analyzed citizens’ levels of agreement with several statements that were included on a
pilot study for the 1984 ANES to come up with the best way to measure egalitarianism and
individualism.53 He identified three statements that provide a valid measure of egalitarianism (see Table
6.4 for the wording of these items). The first two items seem to focus on support for equal opportunity,
whereas the third is more ambiguous. Some people might infer that treating people “more equally”
means ensuring equal results, whereas others might think the statement simply refers to providing
people with equal opportunities. Feldman also identified five statements that provide a valid measure of
individualism (again, see Table 6.4 for the wording of these items). These statements emphasize the
personal effort that is needed for someone to get ahead in life. (Note that agreeing with the first three
statements is the individualistic response, whereas disagreeing with the last two statements is the
individualistic response.) What do you think—do these statements do a good job measuring
individualism and egalitarianism?
Table 6.4 Measuring Egalitarianism and Individualism
“I am going to read several statements. After each one, I would like you to tell me whether you
agree or disagree. I would also like to know whether you agree or disagree strongly or not strongly.”
Egalitarianism Individualism
“Our society should do whatever is necessary to
make sure that everyone has an equal
opportunity to succeed.”
“One of the big problems in this country is that we
don’t give everyone an equal chance.”
“If people were treated more equally in this
country, we would have many fewer problems.”
“Any person who is willing to work hard has a
good chance of succeeding.”
“If people work hard, they almost always get
what they want.”
“Most people who don’t get ahead should not
blame the system; they really have only
themselves to blame.”
“Hard work offers little guarantee of success.”
“Even if people try hard, they often cannot reach
their goals.”
Source: Stanley Feldman, “Structure and Consistency in Public Opinion: The Role of Core Beliefs and Values,” American Journal of
Political Science 32 (1988): 421.
Feldman analyzed the impact of egalitarianism and individualism on citizens’ policy attitudes.54 He
demonstrates that egalitarianism is closely related to citizens’ opinions on a wide range of policies. For
example, egalitarian citizens are more likely to support welfare programs; increased government
spending on health and education; and government efforts to improve the societal position of African
Americans, women, and the poor. Thus, across many different policy areas, egalitarianism leads to
more progressive political views. In contrast, individualism has an effect in only a few policy areas;
nevertheless, its influence is still noteworthy. Individualistic citizens are more likely to oppose welfare
spending and prefer a more limited role for the federal government (compared with state governments)
in handling social and economic problems.
Political values also influence public opinion in the domain of gay rights.55 Egalitarianism has a strong
effect on support for gay rights. Specifically, citizens who are more egalitarian tend to favor laws that
protect gays and lesbians against job discrimination and believe that gays and lesbians should be
allowed to serve in the military. Individualism, especially as this value highlights personal freedom, is
related to support for same-sex marriage and favoring laws allowing same-sex couples to adopt
children. Another influential value—moral traditionalism—is highlighted by Paul Brewer’s research.56
Moral traditionalism refers to citizens’ “underlying predispositions on traditional family and social
organization.”57 Moral traditionalists prefer stable, two-parent—one male and one female, that is—
families and are opposed to changing norms regarding family structure and acceptable lifestyles. See
Table 6.5 for the statements used to measure moral traditionalism.58 (Note that agreeing with the first
and third statements is the morally traditional response, whereas disagreeing with the second and fourth
statements is the morally traditional response.)
Analyzing ANES data from 1992, 1996, and 2000, Brewer found that citizens who are moral
traditionalists are more likely to oppose laws to protect gays and lesbians from job discrimination and to
think that gays and lesbians should not be allowed to serve in the armed forces. This held for all three
years under study, but moral traditionalism was a more influential factor explaining attitudes toward gay
rights in 1992 and 1996 than in 2000.
Table 6.5 Measuring Moral Traditionalism
“Now I am going to read several statements. After each one, I would like you to tell me whether you
agree strongly, agree somewhat, neither agree nor disagree, disagree somewhat, or disagree
strongly with this statement.”
“The newer lifestyles are contributing to the breakdown of our society.”
“The world is always changing and we should adjust our view of moral behavior to those changes.”
“This country would have many fewer problems if there were more emphasis on traditional family
ties.”
“We should be more tolerant of people who choose to live according to their own moral standards,
even if they are very different from our own.”
Source: Paul R. Brewer, “The Shifting Foundations of Public Opinion about Gay Rights,” Journal of Politics 65 (2003): 1218.
Although the primary focus of this section is on values, we would be remiss if we failed to discuss
Brewer’s other important findings regarding support for gay rights. First, Brewer suggests that moral
traditionalism had less of an impact on citizens’ views in 2000 because of people’s reactions to historical
events. He points to the brutal murder of Matthew Shepard as an event that probably shaped how
citizens thought about gay rights. Matthew Shepard was a college student at the University of Wyoming.
In 1998, he was kidnapped, viciously beaten, tied to a fence, and left to die. His death received
extensive media coverage. Why would this tragedy influence public opinion on gay rights? Because
Matthew Shepard was gay. Brewer speculates that this repugnant act might have led some people to
question whether their traditional moral views should be linked to opposing gay rights.
Second, Brewer shows that attitudes toward gays and lesbians are an important predictor of opinion on
gay rights. Citizens who have negative feelings toward gays and lesbians are more likely than those
who have positive feelings to oppose employment rights for the group. This was true each year under
study, although attitudes toward gays and lesbians became a less important predictor of gay rights
opinion across the time period. Again, Brewer links this decrease in the importance of attitudes toward
gays and lesbians to the public responding to historical events. As gays and lesbians became more
visible on television (at the time Ellen and Will and Grace, both with lesbian and gay lead characters,
were popular sitcoms), Brewer argues that people became more familiar with members of the group,
therefore breaking down stereotypes and reducing hostility. Citizens began to see gays and lesbians as
individual people rather than an undifferentiated mass. As a result, citizens were less likely to derive
their opinions on gay rights from their feelings toward gays and lesbians as a group.
More recent research concludes that moral traditionalism continues to influence citizens’ attitudes
regarding employment nondiscrimination policies.59 Moral traditionalism also has an impact on attitudes
toward same-sex marriage and same-sex couple adoption, particularly contentious issues in American
politics in recent years. Citizens who endorse moral traditionalism are significantly more likely to oppose
same-sex marriage and the adoption of children by same-sex couples than are citizens who do not
agree with that value.60 Even though the public has become much more supportive of same-sex
marriage over time (see Chapter 10), it is likely the issue will continue to be visible and vexing for those
who feel a core value is at stake.
In sum, several scholars have demonstrated the important role of values in shaping citizens’ issue
attitudes. Individualism, egalitarianism, and moral traditionalism are all abstract, enduring standards that
influence public opinion. Brewer’s research on attitudes toward gay rights also points toward two
additional factors that are key to understanding public opinion: historical events and group attitudes. We
discuss the effects of historical events in the next section and the influence of groups in the next
chapter.
HISTORICAL EVENTS
Citizens’ attitudes are also shaped by historical events, a phenomenon we touched on earlier. For
example, recall that international crises have an impact on public approval of the president.61 Presidents
typically get a boost in popularity in the wake of a foreign policy–related crisis, known as a rally-round-
the-flag effect (for further detail, see Chapter 4). As noted earlier in this chapter, high-profile events (e.g.,
the opioid crisis) can heighten self-interest considerations, whereas others (e.g., the brutal murder of
Matthew Shepard) can lessen the connection between core values and issue opinions.
The battle over civil rights in the mid-twentieth century also influenced political attitudes. In a
groundbreaking study, Taeku Lee demonstrated the impact of social movements, particularly the civil
rights movement, on public opinion.62 Specifically, Lee argues that events initiated by social movements
can have a more powerful effect on public opinion than events initiated by political elites. He argues that
too often public opinion research focuses on how attitudes are shaped by elite actions and does not
dedicate enough attention to the influence of nonelites, such as movement activists. Furthermore, Lee
makes the case that political scientists rely too heavily on public opinion polls to assess citizens’
attitudes. He maintains that public opinion polls are not well suited for the task of distinguishing between
elite and nonelite influences on the public.
To illustrate the effect of nonelite events on public opinion, Lee conducted an in-depth analysis of the
civil rights movement between 1948 and 1965. Lee shows that political events initiated by movement
activists galvanized the Black public, which then led to a backlash from white Southerners who opposed
the goals of the civil rights movement. This, in turn, led to national elites reacting to the events on the
ground in the South, as well as racially liberal white citizens from across the country responding
sympathetically to the plight of Black citizens.
Lee tracked this chain of events not by monitoring public opinion data from polls but by examining letters
written to the president of the United States during this period. Lee argues that constituency mail
measures activated mass opinion, which is “beliefs and sentiments that are at once salient in the mind
and impel one to political action.”63 He shows that letter writing to the president increased in response to
events initiated by civil rights activists. For example, the Montgomery bus boycott, Freedom Summer,
and especially Bloody Sunday in Selma stimulated an outpouring of letters from citizens.64 In contrast,
the passage of the Civil Rights Act of 1964, the presidential election of 1964, and the Voting Rights Act
of 1965—all elite-driven events typically considered watershed moments in the push for civil rights—did
not inspire citizens to write the president.65
Lee also examined the race of the correspondent, which was identifiable in about 47 percent of the
letters, and the timing of the letter writing.66 He found that Blacks, who were universally sympathetic to
the civil rights activists, were the first to be activated by movement events. This was followed by
sympathetic Northern whites who wrote to the president outraged by the extremely violent backlash of
white Southerners against Blacks participating in the civil rights movement. Next, white Southerners
were motivated to write the president in strong opposition to elite actions at the national level to end
racial segregation. Finally, racially liberal whites from across the nation (except the South) wrote to the
president in support of integration. Based on this pattern of letter writing, Lee does not dismiss the
importance of elite events entirely but concludes that a “movement-initiated, movement-elite interaction”
model best describes the events that influenced public opinion on racial issues during this period.67 In
other words, citizens’ opinions were affected by movement events, and their opinions were shaped by
elite activities in response to movement events.
Lee’s research speaks to one of the most important debates within the study of public opinion: the
capability of citizens to function effectively in a democratic society. Elite democratic theorists, of course,
argue that citizens do not have the ability or the desire to participate in the rough-and-tumble of politics.
But Lee’s important work suggests that is not always the case. Lee concludes that “ordinary citizens,
under appropriately compelling circumstances, will take an active part in crafting politics rather than
merely consuming the political outputs of elite actors.”68 These are heartening words to the ears of
participatory democratic theorists.
Economic events can shape the public’s views as well. For instance, the effects of living through an
economic downturn during young adulthood can ripple throughout one’s life. To study this topic, Paola
Giuliano and Antonio Spilimbergo compared the political attitudes of American adults who had
experienced an economic recession when they were between 18 and 25 years old to those adults who
had not.69 The former were more likely than the latter to identify as Democratic, favor government
policies that assist the poor, and believe that getting ahead in life is due more to luck than hard work.
These results are based on analyzing General Social Survey data collected between 1972 and 2010.
Time will tell whether the same effects emerge for those who were young adults during the Great
Recession of 2008–2009.
So far, we have discussed examples demonstrating that people’s political opinions are influenced by
events that take place during their lifetimes. Can historical events also have lasting effects, shaping the
views of citizens long after the events occurred? Yes, conclude Avidit Acharya, Matthew Blackwell, and
Maya Sen.70 They examine the lingering effects of chattel slavery on the attitudes of contemporary
white Southerners. Central to their analysis is comparing the opinions of whites living in counties where
a high percentage of the residents were enslaved in 1860 versus those residing in historically low-slave
rate counties. The stretch of land where slavery was most prevalent is known as the Black Belt,
originally because of the color of the fertile soil in the region. This area encompasses eastern Virginia
and North Carolina, most of South Carolina, Georgia, Alabama, Mississippi, and Louisiana, as well as
northern Florida, southern Arkansas, and eastern Texas. As it turns out, whites living in these areas as
recently as 2011 are more likely to be Republican, racially intolerant, and oppose policies such as
affirmative action than are whites residing in places where slavery was less prevalent. Such place-based
differences do not emerge among whites for opinions toward nonracial policy issues, however.71
How is it even possible that the attitudes of whites today are affected by slavery, which was abolished in
the United States over 150 years ago? Even Acharya, Blackwell, and Sen admit that such a notion
might seem “outlandish.”72 To explain their findings, the authors propose a theory of behavioral path
dependence. According to this approach, “once we start down a path of development in political culture,
it becomes harder and harder to extract ourselves from that path … [T]he political attitudes of a place or
region—such as the Black Belt—can persist across generations, nurtured by institutions, laws, families,
and communities.”73 Slavery and its immediate aftermath started the path that Acharya, Blackwell, and
Sen explore. When slavery came to an end and Black men were granted the right to vote, Southern
whites’ economic livelihood and political dominance were both threatened. The more prevalent slavery
had been in an area, the more whites felt threatened. Thus, it is in these high-slavery places where laws
and customs that attempted to reestablish the racial hierarchy of slavery were most likely to emerge. Put
another way, white racial superiority became embedded in institutions and habits (such as Jim Crow
laws that mandated racial segregation), which facilitated the transmission of this ideology from
generation to generation.
A number of interesting expectations flow from Acharya, Blackwell, and Sen’s theory, some of which
they test empirically.74 For instance, if political outlooks surrounding race are reflected in Southern
institutions and culture, racial orientations should differ between whites who grew up in the South versus
those who moved there in adulthood. They do. Present-day Southerners’ feelings toward Blacks are
correlated with living in a county that had a high versus low slaveholding rate for those who were raised
in the South but not for people who were raised elsewhere. Furthermore, the effect of slavery on
contemporary views regarding race is smaller in places where agriculture was mechanized earlier. In
such locations, the demand for Black labor was lower, therefore weakening an economic incentive for
maintaining institutions and customs of racial inequality.
Finally, you might be wondering whether more recent historical events, such as the civil rights
movement, left their imprint on Southerners. Acharya and his coauthors wondered this, and so they
examined Southern racial inequality in the post–civil rights era.75 They find that for the years 2008
through 2013, the gap between Blacks’ and whites’ education and income levels did not vary by the
percentage of a Southern county’s residents that had been enslaved in 1860. In contrast, such regional
variation did exist before the civil rights movement: the more that slavery had been prevalent, the larger
the racial gaps. The connection between slavery and contemporary racial inequality, at least when it
comes to income and education, has thus been severed. No such de-linking has occurred for whites’
political attitudes, however. The pattern noted earlier, whereby whites living in counties that had higher
slaveholding rates are more racially conservative, exists for people born before 1965 and those born
after 1965.76 These results buttress Taeku Lee’s research. As you might recall, Lee demonstrated that
in the midst of the civil rights movement, white Southerners displayed more opposition to racial equality
than did Northern whites and Blacks from across the nation. Acharya, Blackwell, and Sen conclude that
these racial and geographic differences have persisted, particularly when we isolate the views of whites
living in the Black Belt.
CONCLUSION
Do citizens organize their political attitudes in any kind of coherent way? Yes, the research we have
discussed in this chapter suggests that public opinion is derived from several factors. Therefore, instead
of a single ideological dimension underlying public opinion (à la Converse), it is more appropriate to
think in terms of the pluralistic roots of public opinion. The research on authoritarianism provides
compelling evidence that personality predispositions influence a wide range of attitudes on moral,
political, and racial matters. Further, core personality traits, including openness to experiences and
conscientiousness, shape attitudes toward ideology, partisanship, and policy issues. Self-interest, in
contrast, has a much more limited effect on public opinion. In certain circumstances, however, self-
interest does matter. When the effects of a policy are clear, large, and salient, self-interest can kick in
and influence public opinion. In addition, citizens’ values play an important role in shaping political
attitudes. Egalitarianism, for instance, influences public opinion on a number of political issues, and
moral traditionalism drives opinion toward gay rights policies. Public opinion is also molded by historical
events. The civil rights movement affected citizens’ political attitudes, as has America’s slaveholding
past. War and economic downturns also shape public opinion.
Elite democratic theorists would be surprised by the limited role that self-interest plays in shaping public
opinion. Their justification for the centrality of elites in the political system is based at least in part on the
argument that citizens are too self-interested to be intimately involved in decision making. The research
suggests, however, that their argument is more of a rationalization than a justification. Taeku Lee’s
research also delivers a body blow to the elite democratic theorists. His research shows quite
convincingly that elite actions are not always central to the lives of citizens. Instead, events initiated by
social movement activists have an important effect on public opinion.
Participatory democratic theorists are delighted that the studies discussed in this chapter demonstrate
that citizens are capable of holding reasoned opinions derived from important political factors. The
research on the pluralistic roots of public opinion revives an image of citizens as more sophisticated
than the one suggested by Converse’s research. This research, however, does not put participatory
democratic theorists completely at ease. The opposition of moral traditionalists to gay rights is troubling.
Further, the intolerant reaction of authoritarians to diverse viewpoints undermines the faith that
participatory democratic theorists have in citizens coming together to work on political problems in a
constructive way.
KEY CONCEPTS
activated mass opinion 205
authoritarians 187
behavioral path dependence 206
Big Five 194
chattel slavery 206
egalitarianism 201
historical events 204
individualism 201
libertarians 187
moral traditionalism 202
normative threats 189
personality traits 194
pluralistic roots 186
pocketbook issues 198
self-interest 197
social movements 204
sociotropic concerns 198
values 200
SUGGESTED SOURCES FOR FURTHER READING
Acharya, Avidit, Matthew Blackwell, and Maya Sen. Deep Roots: How Slavery Still Shapes Southern
Politics (Princeton, NJ: Princeton University Press, 2018).
Hajnal, Zoltan, and Michael U. Rivera. “Immigration, Latinos, and White Partisan Politics: The New
Democratic Defection.” American Journal of Political Science 58 (2014): 773–789.
Deep Roots presents a rich theoretical and empirical examination of the continuing influence of slavery
on white Southerners’ political attitudes. Hajnal and Rivera demonstrate that long-term demographic
developments can shape public opinion. They link increased racial and ethnic diversity to white
Americans’ move away from Democratic and toward Republican Party identification.
Adorno, T. W., Else Frenkel-Brunswik, Daniel J. Levinson, and R. Nevitt Sanford. The Authoritarian
Personality. New York: Harper & Brothers, 1950.
Altemeyer, Bob. Enemies of Freedom: Understanding Right-Wing Authoritarianism. San Francisco:
Jossey-Bass, 1988.
Hetherington, Marc, and Jonathan Weiler. Prius or Pickup? How the Answers to Four Simple Questions
Explain America’s Great Divide. Boston: Houghton Mifflin Harcourt, 2018.
Stenner, Karen. The Authoritarian Dynamic. New York: Cambridge University Press, 2005.
In these four books, the authors grapple with the concept of authoritarianism. Adorno and his colleagues
argue that authoritarianism is a deeply rooted set of personality traits stemming from exposure to harsh
child-rearing practices, whereas Altemeyer asserts that authoritarianism develops through social
learning. Stenner emphasizes that authoritarianism has a stronger effect on attitudes when citizens are
faced with normative threats. Hetherington and Weiler argue in their book, which was written for a broad
audience, that differences in worldviews are a key reason why citizens disagree about the most vexing
issues of our day. They do not label this worldview authoritarianism, yet measure it based on responses
to the child-rearing items that are typically used to assess authoritarianism.
Brewer, Paul R. Value War: Public Opinion and the Politics of Gay Rights. Lanham, MD: Rowman &
Littlefield, 2008.
Rhodebeck, Laurie. “Issue-Relevant Values and Opinions about Gay Rights: Beyond Equality and
Morality.” Journal of Homosexuality 65 (2018): 379–405.
Brewer uses a variety of methods (surveys, experiments, and content analysis) to examine how the
public debate over gay rights has shaped citizens’ attitudes. In particular, he highlights the importance of
values in molding citizens’ beliefs about gay rights. Rhodebeck also takes a multi-methodological
approach to identify a wide range of values that are associated with gay rights policy, in media and
interest group discourse as well as public support.
Caprara, Gian Vittorio, and Michele Vecchione. “Personality Approaches to Political Behavior.” In Oxford
Handbook of Political Psychology, 2nd ed., ed. Leonie Huddy, David O. Sears, and Jack S. Levy.
Oxford: Oxford University Press, 2013.
Johnston, Christopher D., Howard G. Lavine, and Christopher M. Federico. Open versus Closed:
Personality, Identity, and the Politics of Redistribution. Cambridge: Cambridge University Press, 2017.
In their chapter, Caprara and Vecchione discuss traits and authoritarianism as well as other personality
characteristics that are related to political attitudes and behaviors. Johnston and coauthors focus on the
personality characteristic of openness, linking it to attitudes regarding redistributive economic policies.
Sears, David O., and Jack Citrin. Tax Revolt: Something for Nothing in California. Enlarged ed.
Cambridge, MA: Harvard University Press, 1985.
Weeden, Jason, and Robert Kurzban. “Self-Interest Is Often a Major Determinant of Issue Attitudes.”
Advances in Political Psychology 38 (2017): 67–90.
Sears and Citrin’s classic book examines citizens’ opinions on taxes and government spending in
California during the “tax revolt” of the late 1970s. The authors demonstrate that both self-interest and
symbolic racism influenced support for the tax revolt. Weeden and Kurzban argue that self-interest is
actually more relevant for policy opinions than most public opinion scholars acknowledge, in part
because of how self-interest is commonly defined.
SDA: Survey Documentation and Analysis, http://sda.berkeley.edu
This website allows you to easily analyze survey data from the American National Election Studies and
the General Social Survey. Let’s say you are interested in whether authoritarians were more likely than
nonauthoritarians to vote for Donald Trump in 2016. You could conduct that analysis using the American
National Election Study 2016, which is available in the SDA Archive. This example is just the tip of the
iceberg. The SDA website allows you to do all kinds of fun and informative data analysis. Check it out.
http://sda.berkeley.edu/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists percentage and ranges from 0 to 100 in increments of 20. The vertical axis lists
political attitudes. Approximate data from the graph are tabulated as follows:
Attitude
Percentage under minimum
authoritarianism who are in
agreement
Percentage under maximum
authoritarianism who are in
agreement
Support same-sex marriage 87 45
Believe use of wiretaps has gone
too far
58 30
Support building a wall on U.S.
border with Mexico
10 42
Believe Blacks should work their
way up without special favors
25 67
Support building a wall on U.S.
border with Mexico, whites only
10 58
Believe blacks should work their
way up without special favors
(whites only)
25 79
CHAPTER 7 ROOTS OF PUBLIC OPINION: THE
CENTRAL ROLE OF GROUPS
PUBLIC OPINION has many roots.1 Personality, self-interest, values, and historical events shape policy
attitudes, as we discussed in Chapter 6. Groups also play a central role in opinion formation, which
thrills pluralists because these theorists view groups as fundamental to political life in a democratic
society.
Political scientists have paid particular attention to two ways in which groups influence public opinion.
First, scholars have examined whether political attitudes vary by group membership. “Group
membership refers to the assignment of an individual into a particular group based on characteristics
that are specific to that group, in accordance with widely held intersubjective definitions.”2 In other
words, group membership is based on ascription, that is, how others describe a person based on social
norms and legal definitions, rather than on a person’s identification with a group. Public opinion scholars
have investigated the views of members of many different groups. For instance, researchers have
examined and compared the attitudes of members of different racial and ethnic groups.
The second way in which scholars have tackled the influence of groups on public opinion is by
examining the impact of attitudes toward groups on citizens’ issue opinions. Attitudes toward groups is
a broad concept that includes citizens’ prejudice toward and stereotypes about groups, as well as their
identification with groups. These attitudes play an incredibly important role in shaping political views
because many public policies can be thought of in terms of the groups that are affected by the policy.3
Here are some examples: equal pay policy has clear implications for women; welfare policy has a direct
effect on poor people; and Social Security policy has an impact on elderly people. In each of these
instances, attitudes toward the groups involved drive opinion on these issues.4 Scholars have dedicated
significant effort to understanding how group attitudes influence policy opinions.
It is interesting, of course, to analyze how attitudes differ by group membership, and we do that in this
chapter by examining the attitudes of different racial and ethnic groups and assessing gender
differences in opinion. It is more enlightening, however, to understand why groups differ. To get a handle
on that, it is important to consider attitudes toward groups. Black–white differences in public opinion, for
example, can largely be explained by group identity and racial prejudice among some whites and high
levels of group identity among Blacks. We also explore the power of a rural consciousness to shape
political attitudes. Another example is gender differences in opinion, which can partially be explained by
a feminist consciousness among some women.
RACE, ETHNICITY, AND PUBLIC OPINION
Race is an enduring cleavage among American citizens. As a result, racial groups often differ in their
policy opinions and party preferences. Public opinion scholars have examined Black–white differences
in opinion, successfully identifying key factors, including racial prejudice and identity, that lead the two
groups to diverge. Less is known about public opinion among Latinos and Asian Americans, although
recent scholarship has begun to fill this gap in our knowledge.
Black–White Differences in Party Identification and Issue Opinions
In the United States, Blacks and whites hold very different opinions, especially when it comes to party
identification and racial issues. In the 1960s, the partisanship of Blacks was heavily influenced by the
national Democratic Party’s support of the civil rights movement and the party’s push for legislation to
end discrimination against Blacks. As a result, African Americans who came of age during the civil rights
movement have the strongest identification with the Democrats. Younger cohorts of African Americans
who were socialized after the civil rights movement have demonstrated a small yet significant move
toward identification with the Republican Party.5 Nevertheless, in 2016, an overwhelming majority of
Blacks (81 percent) continued to identify with the Democratic Party (see Figure 7.1).6 As we discussed
in Chapter 2, the civil rights movement also influenced the party identification of some white citizens as
they opposed the pro-equality position of the national Democratic Party and changed their allegiance to
the Republican Party. In 2016, just 38 percent of whites identified with the Democratic Party.
In addition to party identification, Blacks and whites also differ in their issue opinions, as you can see in
Figure 7.1. Blacks are more supportive of social welfare programs than whites. Forty-four percent of
Blacks favored more government services even if it meant an increase in spending, whereas only 27
percent of whites endorsed that position in 2016. There was also a difference between Blacks and
whites on support for military spending. Thirty-three percent of Blacks favored increased defense
spending, whereas 46 percent of whites held that view. Blacks and whites tend to differ a great deal
when the issue domain is racial policies. For instance, 53 percent of Blacks supported government
efforts to improve the social and economic position of their group in 2016; only 20 percent of whites
agreed the government should make such efforts.
Description
Figure 7.1 Black–White Differences in Party Identification and Issue
Opinions, 2016
Source: The American National Election Studies (www.electionstudies.org), The ANES Guide to Public
Opinion and Electoral Behavior.
Notes: Respondents placed themselves on 7-point scales to answer these questions. The bars represent
the percentage of respondents who placed themselves on the Democratic, more government services, more
military spending, and more effort side of the scales.
What explains these Black–white differences in issue opinions? A critical reason why Blacks and whites
disagree on racial and social welfare policies is because they hold different political values dear.7 In
particular, Blacks are substantially more likely to endorse egalitarianism than whites. For example, 58
percent of African Americans strongly agreed that “if people were treated more equally in the country,
http://www.electionstudies.org/
we would have many fewer problems,” whereas only 25 percent of whites strongly agreed with that
value in 2016.8
These racial differences in support for egalitarianism are extremely important because they translate
into differences in policy opinions between Blacks and whites. Greater egalitarianism among Blacks
leads them to be more supportive than whites of a wide variety of racial and social programs, including
government programs that assist Blacks as well as federal spending on education, aid for college
students, the poor, the homeless, and the unemployed. Egalitarianism is a key factor that creates this
racial gap in opinion, but other factors, such as racial prejudice among whites and racial identification
among Blacks, also help explain the divergent views between the two groups.9 These are topics we
discuss next.
White Racial Prejudice and Race-Targeted Policy Opinion
One of the most controversial areas of public opinion research is the study of racial prejudice and its
impact on policy attitudes. Racial prejudice is a complex concept, and thus debates over its definition
and measurement have raged for several decades. One school of thought among public opinion
scholars is that racism is significantly different today from what it once was. Researchers from this camp
argue that there is a distinction between old-fashioned racism and new racism. Old-fashioned racism
is a set of beliefs about the innate inferiority of Black Americans. These beliefs served as a justification
for segregating Blacks and whites. Old-fashioned racism led to separate blood supplies for Blacks and
whites, white fears of miscegenation, the segregation of public accommodations (such as restrooms,
hotels, and restaurants), and legal discrimination against Black Americans. Over time, these beliefs
about Black inferiority were discredited by scientific research (which is ironic because for many years
“scientific” arguments were used to justify old-fashioned racism). Consequently, white citizens are much
less likely to endorse old-fashioned racism these days than they were fifty years ago. For example, 84
percent of white Americans now say they approve of marriage between Blacks and whites, whereas
only 17 percent approved of interracial marriage in 1969.10
Although old-fashioned racism has decreased considerably, many scholars contend that a different,
more subtle form of racial prejudice has emerged.11 Donald Kinder and Lynn Sanders argue that a “new
form of prejudice has come to prominence, one that is preoccupied with matters of moral character,
informed by the virtues associated with the traditions of individualism. At its center are the contentions
that Blacks do not try hard enough to overcome the difficulties they face and that they take what they
have not earned.”12 Kinder and Sanders label this new form of prejudice racial resentment.13 Instead
of blaming biological or genetic factors for Black inferiority, as old-fashioned racism does, racial
resentment blames a lack of work ethic for the continued inequality between Blacks and whites in our
society.
Kinder and Sanders make the case that racial resentment developed among whites in response to the
success of the civil rights movement. As discriminatory laws were knocked down by the Supreme Court
and Congress passed civil rights legislation, white Americans began to think that racial problems were
solved. Yet five days after the 1965 Voting Rights Act was signed by President Lyndon B. Johnson, riots
broke out in Watts, a neighborhood in South Central Los Angeles. The uprising went on for days, and
other riots followed in major cities across the country over the next few years. This social disorder
appalled Americans across the racial divide, but many Blacks understood the uprisings as legitimate
protests based on long-standing grievances, whereas many whites saw the riots as inexcusable criminal
behavior. Many whites felt that Blacks should be grateful for government efforts on their behalf and
should take advantage of the opportunities now open to them rather than continuing to complain about
discrimination.
Kinder and Sanders argue that racial resentment is fairly widespread among white citizens and that it
has a strong influence on policy opinions dealing with race. Before we discuss the impact of racial
resentment on policy attitudes, we need to review the way Kinder and Sanders measured racial
resentment. They used four survey statements (included on the ANES) to tap into this concept (see
Figure 7.2 for the wording of these survey items). Because racial resentment is a subtle form of
prejudice, Kinder and Sanders asserted that a “roundabout” approach is the most appropriate way to
measure the concept.14 Thus, their survey statements did not require whites to agree that Blacks are
outright lazy or that Blacks are simply hucksters trying to con white America into giving them something
they don’t deserve. No, nothing of the sort. Instead, the statements used subtle language to get whites
to reveal their bias. And, indeed, many whites demonstrated racial resentment, according to Kinder and
Sanders’s research.
In 1986, for example, roughly 60 percent of whites endorsed racially resentful responses when asked
whether (1) Blacks should work their way up without special favors, (2) Blacks should try harder to be as
well off as whites, and (3) Blacks have gotten less than they deserve (see Figure 7.2). The only question
on which a majority of whites did not give the resentful response was with regard to the lingering effects
of slavery and discrimination. Just over 30 percent were resentful on that question.15
Kinder and Sanders’s data are from thirty years ago, so perhaps levels of racial resentment have
decreased since the mid-1980s? Compared with 1986, survey data from the 2016 ANES show that
white citizens were less resentful on three of the four items. The one increase is in the percentage who
disagreed that slavery and discrimination have created conditions making it difficult for Blacks to
succeed, which increased from 32 to 46 percent. Overall, racial resentment is lower than thirty years
ago, although it is still prevalent among roughly half of white Americans.
Let’s turn to the impact of racial resentment on white citizens’ attitudes toward race-targeted policies.
Race-targeted policies are those designed to specifically aid Black Americans, such as affirmative
action and government steps to ensure fair treatment in employment. Kinder and Sanders analyzed the
influence of racial resentment on race-targeted policy attitudes using ANES data from 1986, 1988, and
1992. They found that white citizens who were more racially resentful were more opposed to a variety of
race-targeted policies. For example, racially resentful whites were significantly more likely to oppose
affirmative action in college admissions and employment than whites who did not hold such attitudes.16
More recent research confirms that racial resentment continues to shape citizens’ attitudes toward race-
targeted policies in contemporary times.17
Description
Figure 7.2 Racial Resentment among Whites, 1986 and 2016
Sources: Data for 1986 from Donald R. Kinder and Lynn M. Sanders, Divided by Color (Chicago: University
of Chicago Press, 1996), 107. Data for 2016 from analysis of American National Election Studies 1948–
2016 Cumulative Datafile.
Note: Bars represent the percentage of white respondents who gave racially resentful responses. For the
first two statements, the racially resentful response is agreeing “strongly” or “somewhat” with each
statement. For the last two statements, the racially resentful response is disagreeing “strongly” or
“somewhat” with each statement.
Ultimately, it is not that surprising that attitudes toward Blacks influence white citizens’ opinions on race-
targeted policies. Pluralists, in particular, would view this as entirely appropriate given the emphasis
these theorists put on groups as the primary organizers of political life in the United States. Participatory
democratic theorists, in contrast, would not be thrilled with the way in which prejudicial group attitudes
drive the policy positions of whites. Because the participatory democratic theorists care a great deal
about equality, they would be very concerned about the level of racial resentment among whites and the
extent to which racial resentment influences white policy opinions. Also troubling to participatory
democratic theorists is the power of racial prejudice to shape attitudes on policies beyond those directly
tied to African Americans. We turn to research on that topic next.
Race-Neutral Policy Opinion or Race-Coded?
A distinction is often made between social policies that are race-targeted and social policies that are
race-neutral. As we just discussed, race-targeted policies are ones that directly affect African
Americans. In contrast, race-neutral policies are ones that affect citizens regardless of race. For
example, government assistance to help Blacks is a race-targeted policy, whereas government
assistance to help poor people is a race-neutral policy. Research by Martin Gilens, however, leads us to
consider another possibility—race-coded policies.18 Race-coded policies are race-neutral, yet have
become linked with Black Americans in the minds of white citizens. Gilens argues that welfare is an
example of such a policy.19 Gilens maintains, “Although political elites typically use race-neutral
language in discussing poverty and welfare, it is now widely believed that welfare is a ‘race-coded’ topic
that evokes racial imagery and attitudes even when racial minorities are not explicitly mentioned.”20
Thus, because welfare is linked to race, Gilens hypothesizes that opposition to welfare policy among
whites largely stems from their racial attitudes.
To test this hypothesis, Gilens analyzed data from the 1991 National Race and Politics Study, a
nationally representative survey. To measure attitudes toward welfare spending, Gilens used this
question: “suppose you had a say in making up the federal budget, would you prefer to see more spent,
less spent, or the same amount of money spent on welfare as it has been?”21 To measure racial
attitudes, Gilens focused on the stereotype that Blacks are lazy. Specifically, Gilens asked respondents
to assess the work ethic of Blacks by indicating whether terms such as lazy and hardworking
characterize the group. Gilens found that white citizens who endorsed the view of Blacks as lazy and
not hardworking were significantly more opposed to welfare spending than white citizens who rejected
that stereotype.
How did a seemingly race-neutral issue like welfare become a race-coded issue? Gilens argues that
news media coverage of poverty and welfare disproportionately focuses on Black Americans, which
leads to this linkage between race and poverty in the minds of white citizens.22 To provide evidence to
support this contention, Gilens analyzed newsmagazine stories on poverty between 1950 and 1992. He
identified every poverty-related story in Time, Newsweek, and U.S. News & World Report during this
period. Then, he examined every picture within the poverty stories. He coded whether the race of each
poor person pictured was Black, non-Black, or undeterminable; he also coded whether the topic of the
poverty story was sympathetic (such as old-age assistance) or unsympathetic (such as the underclass,
which refers to people who belong to the lowest socioeconomic class and who are often stereotyped as
deviant).
Gilens found that whites were the dominant image of poverty in the 1950s and early 1960s. In 1965,
however, the racial makeup of the poor in these magazine stories on poverty changed quite
dramatically. The representation of Blacks jumped from 27 percent of the poor in 1964 to 72 percent in
1967. From 1967 on, Blacks tended to dominate the coverage of poverty, averaging 57 percent of the
poor portrayed by these newsmagazines between 1967 and 1992. To illustrate the skewed nature of this
media coverage, Gilens compared the true percentage of Blacks among the poor (based on data from
the U.S. Census Bureau) with the racial makeup of the poor depicted in the newsmagazines. In reality,
Blacks only made up about 29 percent of the poor during this period. Thus, the magazines’ depiction
essentially doubled the proportion of Blacks among the poor.
Gilens also found that the overrepresentation of Blacks became even more extreme when
unsympathetic topics were covered. For example, Blacks made up 100 percent of the poor people
pictured in stories on the underclass and 84 percent of the poor in stories on urban problems. In
contrast, there were no Blacks pictured in stories on old-age assistance and only about one-quarter of
the poor were Black in stories on hunger and medical care.23
Rosalee Clawson and her colleagues have updated and extended Gilens’s study by examining the
representation of poverty in five newsmagazines between 1993 and 2012.24 They analyzed the same
three magazines as Gilens, as well as two additional newsmagazines, the New York Times Magazine
and Business Week. They found that Blacks continued to be grossly overrepresented in pictures of the
poor, making up well over 50 percent of the magazine poor during this period. Another study
demonstrates that Hispanics are underrepresented in newsmagazine pictures of poverty, a finding
attributed to the common stereotype that Hispanics have a stronger work ethic than other racial groups,
whites included.25 This enduring racialization of poverty in the news media helps us understand why
stereotypes about the work ethic of Blacks continue to shape citizens’ welfare attitudes.26
Even more problematic is a recent experimental study of how people evaluate applicants for welfare
assistance. This research demonstrates that white applicants who are described as “excellent” workers
are rewarded with much greater financial assistance than equivalent Black applicants who are described
that way.27 At the same time, whites who are characterized as “poor” workers are penalized only a small
amount, whereas Blacks described that way are punished a great deal. The experimental subjects who
held racially resentful attitudes were those most likely to allocate money to white applicants and withhold
funds from Black applicants. “In an age when some argue that we have moved beyond the importance
of race, this article shows just how much race still colors Americans’ beliefs about work ethic and the
deserving poor. While some may be opposed to welfare spending on purely fiscal and race-neutral
grounds, the average American still believes Blacks deserve less than whites, all else equal.”28
An issue that recently became race-coded is health care policy, according to Michael Tesler.29 Tesler
argues that health care is now a racialized issue because it was linked with President Obama in the
minds of American citizens, a phenomenon he refers to as the spillover of racialization.30 Because
President Obama is Black, racialized considerations were activated in discussions of health care reform.
As a result, racial prejudice shaped white citizens’ views on health care policy proposals. Using survey
data between 1988 and 2009, Tesler demonstrated that racial resentment became a powerful predictor
of white opposition to health care reform after the policy became associated with President Obama. In
fact, racial resentment came to trump other considerations, including party identification and ideology, in
white citizens’ thinking about Obamacare.31
Racial resentment is so powerful and pervasive that it influences white attitudes toward sports as well.
There has been a lot of discussion recently about paying college athletes and letting them make money
from their names and images. For example, in October 2019, the governor of California signed into law
the Fair Pay to Play Act, which will allow California’s college athletes to earn money for their efforts
starting in 2023. Public opinion on this issue differs by race and racial resentment. In 2018, 70 percent
of African Americans supported allowing athletes to earn money off their names and images, whereas
56 percent of whites agreed with such a policy.32
Furthermore, many whites overestimate the proportion of college athletes who are Black. When whites
believe that Blacks are a small proportion of college athletes, their racial views do not shape their
attitudes toward these policies. However, when whites perceive Blacks as a large proportion of college
athletes, racially resentful whites become more opposed to allowing athletes to earn profits from their
efforts, whereas less racially resentful whites become more supportive of such policies.33
Similarly, there are large differences between whites and Blacks on whether they support the rights of
athletes to kneel during the singing of the national anthem. Sixty-one percent of Blacks support the right,
whereas only 28 percent of whites do. Furthermore, those whites who hold negative stereotypes of
Blacks are the ones most likely to strongly oppose the right of athletes to kneel.34 A study of young
Americans (ages 18–30) shows a similar pattern with 47 percent of whites approving of athletes
protesting racism by sitting or kneeling during the national anthem. In contrast, 76 percent of Blacks, 61
percent of Latinos, and 54 percent of Asian Americans approve protesting racism in this way. Racially
resentful whites are more likely to oppose athletes protesting during the national anthem than those
whites who are less racially resentful. In fact, racial resentment is the most powerful factor shaping the
attitudes of young white adults toward anthem protests.35 In the next section, we elaborate further on
the racial attitudes of young adults.
Race, Ethnicity, and Public Opinion among Millennials
Millennials are more progressive than older generations. But is this a generation effect, or are young
people more progressive because they are more diverse? One study of racial attitudes shows that white
Millennials have attitudes more similar to white non-Millennials than to their nonwhite counterparts. Race
trumps generation in other words, which bolsters the argument that young people are more progressive
because of their racial and ethnic diversity.36
The GenForward survey was specifically designed to explore the political attitudes of young adults of
different racial and ethnic backgrounds. To ensure representative samples of all groups, the
GenForward survey oversamples African Americans, Asian Americans, and Latinxs (which is a gender
neutral way to refer to a person of Latin American origin or descent) between the ages of 18 and 36.
The GenForward survey is bimonthly and assesses Millennials’ attitudes on a wide range of political
topics.37
Figure 7.3 displays Millennials’ attitudes broken down by race and ethnicity for six political issues at the
forefront of our politics.38 A large majority of African Americans and a majority of Asian Americans and
Latinxs indicate that police killings of Black people are an extremely or very serious problem, whereas
only 47 percent of whites endorse that view. A majority of young adults in every group strongly or
somewhat support the Black Lives Matter movement, with African Americans the most supportive at 81
percent. African Americans and Latinxs are more likely to perceive sexual harassment as a very serious
issue than Asian Americans and Whites.
On climate change and restoring voting rights for people who have been incarcerated, there are only
small racial and ethnic differences in the proportion of respondents who say they agree strongly or
somewhat with restoring voting rights (although note that African Americans are much more likely than
the other groups to agree strongly). Regardless of race or ethnicity, a large majority of young adults
strongly or somewhat support taxing the wealthy, with Asian American youth demonstrating the highest
level of support at 87 percent. Overall, this survey of young adults suggests there are some issues on
which racial and ethnic groups are much more alike than different, and other issues, especially those
concerning race, on which they diverge quite a bit.
Description
Figure 7.3 Political Attitudes of Millennials by Race and Ethnicity
Sources: Data from GenForward Surveys: Millennial and Sexual Harassment, June 2018; The Nature of
Work, March 2019; Climate Change, May 2019; Criminal Justice by Race and Ethnicity, July 2019,
https://genforwardsurvey.com/reports/
https://genforwardsurvey.com/reports/
White Identity
We have discussed white prejudice toward African Americans and its impact on policy attitudes, but
what about whites’ racial identification with their own group? In a pathbreaking study, Ashley Jardina
argues that “the growing non-white population, the pending loss of whites’ majority status, and the
increasing political and economic power of people of color in the United States” have made racial
identity salient to some whites.39 Jardina refers to this sense of commonality, attachment, and solidarity
as white identity and demonstrates that it has a major impact on political attitudes and behavior. White
identity is activated when whites feel anxious about and threatened by their changing political, social,
and economic circumstances. They feel a loss of privilege, and they don’t like it.
Table 7.1 Measuring White Identity
Racial Identity Importance: “How important is being white to your identity?”
Racial Pride: “To what extent do you feel that white people in this country have a lot to be proud of?”
Racial Commonality: “How much would you say that whites in this country have a lot in common
with one another?”
Source: Ashley Jardina, White Identity Politics (Cambridge: Cambridge University Press, 2019), 58.
To measure white identity, Jardina uses three survey questions that tap into (a) the importance of being
white, (b) pride in being white, and (c) a sense of commonality with other whites. See Table 7.1. The
racial identity importance item was included on several national surveys in 2016. Between roughly 30
percent and 40 percent of white respondents said that being white is very or extremely important to their
identity.40 The racial pride and racial commonality items were included on one national survey in 2016.
Fifty-four percent of whites said that whites have a lot or a great deal to be proud of in this country, and
41 percent of whites said that whites have a lot or a great deal in common with one another.41
Jardina also explores who is more likely to possess this identity. She finds that older individuals, the less
educated, the more authoritarian, and people in rural areas are more likely to have a strong white racial
identity. Working-class whites are only slightly more likely than the middle class to have a strongly racial
identity, whereas those who are economically vulnerable are no more likely than the economically
secure to have a strong white identity. In responses to open-ended questions about their racial identity,
whites with high levels of racial identity recognized the political and economic power of their group. They
understood their privilege and wanted to maintain it all the while feeling their group was not getting its
fair share.
White racial identity also influences policy attitudes. Jardina makes the case that the “history of
immigration in the United States is intricately tied to the notion of whiteness.”42 Immigrants, with their
new foods, language, and culture, threaten white dominance and privilege. For this reason, she expects
white racial identity to be an important factor shaping attitudes toward immigration. Indeed, whites with a
strong sense of attachment to their group are more likely to favor restrictive immigration policies and
believe that immigration has negative consequences for the country. Jardina argues that when political
elites use anti-immigrant rhetoric, as Donald Trump has done, whites’ racial identity becomes an even
greater factor shaping immigration policy attitudes.
White racial identification also predicts how citizens evaluate political candidates. A Black president is
particularly threatening to white dominance, so not surprisingly whites who identified strongly with their
own racial group were much less likely to vote for Barack Obama than those low in racial identification.
Donald Trump, in contrast, campaigned to Make America Great Again, which many interpreted as Make
America White Again.43 As a result, whites with a strong sense of racial identification were more likely to
view candidate Trump favorably.
Racial resentment also shaped how the public evaluated Obama and Trump. Racially resentful whites
opposed Obama and supported Trump.44 Therefore, both ingroup identification and outgroup hostility
influenced whites’ evaluations of political candidates, an unfortunate and combustible combination.
Many pundits suggested that Donald Trump’s victory was a result of appeals to white voters’ economic
anxieties and fears. The pundits were correct that Trump appealed to these emotions, but it was their
anxiety and fear over losing their privileged place in American society and their animus toward African
Americans that primarily drove their support for Trump, not economic considerations.
Group Consciousness among Blacks
Attitudes toward groups also shape the opinions of minority citizens. Here the research has primarily
focused on the effects of group consciousness on Black public opinion.45 Group consciousness refers
to a politicized awareness of how membership in a particular group shapes the “lives and fortunes” of
individuals.46 Many scholars agree that group consciousness is made up of several components,
including group identity and perceived discrimination.47
In separate research, two prominent scholars, Michael Dawson and Katherine Tate, provide insight into
the critical role of racial group identity in shaping Black public opinion. They demonstrated that racial
identity is an important determinant of policy opinions. Blacks who have higher levels of racial identity,
as measured by the extent to which they perceive their fate is directly tied to the fate of their race, are
more likely to support social welfare policies and government programs to assist Blacks. They are also
more likely to favor the drawing of congressional districts in such a way that increases the likelihood a
minority will be elected.48
Dawson refers to this belief that people’s life chances are tied to the fate of their group as linked fate.
He argues that Blacks use information about how their racial group as a whole is faring to make sense
of their own position in society. Therefore, Blacks use racial group interests as a cue to determine their
own issue opinions and party preferences. Dawson calls this phenomenon the Black utility heuristic
and argues that “as long as race remains dominant in determining the lives of individual Blacks, it is
‘rational’ for African Americans to follow group cues in interpreting and acting in the political world.”49
Not all Blacks agree, of course, that their fate is tied to their group, but most do think that is the case. In
three national surveys of African Americans conducted in the 1980s and 1990s, roughly 80 percent
agreed that “what happens generally to Black people in this country will have something to do with what
happens” in their own lives.50 In the early 2000s, 54 percent of African Americans said they felt “very
close” to Black people in the United States, and another 37 percent said they felt “fairly close.”51 The
2016 Collaborative Multiracial Post-Election Survey shows that 66 percent of Blacks think that what
happens generally to Black people in this country will have something to do with what happens in their
lives. Of those, 34 percent said it will affect them “a lot.”52
One question that arises with the study of group consciousness is whether minorities with high levels of
consciousness are hostile toward whites. Some observers argue that race consciousness leads to
separatism and a pursuit of divisive politics by minority citizens, whereas others argue that race
consciousness is empowering to minorities and has very little to do with their attitudes toward whites.
One study examined several measures of racial identity and found little evidence to support the claim
that high levels of racial identification are closely associated with hostility toward whites.53 For example,
there was no relationship between Blacks feeling close to their racial group and the belief that whites
keep Blacks down.
Melanye Price argues that to understand differences of opinion within the Black community, political
scientists need to move beyond studies of linked fate.54 She views linked fate as essentially a constant
and therefore argues the concept does little to explain why Blacks differ on a range of policies. Using
focus groups conducted in Columbus, Ohio, and national survey data, she explores Black nationalism.
Black nationalism is best defined as an identity based on adherence to four principles: support for Black
self-determination; support for Black economic and social independence in the form of self-help
programs; distance from whites and white supremacy; and support for a global Black community (also
see Dawson’s definition of Black nationalism in Chapter 5, Table 5.1).55 Only a minority of Blacks fully
endorse Black nationalism. Those who do, however, hold more negative feelings toward white
Americans than those who reject the identity. Price argues that many Black Americans endorse certain
tenets of Black nationalism, such as self-help, yet they don’t only blame whites and systemic
discrimination for their group’s economic and social situation. In focus group conversations, statements
of Black blame are often intermingled with a discussion of obstacles faced by Black Americans. Price
argues that as long as Blacks continue to face discrimination and inequality, Black nationalism has the
potential to be a potent political force. Similarly, other scholars show that Black citizens who have felt
discrimination against themselves because of their race are more likely to identify as Blacks first rather
than Americans first.56
Latino Public Opinion
Historically scholarship on race and public opinion has focused on Blacks and whites, constructing a
dichotomous political world that largely ignored other minority groups. This is problematic because
Latinos are increasingly influential in national politics and they constitute the largest minority group in
the United States.57 Latinos now make up 18 percent of the U.S. population.58 In recent years, national
surveys of Latinos have provided insight into their political attitudes.
The first question to answer with regard to Latinos is whether it even makes sense to talk about them as
a coherent political group. Latinos in the United States may have arrived yesterday, or their families may
have been here since long before the United States was the United States. Latinos come from more
than twenty different national origin groups, with individuals of Mexican, Puerto Rican, Cuban,
Dominican, and Salvadoran descent making up the largest proportion of Latinos in the United States.59
Over two-thirds are proficient in English, many speak Spanish, and others are bilingual.60
Despite this variety, national public opinion surveys suggest that Latinos do have a shared, although
certainly not uniform, view of the political world. In the 2016 Collaborative Multiracial Post-Election
Survey, 57 percent of Latinos agreed that what happens generally to Latinos in this country will have
something to do with their lives. Of those who felt this sense of linked fate, 33 percent said what
happens to other Latinos will affect them “a lot.”61
In 2016, large majorities of Latinos supported Obamacare, the wealthy paying higher taxes to allow for a
middle-class tax cut, increased federal spending on public education, and government efforts to combat
the effects of climate change. Immigration reform, health care, and improving the economy were the
issues Latinos identified as most important to their community. Eighty-two percent of Latino voters said
that undocumented or illegal immigrants should be allowed to stay in their jobs and apply for citizenship.
Only 33 percent believed that federal spending to tighten border security should be increased.62 In
2018, 25 percent of Latinos said there were too many immigrants living in the United States today, which
is down dramatically from 2002 when 49 percent felt that way.63
Table 7.2 Latino and Asian American Party Identification among Registered Voters, 2018
Party identification Latinos Party Identification AsianAmericans
Republicans (including
leaners)
23% Republicans (including leaners) 22%
Democrats (including leaners) 56 Democrats (including leaners) 38
Other/refused to lean 18 Independents 30
Don’t know/refused 3 Nonidentifiers/Other Party/Don’t
Know
10
Sources: Latino data from 2018 National Survey of Latinos, Pew Research Center, Hispanic Trends, Washington, DC, September 9,
2018, https://www.pewresearch.org/hispanic/datasets/. Asian American data from the 2018 Asian American Voter Survey, Asian and
Pacific Islander American Vote and AAPI Data, Washington, DC, October 9, 2019, https://www.apiavote.org/sites/default/files/2018-AA-
Voter-Survey-report-Oct9_0
https://www.pewresearch.org/hispanic/datasets/
https://www.apiavote.org/sites/default/files/2018-AA-Voter-Survey-report-Oct9_0
On speaking up for Latinos, disadvantaged groups in society, and American principles and values,
Latinos say that the Democratic Party will do a better job than the Republican Party. Latinos also see
Democrats as the party that has the better approach to handling immigration policy.64 Furthermore,
Latinos identify with the Democratic Party more than the Republican Party (see Table 7.2). Fifty-six
percent of Latino registered voters identify as Democrats.65 Seventy-six percent of Latinos view Donald
Trump unfavorably.66 This is not surprising since he launched his 2016 presidential campaign by
referring to Mexicans as rapists and has continued this vile rhetoric toward immigrants throughout his
presidency.
Asian American Public Opinion
Asian Americans are the fastest-growing racial or ethnic group in the United States.67 They make up a
relatively small proportion of the U.S. population (about 6 percent), but their concentration in particular
geographical areas makes them a powerful political group in certain counties, congressional districts,
and states.68 Asian Americans are becoming increasingly dispersed, which means their potential for
influence across the nation is growing.69 Like Latinos, Asian Americans are a diverse group. Some are
recent immigrants; others are in families who have been in the United States for generations. Asian
Americans come from many different national origin groups, ranging from China and Japan to India and
Pakistan. Unlike Latinos, who primarily speak English or Spanish, Asian Americans speak a wide range
of languages.
Do Asian Americans form a cohesive political group and have political attitudes in common? The
evidence is somewhat mixed based on recent surveys of Asian American voters. In 2016, 59 percent of
Asian Americans said that what happens generally to other Asians in the United States will have
something to do with what happens in their lives.70 Of those who felt this sense of linked fate, 19
percent said that what happens to other Asian Americans will affect them “a lot.” This sense of shared
destiny has political implications. For example, Asian Americans who have high levels of linked fate with
their own group are also more supportive of the Black Lives Matter movement.71
Asian Americans were divided when it came to the appropriate size of government with 44 percent
preferring bigger government and more services, 24 percent favoring smaller government and fewer
services, and 26 percent saying it depends.72 Yet over two-thirds supported tax increases on the
wealthy to provide the middle class with a tax cut, increased spending on public education, Obamacare,
and government efforts to combat the effects of climate change. Health care and improving the
economy were the issues Asian Americans identified as most important to their community.73
In 2016, 64 percent of Asian Americans favored a path to citizenship for undocumented immigrants.74 In
2008, only about one-third of Asian Americans held that position.75 This shift toward a more liberal
position on immigration is likely to continue as leaders in the Republican Party speak about immigrants
in derogatory terms. In 2015, for instance, during the Republican presidential nomination process,
candidate Jeb Bush said that Asian people were coming to the United States to have “anchor babies” to
gain U.S. citizenship for their children.76 And of course Donald Trump has made anti-immigrant rhetoric
a staple of his presidency. Sixty-eight percent of Asian Americans view Trump unfavorably.77
Perhaps not surprisingly, Asian Americans prefer the Democratic Party over the Republican Party (see
Table 7.2).78 Particularly noteworthy about Asian Americans’ partisanship, however, is the large
proportion of people who did not identify with either party in the 2018 Asian American Voter Survey.
Thirty percent identified as Independents and another 10 percent said they do not think in terms of
political parties, identified with a different party, or simply didn’t know. This suggests there is room for
either party to speak to the concerns of over one-third of Asian Americans and pull them into their
partisan fold. The party leaders who realize this and conduct language-targeted and culturally
sophisticated outreach efforts will reap the benefits of support from this fast-growing group.79 Overall,
the current picture of Asian Americans is of a political group that shares many important political
attitudes while exhibiting intragroup differences as well.
Box 7.1 Attitudes toward Immigrants around the World
Public Opinion in Comparative Perspective
Immigration has received a lot of attention around the world in recent years. In 2018 the Pew
Research Center conducted a Global Attitudes survey in 18 nations to assess people’s attitudes
toward immigrants.1 These 18 nations contain just over half of the world’s migrant population.
The countries include several that receive very large numbers of migrants each year, such as the
United States, Germany, the United Kingdom, France, Canada, and Australia.
Description
Views on Immigrants around the World
Source: “Around the World, More Say Immigrants Are a Strength Than a Burden,” Pew Research
Center, Washington, DC, March 14, 2019, https://www.pewresearch.org/global/2019/03/14/around-
the-world-more-say-immigrants-are-a-strength-than-a-burden/
The survey asked respondents to choose between two statements: immigrants today make our
country stronger because of their work and talents, or immigrants today are a burden on our
country because they take our jobs and social benefits. As you can see in the figure, attitudes
vary quite a bit across the 18 countries. People view immigrants most favorably in Canada with
https://www.pewresearch.org/global/2019/03/14/around-the-world-more-say-immigrants-are-a-strength-than-a-burden/
68 percent endorsing the statement that immigrants make our country stronger. In sharp
contrast, a mere 5 percent of Hungarians choose that statement.
Despite the negative and false rhetoric about immigrants coming from the Trump administration,
59 percent of Americans view immigrants as making the country stronger. In response to a
question about immigrants and crime, 77 percent of Americans said that “immigrants in our
country today are no more to blame for crime than other groups.” Majorities in Canada, France,
the United Kingdom, Spain, Australia, Mexico, Netherlands, and Japan share that belief.
Within each nation, views on immigrants also vary. For example, in most countries people on the
ideological left are substantially more positive toward immigrants than people on the right. Young
adults and those with higher education and higher incomes also tend to see immigrants more
favorably.
1 Ana Gonzalez-Barrera and Phillip Connor, “Around the World, More Say Immigrants Are a
Strength Than a Burden,” Pew Research Center, Washington, DC, March 14, 2019,
https://www.pewresearch.org/global/2019/03/14/around-the-world-more-say-immigrants-are-a-
strength-than-a-burden/
https://www.pewresearch.org/global/2019/03/14/around-the-world-more-say-immigrants-are-a-strength-than-a-burden/
RURAL CONSCIOUSNESS
Much of what we know about group consciousness is based on survey data from nationally
representative samples and is focused on consciousness stemming from racial identity. Katherine
Cramer takes a different approach.80 She investigated rural consciousness by listening in on
conversations among friends and colleagues in diners, fast food restaurants, and even gas stations in
twenty-seven communities in Wisconsin between 2007 and 2012. Her primary objective was to gain a
better understanding of how people think about politics.81
Based on listening to these conversations, she argued that people living in rural areas have a social
identity rooted in place, class, and a sense they do not get their fair share of resources. This rural
consciousness affects how they perceive and understand the political world around them. These
ordinary rural citizens feel ignored by policymakers. As a result, they resent government, including the
people who work for the government. They perceive themselves as hard-working and deserving,
whereas public employees are not. They also believe their culture and lifestyle are disrespected by
people in urban areas who don’t work hard and don’t deserve the resources they get. This resentment
leads them to focus their ire on public employees and urbanites rather than on the structural forces that
cause economic inequality.
Rural individuals resent government regulations made in cities that affect how they live their daily lives.
The Department of Natural Resources, for example, was pointed to by many of the rural residents
Cramer listened to as a prime example of government interference. Rural people perceived regulations
on fishing, hunting, and the environment as intrusive, disrespectful, and ignorant of their way of life.
Furthermore, because they view the government and public employees as anti-rural, they are not
sympathetic to redistributive policies that might actually benefit them and their rural communities. From
their rural consciousness perspective, less government is better than more. These rural citizens
supported then-Wisconsin governor Scott Walker’s efforts to reduce the power and compensation of
public employees. And in Donald Trump, they found a presidential candidate who spoke directly to their
resentment.
GENDER AND PUBLIC OPINION
Do women and men differ in their political attitudes? The quick answer is, it depends. On some issues,
women and men consistently differ, but on other issues, gender has little effect on opinions. When
women and men do differ, those differences are often fairly small, especially compared to differences in
opinion based on membership in other groups such as racial groups. In this section, we examine the
effects of gender on public opinion and discuss explanations for why women and men differ in some
instances but not in others.
Gender Differences in Party Identification and Issue Opinions
Since the early 1980s, there has been a modest but consistent gender gap in party identification, with
women leaning more Democratic than men.82 As you can see in Figure 7.4, the difference between
women and men was 7 percentage points in 2016. That is not a huge difference, and it is dwarfed when
compared to the partisan difference between whites and other racial or ethnic groups (refer to Figure 7.1
and Table 7.2). Nevertheless, the difference between men and women is meaningful, particularly when
we consider the large impact that party identification has on vote choice. In fact, since the 1980s,
women have regularly voted for Democratic presidential and congressional candidates at higher rates
than men.83
Description
Figure 7.4 Gender Differences in Party Identification and Issue Opinions,
2016
Sources: Data for Party identification, Government services, Public schools, and Military spending are from
the American National Election Studies (www.electionstudies.org), The ANES Guide to Public Opinion and
Electoral Behavior. Data for Marijuana and Man works are from analysis of American National Election
Study 2016 Time Series. Data for Gun permits are from GSS Cumulative Datafile 1972–2018—Quick
Tables, SDA: Survey Documentation and Analysis, http://sda.berkeley.edu/archive.htm
Notes: Respondents placed themselves on 7-point scales for the Party identification, Government services,
Military spending, and Man works questions. The bars represent the percentage of respondents who placed
themselves on the Democratic, more-spending, and family-better-if-man-works side of the scales. For Public
schools, respondents were asked whether federal spending on schools should be increased, decreased, or
kept the same. The bars represent the percentage of respondents who favored more spending. For
Marijuana, respondents were asked whether they favor or oppose the use of marijuana being legal. The
bars represent the percentage of respondents who favored legalization. For Gun permits, respondents were
asked whether they favor or oppose a law requiring people to obtain gun permits before buying a gun. The
bars represent the percentage of respondents who favored gun permits.
http://www.electionstudies.org/
http://sda.berkeley.edu/archive.htm
In terms of policy opinions, women and men tend to differ on what political scientists call compassion
issues.84 Compassion issues are social welfare policies that help others, including programs to aid
children, the elderly, the homeless, and the poor. Generally women are several percentage points more
likely to support these policies than men.85 In Figure 7.4, for example, we see that 32 percent of women
favored more government services in 2016, even if it meant an increase in spending, whereas 29
percent of men agreed with that position. When it comes to increasing government spending for
schools, 73 percent of women wanted more money spent on public education compared to 68 percent
of men.
There tend to be larger differences between women and men on use-of-force issues.86 Women are
more opposed to war than men, and they are less supportive of the death penalty. Women are also
more likely to support gun control measures. In 2016, for instance, 75 percent of women favored
requiring a person to obtain a police permit before buying a gun, whereas 66 percent of men supported
that position. Historically women have been less supportive of military spending than men. This holds
true in 2016, although the difference is small. Forty-four percent of men support increased military
spending compared to 40 percent of women (see Figure 7.4).
A new use-of-force issue on which large gender differences have emerged is drone strikes. In a 2015
Pew Research Center survey, 50 percent of women approved of U.S. drone strikes, whereas 67 percent
of men did so. Also, women were more concerned than men about whether the strikes would endanger
civilian lives, lead to retaliation from extremists, were conducted legally, and would damage America’s
reputation.87 This large gap between women and men in the United States is mirrored in countries
around the world.88 What is different globally, however, is that a majority of citizens in most countries
surveyed by the Pew Research Center oppose U.S. drone strikes. The United States is one of only
three countries where a majority of citizens approve U.S. drone strikes. The other two countries are
Israel and Kenya. In most countries, drone strikes are “widely unpopular,” and even more so among
women.
On compassion issues and use of force, women take more liberal positions than men, but on at least
some moral issues, women are more conservative than men.89 Women are more supportive of
government restrictions on pornography, more supportive of prayer in school, and less likely to support
marijuana legalization than men. In 2016, for example, 43 percent of women favored legalizing the use
of marijuana, whereas 48 percent of men supported legalization (see Figure 7.4).
Although these gender gaps in public opinion are certainly worth noting, we should not overstate the
differences between women and men. Women and men overlap in their views much more than they
diverge.90 Also, gender differences are not constant over time, and subgroups of women (based on
race, class, age, religion, marital status, or sexual orientation) often vary in their attitudes as much as, if
not more than, women and men. For instance, the party identification of white women looks much more
similar to white men than it does to Black women. And there are gender differences within the
Republican Party. Republican women are more moderate than Republican men on government
spending on child care, education, and health care (although not welfare for the poor); Republican
women also favor a tax on millionaires at higher rates than Republican men.91
When it comes to women’s issues on which we might think gender differences would be quite large,
men’s and women’s opinions tend to be fairly similar. An examination of abortion attitudes over time, for
example, shows that roughly 10 percent to 15 percent of both women and men believe abortion should
never be permitted.92 In Figure 7.4, we see just a 4 percentage point difference between men and
women on whether they think it is better for the family as a whole if the man works outside the home and
the woman takes care of the home and family.
On whether the country has done enough “when it comes to giving women equal rights with men,” we
do see substantial gender differences.93 Fifty-seven percent of women say the country “hasn’t gone far
enough,” whereas 42 percent of men share that sentiment. These data are from a survey conducted by
the Pew Research Center in August and September of 2017, two months before the Harvey Weinstein
sexual harassment scandal broke in the media. The Pew Research Center also conducted a survey
assessing concerns about sexual harassment at work in February and March of 2018 after sexual
harassment and the #MeToo movement had become topics of conversation across the United States
and the world. When respondents were asked whether men getting away with sexual harassment and
assault in the workplace today is a major problem, 55 percent of women said it was, compared to 44
percent of men.94
Explaining Gender Differences in Policy Opinions
Why is it that men and women differ on compassion, use-of-force issues, and moral issues? Scholars
have provided a variety of explanations for these gender differences.95 First, girls and boys are
socialized differently. Anyone who has ever walked into a store and looked at children’s toys knows that
gender norms are heavily enforced still today. Girls are encouraged to play with dolls and dress up like
princesses, whereas boys are encouraged to play with balls and fight with lightsabers. Gendered
socialization practices in childhood may result in differences between men and women in adulthood.
Girls are taught to be caring and nurturing, which may translate into greater support for compassion
issues among women; boys are taught to be competitive and aggressive, which may evolve into greater
support for the use of force among men.
A second explanation for gender differences in public opinion emphasizes that women and men often
play different roles in society and have different economic interests. Women are expected to be
caregivers for children and other family members, and they are more likely to work in low-paying
service-sector jobs than men. As a result, women are in poverty and make use of social welfare
programs at higher rates than men. All this may lead women to be more supportive of social welfare
policies than men. Further, women’s caregiver roles may make them more concerned about protecting
their families from moral decay, which may lead to their greater conservatism on moral issues.
A third explanation for gender differences focuses on how a feminist consciousness may shape citizens’
political attitudes, especially party identification. The Democratic Party has been the party of women’s
rights for four decades now, which leads feminists to identify as Democrats more than Republicans.96 A
feminist consciousness can also shape women’s (but not men’s) attitudes toward war.97 For example,
feminist women are less opposed to war when the stakes are humanitarian than when military action is
taken for economic reasons.
Overall, these three factors—socialization, roles and interests, and a feminist consciousness—help
explain the gender gaps in public opinion. Much more research is needed, however, to fully understand
the differences and similarities between women and men. Finally, it is important to note that, although
gender differences in public opinion tend to be fairly modest, they can still have a big impact because
women make up more than half of the U.S. population and they turn out to vote at higher rates than
men.
CONCLUSION
As you learned in Chapter 5, many citizens do not structure their issue attitudes along ideological lines.
Yet that does not mean they are hapless in their political thinking. Group memberships, attachments,
and antipathies help citizens make sense of politics. Racial resentment and stereotypes among white
citizens shape their views on race-targeted and race-coded policies. Racial identification is a key factor
driving the political views of whites and Blacks. Rural consciousness also shapes how ordinary citizens
in rural areas perceive and make sense of the political world around them. Latinos tend to prefer the
Democratic Party and generally take liberal positions on issues. Asian Americans also lean Democratic,
although a large minority do not think of themselves in partisan terms at all. Women identify as
Democrats at higher rates than men and are generally more liberal on compassion and use-of-force
issues. At the same time, they tend to be more conservative on moral issues. Interestingly, women and
men do not differ as much as one might expect on at least some women’s issues.
Overall, these findings about the central role that groups play in organizing citizens’ political attitudes
are not at all surprising to pluralists. According to the pluralist perspective, politics is fundamentally
about groups wrangling with each other to get their points of view turned into public policy. Debate,
competition, and compromise among groups are what make a democracy function. Factors other than
groups, such as personality, values, historical events, and self-interest, also shape citizens’ attitudes,
but group attitudes constitute the most profound root of public opinion. In addition to the groups
discussed in this chapter, we reinforced this conclusion with our discussion of party identification in
Chapter 5.
Whereas pluralists are delighted with the critical role of groups in U.S. politics, participatory democrats
are more ambivalent. On the one hand, public opinion differences based on group membership and
group attitudes confirm participatory democratic theorists’ views that citizens are able to organize their
opinions in a meaningful way. Reliance on group cues, such as the Black utility heuristic, provides a
reasonable means by which citizens can navigate the political world. On the other hand, group-based
thinking results in powerful stereotypes, prejudice, and identity shaping some white citizens’ views on
key political issues, a situation that is not compatible with participatory democratic theorists’ desire for
greater equality among citizens.
KEY CONCEPTS
attitudes toward groups 212
Black nationalism 225
Black utility heuristic 225
compassion issues 233
egalitarianism 214
gender gap in party identification 231
group consciousness 224
group membership 212
linked fate 224
moral issues 233
old-fashioned racism 215
race-coded policies 218
race-neutral policies 218
race-targeted policies 216
racial identity 224
racial resentment 215
racialization of poverty 219
rural consciousness 230
use-of-force issues 233
white identity 223
women’s issues 234
SUGGESTED SOURCES FOR FURTHER READING
Barreto, Matt, and Gary M. Segura. Latino America: How America’s Most Dynamic Population Is Poised
to Transform the Politics of the Nation. New York: Public Affairs, 2014.
Wong, Janelle, S. Karthick Ramakrishnan, Taeku Lee, and Jane Junn. Asian American Political
Participation: Emerging Constituents and Their Political Identities. New York: Russell Sage Foundation,
2011.
These books examine Asian American and Latino political behavior.
Carter, Niambi Michele. American While Black: African Americans, Immigration, and the Limits of
Citizenship. Oxford: Oxford University Press, 2019.
Masuoka, Natalie, and Jane Junn. The Politics of Belonging: Race, Public Opinion, and Immigration.
Chicago: University of Chicago Press, 2013.
Pérez, Efrén O. Unspoken Politics: Implicit Attitudes and Political Thinking. Cambridge: Cambridge
University Press, 2016.
Schildkraut, Deborah J. Americanism in the Twenty-First Century: Public Opinion in the Age of
Immigration. Cambridge: Cambridge University Press, 2011.
These books tackle public opinion and immigration from a variety of perspectives.
Dawson, Michael. Behind the Mule: Race and Class in African-American Politics. Princeton, NJ:
Princeton University Press, 1994.
Frasure-Yokley, Lorrie, Natalie Masuoka, and Matt A. Barreto. “Introduction to Dialogues: Linked Fate
and the Politics of Groups and Identities.” Politics, Groups, and Identities 7 (2019): 610–614.
In Dawson’s classic book, he argues that many Black Americans see their own interests as closely tied
to the interests of their racial group because of historical and current experiences with discrimination.
The “Introduction to Dialogues” piece kicks off a set of articles in recognition of the twenty-fifth
anniversary of Dawson’s book. These articles assess the impact of the concept of linked fate on the
study of public opinion.
Lizotte, Mary-Kate. Gender Differences in Public Opinion: Values and Political Consequences.
Philadelphia: Temple University Press, 2020.
The author examines gender differences in public opinion using data from the American National
Election Study.
AAPI DATA: Demographic Data and Policy Research on Asian Americans and Pacific Islanders,
www.aapidata.com
This website provides extensive data on Asian American and Pacific Islander politics and public opinion.
Latino Decisions, www.latinodecisions.com
This website provides extensive data on Latino public opinion.
Center for American Women and Politics, https://cawp.rutgers.edu/
This website provides extensive data on women in politics.
https://cawp.rutgers.edu/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists percentage and ranges from 0 to 100 in increments of 20. The vertical axis lists
different opinions. Approximate data from the graph are tabulated as follows:
Opinion Percentage of Black Percentage of White
Party Identification 82 38
Government services 45 28
Military spending 32 47
Aid to Blacks 52 20
Back to Figure
The horizontal axis lists fragments of statements. The vertical axis lists percentage and ranges from 0 to
100 in increments of 20. Approximate data from the graph are tabulated as follows:
Statement Percentage of response in1986
Percentage of response in
2016
“Irish, Italians…” 65 60
“It’s really a matter of
time…”
60 42
“Generations of slavery…” 35 48
“Over the past few years…” 60 50
Back to Figure
The horizontal axis lists percentage of agreements and ranges from 0 to 100 in increments of 10. The
vertical axis lists different opinions. Approximate data from the graph are tabulated as follows:
Opinion Percentageof Whites
Percentage of
Asian Americans
Percentage
of Latinxs
Percentage of
African Americans
Opinion Percentageof Whites
Percentage of
Asian Americans
Percentage
of Latinxs
Percentage of
African Americans
Seriousness of police
killings of black people
45 55 65 78
Support for Black Lives
Matter movement
58 68 70 81
Seriousness of sexual
harassment
40 48 59 63
Restore voting rights for
formerly incarcerated
69 72 70 77
Prioritize climate change 39 41 43 33
Support wealth tax 71 88 75 70
Back to image
Data from the graph are tabulated as follows:
Country Percentage who think immigrants are aburden on our country
Percentage who think immigrants
make our country stronger
Canada 27 68
Australia 31 64
UK 29 62
Sweden 32 62
Japan 31 59
U.S. 34 59
Germane 35 59
Country Percentage who think immigrants are aburden on our country
Percentage who think immigrants
make our country stronger
Mexico
37 57
Spain 37 56
France 39 56
Netherlands 42 50
South
Africa
62 34
Israel 60 26
Poland 50 21
Russia 61 18
Italy 54 12
Greece 74 10
Hungary 73 5
18-country
median
38 56
Back to Figure
The horizontal axis lists percentage and ranges from 0 to 80 in increments of 20. The vertical axis lists
opinions. Approximate data from the graph are tabulated as follows:
Opinion Percentage of men Percentage of women
Opinion Percentage of men Percentage of women
Party identification 42 50
Government services 29 32
Public schools 67 73
Military spending 45 40
Gun permits 65 75
Marijuana 48 43
Man works 38 33
PART IV DO CITIZENS
ENDORSE AND DEMONSTRATE
DEMOCRATIC BASICS?
CERTAIN FEATURES of a democratic society place expectations on
the citizenry. Popular sovereignty, for example, puts some decision
making in the hands of the citizens. To exercise this duty, must
democratic citizens be knowledgeable about and attentive to
politics? A democratic society also guarantees a number of
freedoms for citizens, such as freedom of speech and religion, and
ensures that citizens are equal before the law. These are known as
civil liberties and civil rights, respectively. In a democracy, however,
how important is it for citizens to support civil liberties and civil
rights?
Not surprisingly, the answers to these two questions differ depending
on which type of democratic theorist is responding. The normative
debates over these key democratic basics (knowledge, interest,
attentiveness, support for civil liberties, and support for civil rights)
are profiled in Part IV. In the chapters that follow, we also pay
particular attention to whether citizens endorse these democratic
basics in the abstract and demonstrate them in practice.
CHAPTER 8 KNOWLEDGE, INTEREST, AND
ATTENTION TO POLITICS
DO YOU KNOW why the water in Flint, Michigan, is unsafe to drink? Do you know who Neil Gorsuch is?
Can you identify the president of France? If you are like nearly three-quarters of Americans, you know
the water in Flint is unsafe to drink due to lead. It is likely that just less than half of you know Neil
Gorsuch is a Supreme Court justice. And probably only about one-third of you know that Emmanuel
Macron is the president of France.1 Notably, men are more likely than women to know Gorsuch and
Macron: 53 percent of men identify Gorsuch as a Supreme Court justice, whereas only 37 percent of
women do so. Similarly, 43 percent of men can correctly identify the president of France compared to 32
percent of women. Such a gender gap in knowledge is not unusual, although men and women are just
as likely to know why the water in Flint is unsafe to drink.2 See Table 8.1 to learn more about what
citizens do and do not know about politics, according to a Pew Research Center poll conducted in June
2017.
Political scientists often define political knowledge as “the range of factual information about politics
that is stored in long-term memory.”3 Does it matter that many citizens are not highly knowledgeable
about key political figures and facts? Participatory democratic theorists say yes. “Political information is
to democratic politics what money is to economics: it is the currency of citizenship.”4 Thus, without
knowledge, citizens cannot function effectively. From this perspective, democracy requires informed
citizens, as well as interested and attentive ones.
Other scholars argue that citizens can make reasonable decisions without being knowledgeable about
or particularly interested in and attentive to politics. Citizens can use informational shortcuts, such as
cues from individuals or groups they trust, to form their political opinions.5 Because many of the cues
citizens rely on come from elites, this perspective is in line with the thinking of elite democratic theorists.
A more extreme position taken by some elite democratic theorists is that citizen ignorance actually
allows democracies to flourish.6 Clearly, elite democratic theorists and participatory democratic theorists
have very different views on how engaged the public should be.
In this chapter, we discuss the competing visions of elite democratic theorists and participatory
democratic theorists regarding levels of citizen knowledge, interest, and attention to politics. Next, we
examine how knowledgeable citizens actually are about politics. As with so many concepts, there is
controversy over how to measure political knowledge; we discuss that debate. Then we turn to an
analysis of why some citizens are more knowledgeable than others. We consider the effect of individual
characteristics and political context on knowledge. We also examine the consequences of political
knowledge and ignorance. Finally, we discuss how interested citizens are in politics and whether some
individuals might best be considered bystanders to democracy. We conclude the chapter by considering
whether citizens are knowledgeable enough, interested enough, and attentive enough to function
effectively in a democracy.
Table 8.1 Political Knowledge, June 2017
What the public does and does not
know
Answered
correctly
Answered
incorrectly
Don’t know/ refused
to answer
Zika is spread primarily by? (mosquitoes) 86% 9% 5%
Water in Flint, MI, is unsafe due to? (lead) 72 17 11
What the public does and does not
know
Answered
correctly
Answered
incorrectly
Don’t know/ refused
to answer
Speaker of the House? (Paul Ryan) 62 19 19
Country leaving the EU? (United Kingdom) 60 22 17
Department of Justice lead Russia
investigator? (Mueller)
47 32 22
Neil Gorsuch is a? (Supreme Court justice) 45 26 30
Conservative Republican group in House
is? (Freedom Caucus)
45 22 33
Secretary of State? (Tillerson) 44 35 22
President of France? (Macron) 37 34 29
Unemployment rate? (4%) 37 52 12
Source: Data from “From Brexit to Zika: What Do Americans Know?” Pew Research Center, Washington, DC, July 25, 2017,
https://www.people-press.org/2017/07/25/from-brexit-to-zika-what-do-americans-know/#survey-report
Note: Correct answers are in parentheses. Due to rounding, some rows do not add to 100%.
https://www.people-press.org/2017/07/25/from-brexit-to-zika-what-do-americans-know/#survey-report
HOW KNOWLEDGEABLE, INTERESTED, AND ATTENTIVE SHOULD
CITIZENS BE IN A DEMOCRACY?
Democratic theorists debate about how much knowledge, interest, and attention to politics is necessary
for citizens to function effectively in a democracy.7 Participatory democratic theorists argue that citizens
should have high levels of knowledge and be actively engaged in the political world so they can hold
elites accountable. Citizens should be well informed so they can (1) figure out their interests, (2)
recognize which policies serve those interests, and (3) identify which political parties, interest groups,
and politicians are pursuing those policies. These theorists argue that citizens need to be on the ball
because elites may try to misrepresent the facts or may not even know the facts to begin with.
Participatory democratic theorists assume that all citizens have the ability to understand what is going
on in the political arena. Thus, if citizens are not knowledgeable, interested, and attentive, there must be
barriers that keep citizens from engaging. For example, participatory democratic theorists might point to
the shortcomings of the media as an obstacle to citizens acquiring the knowledge they need to
successfully evaluate elite behavior; or, they might raise concerns about whether a president who
systematically lies about facts ranging from the size of his inauguration crowd to the accuracy of a
whistleblower complaint will lead to a misinformed public.
Elite democrats, on the other hand, are not particularly concerned about levels of citizen knowledge,
interest, and attention to politics. In fact, some scholars argue that it is irrational for citizens to spend
time on politics.8 Why is that the case? First, there are substantial costs to acquiring information about
politics. It takes significant time for citizens to arm themselves with the knowledge necessary to form
opinions, evaluate policy proposals, and understand the political process. Staying informed also
requires financial resources. Subscribing to newspapers or traveling to public meetings, for example,
entails a monetary commitment on behalf of citizens. Second, citizens can expect very little in return for
becoming knowledgeable. The voice of a citizen would be just one of many; thus, it is extremely unlikely
that a single person could influence the adoption of a particular policy or the outcome of an election.9 If
citizens were assured a payoff from becoming informed, it would be rational for them to invest in
gathering information. But because that is not the case, it does not make sense for them to spend time
and money becoming knowledgeable.
Even though it is irrational to become knowledgeable about politics, citizens can still function in a
democracy because they can rely on cues in their environment to make reasonable political
judgments.10 Citizens can use heuristics (shortcuts), such as party identification and ideology, to figure
out where they stand on issues. They can take cues from trusted groups and individuals to determine
which politicians best represent their interests. Scholars refer to this as low-information rationality.
From this perspective, citizens can get by with fairly minimal levels of knowledge. As long as they know
enough to take cues from political parties, interest groups, and other elites, citizens can function
effectively. Note that taking cues from interest groups is consistent with the pluralist vision of a
democratic society.
Other scholars emphasize that citizens use heuristics not because it is irrational to become informed but
because it is impossible for citizens to become fully informed. The political world is extremely complex,
and citizens are limited information processors. They simply do not have the cognitive abilities to
systematically process the vast amounts of political information out there. Instead, they rely on whatever
shortcuts are available to simplify complex material and thus function efficiently, if not always
accurately.11 Indeed, a recent study found that inaccurate heuristics are particularly problematic for
those who are most informed. The most informed are typically able to make the best use of heuristics
due to their preexisting knowledge. For example, well-informed citizens know the party identification of
various politicians and can therefore use that information to predict their votes on issues. But relying on
that heuristic can lead the most informed astray when a representative votes against his party.12
Hard-core elite democrats make a different argument about citizen knowledge, interest, and attention to
politics. They believe that democratic nations are actually better off when citizens are apathetic and
ignorant. From this perspective, elites do the heavy lifting, and citizen engagement would just
complicate matters. These theorists remind us that many features of the U.S. government, such as the
Electoral College and the Supreme Court, were specifically intended to keep citizens at bay. Moreover,
because the institutions of government were designed to check and balance each other, it is not
necessary for citizens to hold elites accountable. Some citizens might be political junkies, but most
people are not interested in, and many are even turned off by, the conflict and compromise that are part
and parcel of democratic decision making.13 And that’s a good thing according to some elite democratic
theorists.
ARE CITIZENS KNOWLEDGEABLE ABOUT POLITICS?
“Nothing strikes the student of public opinion and democracy more forcefully than the paucity of
information most people possess about politics.”14 In this section, we discuss the empirical evidence
that speaks to this claim. Is it the case that citizens are uninformed about politics? Also, are citizens
misinformed about politics? Have citizens become more or less knowledgeable over time?
Are Citizens Informed or Uninformed?
As we briefly discussed in the introduction to this chapter, citizens are not deeply informed about politics.
Less than 40 percent of citizens know the unemployment rate and fewer than one-half can identify the
U.S. secretary of state or the conservative group in the House of Representatives, the Freedom Caucus
(refer back to Table 8.1). Yet 86 percent of Americans know that Zika is spread primarily by mosquitoes,
and about two-thirds of citizens can identify the Speaker of the House and know the United Kingdom is
the country leaving the European Union. Despite significant media attention to the Department of
Justice investigation of Russian involvement in the 2016 election, only 47 percent could identify Robert
Mueller as leading that effort. Overall, the 2017 Pew Research Center poll results presented in Table 8.1
show that citizens are well informed about a few political facts, but there are many gaping holes in their
knowledge.
To delve further into this topic, we turn to the classic research of Michael X. Delli Carpini and Scott
Keeter, who define political knowledge as “the range of factual information about politics that is stored in
long-term memory.”15 To be knowledgeable about politics, Delli Carpini and Keeter argue citizens
should be well informed in three areas: (1) the rules of the game, (2) the substance of politics, and (3)
the people and players (including parties and groups).16 To assess citizen knowledge in these areas,
Delli Carpini and Keeter pulled together fifty-four years of national survey data. They examined
numerous polls conducted between 1940 and 1994, which included over 2,000 questions regarding
factual knowledge of politics.
Overall, Delli Carpini and Keeter found a mixed bag when it comes to levels of citizen knowledge.17 On
the one hand, citizens clearly did not live up to the standards set by participatory democratic theorists.
“Only 13 percent of the more than 2,000 questions examined could be answered correctly by 75 percent
or more of those asked, and only 41 percent could be answered correctly by more than half the
public.”18 Furthermore, many citizens did not even know enough to use heuristics. For example, bare
majorities of the public could identify party positions on key issues. Thus, for those who didn’t know
party positions, it would be impossible to use party as a cue. On the other hand, citizens were not as
ignorant of politics as suggested by some observers. Most citizens were aware of some basic facts,
such as the length of a presidential term (96 percent), the name of their governor (86 percent), that
there is no religious test for political candidates (81 percent), that Cuba is a communist country (82
percent), and that Social Security does not provide job training (89 percent).
Citizens did best at answering questions about the rules of the game. This is probably due to the fact
that the institutions and processes of politics rarely change over time. Further, this is the type of
knowledge citizens are exposed to in high school civics classes. For example, in 1985 virtually everyone
(96 percent) knew that the United States was a member of the United Nations.
Citizens were a bit less knowledgeable when it came to the substance of politics and the people and
players. To illustrate, only about one-half of the public knew that Black South Africans could not vote in
1985, and a mere 22 percent knew in 1972 what Watergate was about. About one-half could identify
Andrew Young as U.S. Ambassador to the United Nations, and only 3 percent could name the president
of Mexico in 1991.
Are Citizens Misinformed?
James Kuklinski, Paul Quirk, Jennifer Jerit, David Schwieder, and Robert Rich make an important
distinction among people who are informed, uninformed, and misinformed.19 They explain the
differences this way:
To be informed requires, first, that people have factual beliefs, and second, that the beliefs be
accurate. If people do not hold factual beliefs at all, they are merely uninformed. They are, with
respect to the particular matter, in the dark. But if they firmly hold beliefs that happen to be
wrong, they are misinformed—not just in the dark, but wrongheaded.20
Using survey data from a representative sample of Illinois residents, Kuklinski et al. demonstrated that
citizens were largely misinformed about welfare policy.21 For example, citizens were asked what
percentage of the federal budget is spent on welfare. The survey used a multiple-choice format and
provided the following options: 1 percent, 5 percent, 8 percent, 11 percent, or 15 percent. The correct
answer is 1 percent, but a whopping 90 percent of the respondents selected one of the other options.
Although citizens were the most misinformed about this fact, more than a majority of citizens
demonstrated they were “not just in the dark, but wrongheaded” on five other questions regarding
welfare. Furthermore, when asked how confident they were of their answers, a majority of citizens said
they were very or fairly confident about their responses. Thus, many citizens were not just wrong, they
were quite confident in their wrongheadedness.
In general, citizens tended to be misinformed in an antiwelfare direction. For example, believing that a
much higher percentage of the budget is spent on welfare than actually is would probably make people
less supportive of welfare. Indeed, Kuklinski et al. found that misinformation led people to be more
opposed to welfare spending than they otherwise would be. Because Kuklinski et al.’s survey
respondents were from Illinois, the authors were careful not to generalize their findings to American
citizens in general. Nevertheless, their research raises significant concerns about the public’s ability to
understand and evaluate important public policies.
In an important study of misperceptions regarding the war in Iraq, Steven Kull, Clay Ramsay, and Evan
Lewis found that many citizens were confused about facts pertaining to the war.22 These scholars
collected survey data from a random sample of 1,362 respondents between June and September 2003.
They were interested in whether citizens had misperceptions about the war and whether citizens
receiving news from particular sources were more likely to hold those misperceptions. Kull, Ramsay,
and Lewis crafted three questions to measure misperceptions:
Is it your impression that the US has or has not found clear evidence in Iraq that Saddam
Hussein was working closely with the al Qaeda terrorist organization?
Since the war with Iraq ended, is it your impression that the US has or has not found Iraqi
weapons of mass destruction? (Note that the end of the war wording here refers to the end of
major combat operations, as declared by President George W. Bush in May 2003.)
Thinking about how all the people in the world feel about the US having gone to war with Iraq,
do you think: the majority of people favor the US having gone to war; The majority of people
oppose the US having gone to war; or Views are evenly balanced.23
Because the United States did not find a clear link between Saddam Hussein and al-Qaeda or weapons
of mass destruction in Iraq (despite occasional Bush administration statements implying otherwise),
answering “has” to the first two questions were coded as misperceptions. Saying world opinion favored
the war in response to the third question was also coded as a misperception because public opinion
polls showed that most people around the world opposed the war. Kull et al. found that 60 percent of
Americans held one or more of these misperceptions about the war.
Next, Kull et al. examined whether citizens’ levels of misperceptions varied with their news source.
Indeed, they found stark differences in beliefs about the war based on where citizens tended to get their
news (see Figure 8.1). Viewers of Fox News were particularly likely to believe one or more of the
misperceptions. More than a majority of CBS, ABC, CNN, and NBC viewers held one or more of these
misperceptions; slightly less than a majority of print media users were confused about the facts. Three-
fourths of the NPR and PBS audience did not hold any misperceptions. The relative accuracy of the
NPR and PBS audience demonstrates the importance of noncommercial outlets in a media environment
dominated by conglomerates.
Description
Figure 8.1 Misperceptions by News Source
Source: Data from Steven Kull, Clay Ramsay, and Evan Lewis, “Misperceptions, the Media, and the Iraq
War,” Political Science Quarterly 118 (2003–2004): 582.
Note: Bars represent the percentage of people holding at least one misperception about the Iraq War by
primary news source.
What makes the high level of misperceptions among the public particularly troublesome is that the more
misperceptions citizens held, the more they supported the war in Iraq. For example, only 23 percent of
citizens who had no misperceptions supported the war, whereas 86 percent of those with three
misperceptions supported it. Kull et al. did not demonstrate that the news sources caused the
misperceptions or that the misperceptions caused support for the war. Nevertheless, the associations
between news sources and misperceptions and between misperceptions and support for the war raise
serious questions about the sources, extent, and effect of misinformation in a democratic society.
Clay Ramsay and his colleagues conducted a similar survey of citizens’ misperceptions shortly after the
2010 midterm elections.24 They found that many citizens were misinformed on issues debated during
the campaign. Take Obama’s place of birth, for example. Fifty-six percent of citizens knew that Obama
was born in the United States, but 27 percent said it was not clear whether he was born in the United
States and another 15 percent said he was not born in the United States. This is a striking level of
ignorance about the president of the United States. Moreover, people who watched Fox News almost
daily were significantly more likely to be misinformed than those who never watched it. To illustrate, 63
percent of daily Fox News viewers did not know that President Obama was born in the United States,
compared with 32 percent of citizens who never watched Fox News.25
In another study, Josh Pasek, Gaurav Sood, and Jon Krosnick expressed concerns about
misperceptions of the Affordable Care Act (ACA), also referred to as Obamacare.26 Using an online
survey, they asked a nationally representative sample of Americans to identify whether a list of eighteen
provisions were or were not included in the ACA. Twelve of the provisions truly were in the ACA, and six
were not. They also asked respondents to indicate how sure they were in their judgment. Similar to
Kuklinski et al., they argue that it is important to make the distinction between (1) simply being incorrect
and unsure and (2) being incorrect and certain about that belief. They found that many Americans were
unsure about aspects of the ACA. For example, slightly less than one-half of respondents correctly
identified that the ACA required companies that sell health insurance to pay new fees to the federal
government each year, and most of the people who answered incorrectly indicated they felt unsure
about their response. What made Pasek, Sood, and Krosnick concerned was that some people were
incorrect and certain on a set of items about which opponents of the law had spread false information.
For example, 53 percent of Americans incorrectly believed that the ACA required doctors and hospitals
to treat illegal immigrants free of charge if they cannot afford to pay. Of those 53 percent, 18 percent
held that incorrect view with certainty.
Several scholars have raised concerns about the effects of misinformation in a democratic society. They
fear that misinformation distorts public deliberation and undermines the ability of people to make
reasoned judgments.27 Scholars also worry about whether efforts to correct misperceptions can be
successful. When citizens process information with directional goals in mind, meaning they are trying to
reach a desired conclusion, they will seek information that confirms their predispositions, counterargue
information that conflicts with their opinions, and perceive consistent information as more convincing.
People often engage in directionally motivated reasoning when it comes to politics, which makes it
very difficult to correct misperceptions. Indeed, efforts to correct misperceptions may even lead to
“backfire effects” whereby misperceptions actually increase among ideological groups who are
motivated to believe the inaccurate information.28
Have Levels of Political Knowledge Changed over Time?
To examine whether citizens have become more knowledgeable over time, Delli Carpini and Keeter
collected national survey data from randomly selected respondents in 1989. Their survey included
political knowledge questions matching those asked on surveys from the 1940s and 1950s. Thus, Delli
Carpini and Keeter were able to compare levels of political knowledge across time. Before we turn to
their results, let’s discuss why we might expect citizens to be more informed in 1989 than fifty years
earlier as well as why we might expect citizens to be less informed.
There are certainly reasons to believe that citizens would be significantly more informed in 1989.29 First,
citizens’ ability to understand politics should have increased because levels of formal education have
risen dramatically over the years. Second, citizens’ motivation to understand politics should have
increased because the government has become significantly “bigger” over the past few decades. Both
domestically and on the world stage, the U.S. government plays a much larger role today than it did in
the 1940s. Thus, citizens should be more concerned with all the ways the government affects their daily
lives and how U.S. power is wielded around the world. Third, citizens should have more of an
opportunity to learn about politics given the huge leaps in communication that have occurred since the
1940s.
But let’s not get too far ahead of ourselves. There are also reasons to suspect that citizens would be
less knowledgeable in 1989 than in the 1940s.30 Although levels of formal education have increased,
that does not necessarily mean a corresponding increase in civics education. Moreover, some observers
have lamented a decline in the quality of education in schools. More students may be graduating, but
with fewer skills and knowledge than once was the case. As for motivation, it may be that a larger and
more visible government turns citizens off rather than encourages engagement. And certainly changes
in communication technology provide citizens greater access to news, but they also provide greater
access to entertainment that draws attention away from the politics of the day.
So there are arguments in both directions regarding whether citizens should be more knowledgeable in
1989 than several decades ago, but ultimately this is an empirical question. By comparing levels of
knowledge in 1989 to roughly fifty years earlier, Delli Carpini and Keeter found that “the level of public
knowledge has remained remarkably stable.”31 For example, over 90 percent of the public knew that a
presidential term lasts four years, and roughly two-thirds of citizens knew which party controlled the
House in both 1947 and 1989. There were some questions on which people demonstrated significant
increases or decreases in knowledge, but many of these changes seem understandable given varying
patterns of media coverage and elite emphasis over time.
More than three decades have passed since Delli Carpini and Keeter conducted their 1989 study, and
during that time a technological revolution has occurred, providing citizens with unprecedented access
to information. With a few keystrokes on the computer or a smartphone in hand, citizens can hunt down
political information at the drop of a hat. Have citizens become more knowledgeable as a result?
Perhaps Delli Carpini and Keeter’s conclusions about the stability of political knowledge are timebound.
We turn to survey data collected by the Pew Research Center to address this issue. Pew conducted
knowledge surveys in 1989 and 2007 that included a set of identical or comparable questions.32 The
Pew data do not show that citizens have become markedly more knowledgeable; instead, citizens were
more knowledgeable about some topics in 2007 but less informed about others (see Table 8.2). Citizens
were less likely to correctly name their state’s governor, the president of Russia, and an administration
official involved in a scandal (Scooter Libby in 2007, who was in the midst of a trial for perjury and
obstruction of justice in a case involving the outing of a CIA agent, and John Poindexter in 1990, who
was in the midst of a trial for involvement in the Iran-Contra scandal). On the upside, the public was
more aware of which party controlled the House and the ideology of the chief justice. They were also
more likely to identify the Speaker of the House. Overall, the Pew surveys suggest citizens have not
become more knowledgeable as a result of easier access to information. Therefore, it seems that Delli
Carpini and Keeter’s research findings are not timebound; levels of citizen knowledge have stayed
roughly the same over time.
Table 8.2 Political Knowledge, 1989–2007
Political knowledge 1989 2007 Difference
Percentage who could correctly name:
Their state’s governor 74 66 –8
The president of Russiaa 47 36 –11
Percentage who knew:
The party controlling the House 68 76 +8
The chief justice is conservative 30 37 +7
Percentage who could correctly identify:
Tom Foley/Nancy Pelosi 14 49 +35
John Poindexterb/Scooter Libby 60 29 –31
Source: Data from “Public Knowledge of Current Affairs Little Changed by News and Information Revolutions,” Pew Research Center,
Washington, DC (April, 2007), http://www.people-press.org/2007/04/15/public-knowledge-of-current-affairs-little-changed-by-news-and-
information-revolutions/
a Data in first column from 1994.
b Data in first column from 1990.
In a recent study, however, Mona Kleinberg and Richard Lau argue that we need to rethink how we
understand political knowledge in our current technological age, especially for younger Americans who
often get their news online.33 You will recall that Delli Carpini and Keeter define political knowledge as
factual information about politics stored in long-term memory.34 But what happens when citizens can
look up political information any time they want online? Perhaps people can simply rely on the Internet
rather than their brains to store information for them. “In other words, when it comes to politics, is
googling as good as knowing?”35 Kleinberg and Lau say yes. They demonstrate that using the Internet
to follow politics can substitute for lower political knowledge among young Americans. Even with lower
stores of political knowledge than older Americans, young people can gather political information online
that allows them to have accurate perceptions of where candidates stand on issues and choose
candidates consistent with their interests and values.
http://www.people-press.org/2007/04/15/public-knowledge-of-current-affairs-little-changed-by-news-and-information-revolutions/
MEASURING POLITICAL KNOWLEDGE
When it comes to measuring political knowledge, a key question is whether citizens are generalists or
specialists. A generalist would be knowledgeable across all political topics. In contrast, a specialist
would have knowledge on some topics but not others. For example, a specialist might be well informed
about foreign affairs but pay little attention to domestic politics.
If it is the case that people specialize, then measuring political knowledge would be quite difficult. It
would require lots of survey questions covering a wide range of political topics. Survey researchers
would need to include multiple questions in both foreign and domestic domains on the substance of
policy, rules of the game, and people and players. If, on the other hand, citizens are generalists, then a
relatively small set of survey questions would provide a valid and reliable measure of political
knowledge.
Delli Carpini and Keeter analyzed a variety of survey data to determine whether citizens are generalists
or specialists.36 They found that citizens can best be characterized as generalists.37 In other words, the
same person who can name the president of Russia also tends to know which party controls the House,
who declares war, what a recession is, and whether the United States has a trade deficit. Likewise, a
person who does not know the Russian president doesn’t know other political facts either. As a result, it
is possible to successfully measure political knowledge using a small set of items. Based on extensive
analyses, Delli Carpini and Keeter recommended using five questions to measure political knowledge.38
Those items, as well as the introduction they suggested, are presented in Table 8.3.
Table 8.3 Measuring Political Knowledge
“Here are a few questions about the government in Washington, DC. Many people don’t know the
answers to these questions, so if there are some you don’t know, just tell me and we’ll go on.”
“Do you happen to know what job or political office is now held by [insert name of current vice
president]?”
“Whose responsibility is it to determine if a law is constitutional or not … is it the president, the
Congress, or the Supreme Court?”
“How much of a majority is required for the U.S. Senate and House to override a presidential veto?”
“Do you happen to know which party had the most members in the House of Representatives in
Washington before the election this/last month?”
“Would you say that one of the parties is more conservative than the other at the national level?
Which party is more conservative?”
Source: Michael X. Delli Carpini and Scott Keeter, What Americans Know about Politics and Why It Matters (New Haven, CT: Yale
University Press, 1996), 305–306.
There are two noteworthy exceptions to the argument that citizens are generalists. First, Delli Carpini
and Keeter’s research shows that some people specialize in state or local politics.39 These citizens are
knowledgeable about political events and issues in their states or in their local communities but are not
as aware of national politics. Thus, their measure is not a good indicator of state and local political
knowledge; instead, it assesses “national political competence rather than citizen competence.”40
Second, some people may specialize in knowing about issues and people particularly pertinent to them
or groups to which they belong. Delli Carpini and Keeter’s research, for example, shows that Blacks
have lower levels of political knowledge than non-Blacks, in general, but are just as knowledgeable on
race-related issues.41 Similarly, in a 2007 poll, the Pew Research Center found that Blacks identified
then–senator Barack Obama and then–secretary of state Condoleezza Rice at the same level as whites
but that Blacks were less informed than whites in other areas.42 A study conducted a few months after
the al-Qaeda attacks on the World Trade Center and the Pentagon showed that citizens living in the
Northeast were more knowledgeable about the events of 9/11 than people living in other parts of the
country but not better informed about politics in general.43 Overall, then, the research demonstrates that
some citizens have domain-specific knowledge in areas that are particularly important to them.
Delli Carpini and Keeter’s method of measuring political knowledge has greatly influenced how political
scientists study the concept. Not all scholars agree, however, that their conceptualization and
measurement are the best. Indeed, research conducted by Kathleen Dolan raises concerns about Delli
Carpini and Keeter’s narrow definition of politics and the gender differences in knowledge that result
from that limited understanding of politics.44 In addition, several scholars have demonstrated that levels
of citizen knowledge are underestimated by the open-ended type of questions recommended by Delli
Carpini and Keeter.45 Yet other scholars argue that Delli Carpini and Keeter’s emphasis on factual
knowledge measured through surveys is misplaced.46 We discuss these critiques next.
Gender-Relevant Domains of Knowledge
Noting that many studies have documented that men are more knowledgeable about politics than
women, Kathleen Dolan argues that scholars should broaden their conceptualization of political
knowledge to include “gender relevant domains of knowledge.”47 Delli Carpini and Keeter’s
understanding of politics emphasizes electoral politics and the institutions of government, but those are
not the aspects of politics most important to women. Women are more likely to be interested in
government programs, women’s issues, and female leaders. Therefore, questions that measure
knowledge about these topics should be included on surveys, which would reduce, if not reverse, the
gender gap in political knowledge.48
To test this proposition, Dolan conducted an online survey of a random sample of 1,039 adults. The
survey included several items designed to measure gender-relevant knowledge, including asking
respondents the name of the Speaker of the House (who was a woman at the time), the name of any
woman serving in the U.S. Congress, the number of women Supreme Court justices, and the
percentage of women serving in Congress. For comparison purposes, the survey also included one
traditional political knowledge item that asked respondents which party held a majority in the U.S. House
of Representatives. In Table 8.4, we see that women were significantly less informed about who
controlled the House but comparable to men in terms of naming Nancy Pelosi as House Speaker,
naming a woman House member, and knowing how many women were on the Supreme Court. Further,
the gender gap was reversed when it came to knowing the percentage of women serving in Congress.
Women members made up 16 percent of Congress at the time, so Dolan coded any answer between 15
and 20 percent as correct. Women were significantly more likely than men to answer correctly, whereas
men were more likely than women to underestimate the presence of women in Congress. Dolan’s point
that survey researchers should conceptualize and measure political knowledge in a way that represents
the experiences of men and women is well taken; nevertheless, even with gender-relevant items
included, the conclusion is still that political knowledge among the American public is modest at best
and woefully low at worst.
Table 8.4 Traditional and Gender-Relevant Political Knowledge
Answered correctly
Women Men
Traditional political knowledge
Majority in House 58% 68%
Gender-relevant political knowledge
Speaker of House 39 39
Woman member of Congress 51 54
Women on Supreme Court 43 47
Answered correctly
Women Men
Percentage women in Congress 34 28
Source: Data from Table 1 of Kathleen Dolan, “Do Women and Men Know Different Things? Measuring Gender Differences in Political
Knowledge,” Journal of Politics 73 (2011): 102.
Note: The boldface percentages indicate statistically significant gender differences.
The good news is that a recent study demonstrates it is possible to reduce the gender gap in political
knowledge by providing citizens with information.49 When men and women are exposed to political
information, learning occurs and the gender gap decreases. This may seem like common sense, but it is
important because it shows that gender gaps in knowledge are not inevitable.
Recall versus Recognition
Another concern with the measurement of political knowledge is the use of short-answer questions.50
People are more likely to say “don’t know” to these questions than to multiple-choice questions.
People who are semiconfident may not be willing to toss out an answer with a short-answer question but
will articulate a response once they hear the option mentioned in a multiple-choice format. Further,
multiple-choice questions may jog people’s memories, allowing them to recognize the correct answer.
Box 8.1 Gender and Political Knowledge in Canada
Public Opinion in Comparative Perspective
Canadian political scientists have also raised concerns about traditional measures of political
knowledge. Similar to Kathleen Dolan, Dietlind Stolle and Elisabeth Gidengil argue that the
conventional conceptualization and measurement of knowledge are too narrow because of the
focus on electoral politics and government institutions.1 They prefer a broader measure that
assesses citizens’ knowledge of government programs and services. Such knowledge is
practical and useful; it helps citizens “know how to access welfare and other government
services that are essential to their own well-being and that of their families.”2 This knowledge is
especially critical for women, who use government services at a higher rate, are more likely to
have jobs in the public sector, and support welfare state policies more than men.
Because this information is so relevant to women, Stolle and Gidengil argue that women are as
knowledgeable as men, if not more informed, about important government services and benefits.
Thus, the typical gender gap in knowledge will disappear when a wider range of questions is
used to measure political knowledge. To investigate this issue, Stolle and Gidengil conducted a
survey of Canadian citizens living in Montreal and Toronto. The survey included traditional
questions about key political leaders and political parties, but it also included questions
measuring knowledge of several government programs, such as legal aid, the availability of free
health-screening tests, and where to turn if someone knew of a child being abused.
Answered correctly
Women Men
Traditional political knowledge
Prime minister 91% 92%
Party forming official opposition 56 73
Female cabinet minister 38 44
Knowledge of government programs
Legal aid 83 81
Health-screening tests 76 50
Child abuse 68 62
Source: Data from Dietlind Stolle and Elisabeth Gidengil, “What Do Women Really Know? A Gendered Analysis of Varieties of
Political Knowledge,” Perspectives on Politics 8 (2010): 99.
Note: The boldface percentages indicate statistically significant gender differences.
Stolle and Gidengil found that men and women were comparable in their knowledge of high-
profile political leaders, such as the prime minister, but that women tended to be less
knowledgeable than men on the other measures of conventional political knowledge. In sharp
contrast, women generally outperformed men when it came to knowledge of government
services and benefits. For example, women were much more likely to know about free health-
screening tests available through their health care plans than men.
Stolle and Gidengil acknowledge that knowledge of conventional politics is important but argue
quite persuasively that practical knowledge is also essential in democratic societies. Thus, they
recommend including questions that measure both types of information on surveys assessing
political knowledge.
1 Dietlind Stolle and Elisabeth Gidengil, “What Do Women Really Know? A Gendered Analysis of
Varieties of Political Knowledge,” Perspectives on Politics 8 (2010): 93–109.
2 Ibid., 94.
Table 8.5 presents survey data from the Pew Research Center that illustrate the differences between
short-answer and multiple-choice questions.51 Note the data presented here are not from a split-half
survey in which respondents were randomly assigned to receive one type of question or another.
Instead, they are drawn from two nationally representative surveys, one conducted in February and the
other in March 2007. Thus, it could be the case that factors other than the question format affected
levels of political knowledge. We should be cautious in interpreting these results; nevertheless, the data
suggest that format matters. On both items, citizens appear more knowledgeable with a multiple-choice
question. And notice the dramatic difference on the question about the president of Russia—citizens go
from being fairly ignorant to fairly well informed with the switch in format.
Table 8.5 The Perils of Measuring Political Knowledge: Short-Answer versus Multiple-Choice Questions
Short-answer formatFebruary 1–13, 2007 Multiple-choice format March 9–12, 2007
“Can you tell me who
Robert Gates is?”
“Is Robert Gates the U.S. Secretary of Defense? A senator
from Michigan? The chairman of General Motors? Or is he
something else?”
U.S.
secretary of
defense
21% 37%
Anything
else/don’t
know/refused
79 63
Short-answer formatFebruary 1–13, 2007 Multiple-choice format March 9–12, 2007
“Can you tell me the
name of the president
of Russia?”
“Can you tell me who is the president of Russia? Is it Boris
Yeltsin? Vladimir Putin? Mikhail Gorbachev? Or is it
someone else?”
Vladimir
Putin
36% 60%
Anything
else/don’t
know/refused
64 40
Source: “Public Knowledge of Current Affairs Little Changed by News and Information Revolutions,” Pew Research Center, Washington,
DC, April 15, 2007, 23–33, http://people-press.org/2007/04/15/public-knowledge-of-current-affairs-little-changed-by-news-and-
information-revolutions
http://people-press.org/2007/04/15/public-knowledge-of-current-affairs-little-changed-by-news-and-information-revolutions
Too Much Emphasis on Factual Knowledge?
Doris Graber raises an even more fundamental concern with Delli Carpini and Keeter’s measure of
political knowledge.52 She argues that political scientists should focus on what citizens need to know to
function in a democracy rather than emphasizing “precisely remembered factual knowledge about
historically important past and current events.”53 Why? Because citizens are limited information
processors. It is simply not possible for citizens to remember every little detail about politics because our
brains are not designed to do so. Moreover, our brains are not set up to quickly recall political
knowledge in the context of a survey.54
This does not mean, however, that citizens are clueless about politics. Instead, they pay attention to
issues that are relevant to them. People are interested in understanding the impact of political events on
their lives and on the well-being of the country, not on memorizing the names of politicians or
constitutional rights. Citizens want useful information, not factoids. Graber argues, for example, “There
may be areas of knowledge where the poor, trained in the school of hard knocks, may excel. But scores
of streetwise knowledge are not usually gathered and reported.”55 To substantiate her claims, Graber
analyzed transcripts of nine focus groups conducted in the Chicago area.56 The ninety-eight focus
group participants ranged from suburban voters to city voters to young people to homeless people. The
focus groups began with the moderator asking, “What are the issues that are most important to you in
your community, however you define that? What would you tell an elected official?”57 Graber coded for
whether participants’ responses were “simple,” such as statements of fact or description, or “complex,”
such as statements that showed understanding of a variety of perspectives and drew connections
among different ideas.
Here is an example of a Black voter discussing economic problems using complex statements: “I would
like to hear a political person say that one of the viable alternatives to crime in our neighborhoods is
really lobbying for minimum wage standards … not just talking about minimum wage but how do we get
people who have smaller stores to expand and employ more people. … How do we talk to Sears about
having part-time staff with no benefits?”58 Overall, Graber found that people often made complex
statements or a mix of complex and simple statements when talking about political topics. Simple
statements dominated the discussion in only 18 percent of the issue areas discussed by the focus
groups.
Graber admits these people would not score well on Delli Carpini and Keeter’s measure of political
knowledge, but she does not expect them to given the inherent limitations of the brain. Instead, Graber
argues that the focus group participants “possess reasonably sophisticated, politically useful knowledge
about current problems that confront them and that the issue areas covered by this knowledge are
generally quite well suited to carrying out the actual tasks of citizenship that most Americans perform.”59
Katherine Cramer and Benjamin Toff also challenge the fact-based approach to measuring political
knowledge.60 They propose an expanded model of civic competence that emphasizes how citizens
interpret politics through the lens of their own personal experiences. Cramer and Toff observed citizens
in Wisconsin talking about politics in natural settings, such as in a diner or at a gas station. They found
that people often engage in storytelling and interpret factual information through their lived experiences.
For example, when a jobs report was mentioned in one conversation, a small business owner
responded by linking the jobs outlook to his own negative experiences with government regulations.61
Based on their analysis of such political talk, Cramer and Toff argue that informed citizens are those who
can listen to and understand the experiences of others. The conversations among citizens is what
matters rather than knowing specific pieces of objective information. This view of democracy “values the
ability of citizens to interact with one another and share experiences as a necessary condition for
collectively governing each other and shaping each other’s futures in a just manner.”62
WHY ARE SOME CITIZENS MORE KNOWLEDGEABLE THAN
OTHERS?
In this section, we turn our attention to understanding why some citizens are more informed than other
citizens. Citizens vary in political knowledge based on their abilities, motivations, and the opportunities
available. We begin with a discussion of the relationship between demographic characteristics and
political knowledge. Next we examine the role of motivation. Then we turn to a discussion of contextual
factors that affect political knowledge.
Demographic Groups and Political Knowledge
Several studies have demonstrated substantial differences in political knowledge across demographic
groups.63 We have already discussed the gender gap in political knowledge at some length and
mentioned the racial gap in political knowledge on issues that are not race related. It is also important
to mention that these gender and racial gaps extend to young people. A study of students at six middle
schools in Maricopa County, Arizona, showed that white adolescents were significantly more
knowledgeable about politics than African American, Latino, and Native American adolescents.64
Further, a nationally representative survey of high school seniors demonstrated that boys have slightly
more civics knowledge than girls and that whites have substantially more civics knowledge than
Hispanics and African Americans.65 A study of European youth, however, demonstrates that boys do
better when asked about facts but that girls show greater knowledge when asked to critically analyze
and make a judgment about a political matter. Again, this suggests the way in which we measure
political knowledge influences our conclusions.66
Levels of political knowledge also differ across age groups. In general, older folks are better informed
than younger ones. This appears to be due to both life cycle and generational effects. As people
progress through the life cycle, they are simply exposed to more political information. Older citizens
have experiences, such as buying a house and paying property taxes, that younger people are less
likely to have. These experiences result in an increase in political knowledge. Also, older people were
raised in a more politically engaged era than today’s young people. Accordingly, this early socialization
put the older generation on a track to pay attention to, and thus be knowledgeable about, politics
throughout their lives. A recent study of Americans and digital knowledge, however, shows eighteen- to
twenty-nine-year-olds are more knowledgeable than those sixty-five and older about such things as net
neutrality and Facebook ownership of WhatsApp and Instagram. Here experience and socialization
work in favor of young people.67
Income and education also influence political knowledge. Wealthy citizens are better informed than their
poorer counterparts. Similarly, citizens with higher educations are substantially more knowledgeable
than those with a minimal education, especially when it comes to rules of the game and people and
players.68 Education is critical, of course, because it improves citizens’ cognitive abilities to learn about
politics. But it also boosts interest in politics, which motivates citizens to become more knowledgeable.
In addition, education affects the opportunities citizens have to become informed; formal education
allows people to obtain jobs in which politics matter on a regular basis (such as lawyers, business
executives, and political science professors) and places people in social networks that value political
knowledge. Indeed, “formal education is the single most important factor differentiating those who know
more about politics from those who know less. Citizens who spend more years in school simply know a
lot more about politics.”69
In Table 8.6, we illustrate demographic differences in political knowledge by examining citizens’
familiarity with foreign interference in the 2016 election. The June 2017 poll conducted by the Pew
Research Center asked respondents to identify who (from four choices) led the Justice Department’s
investigation into Russian involvement in the 2016 election. We see that 47 percent of citizens correctly
identified Robert Mueller as leading the investigation. There is striking variation, however, across sex,
race, age, and education. Men, whites, older citizens, and the better educated were more likely to
answer correctly. Notice in particular the very large impact of education on political knowledge.
Motivation and Political Knowledge
People also differ in how motivated they are to learn about politics. Citizens who are interested in and
attentive to politics are significantly more knowledgeable than those who just don’t care.70 This finding
extends to young people. Among high school seniors, students who expressed an interest in studying
government were substantially more likely to have a grasp of civics knowledge than other students.
Political discussion is also a factor. A study of citizens in the United States and Great Britain found a
significant link between discussing politics with family and friends and being well informed about
politics.71
Table 8.6 Demographic Differences in Political Knowledge
“Who is leading the Justice Department’s investigation into Russian involvement in the 2016
election?”
Characteristics Answered correctly
Total 47%
Sex
Male 54
Female 41
Race
White 51
Black 31
Age (in years)
18–29 37
30–49 40
50+ 57
Education
High school or less 35
Some college 43
College graduate 68
Sources: Data from “From Brexit to Zika: What Do Americans Know?” Pew Research Center, Washington, DC, https://www.people-
press.org/2017/07/25/from-brexit-to-zika-what-do-americans-know/; Analysis of June 22-25, 2017 Weekly Survey data file, Pew
Research Center, Washington, DC, July 25, 2017, https://www.people-press.org/dataset/june-22-25-2017-weekly-survey/
An experiment demonstrated that citizens who had the opportunity to interact in an e-townhall with their
member of Congress and other constituents about a controversial issue—immigration—became more
knowledgeable about the issue.72 Having the chance to interact with their representative motivated
citizens to seek out information and to discuss the issue with additional people. Thus, motivation should
not be considered a static characteristic of individuals; instead, citizens have the capacity to learn about
politics and can be motivated to do so through electronic deliberation with elected officials and fellow
citizens.
https://www.people-press.org/2017/07/25/from-brexit-to-zika-what-do-americans-know/
https://www.people-press.org/dataset/june-22-25-2017-weekly-survey/
Context and Political Knowledge
Political knowledge is not simply a function of individual characteristics and motivation. To become
knowledgeable about politics, citizens must also have the opportunity to do so.73 Some contexts offer
greater opportunities for citizens to become informed than others. Richard Niemi and Jane Junn
emphasize the influence of civic education on political knowledge.74 They analyzed survey data from a
nationally representative sample of high school seniors and found that students who had more civics
classes and more recent civics instruction were more knowledgeable than others. Furthermore, the
particular type of curriculum mattered—students who were exposed to a wide variety of topics and
discussed current events in their classes were more likely to be knowledgeable about politics.
The political structure also influences citizens’ levels of political knowledge. Because political
structures differ across nations, scholars can compare the effects of different institutional arrangements
and electoral systems on political knowledge cross-nationally.75 For example, in a nation with
competitive elections, parties are motivated to provide information to citizens, and citizens find the
information useful. In such a political context, citizens are more knowledgeable about politics. Similarly,
in the United States, competitive House races lead to more local news coverage, which then results in
citizens being better informed about the congressional election.76 Unfortunately there are many
uncompetitive congressional districts in the United States, so this is a significant hindrance to citizen
knowledge.
Jennifer Jerit, Jason Barabas, and Toby Bolsen examined how the information environment affects
citizens’ political knowledge.77 They argue that citizens will be better informed about politics when the
information environment is rich. In other words, when the media pay a great deal of attention to political
issues, the public becomes more knowledgeable as a result. An information-rich environment, however,
helps the better educated more than their less-educated counterparts. Thus, information does not level
the playing field; instead, well-educated citizens are better equipped to integrate new knowledge into
their existing stores of information, which allows them to move even further ahead of those who are less
educated. Thus, the knowledge gap between the less educated and better educated is exacerbated by
an information-rich environment.
Does it matter whether the information-rich environment is created by news stories in print or on
television? Yes, it does. Jerit et al. examined the relationship between different types of media coverage
and political knowledge of forty-one issues, ranging from understanding how the West Nile virus is
spread to knowledge of the Supreme Court’s ruling on partial-birth abortion to knowing about President
George W. Bush’s drug plan. The researchers found that on issues with low levels of media attention
(such as Bush’s drug plan), highly educated people were only slightly more likely to answer questions
correctly than those with less education, whereas there was a large knowledge gap on issues receiving
a great deal of media attention (such as how the West Nile virus is spread). Furthermore, this
knowledge gap was exacerbated when the issue received extensive coverage in the print media. In
contrast, for issues receiving extensive television coverage, both less-educated and highly educated
people gained. A knowledge gap remained, of course, but the important point is that the gap was not
increased by the television coverage.78 Media critics have complained about the simplicity and inanity of
television news compared with print, but these findings suggest that television presents information in a
way that benefits those who are least knowledgeable. What about obtaining information online? A
recent summary of several studies suggests that online media use widens the knowledge gap between
those who are more and less educated. Thus, unfortunately, the Internet does not seem to level the
playing field.79
The political context in the United States is dominated by men, which has an effect on levels of political
knowledge among women. A study by Nancy Burns, Kay Schlozman, and Sidney Verba asked the
question, “What if politics weren’t a man’s game?”80 Burns et al. found that women living in states with a
female senator or a female Senate candidate were substantially more knowledgeable about politics than
women living in states without those characteristics. Specifically, only 51 percent of women from male-
dominated states could name one of their U.S. senators, whereas 79 percent of women could do so
when the political context included powerful female politicians.81 This is a huge difference and suggests
that the knowledge gap between men and women could close if more women served in elective office.
There is also evidence that women’s political participation would be higher if there were more women
senators because more knowledgeable citizens also tend to be more involved citizens.82
Overall, these studies suggest the political context matters a great deal. Citizens are more
knowledgeable when the political context is a favorable one. This is consistent with the views of
participatory democratic theorists. If you are interested in boosting levels of political knowledge, as
participatory democratic theorists are, then it makes sense to structure the political environment in a way
that encourages and enables citizens to easily and efficiently acquire information.
WHAT ARE THE CONSEQUENCES OF POLITICAL KNOWLEDGE?
Does it matter whether citizens are politically knowledgeable or ignorant? This is a normative question to
be sure. As we have already discussed, different democratic theorists have different views on the
consequences of political ignorance. But it is also an empirical question. By examining the differences
between those who are knowledgeable and those who are not, we can get a handle on the
consequences of political ignorance.
Delli Carpini and Keeter identify a number of ways in which knowledgeable citizens differ from others.83
First, well-informed citizens are more likely to demonstrate political tolerance, a fundamental norm in a
democratic society. Second, knowledgeable citizens participate at higher rates than their counterparts.
Third, knowledgeable citizens are more likely to have stable issue opinions and structure their opinions
along a liberal-conservative continuum. Fourth, informed citizens are more likely to recognize their
interests, bring their issue positions into line with their party identification, and vote accordingly. Fifth,
knowledge begets knowledge. In other words, knowledgeable citizens are able to handle new
information with ease; existing knowledge allows citizens to efficiently incorporate new tidbits into their
belief systems. All these things are critical in a democracy, according to Delli Carpini and Keeter.
“Because so many of these differences bear on the issue of political power, a key implication of these
findings is that the maldistribution of political knowledge has consequences: it threatens the basic
democratic principle of political equality among citizens.”84 Delli Carpini and Keeter find this
maldistribution particularly alarming because those who are already less powerful in society (such as
women, minorities, the poor, the less educated, and young people) are the ones who are less
knowledgeable.
Political ignorance also has an effect on policy attitudes and vote choice. Recall the studies on
misinformation we discussed earlier in the chapter.85 The Kuklinski et al. study showed that incorrect
beliefs led citizens to be more opposed to welfare spending than they otherwise would be, and the Kull
et al. research demonstrated that citizens who held more misperceptions about the war in Iraq were
more likely to support the war. Similarly, one study revealed that the less accurate Americans were at
locating Ukraine on a map, the more likely they were to support U.S. military intervention in the
country.86 Another study used an experimental design to show that when citizens were provided with
accurate information about the decrease in crime in recent years, they were more likely to say the
government was spending too much money on building prisons.87 Other scholars have demonstrated
that uninformed citizens would vote for different candidates if they were fully informed.88 Thus, this
research highlights the important, and often quite worrisome, policy and electoral impact of citizens’ lack
of knowledge.
ARE CITIZENS INTERESTED IN AND ATTENTIVE TO POLITICS?
Are citizens interested in politics and political campaigns? Do they pay attention to politics, or are they
disengaged? These are important questions because citizens who are interested in and attentive to
politics tend to have higher levels of political knowledge than citizens who are not engaged.89 Political
interest allows citizens to join “the self-governing class.”90 In other words, political interest provides the
motivation and knowledge for citizens to fully and effectively participate in our democratic system.
Interest in Politics and Political Campaigns
Let’s first examine public interest in politics, or the degree to which people find politics appealing. In a
2016 ANES survey, 19 percent of citizens said they always pay attention to what’s going on in
government and politics; 34 percent said most of the time; 21 percent said about half of the time; 24
percent said some of the time; and only 3 percent said never.91
How does this compare to prior years? As you can see from the note to Figure 8.2, the interest-in-
politics question wording and response options were different in 2012 and 2016 than in prior years. One
of the most important changes was that respondents in 2012 and 2016 were offered five response
options, whereas previously they were offered only four. Unfortunately that makes the across-time
comparison problematic. To make that clear in the figure, the interest in politics data for 2012 and 2016
are connected to each other, but not to the longer time series lines.
In the 2008 ANES, 26 percent of citizens said they “follow what’s going on in government and politics
most of the time.” Another 37 percent indicated they were interested “some of the time” (see the solid
lines in Figure 8.2). Despite the different question wordings, it seems that modestly interested is the best
way to characterize citizens’ interest in politics in recent years.
What about interest in political campaigns? In 2016, citizens were quite interested.92 Forty-nine percent
said they were “very much interested” and another 37 percent said they were “somewhat interested”
(see the dotted lines in Figure 8.2). Note that 2016 was the highest level of interest in political
campaigns recorded since the survey’s inception in 1952, which may reflect public interest in Donald
Trump’s unconventional campaign as well as Hillary Clinton’s historic candidacy as the first woman to
capture a major party nomination for the presidency. Overall, then, citizens seem to perk up a bit during
the presidential campaign season but are somewhat less engaged when it comes to day-to-day politics.
Description
Figure 8.2 Interest in Politics and Current Campaign, 1964–2016
Sources: The American National Election Studies (www.electionstudies.org), The ANES Guide to Public
Opinion and Electoral Behavior; Analysis of American National Election Study 2012 Time Series; User’s
Guide and Codebook for the ANES 2016 Time Series Study.
http://www.electionstudies.org/
Notes: Interest in Politics question wording: In 2012 and 2016: “How often do you pay attention to what’s
going on in government and politics? Always, most of the time, about half the time, some of the time, or
never”; 1964–2008: “Some people seem to follow (1964: think about) what’s going on in government and
public affairs most of the time, whether there’s an election going on or not. Others aren’t that interested.
Would you say you follow what’s going on in government and public affairs most of the time, some of the
time, only now and then, or hardly at all?”
Interest in Current Campaign question wording: “Some people don’t pay much attention to the political
campaigns. How about you, would you say that you have been/were very much interested, somewhat
interested, or not much interested in (1964–2008: following) the political campaigns (so far) this year?”
A recent study by Markus Prior examined several measures of political interest during different time
periods and in different countries. Based on his exhaustive analysis, Prior concludes that the stability of
political interest is most notable.93 Despite differing political contexts, aggregate levels of political
interest stay fairly constant over decades. And at the individual level, citizens tend to remain either
interested or not once they reach adulthood. So, although there was an upward blip in interest in the
2016 campaign apparent in the ANES data, Prior argues that the overall story is one of people either
being interested in the political world around them or not, regardless of the particular situation at hand.94
Despite Prior’s emphasis on the stability of political interest, there are some instances in which political
context matters.95 During the 2008 presidential campaign, for example, 51 percent of Blacks said they
were very much interested in the campaign compared with only 44 percent of whites. Further, in 2004,
only 32 percent of Blacks said they were very much interested in the campaign. Having Barack Obama
on the ballot stimulated a great deal of interest among Black Americans in 2008, a pattern that
continued in 2012.96 Similarly, in states where females hold statewide offices (such as senator or
governor), women are significantly more interested in politics than in states without such female
representation.97 Likewise, women in states with competitive female candidates are more likely to
discuss politics and try to convince others to support a political candidate than women in states without
viable female contenders.98 The political context is also important for adolescents. Girls, for example,
are more likely to anticipate future political involvement when they see women run high-profile, viable
political campaigns.99
As with political knowledge, levels of interest vary across demographic groups.100 Men and whites are
generally more interested in politics. Likewise, those with more education and higher incomes are more
concerned with political happenings. Age, in particular, has a strong relationship with political interest;
older individuals follow politics more than young people. Partisans and ideologues are also much more
likely to follow politics than independents or moderates.
What do democratic theorists have to say about levels of political interest? Participatory democratic
theorists would be concerned with the demographic differences in political interest, and they would
argue that changes in the political context are necessary to ensure all citizens are interested in politics.
Some elite democrats, on the other hand, would wonder why so many citizens are interested in politics.
From their point of view, it is irrational for citizens to spend time on politics. The likelihood that citizens,
even interested and informed ones, could make a difference is so small that it simply doesn’t make
sense for them to devote resources to such an endeavor.
Bystanders
Do citizens pay attention to politics? According to a 2017 survey conducted by the Pew Research
Center, 83 percent of the public say they “follow what’s going on in government and public affairs” most
or some of the time. Some citizens, however, are inattentive to politics. Indeed, the Pew data
demonstrate that 8 percent of the public are best characterized as bystanders.101 Bystanders are not
registered to vote and rarely follow public affairs; they are politically disengaged. In the words of the Pew
Research Center, bystanders “don’t give a hoot about politics.”102 Instead, they are interested in health,
celebrities, and entertainment. They are more likely to have a high school education or less and be
young, Hispanic, and foreign born than citizens who show at least some interest in politics.
What do we make of the fact that 8 percent of the American public are totally disengaged from politics?
Should we look at this as only 8 percent are inattentive so it’s no big deal, or should we be concerned
that a small minority of citizens are interested in Kim Kardashian but not Kim Jong-un? Participatory
democrats would certainly have higher expectations for the public and would be especially concerned
that bystanders are disproportionately uneducated, young, Hispanic, and foreign born. Elite democrats
would be less concerned yet would be bothered that members of this group are not even registered to
vote, a minimal threshold these bystanders do not cross.
CONCLUSION
Are citizens highly knowledgeable about politics? No, many people are not particularly well informed,
and some citizens are outright misinformed about important political issues. Nevertheless, people are
not as ignorant about politics as some observers have suggested. Further, citizens have not become
more ignorant over time; levels of political knowledge have been remarkably stable over the past eighty
years. There is a controversy about how to measure political knowledge, with several scholars arguing
that Delli Carpini and Keeter’s method is too narrow or that it underestimates citizen competence.
Why are some citizens more knowledgeable than others? Citizens differ in ability, motivation, and
opportunity. Formal education, for example, provides citizens with the cognitive ability to become
informed about politics. Other demographic characteristics matter as well; men, whites, higher-income,
and older people are more likely to be knowledgeable. Motivational factors, such as interest, attention,
and discussion, lead citizens to be informed. The political context also acts as either a facilitator or
inhibitor of political knowledge.
Are there consequences of political knowledge? Yes, indeed there are. Citizens who are knowledgeable
are more likely to endorse democratic values and participate in political activities. They have opinions on
issues that are more stable and more closely linked to their ideology and party identification. In addition,
knowledgeable citizens are capable of dealing with new information in an efficient manner, making it
easier for them to learn even more about politics. And perhaps most important, informed citizens hold
different issue opinions than they otherwise would. Thus, political knowledge has significant
ramifications for policy preferences among the public.
Are citizens interested in and attentive to politics? Citizens demonstrate modest interest in politics and
fairly high interest in political campaigns. A small subset of the population indicate they basically do not
pay attention to politics at all. They are bystanders to democracy.
Overall, do people live up to the democratic ideal of knowledgeable, engaged citizens? Participatory
democratic theorists would say no. They would argue that citizens should be more knowledgeable so
they can play an active role in the democratic process. Informed and engaged citizens are at the heart
of a democracy. They are necessary to influence elite behavior and hold elites accountable.
Participatory democratic theorists find the demographic differences in political knowledge especially
troubling because they are concerned with inequalities in society. These theorists do not blame
individuals for their lack of political savvy and interest, however. Instead, they point to the barriers that
limit citizen engagement and knowledge.
In sharp contrast, some elite democratic theorists are curious about why citizens are interested and
knowledgeable at all, given the low odds that their voices will be influential. In their minds, the puzzle is
not why so many people lack knowledge but why so many people are informed. Elite democrats who
think government works best when the public interferes least would probably be satisfied with the low
levels of knowledge and modest levels of political interest. Scholars who emphasize that citizens can
make sensible political decisions by using heuristics would be troubled by the lack of knowledge about
such things as party stances on the issues. Some citizens can use cues to effectively muddle their way
through, but many people do not even have those basic insights.
You are an expert now on political knowledge, interest, and attention. You understand the different
normative approaches to citizen competence, and you have a good grasp of the empirical research
findings. So what do you think? Are citizens knowledgeable enough, interested enough, and attentive
enough to function effectively in a democracy?
KEY CONCEPTS
attention to politics 268
bystanders 269
civic education 263
directionally motivated reasoning 249
expanded model of civic competence 259
focus groups 259
gender gap in political knowledge 254
generalist 252
heuristics 243
information environment 263
informed 246
interest in politics 266
limited information processors 244
low-information rationality 244
misinformed 246
multiple-choice questions 255
people and players 246
political knowledge 241
political structure 263
racial gap in political knowledge 260
rules of the game 246
short-answer questions 255
specialist 252
substance of politics 246
timebound 251
uninformed 246
SUGGESTED SOURCES FOR FURTHER READING
Davis, Darren W., and Brian D. Silver. “Stereotype Threat and Race of Interviewer Effects in a Survey on
Political Knowledge.” American Journal of Political Science 47 (2003): 33–45.
McGlone, Matthew S., Joshua Aronson, and Diane Kobrynowicz. “Stereotype Threat and the Gender
Gap in Political Knowledge.” Psychology of Women Quarterly 30 (2006): 392–398.
These scholars argue that stereotype threat explains at least part of the racial and gender gap in
political knowledge. Stereotype threat occurs when Blacks and women feel pressure to perform well on
political knowledge questions because they are aware of negative stereotypes about their intellectual
abilities. As a result of stereotype threat, Blacks and women do worse on political knowledge tests—not
because they are less knowledgeable but because the stress of the situation interferes with their ability
to answer correctly.
Hochschild, Jennifer L., and Katherine Levine Einstein. Do Facts Matter? Information and
Misinformation in American Politics. Norman, OK: University of Oklahoma Press, 2015.
Hochschild and Einstein discuss several case studies of misinformation and its pernicious effect on
American democracy.
Levitsky, Steven, and Daniel Ziblatt. How Democracies Die. New York, NY: Broadway Books, 2018.
Levitsky and Ziblatt argue it is critical for American citizens to have knowledge of how democracies die
so that they will know what to do, and what not to do, to preserve democracy in the United States.
Prior, Markus. Hooked: How Politics Captures People’s Interest. New York, NY: Cambridge University
Press, 2019.
Prior’s book is an in-depth exploration of interest in politics in the United States, Germany, Switzerland,
and Great Britain.
Pew News IQ Quiz, http://pewresearch.org/newsiq
Check out this website for an online news quiz. Take the quiz and then compare your performance with
that of other American citizens.
http://pewresearch.org/newsiq
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists news sources. The vertical axis lists percentage and ranges from 0 to 100 in
increments of 20. Approximate data from the graph are tabulated as follows:
News source Percentage
Fox 80
CBS 70
ABC 61
CNN 55
NBC 55
Print media 47
NPR or PBS 23
Back to Figure
The horizontal axis lists years and ranges from 1964 to 2016 in increments of 4. The vertical axis lists
percentage and ranges from 0 to 100 in increments of 20. Approximate data from the graph are
tabulated as follows:
Year
Interest in politics:
Always in 2012 and
2016, most of the time
between 1964 and
2008
Interest in politics: Always or
most of the time in 2012 and
2016, most or some of the
time between 1964 and 2008
Interest in
current
campaign:
Very much
interested
Interest in
current
campaign: Very
much or
somewhat
interested
1964 30 72 38 75
1968 33 63 39 80
1972 38 72 30 75
Year
Interest in politics:
Always in 2012 and
2016, most of the time
between 1964 and
2008
Interest in politics: Always or
most of the time in 2012 and
2016, most or some of the
time between 1964 and 2008
Interest in
current
campaign:
Very much
interested
Interest in
current
campaign: Very
much or
somewhat
interested
1976 38 70 38 80
1980 25 61 30 75
1984 25 62 27 75
1988 20 60 27 75
1992 24 65 40 83
1996 21 61 24 75
2000 20 59 24 75
2004 25 65 41 83
2008 25 63 43 83
2012 15 42 48 82
2016 15 53 48 83
CHAPTER 9 SUPPORT FOR CIVIL LIBERTIES
SHOULD ELEMENTARY AND HIGH SCHOOL STUDENTS be permitted to wear armbands indicating
their political views? Should members of the National Socialist Party of America (Nazis) be permitted to
march through a town that is home to many Jewish residents? Should burning the American flag be
illegal? Should journalists who write news articles critical of the government be censored or have their
access to government officials restricted? Should white supremacists be allowed to hold a public rally?
What about antiwar or anticapitalist activists? Should people holding controversial views be permitted to
speak on college campuses? These, of course, are not hypothetical questions. The first three were
raised in response to, respectively: five Des Moines, Iowa, students protesting the Vietnam War (1965);
a proposed march in Skokie, Illinois (1977); and flag burning at a protest during the Republican National
Convention in Dallas, Texas (1984). The topic of how far the freedom of the press should extend has
popped up repeatedly, from the Sedition Act of 1798 to the Trump administration revoking White House
press passes of some journalists.1 And, questioning the rights of individuals to espouse their views,
whether in a town square or on a campus, is a long-standing practice in the United States.
Alongside the principle of majority rule, democracies guarantee the political rights of those individuals or
groups holding unpopular viewpoints. In fact, this protection of minority views is one key characteristic of
democratic nations. It was the principle that was upheld by courts in the three historical examples
mentioned earlier: the Nazis did hold their march, and both student armband wearing and flag burning
were declared constitutionally protected speech.2 In the United States, the rights of freedom of speech,
freedom to assemble, freedom of religion, freedom to petition the government, and freedom of the press
are protected by the First Amendment of the Constitution. These rights are known as civil liberties,
which are defined as rights granted to citizens that are protected from government suppression. Public
support for civil liberties is known as political tolerance. Yet we do not commonly refer to someone as
tolerant if he supports the exercise of civil liberties only for those who share his political beliefs. More
typically, tolerance refers to extending freedoms to individuals or groups with whom one disagrees. As
Samuel Stouffer, the author of a groundbreaking book on political tolerance, put it, “If I am tolerant, there
are rights which I will grant, within the law, even to those whom I most condemn.”3
For classical and participatory democratic theorists, a key characteristic of democracy is the active
engagement of the public. Citizens should debate issues and share their preferences with politicians,
and the policy that results should reflect the wishes of the people. Political equality is also important to
these theorists; all citizens should be able to share their views in the public forum regardless of the
content of these views. Equality is important because, among other matters, only when there is such a
free marketplace of ideas will leaders be able to discern what is in the public’s interest. For such
equality to exist in practice, support for freedom of expression must be high among the public. Although
these theorists do not necessarily assume that everyone will be tolerant, high levels of public intolerance
would be counter to the democratic principles they espouse. Participatory democratic theorists have
focused particularly on the importance of civil liberties for democracy. They further believe that with
more formal education, greater involvement in politics, and more exposure to democratic principles,
citizens can learn to tolerate the political expressions of all, even those with whom they vehemently
disagree. Thus, participatory democrats value educational and political opportunities for the public to
learn key democratic norms.
In contrast, elite democrats are not nearly as optimistic about the citizenry. In fact, given these theorists’
beliefs that the public is uninformed about and uninterested in politics and their assumption that when
citizens do participate they do so only to further their own interests, democratic elitists expect the public
to be intolerant of political viewpoints that differ from their own. How, then, can a democracy that values
minority rights continue to function if there are high levels of public intolerance? To answer this question,
democratic elitists, and also pluralists, turn to elites. Elites, they argue, strongly support minority rights
and ensure that these rights are upheld. The elites, these theorists contend, are aware of the
importance of civil liberties and further understand the consequences to democracy if civil liberties were
to be eroded. Democratic regimes will be stable, even in the face of widespread public intolerance, as
long as the political elites support civil liberties and ensure that the principle of minority rights is applied
to all.
Not all variants of democratic elitism and pluralism have such faith in the good judgment of elites,
however. In particular, James Madison, at the time of the writing and ratification of the U.S. Constitution,
made a forceful case that all people, elites included, could not be trusted to welcome the political
expression of all viewpoints.4 Madison’s remedy to this problem was the creation of a political system
that, first, divided power among three branches of government and, second, ruled over such a large
population that diverse interests would be represented in the national government. These characteristics
of the constitutional framework would ensure that leaders would need to compromise with others—
others who were representing a multitude of different interests—in the policymaking process. Interacting
with their opponents repeatedly would instill in these leaders an appreciation for the value of tolerating
diverse viewpoints. Occasionally, elites might even need to join in coalition with individuals who hold
very different views than they do. Madison’s was thus a pragmatic approach to ensuring that in America
civil liberties would be ensured, rooted as they were in the political system rather than in the natural
virtue of either the public or the elites.
In this chapter, we discuss attitudes toward civil liberties and tie this evidence back to the normative
debates about support for civil liberties in a democracy. We begin by discussing a topic that has been
the focus of political discourse since Donald Trump was elected: support for democratic principles. Then
we move on to explore levels of tolerance in the United States, sources of tolerance, and differences
between elites and citizens in their support for civil liberties. Much of our attention will be on views
toward the freedoms of expression contained in the First Amendment. The first ten amendments of the
Constitution (the Bill of Rights) list a number of other civil liberties such as the right to bear arms and
some rights of people accused of crimes, including the rights against self-incrimination and
unreasonable searches and seizures. Historically, less scholarly attention has focused on public support
for these liberties. That changed, however, in the wake of the September 11, 2001, al-Qaeda attacks
and increased national attention on how to effectively capture terrorists. That debate has swirled around
the proper balance between protecting civil liberties and promoting national security. The implications of
this changed climate on public support for civil liberties since the 9/11 attacks are discussed later in the
chapter.
SUPPORT FOR DEMOCRATIC PRINCIPLES
Is democracy dying? This provocative question appeared on the cover of the October 2018 issue of The
Atlantic. Inside the magazine were essays on a range of topics, such as democratic backsliding in
Europe, the threat to democracy of technology, the dwindling practice of democratic activity among the
American public, and how hyper-partisanship and racism undermine democratic principles. It is not only
The Atlantic where discourse about the state of American democracy has appeared in the years since
Trump entered the White House, much of which centers on his antidemocratic impulses and actions. As
summarized by a group of political scientists, “Trump openly derides core institutions of democratic
governance: the independent press, the judiciary, the bureaucracy, the validity of elections, the
legitimacy of democratic contestation, and the centrality of facts to political discourse … In international
affairs, he clashes with America’s strongest democratic allies and obviously admires autocratic rulers.”5
Does the public hold views similar to Trump? Or do Americans display stronger support for democratic
principles?
An overwhelmingly large portion of the public supports democracy as a form of government. Nearly 90
percent of Americans stated that it is very good or somewhat good that the nation be governed by “a
democratic system where representatives elected by citizens decide what becomes law” in the early
2000s and again in 2017.6 When it comes to support for specific democratic principles, a similar picture
emerges. In the late 1930s and early 1940s, for example, over 95 percent of survey respondents
answered yes to the following question: “Do you believe in freedom of speech?”7 Across the decades,
similarly high percentages of citizens have agreed that “people in the minority should be free to try to
win majority support for their opinions,” agreed that “[no] matter what a person’s political beliefs are, he
is entitled to the same legal rights and protections as anyone else,” and believed in “free speech for all,
no matter what their views might be.”8
Recent polling conducted by the Pew Research Center demonstrates Americans’ continued high
support for a number of democratic principles, such as judicial independence, freedom of the press, and
freedom of expression. In particular, the Pew survey asked people whether certain values are important
for the nation. The lighter colored bars in Figure 9.1 represent the percentage of the public indicating
that the principle is very important; the darker color bars are the percentage of those considering the
principle very or somewhat important. A majority of the public view all five of these principles as very
important, although the percentages vary from 62 for “views of those not in the majority on issues are
respected” to 84 for “rights and freedoms of all people are respected.” When we add in those members
of the public who consider a principle to be somewhat important, however, the percentages indicating
that the principle is important rise to at least 92 percent for all. The American public thus values many
democratic principles, but do they think that Donald Trump does? The same Pew survey asked
respondents how much respect Trump has “for this country’s democratic institutions and traditions.”
Forty-five percent replied that he has a great deal or a fair amount of respect versus 54 percent who
said he has not too much respect or none at all.
Description
Figure 9.1 Importance of Democratic Principles
Source: Data from “The Public, the Political System and American Democracy,” Pew Research Center,
Washington, DC, April 26, 2018, https://www.people-press.org/2018/04/26/the-public-the-political-system-
and-american-democracy
Note: Bars represent the percentage of the public that considers a principle important for the country.
https://www.people-press.org/2018/04/26/the-public-the-political-system-and-american-democracy
ARE AMERICANS TOLERANT?
As we discussed in the previous section, public support for civil liberties in the abstract is very high. Yet
support for the application of these principles in specific circumstances tends to be lower, sometimes
substantially so.9 For one, the public is not equally tolerant of all types of political acts. Support for
certain acts, such as blocking entrances to government buildings or violent demonstrations, is quite
low.10 Second, when it comes to freedom of speech on college campuses, some people—college
students and nonstudent adults alike—express willingness to place restrictions on specific types of
speech. Broadly speaking, 22 percent of college students and 28 percent of adults view it important for
fostering a conducive learning environment to prohibit speech that is “offensive or biased against certain
groups of people.”11 As for which types of speech colleges should be able to limit, college students are
much more supportive of policies that would restrict “slurs and other language … intentionally offensive
to certain groups” versus “political views that are upsetting or offensive to certain groups” (69 vs. 27
percent support, respectively). Third, tolerance of civil liberties for political extremists tends not to be
very high among the American public, although such tolerance has increased over the decades. Or has
it? Drawing conclusions about trends over time is not so easy and, as we demonstrate on the following
pages, depends in part on how tolerance is assessed.
Box 9.1 Support for Democracy around the Globe
Public Opinion in Comparative Perspective
“A deepening anxiety about the future of democracy around the world has spread over the past
few years. Emboldened autocrats and rising populists have shaken assumptions about the future
trajectory of liberal democracy, both in nations where it has yet to flourish and countries where it
seemed strongly entrenched.”1 So began a 2017 Pew Research Center report of survey data
collected in 38 nations. One of the main conclusions is that large majorities of people prefer that
their nation be governed via either representative or direct democracy. Support for representative
democracy tends to be higher in nations that are wealthier and have well-functioning
democracies and is lower in South America compared to other regions of the world.
On the other hand, support for nondemocratic governing measures was evident. For instance,
across these nations, an average of 49 percent of the public express support for their country to
be governed by a panel of experts rather than elected representatives. Approximately one-
quarter of these national publics believe that governing by a strong leader (who can make
decisions without other political institutions interfering) or by the military is good. Support for all
three measures was higher than average in Asia-Pacific and African nations. Finally, even
though people living in wealthier nations display high support for representative democracy,
preferences for other governing models is not insignificant in some of these countries. In the
figure on the next page, note the relatively high preference for strong leader rule in Japan, Italy,
and the United Kingdom as well as the percentages of people favoring military rule in Chile,
France, Italy, and the United States.
Description
Support for Nondemocratic Governing Systems in High-Income Nations
Source: Data from “Globally, Broad Support for Representative and Direct Democracy,” Pew
Research Center, Washington, DC, October 16, 2017,
https://www.pewresearch.org/global/2017/10/16/globally-broad-support-for-representative-and-
direct-democracy/
Note: Numbers represent the percentage of people in each nation who responded that “rule by
[military/strong leader/experts] would be a good way to govern our country.”
1 Richard Wike, Katie Simmons, Bruce Stokes, and Janell Fetterolf, “Globally, Broad Support for
Representative and Direct Democracy,” Pew Research Center, Washington, DC, October 16,
2017, https://www.pewresearch.org/global/2017/10/16/globally-broad-support-for-representative-
and-direct-democracy/
https://www.pewresearch.org/global/2017/10/16/globally-broad-support-for-representative-and-direct-democracy/
https://www.pewresearch.org/global/2017/10/16/globally-broad-support-for-representative-and-direct-democracy/
Stouffer’s Classic Study
In one of the very first empirical analyses of public support for civil liberties, Samuel Stouffer assessed
tolerance toward individuals holding unconventional political views.12 It was the 1950s, and Stouffer was
interested in probing a variety of public attitudes toward communism. A primary goal of his analysis was
determining levels of tolerance toward communists and two other groups, socialists and atheists. During
this era, known as the Red Scare, political leaders at both the national and state levels, prominently led
by U.S. senator Joseph McCarthy, had taken actions to restrict the rights of communists. Suspected
communists were jailed, blacklisted, and, in some cases, deported. Stouffer wanted to explore whether
the public was intolerant of communists and, if so, what measures might be taken to increase public
tolerance.
Stouffer’s conclusions are based on results from two public opinion polls conducted in 1954. Survey
respondents were selected from a national probability sample. Their responses, therefore, can be
generalized to the entire nation. For each of the three nonconformist groups under study, Stouffer
queried respondents’ willingness to tolerate specific activities performed by group members. The survey
questions used by Stouffer to assess tolerance toward these groups have been repeated over the years
by many others researching civil liberties. The questions have also come under criticism, as we will soon
see. Thus, it is important that we consider how these questions were worded. Table 9.1 presents the
questions regarding atheists. As you can see, the term atheist does not appear in the question. Instead,
an atheist’s beliefs were presented (specifically, “somebody who is against all churches and religion”).
Furthermore, these questions assessed tolerance for three specific activities: making a speech, keeping
a published book in the public library, and teaching at a university. The first two are clearly examples of
freedom of expression, a civil liberty protected by the Bill of Rights. Following these were questions
gauging tolerance of the same three activities for the other two groups. As with atheists, Stouffer
described the views of socialists (those favoring “government ownership of all the railroads and all big
industries”) rather than using the group name. In contrast, he referred to communists directly by name,
instead of presenting their views. Presumably citizens were more familiar with communists than with the
other two groups, given the national attention devoted to communism during the 1950s.
Table 9.1 Assessing Public Tolerance of Atheists: Stouffer’s Survey Questions
“There are always some people whose ideas are considered bad or dangerous by other people. For
instance, somebody who is against all churches and religion.”
“If such a person wanted to make a speech in your city (town, community) against churches and
religion, should he be allowed to speak, or not?”
“Should such a person be allowed to teach in a college or university, or not?”
“If some people in your community suggested that a book he wrote against churches and religion
should be taken out of your public library, would you favor removing this book, or not?”
Source: Samuel A. Stouffer, Communism, Conformity, and Civil Liberties: A Cross-Section of the Nation Speaks Its Mind (Garden City,
NY: Doubleday, 1955), 252.
Levels of public tolerance toward these groups in 1954 are presented in Figure 9.2. The bars represent
the percentage of the public who would tolerate the specific act for each group. At first glance, it is clear
that levels of tolerance toward these groups were not very high. In fact, the public might best be
described as intolerant, especially toward communists and atheists. At most, only about one-quarter of
the public supported any of the three rights for communists. Public support for civil liberties toward
atheists was somewhat higher, with just over one-third feeling that atheists should be able to make
speeches or have their books kept in the local library. Tolerance was highest for socialists, with a slight
majority of the public expressing support for two of the rights.
Overall, then, although public tolerance was quite low in 1954, citizens did make distinctions across the
political groups. Support for the civil liberties of socialists was highest, possibly because atheists and
communists were perceived to be further from the mainstream and also because communism was
viewed by many to be a threat to the American way of life. Levels of tolerance also varied across the
specific acts, with tolerance consistently higher for the two freedoms of expression than for holding a
teaching job. University teaching, with its potential to indoctrinate young adults, was thought to be more
risky for society than speechmaking or book publishing; hence it was less tolerated.
Description
Figure 9.2 Tolerance of Political Minorities, 1954
Source: Data from Samuel A. Stouffer, Communism, Conformity, and Civil Liberties: A Cross-Section of the
Nation Speaks Its Mind (Garden City, NY: Doubleday, 1955), Chap. 2.
Trends in Political Tolerance
Have Americans become more or less tolerant since the 1950s? Fortunately, we can directly compare
levels of tolerance in that decade with more recent years because Stouffer’s tolerance questions have
appeared on the General Social Survey (GSS) many times since this survey began in 1972. Every year
at first, then biennially since 1994, a national sample of adults has been asked their attitudes toward a
variety of social and political objects.13 The GSS is a cross-sectional study, which means that different
people are surveyed each year. Many of the same questions have been asked over the years, so we
can compare the responses of one year with responses in other years to determine whether the
attitudes of Americans have changed. Keep in mind that the GSS is not a panel study; in other words,
the same people are not surveyed every time. Instead, each time the GSS is conducted, a new national
probability sample of adults is contacted. Based on this sampling design, we can assume that the
results for a given GSS represent the opinions of all American adults for the year of the specific study.
Over the years, the GSS has assessed tolerance of the speechmaking rights of a variety of political
minority groups (see Figure 9.3). First, let’s consider the three groups that Stouffer examined. Public
levels of tolerance toward these groups were much higher in 1972 compared with 1954. For communists
and atheists, tolerance increased further through 1990 and has remained relatively stable for
communists since then. Tolerance of atheists has gradually increased over recent decades, reaching its
highest level recorded by the GSS in 2018, when 81 percent of the public supported the speechmaking
rights of this group.14 It is very possible that tolerance of socialists has also increased, but because the
GSS stopped asking questions about this group in 1974, we cannot be certain about this.
Beginning in the 1970s, the GSS also queried the public about its tolerance of three other groups:
racists (described as “a person who believes that Blacks are genetically inferior”), militarists (“a person
who advocates doing away with elections and letting the military run the country”), and male
homosexuals. Levels of tolerance for these three groups appear in Figure 9.3 as dashed lines.
Tolerance gains have been largest for homosexuals. In 2018, 90 percent of citizens supported
speechmaking by homosexual men, making tolerance of this group higher than the other groups
included on the GSS. The public is also more supportive of civil liberties for militarists now than was the
case a few decades ago. In contrast, tolerance of racists has not increased. Slightly more people
tolerated racists in 1990 compared to the 1970s, but since 1990 public support for racists’
speechmaking has actually declined.15 Finally, in 2008 the GSS began assessing tolerance toward
Muslim extremists (specifically, “a Muslim clergyman preaching hatred of the United States”).
Tolerance toward this group is significantly lower than for the others.16 Only 46 percent would allow
such a Muslim clergyman to make a speech in their communities in 2018, five percentage points higher
than in 2010.
Description
Figure 9.3 Tolerance of Speechmaking, 1954–2018
Sources: Data for 1954 from Samuel A. Stouffer, Communism, Conformity, and Civil Liberties: A Cross-
Section of the Nation Speaks Its Mind (Garden City, NY: Doubleday, 1955), Chap. 2. Data for all other years
from an analysis of the General Social Survey Cumulative Data File, 1972–2018.
Note: 1972 data for racist and militarist are actually from 1976; 1972 data for homosexual man are actually
from 1973.
It is now the case that majorities of citizens, sometimes more than 70 percent, support the practice of
civil liberties for some individuals professing unconventional views, those that are outside of the political
mainstream. Yet tolerance does vary across groups, and for no group or act is support unanimous.
Thus, is it accurate to conclude that U.S. citizens are generally a tolerant lot? Certain democratic
theorists, especially participatory democrats, would be reluctant to offer such a rosy conclusion. They
would be pleased to see the increasing levels of tolerance toward some groups but would probably
prefer even more citizens to possess tolerant attitudes. Furthermore, the empirical evidence
demonstrating increases in tolerance since Stouffer’s original analysis has come under criticism, most
notably by John Sullivan, James Piereson, and George Marcus.
Sullivan, Piereson, and Marcus: Tolerance of Least Liked Groups
To be tolerant, people must allow the behavior of those whom they oppose. To Sullivan, Piereson, and
Marcus, this is a crucial component of tolerance, yet it is one that is not captured by Stouffer’s survey
questions.17 Although many Americans probably disagree with communists or atheists, Stouffer’s
questions do not directly probe for whether the public dislikes these groups. To compensate, Sullivan et
al. used an alternative method for assessing public tolerance, one that they argue more faithfully
represents a definition of tolerance as support for the rights of those holding repugnant political views.
Sullivan et al.’s measures of tolerance have come to be called the “least liked group” approach, based
on the wording of their survey questions. Rather than listing specific groups that they assumed the
public would not support, they asked survey respondents to identify which political group they liked the
least from a list of potentially unpopular groups. Individuals were also allowed to suggest a group that
was not on the list, in the hope that respondents would select the group that they most strongly disliked.
As the respondents perused a printed list of the groups, survey interviewers asked this question:
I am giving you a list of groups in politics. As I read the list please follow along: socialists,
fascists, communists, Ku Klux Klan, John Birch Society, Black Panthers, Symbionese
Liberation Army, atheists, pro-abortionists, and anti-abortionists. Which of these groups do you
like the least? If there is some group that you like even less than the groups listed here, please
tell me the name of that group.18
In addition to the three groups analyzed by Stouffer, included in this list are groups from the political left
(Black Panthers and Symbionese Liberation Army) and right (fascists, Ku Klux Klan, and John Birch
Society), as well as two groups on opposite sides of the abortion debate. This question was first used by
Sullivan and colleagues in the 1970s, so the groups selected are ones that were salient at that time. For
example, the Symbionese Liberation Army, perhaps best known for kidnapping newspaper heiress Patty
Hearst in 1974, combined urban guerilla tactics with calls for a Black revolution. On the other end of the
political spectrum, the John Birch Society is a very conservative organization that promotes individual
freedom and a limited government, and that was formed to protect the Constitution from communism.
More contemporary research conducted by James Gibson contains some of the same groups that
Sullivan and his colleagues did but also includes new ones (such as gay rights activists and radical
Muslims).
Which groups does the American public like the least? Sullivan et al. included their original list of groups
on a national survey of the American public in 1978. Gibson did so in 2005 and again between 2007 and
2011.19 Because the results were similar across the two Gibson studies, we present results from only
one, 2005 (see Table 9.2). In 1978, the most commonly disliked group was communists, followed closely
by the Ku Klux Klan (KKK). By 2005, the KKK was far and away the most disliked group in the nation.
Whereas 44 percent of Gibson’s respondents selected the KKK as their most disliked group, the next
two groups (radical Muslims and Nazis) were selected by only 13 and 12 percent. Another notable
change across these years is that the public was much more likely to select groups on the political right
than the left as their least liked in 2005. The opposite was true in 1978. Focusing on Stouffer’s primary
groups, two results are noteworthy. First, although communists were the most disliked group in 1978,
only 29 percent disliked this group the most. In other words, over 70 percent of people did not choose
communists as the group they liked the least. Second, by 2005, only 10 percent selected one of
Stouffer’s three groups.
Table 9.2 Least Liked Political Groups, 1978 and 2005
Political group Percentage selecting group as their least liked
1978 2005Political group Percentage selecting group as their least liked
1978 2005
Stouffer’s groups
Communists 29 4
Atheists 8 6
Socialists 1
Other groups on political left
Symbionese Liberation Army 8
Black Panthers 6
Gay rights activists
4
Groups on political right
Ku Klux Klan 24 44
Fascists (1978)/Nazis (2005) 5 12
John Birch Society 1
Radical Muslims
13
Militarists
6
Christian fundamentalists
1
Abortion groups
Political group Percentage selecting group as their least liked
1978 2005
Pro-abortionists 4 3
Anti-abortionists 2 2
Other group 2 N/A
Don’t know 10 N/A
Sources: Data from John L. Sullivan, James Piereson, and George E. Marcus, “An Alternative Conceptualization of Political Tolerance:
Illusory Increases 1950s–1970s,” American Political Science Review 73 (1979): 790; James L. Gibson, “Intolerance and Political
Repression in the United States: A Half Century after McCarthyism,” American Journal of Political Science 52 (2008): 102.
Although this examination of which groups Americans dislike most is interesting, the heart of Sullivan et
al.’s and Gibson’s research focused on whether the public was tolerant of those groups that they
disliked. Thus, after being asked which group they disliked most, both sets of survey respondents were
asked whether they would support certain activities performed by members of the selected group, such
as holding rallies and making speeches. All participants were also asked whether they would tolerate
the following activities for communists and atheists: speechmaking (in 1978) and holding rallies (in
2005). This clever research design allows us to compare public levels of tolerance toward least liked
groups with two of the groups that were the focus of Stouffer’s research.
For both activities and in both 1978 and 2005, we see that public tolerance is higher for communists and
atheists than it is for people’s self-selected least liked group (see Figure 9.4).20 Regardless of which
group was selected, in 1978, only 50 percent of the public would support speechmaking for their least
liked group, whereas 63 percent and 65 percent would support such rights for communists and atheists,
respectively. By 2005, tolerance of speechmaking for least liked groups had declined to 36 percent. A
similar pattern exists for tolerance of rallies: less tolerance toward least liked groups in 2005 versus
1978 and more tolerance toward communists and atheists versus least liked groups.
Sullivan et al.’s new measures of tolerance significantly altered research on political tolerance and
changed many people’s conclusions about how tolerant Americans are. Their results, and Gibson’s
more recent work, seem to indicate that the public has not become significantly more tolerant over time,
in contrast to the conclusions of those who updated Stouffer’s analysis beginning in the 1970s. Such
apparent increases in tolerance were misleading, argue Sullivan and colleagues, and were the result of
less negative attitudes toward the groups under study rather than the result of more support for civil
liberties among the public. During the Cold War of the 1950s, the political times produced especially
high levels of intolerance toward communists and groups with similar ideologies. Yet, as time passed
and the threat of communism receded in many people’s minds, the public was more likely to support
freedom of expression for these groups. Concurrently, other political groups became the focus of public
intolerance. By assessing public attitudes toward a variety of groups, and especially by allowing
individuals to select which groups they strongly disliked, Sullivan et al. were able to pick up public
intolerance that other researchers missed.
These high levels of intolerance would certainly disappoint classical democratic and participatory
democratic theorists because the results indicate majorities of the public would restrict the rights of
those individuals whose opinions they oppose. Further, when Sullivan and his colleagues asked whether
least liked groups should be outlawed, over two-thirds of the public felt their most disliked group should
not even be allowed to exist. Democracy, these theorists argue, requires the free exchange of ideas and
the opportunity for political minorities to have their views represented in public forums.
Description
Figure 9.4 Tolerance of Least Liked Groups, Communists, and Atheists,
1978 and 2005
Sources: Data from John L. Sullivan, James Piereson, and George E. Marcus, “An Alternative
Conceptualization of Political Tolerance: Illusory Increases 1950s–1970s,” American Political Science
Review 73 (1979): 787; James L. Gibson, “Intolerance and Political Repression in the United States: A Half
Century after McCarthyism,” American Journal of Political Science 52 (2008): 104.
Sullivan et al. did not view their findings quite as pessimistically for the prospect of democracy. Rather
than leveling intolerance toward only one group, they concluded that pluralistic intolerance best
describes Americans’ attitudes toward civil liberties. Collectively, the public selects a number of groups
to not tolerate. With intolerance spread around, it is not the case that only one group’s rights are
undermined by the tyranny of the majority. In other words, “the diversity of the targets selected works
against and mitigates to some extent the rather high levels of intolerance.”21 On the other hand, Gibson
draws quite a different conclusion about the effects of pluralistic intolerance. With more groups being
disliked, it is more likely that any one person will think her civil liberties are under attack. As Gibson puts
it, “Because some Communists are not tolerated and do not feel free, and because some Religious
Fundamentalists are not tolerated and do not feel free, and because some of those sympathetic to other
groups are not tolerated and do not feel free, the cumulative effect is more widespread feelings of lack
of freedom today than in the McCarthy era.”22
SOURCES OF TOLERANT ATTITUDES
There are many reasons why some people are more tolerant than others. Chief among these is a
person’s years of formal education. Just as theorists of participatory democracy hope and predict, the
more education people have, the more likely they will be tolerant toward those with unpopular political
views. This relationship between education and tolerance has been uncovered by many researchers,
beginning with Stouffer’s study and continuing since then.23 Furthermore, some researchers have even
concluded that education is the strongest predictor of how tolerant a person is.24 Although there has
been the rare challenge to this consensus, the overwhelming evidence is that more education leads to
greater tolerance.25 Why would this be? It is generally assumed that cognitive development
accompanies formal education. More years of education can result in “greater acquaintance with the
logical implications of the broad democratic principles” as well as higher levels of related knowledge and
a greater ability to process new information.26 In two of the few analyses to directly examine this
assumption, people with more years of education were found to be more cognitively sophisticated (more
intellectually engaged and more able to reason about complex concepts) and thus more tolerant of
diverse viewpoints.27
Greater exposure to different ideas and people is another presumed outcome of more education. With
exposure to diversity often comes greater acceptance of difference, as manifested in greater levels of
tolerance.28 For example, one study demonstrates that urban dwellers and white-collar workers were
more tolerant than, respectively, rural residents and blue-collar employees, presumably because the
former are more likely than the latter groups to come into contact with people unlike them.29 More
recently, Diana Mutz explored the influence of exposure to diverse political perspectives on tolerance.
She finds that people who have political discussion partners who hold views different than their own are
more tolerant than are those who discuss politics only with like-minded individuals.30 Repeated
exposure to political difference, Mutz argues, translates into more support for the free expression rights
of political groups, even those groups that an individual dislikes.
Other characteristics that predict whether a person will be tolerant include psychological features, most
especially personality traits. People with low levels of self-esteem tend to be less tolerant, presumably
because such negative attitudes toward self interfere with the learning of social norms, including
tolerance. Further, dogmatic individuals are less likely to be tolerant, especially when compared with
those who are more open-minded. In the words of Sullivan and his colleagues, “Persons with flexible,
secure and trusting personalities are much more likely to be tolerant.”31 Emotions are also linked to
tolerance. Individuals who express anger, hatred, or fear toward groups whose views they dislike
demonstrate more intolerance toward these groups, although the role of emotional reactions is less
strong than other predictors of tolerance.32
Certain political attitudes, such as support for democratic norms, are also related to tolerance. People
who tend to support civil liberties in the abstract, such as the principle of freedom of speech, and to
support democratic procedural norms, such as majority vote, are more tolerant toward specific groups or
acts.33 Political expertise is also relevant. The more knowledgeable about and interested in politics
people are, the more tolerant they will be.34
Last but by no means least, threat perceptions are a particularly strong predictor of tolerance. Many
researchers, beginning with Stouffer, have found that the more a person believes that a particular group
is threatening, the less tolerant she is toward the group.35 She might feel that the group threatens the
American way of life and would threaten the stability of the government if the group becomes too
powerful, engages in violent tactics and thus could threaten people and property, or is personally
threatening because of the nature of the group’s political views. Regardless of the exact nature of the
threat, however, perceiving that a particular group is threatening can significantly reduce a person’s
tolerance of group members.
These predictors of tolerance can help us to understand one demographic group difference: women
tend to be less tolerant than men. To explain why, Ewa Golebiowska compared the sources of women’s
tolerance judgments with the sources of men’s.36 Compared to men, women display lower support for
democratic norms, in part because women are less likely to be politically knowledgeable and engaged.
Furthermore, women are more likely than men to feel threatened by disliked groups. Both heightened
threat perceptions and lessened support for democratic norms explain why women’s support for free
expression rights is lower than men’s.
CONTEXTUAL INFLUENCES ON TOLERANCE JUDGMENTS
You have just learned that there are many individual-level attributes and attitudes that influence people’s
tolerance for political diversity. Certain features of the context external to individuals also affect whether
someone is tolerant. Contemporary events, both national and global, are relevant, as nicely illustrated
by an examination of tolerance for communists.37 Shortly before World War II, at a time of alliance
between the Nazis in Germany and the communist-led government of the Soviet Union, tolerance for
communists was fairly low. This was probably due to Americans feeling their democracy was especially
threatened by both the Nazis and the Soviets. During the war, the Nazi-Soviet pact collapsed, the Soviet
Union became an ally of the United States, and tolerance for communists increased. Once the war
ended, tensions between the Soviet Union and the United States escalated. The public began to display
lower support for civil liberties for communists because some citizens perceived an external threat from
the Soviet Union and an internal threat from communists in the United States. As national attention, from
both politicians and the news media, to domestic communism faded in the 1960s and 1970s, tolerance
for the expression of communist beliefs increased.
One of the most systematic analyses of the influence of context on public tolerance was conducted by
George Marcus, John Sullivan, Elizabeth Theiss-Morse, and Sandra Wood, results of which were
published in their book With Malice toward Some.38 Marcus and his colleagues carried out a series of
experiments in which they manipulated two characteristics of groups involved in hypothetical civil
liberties controversies. The first was whether the group violated norms regarding appropriate political
activity. Was the group respectful and law-abiding or reckless and violent? Second, the likelihood of the
group gaining political power was also manipulated. Marcus and colleagues found that participants who
learned the group was likely to violate key norms expressed lower levels of tolerance than those who
were reassured that the group would conduct its rallies peacefully and lawfully. The probability of power
manipulation did not influence tolerance judgments. Therefore, Marcus and his coauthors demonstrate
that some features of the context surrounding a civil liberties controversy can influence tolerance. They
do not conclude, however, that tolerance is related only to this contextual information. In contrast, they
conclude that “some people tend to be tolerant while others tend to be intolerant, but contemporary
information … can elicit judgments that differ from people’s standing decisions.”39 In other words,
people who are predisposed to be tolerant can become less tolerant when presented with certain types
of information.
Support for freedom of expression is an important value in democratic societies, but it is not the only
one. This value can come in direct conflict with others, such as a desire for public order. Significantly,
the broader context can, at times, influence which of these competing values is most dominant. One
source of this influence is the news media. Recall the Nelson, Clawson, and Oxley study about media
framing of a Ku Klux Klan rally that we discuss in Chapter 3. That study demonstrated the effects of
different media frames on levels of tolerance for the KKK. To briefly summarize, Nelson et al. conducted
an experiment and showed that subjects exposed to a media frame emphasizing the disorder that might
result from a KKK rally were significantly more intolerant of the KKK than subjects exposed to a media
frame emphasizing freedom of speech for the KKK. This research demonstrates that journalists’ choices
regarding how to present the news can influence support for civil liberties. A steady stream of stories
stressing groups’ free speech rights would probably increase tolerance levels. Given the media’s
proclivity for airing stories that highlight violence and public disruption, however, a more likely outcome
is lower levels of public tolerance toward all types of groups in the political minority.
Methods to Increase Tolerance
What other contextual features can increase tolerance? Among adolescents, exposure to a tolerance
curriculum can produce greater tolerance of disliked groups. This is particularly likely if the lessons are
designed to link democratic principles and the American legal system with specific examples of speech
by unpopular groups, such as a case study of the Nazi-Skokie controversy.40 In contrast, curricula that
highlight only abstract democratic norms or that focus on the Bill of Rights (such as its history or related
court cases) tend not to change support for civil liberties.41
Being reminded of the ideals of free expression can also increase tolerance. Marcus and his colleagues,
in addition to the experiments already described, also conducted a study in which they controlled
whether the participants were exposed to information extolling the value of free expression. The
societal importance of open public debate was highlighted via statements such as “People should have
access to a variety of ideas, so that they can make up their minds based on full information.”42 Being
exposed to such a statement resulted in higher levels of tolerance. Although this conclusion was drawn
from a laboratory experiment, Marcus and his coauthors conclude that in real-life civil liberties
controversies, such as that which occurred in the 1970s when the Nazis wanted to march in Skokie,
political elites could discuss these democratic principles more often: “Framing the issues in terms of a
marketplace of ideas, stressing the positive role of dissent, and noting historic freedoms in the United
States can increase support for tolerance, even when the group involved is extremely unpopular.”43
ARE ELITES MORE TOLERANT?
Do elites express greater support for civil liberties than citizens do, as assumed by the theory of
democratic elitism? Yes, according to a number of empirical analyses of this topic, beginning with
Stouffer’s important work. In addition to his public opinion poll of adults across the nation, Stouffer
surveyed 1,500 local leaders from cities with populations between 10,000 and 150,000. For this portion
of his research, he selected political elites (such as mayors and the chairs of county Democratic and
Republican Party committees), community activists (such as presidents of women’s clubs and of Parent-
Teacher Associations), and people holding other top positions in these communities (including the
Chamber of Commerce president and the publisher of the local newspaper).44 The local leaders
responded to the same tolerance questions contained on the public opinion poll. Results of the
comparison between these elites and the mass public are clear—the leaders are much more tolerant of
political minorities.45 For example, whereas 84 percent of the leaders supported the right of a socialist to
make a speech in the community, only 58 percent of the public did. Similarly, 51 percent of elites
supported communists’ speechmaking rights, whereas only 27 percent of the public expressed such
support.
This greater support for civil liberties among elites is a very robust finding. It has been demonstrated a
number of other times, and it holds for many different types of elites, including national party activists
and members of Congress.46 Recently, surveys of elites and citizens have been conducted by Bright
Line Watch, a team of political scientists whose research monitors the state of democracy in the United
States. In 2019, they surveyed a representative sample of the public as well as two elite samples: local
political officials and people who donate to candidates running for federal office.47 The former consist of
officials serving in city, county, and town governments across the United States. The latter include the
top 1 percent of donors in terms of the total amount of money contributed in 2016 and are thus best
characterized as economic elites, albeit politically active ones. All three groups were asked whether
specific democratic principles are important or essential to democracy. Across a wide range of these
principles, elites were more likely than the public to value the principle. For instance, donors were at
least 10 percentage points more likely than citizens to indicate the democratic importance of the
following: the right to express unpopular speech, a press free from government interference, the right to
protest peacefully, and judicial independence from other branches of government. Local officials were
also more likely than the public to consider these principles important, although by a smaller margin.
Why would elites be more tolerant of minority viewpoints than citizens? One explanation focuses on
differences in demographic characteristics of leaders versus the mass public. In general, political elites
tend to possess characteristics, such as higher levels of education, that are related to higher levels of
tolerance. A nice illustration of this pattern comes from a reanalysis of Stouffer’s data.48 The community
leaders who responded to Stouffer’s poll were more educated, more male, and less likely to live in the
South than the mass public. Because men, those with more education, and those living outside of the
South are more tolerant than their counterparts, differences in these characteristics of the leader and
public samples accounted for their differing levels of support for civil liberties.
A second explanation for elite-public disparities posits that elites have higher levels of tolerance
because of their direct involvement in politics.49 Leaders experience specific adult socialization effects
through their political activity, learning to become more tolerant. This increased tolerance occurs for a
number of reasons:
Greater and more intimate contact with ideological diversity decreases authoritarianism and
increases tolerance; the necessity to compromise with individuals who disagree strongly with
oneself can lead to a more realistic and less dramatic view of the threat presented by
nonconformist groups and their ideas; the great responsibility of having actually to govern, of
seeing the consequences of one’s view enacted into policy and of shaping others’ lives can
lead to a “sober second thought” about the consequence of one’s own intolerance.50
Support for this political socialization explanation was found in an examination of attitudes of politicians
in four nations: Britain, Israel, New Zealand, and the United States. The elites in the study, all members
of their national legislature, displayed more tolerance than would be expected given their levels of
education, support for democratic norms, and personality traits such as dogmatism. In other words, the
reason why elites are more tolerant is not only because they differ from the general public in certain
characteristics that we know to be related to tolerance (such as having more education or being less
dogmatic) but also because their engagement with politics socializes them differently than other adults.
Hold on a minute, you might be thinking. Didn’t I read, just a few pages ago, that Trump displays little
tolerance toward people and groups he disagrees with? Yes, but do bear in mind that research
concluding elites are more tolerant than citizens focuses on tendencies. It is not the case that all elites
are more tolerant than all citizens; exceptions to this pattern certainly exist. Furthermore, we should
distinguish between attitudes and actions. Elites are more likely than citizens, on average, to say that
they support civil liberties, but this does not mean that all of their actions are consistent with these
views. When the Nazi party requested permission to march in Skokie, for example, it was the city’s
politicians who enacted local laws designed to prevent the march. As we will discuss later in this
chapter, after the 9/11 attacks, it was members of Congress who passed and President George W. Bush
who signed legislation allowing the federal government more latitude in collecting information on
citizens. Many of Bush’s counterterrorism policies continued under the Obama and Trump presidencies.
And even before Trump ran for office, elected elites took action on the issue of immigration in a way that
restricts civil liberties. In April 2010, for instance, Arizona passed a law that would, in the words of one
reporter, “hand the police in the state broad power under state law to check the legal status of people
they reasonably suspect are [undocumented] immigrants.”51 Specifically, the law requires police officers
to request immigration documents from people, including U.S. citizens, they have stopped if they are
suspected of being in the nation illegally. This provision was upheld by the U.S. Supreme Court in
2012.52
CIVIL LIBERTIES POST-9/11
Not surprisingly, in response to the September 11, 2001, al-Qaeda attacks on New York City and
Washington, DC, national leaders made some significant changes to counterterrorism policies
(policies designed to identify and capture suspected terrorists). Along with these changes came
concerns that the new policies were infringing on individual liberty. The ensuing, and continuing, debate
regarding the proper balance between protecting the homeland and upholding civil liberties is
summarized well by public opinion scholar Darren Davis:
Unequivocal support for democratic rights, in the context of terrorism, has been seen as
tantamount to constitutional suicide—that is, the ultimate demise of society because strict
adherence to civil liberties and rights makes it impossible to detect potential threats. And yet
allowing the government and law enforcement to usurp individual liberties may potentially
jeopardize liberty for everyone. Innocent and law-abiding citizens would be threatened not only
by terrorists, but also by their government.53
What has the public thought about balancing civil liberty and national security in the post-9/11 era?
Thanks to pollsters at the Pew Research Center, we can find out. Over nearly fifteen years, Pew asked
respondents the following question: “In order to curb terrorism in this country, do you think it will be
necessary for the average person to give up some civil liberties, or not?” The percentages of the public
thinking it necessary to give up liberties appear in Figure 9.5 (the solid line). In the late 1990s, only
around 30 percent of the public thought Americans would need to sacrifice some civil liberties to fight
terrorism. This figure shot up to 55 percent immediately after the 2001 al-Qaeda attacks. It has declined
fairly steadily since, with only 27 percent of the public holding this view in 2009.
Description
Figure 9.5 Public Opinion: Civil Liberties versus National Security
Sources: Data for “Necessary for people to give up liberties” from “Independents Take Center Stage in
Obama Era: Trends in Political Values and Core Attitudes: 1987–2009,” Pew Research Center, Washington,
DC, May 21, 2009, http://people-press.org/2009/05/21/independents-take-center-stage-in-obama-era; other
data from “15 Years after 9/11, a Sharp Partisan Divide on Ability of Terrorists to Strike U.S.,” Pew Research
http://people-press.org/2009/05/21/independents-take-center-stage-in-obama-era
Center, Washington, DC, September 7, 2016, https://www.people-press.org/2016/09/07/15-years-after-911-
a-sharp-partisan-divide-on-ability-of-terrorists-to-strike-u-s/
Between 2004 and 2016, Pew also asked respondents to assess the government’s policies to combat
terrorism using the civil liberties versus national security yardstick. Specifically, the surveys asked,
“What concerns you more about the government’s anti-terrorism policies: that they have gone too far in
restricting the average person’s civil liberties [or] That they have not gone far enough to adequately
protect the country?” Results for both of these options are presented in Figure 9.5 (dashed lines).
Between 2004 and 2009, Americans were more likely to think that antiterrorism policies were not
protecting the nation than that the policies were restricting civil liberties, with the gap between these two
narrowing somewhat.
Public attitudes have fluctuated since then, in response to national and international events. In January
2010, the gap blew wide open, with over twice as many people more concerned that policies did not
protect the nation than that they restricted liberties (58 vs. 27 percent). Why this change? On Christmas
2009, an alleged al-Qaeda member, Umar Farouk Abdulmutallab, hid explosives in his underwear and
attempted to blow up a Northwest Airlines plane flying from Amsterdam to Detroit. Abdulmutallab failed.
Yet his near success and the fact that he was able to board a plane with explosives on his body raised
concerns about the effectiveness of the nation’s counterterrorism policies. Three years later, and for the
only time since 2004, the public was more concerned with protecting civil liberties than national security.
Credit should go to Edward Snowden for this reversal. Snowden, a former contractor with the National
Security Agency, leaked documents in 2013 revealing that the NSA had been collecting telephone and
Internet data from millions of Americans. Whereas the public had been willing to support many
government actions in the fight against terrorism, many felt this NSA activity was a step too far. Between
2014 and 2016, Americans’ focus shifted to potential terrorist activity, producing a return to favoring
security over civil liberties. In particular, many citizens were worried about the growing power in the
Middle East of ISIS (the Islamic State of Iraq and Syria, also known as the Islamic State of Iraq and the
Levant or ISIL). A militant extremist group, ISIS aims to create an Islamic state. Its tactics have included
kidnappings, beheadings, and mass executions.
Taken together, the results in Figure 9.5 present a portrait of citizens who generally, although not
always, think current terrorism policies do not unduly restrict liberties. Furthermore, the changes over
time in all three lines make sense, given what we know about the role of threat in civil liberties
judgments. Americans tend to be more willing to sacrifice civil liberties when they feel threatened than
when they feel secure.54 This was the case, for example, after Timothy McVeigh bombed the Oklahoma
City federal building in 1995, purportedly in retaliation for federal authorities raiding the compound of a
religious group, the Branch Davidians, near Waco, Texas. The Oklahoma City bombing resulted in the
deaths of 168 people, making it the deadliest domestic terrorist act in the United States. Shortly after the
Oklahoma City bombing, nearly one-half of the public felt that giving up some civil liberties would be
necessary to fight terrorism. Two years later, however, during which time no other acts of terrorism
occurred on U.S. soil, only 29 percent of the public was willing to sacrifice these liberties.55 We see a
similar trend in the years since the 9/11 attacks. When terrorism is in the news, as was the case with the
Christmas 2009 failed airliner bombing, the balance between prioritizing security or liberty tips more
toward security.
Which Americans are most willing to trade civil liberties for protecting the nation from terrorism? Not
surprisingly, people who feel most threatened by and anxious about terrorism. As Bethany Albertson
and Shana Kushner Gadarian explain in their book Anxious Politics, threatening events such as terrorist
attacks can produce anxiety. One way individuals try to manage this anxiety is by supporting policies
they think will make their environment safer. Specifically, Albertson and Gadarian demonstrate that
people who have high levels of anxiety after watching a news story about terrorism are more willing than
those with low levels to give up some civil liberties protections in order to curb terrorism.56
In probing post-9/11 attitudes toward civil liberties, however, Darren Davis and Brian Silver reveal that
threat perception is only one piece of the puzzle.57 Trust in government, especially law enforcement, is
also relevant. Considering only those people who place a lot of trust in the government, support for civil
https://www.people-press.org/2016/09/07/15-years-after-911-a-sharp-partisan-divide-on-ability-of-terrorists-to-strike-u-s/
liberties is lowest for citizens who are very concerned that another terrorist attack might occur. However,
people who display little trust in government demonstrate stronger support for civil liberties whether they
fear another terrorist attack or not. In other words, for people who are distrustful of government, this
distrust overrides their concern about future terrorist activity, and they are less willing to trade civil
liberties for security. Finally, Davis and Silver conclude that personal characteristics, especially political
ideology and race, also influenced how citizens responded to the 9/11 attacks. Liberals were
consistently less willing to support restrictions on civil liberties than were conservatives, and African
Americans expressed stronger pro–civil liberties positions than either whites or Latinos. Davis and Silver
attribute this latter finding to African Americans’ “struggle for civil rights and … distrust of government.
As a result, African Americans may be reluctant to concede rights that they have worked hard to achieve
or to empower a government in which they have little confidence, even for the sake of personal
security.”58
Public Opinion’s “Dark Side”
A variety of specific counterterrorism policies were implemented by the national government in the
weeks and years after the 9/11 attacks. One was the USA Patriot Act (Uniting and Strengthening
America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism). This legislation,
passed in October 2001, made it easier for federal authorities to issue telephone wiretap orders,
conduct searches, and obtain the records (financial, library, Internet usage, and so on) of individuals.
Since 2001, the federal government has been classifying some suspected terrorists as enemy
combatants. Enemy combatants, according to federal policy, can be held without being charged with a
specific crime and can be tried by a military commission without some of the civil liberties afforded to
citizens who are accused of committing crimes. Many of these enemy combatants are still being held at
a U.S. military prison in Guantanamo Bay, Cuba. Furthermore, people with the same racial or religious
characteristics as the 9/11 hijackers have been targeted by airport security screeners and law
enforcement officials. Such racial profiling, which involves selecting someone for questioning or extra
scrutiny on the basis of race or ethnicity, had historically been used most commonly in the United States
to target African Americans. After the 9/11 attacks, however, profiling is also used in an attempt to
identify suspected terrorists.
Does the public support these and other counterterrorism policies? Answering this question is the
primary goal of Clem Brooks and Jeff Manza’s book Whose Rights?59 Indeed, their research is one of
the most comprehensive analyses of public opinion, civil liberties, and counterterrorism policy in the
post-9/11 era.60 Overall, Brooks and Manza conclude that the “attitudes and beliefs of Americans have
a dark side, a willingness to suppress otherwise strong support for civil rights and liberties in the name
of national crisis and perceived threats.”61 Their evidence comes from nationally representative surveys
conducted in 2007, 2009, and 2010. In these surveys, they assessed public support for a wide range of
counterterrorism policies (see Figure 9.6). The bars in this figure represent the average level of support
for each policy, measured on a scale from zero to one. A score of zero represents strong opposition to
the policy, whereas one equals strong support. Results greater than .5 thus demonstrate that the public
is, on average, more approving than disapproving of the policy.
A few aspects of these results are noteworthy. First, Americans are not equally supportive of all policies.
The public is most supportive of the NSA surveillance policy (as of 2009). The lowest support was
registered for two policies: the use of waterboarding on suspected terrorists and torture for the purpose
of gathering information to prevent a future attack. Second, the average level of support was less than
.5 for only those two counterterrorism policies. Third, these public attitudes were assessed not in the
immediate aftermath of the 9/11 attacks but rather nearly ten years later. The high level of support for
most of these policies coupled with the time delay since the 2001 attacks were two factors contributing
to Brooks and Manza’s categorization of these attitudes as representing a dark side of the public.
Furthermore, they speculate that “support for coercive policies may become relatively enduring,
persisting beyond the initial context in which political leaders offered their original justifications.”62
Pushing beyond overall support for counterterrorism policies, Brooks and Manza explore whether public
attitudes vary depending on the identity of the groups targeted by the policies. To do so, they embedded
experiments within their surveys. Survey-based experiments enable researchers to manipulate the
wording of questions on a survey administered to a nationally representative sample of Americans. In
Brooks and Manza’s case, they altered who was targeted by a specific policy. One version of a question
assessing attitudes toward the use of military commissions specified that the policy would apply to
terrorism suspects who are American citizens. Another version asked about foreign nationals while a
third specified American citizens of Middle Eastern background. For the counterterrorism policies of
NSA surveillance and torture, only two groups were examined (foreign nationals were excluded).
Because these policy questions varied only by the identified group and because question versions were
randomly assigned to survey respondents, Brooks and Manza are able to assess whether public
attitudes toward the policy depend on which groups of people are being targeted.
Description
Figure 9.6 Public Support for Counterterrorism Policies, 2009 and 2010
Source: Data from Clem Brooks and Jeff Manza, Whose Rights? Counterterrorism and the Dark Side of
Public Opinion (New York: Russell Sage Foundation, 2013), Chap. 3.
Note: Data for NSA surveillance and Military commissions from 2009; all others from 2010.
Some of the results from Brooks and Manza’s survey experiments appear in Figure 9.7. The public is
more supportive of using military prisons and trials for suspected terrorists who are foreign nationals
versus those who are American citizens. For this policy, support was not significantly different for
American citizens of Middle Eastern background compared to all American citizens. This latter
distinction did matter when it came to NSA surveillance. Average public support was twice as high when
the American citizens being monitored by the government have a Middle Eastern background. In
contrast, the identity of Americans who would be subject to torture in pursuit of preventing future attacks
did not influence public support for this policy. When it comes to torturing American citizens, the public’s
opinions are not very malleable. In other words, people seem to either support torture or oppose torture
no matter the identity of the suspect.
Brooks and Manza’s broad conclusion from across their set of experiments is that the public tends to be
more supportive of policies that are more obviously directed toward foreigners or toward those with the
same ethnic characteristics as the 9/11 hijackers.63 Other scholars have reached similar conclusions,
such as that the public is more supportive of torture as a counterterrorism policy or less tolerant of a
speech criticizing U.S. foreign policy when the alleged suspect/speaker has an Arabic versus a white
American name.64 Given the history of citizens displaying lower levels of tolerance for the civil liberties
of unpopular groups, these results are not unexpected. Finally, you might have noticed that support for
NSA surveillance is much lower in Figure 9.7 than in Figure 9.6. This is likely explained by differences in
question wording. The results presented in Figure 9.6 demonstrate support for monitoring telephone
conversations between U.S. citizens and suspected terrorists who live abroad. The survey experiment
instead queried support for monitoring “telephone conversations, banking transactions, and email
between American citizens [or American citizens of Middle Eastern background] in the United States.”65
Not only is the sweep of government monitoring broader in the second version, but only U.S. citizens
would be targeted. The different results from these two questions likely capture the public’s greater
willingness to support counterterrorism policies that target foreign nationals.66
Description
Figure 9.7 Counterterrorism Policy Opinions Vary by Identity of Target
Source: Adapted from Clem Brooks and Jeff Manza, Whose Rights? Counterterrorism and the Dark Side of
Public Opinion (New York: Russell Sage Foundation, 2013), 103, 121, 137.
Note: Bars represent the average support for each item.
CONCLUSION
Are Americans tolerant of nonconformity? Is the public more or less tolerant than elites? What are the
sources of tolerance and intolerance? Does public support for civil liberties change as political times do?
These questions motivated Samuel Stouffer to write Communism, Conformity, and Civil Liberties nearly
seventy years ago. Although the answers have changed somewhat and certainly have become more
complex since Stouffer’s book was published, these questions continue to be the focus of most political
tolerance research today. Stouffer was also worried about the implications of public intolerance for the
nation and therefore proposed solutions for increasing levels of tolerance. Discussing whether the level
of tolerance is high or low enough in a polity is, of course, at the heart of democratic theorists’ writings
on civil liberties. It is to democratic theory that we now turn, particularly evaluating the assumptions of
both elite and participatory theories of democracy in light of conclusions drawn from political science
studies of tolerance.
Elite Democracy
Elite democrats expect the public to be intolerant toward those whose views they reject, especially
because of the public’s self-interested proclivities and disengagement from political forums where
opposing views are aired. Thus, these theorists would not be terribly surprised by Stouffer’s findings of
widespread public intolerance, similar conclusions drawn by scholars since then, and post-9/11 public
support for restricting the civil liberties of those who are thought to be sympathetic to or engaged in
terrorist activities. Moreover, democratic elitists argue that the leaders of a nation will display more
support for civil liberties than the public, a point that is also supported by the empirical political science
research. Recall that these theorists went one step further, arguing that the leaders’ support for civil
liberties would ensure that these liberties are protected, even in the face of much public intolerance. Yet,
as we discussed earlier in the chapter, even though elites do profess more attitudinal support for civil
liberties than the citizens do, the actions of elected officials do not always coincide with their expressed
support for civil liberties in the abstract. These examples remind us that elite democratic theory has
perhaps overstated its case that elites will ensure the maintenance of civil liberties in a democracy.
Participatory Democracy
Beyond their faith in elites, there is another reason why elite democrats do not worry very much about
high levels of public intolerance. Because, according to these theorists, citizens are generally too
apathetic to engage in political activity, the likelihood that the public will engage in any behavior to
actually limit the civil liberties of others is low. “Apathy … furnishes its own partial corrective by keeping
the [intolerant] from acting upon their differences. In the United States, at least, their disagreements are
passive rather than active, more the result of political ignorance and indifference than of intellectual
conviction.”67 Such a viewpoint maddens theorists of participatory democracy. A nation’s continued
commitment to support minority rights and other civil liberties should not be predicated on the apathy of
a segment of the citizenry, argue participatory democrats. Among the key democratic goals for these
theorists is the full involvement of the citizenry, so they find any argument that applauds the political
uninvolvement of the public worrisome.
In contrast, participatory democrats want the public to be more involved in politics, believing that people
can learn to become more tolerant through political activity. Thus, participatory democrats are quite
pleased with the empirical evidence that formal education leads to more tolerant attitudes and the
conclusion that some political leaders are more tolerant than the public because of the leaders’
exposure to political activity, especially bargaining and compromising with their opponents. These two
findings are promising for increasing citizen tolerance in that they suggest routes to pursue this goal.
Further, the fact that specific types of high school curricula can increase tolerance is especially relevant
to participatory democrats’ arguments and demonstrates the role that the educational system could play:
“If civic education were to include a systematic examination of the role of dissent in a democratic
society, young people might develop a commitment to protect civil liberties that would ultimately
engender a more fully democratic citizenry.”68
As should be clear, most participatory democrats assume that high levels of public tolerance are ideal
for a democracy. It is worth considering, however, whether supporting the free expression rights of some
groups might actually undermine tolerance. Take groups that advocate white supremacy and hatred of
Blacks, such as the KKK or the Proud Boys. If these white supremacists speak out in favor of curtailing
the political rights of Blacks, and if this speech then reduces the participation of Blacks or removes the
viewpoints of Blacks from the marketplace of ideas, is it tolerant to support the expression of their
views? Or take an issue that has been on the agenda in a few states recently: religious freedom. On
their face, the Free Exercise Clause of the U.S. Constitution or religious freedom laws seem to protect
the exercise of religion by preventing the government from interfering with this right. Yet some versions
of state religious freedom laws have come under fire because they appear to legalize discrimination
against groups toward which a person or business might have religious objections.69 If such a law
exists, opponents argue, what’s to stop owners of a reception hall or catering business from refusing to
provide wedding services to same-sex couples in the name of religious freedom? Indeed, such a
situation resulted in a 2018 U.S. Supreme Court decision.70 In that case, the Colorado Civil Rights
Commission had found that a cake maker violated the state’s antidiscrimination law when he refused, on
religious grounds, to bake a cake for the wedding reception of two gay men. The Supreme Court
overturned the Commission’s decision, arguing that they had infringed upon the cake maker’s freedom
to exercise his religious beliefs.
These debates suggest that tolerance might not require supporting the political expression of all
viewpoints in a democracy. Perhaps some viewpoints simply do not deserve constitutional protection. In
other words, are “the citizens of a democracy obliged to tolerate those who, if they prevailed, would
destroy the practice of tolerance? If tolerance is among the highest values in democratic regimes, does
it make sense to tolerate those who threaten this very principle?”71 To some people, however,
particularly those for whom the freedom of expression of all groups is an absolute necessity, raising
such questions is downright intolerant. This debate will exist as long as differing understandings of
tolerance and intolerance exist in society.
KEY CONCEPTS
atheists 280
Bill of Rights 275
civil liberties 273
communists 281
constitutional framework 275
context 291
counterterrorism policies 295
emotions 290
enemy combatants 298
exposure to diversity 289
formal education 289
homosexuals 283
least liked group 284
marketplace of ideas 274
militarists 283
Muslim extremists 284
personality traits 289
pluralistic intolerance 288
political expertise 290
political times 287
political tolerance 273
racial profiling 299
racists 283
socialists 280
support for democratic norms 290
survey-based experiments 299
threat perceptions 290
tolerance curriculum 292
USA Patriot Act 298
value of free expression 292
SUGGESTED SOURCES FOR FURTHER READING
Albertson, Bethany, and Shana Kushner Gadarian. Anxious Politics: Democratic Citizenship in a
Threatening World. New York: Cambridge University Press, 2015.
Davis, Darren W. Negative Liberty: Public Opinion and the Terrorist Attacks on America. New York:
Russell Sage Foundation, 2007.
Merolla, Jennifer L., and Elizabeth J. Zechmeister. Democracy at Risk: How Terrorist Threats Affect the
Public. Chicago: University of Chicago Press, 2009.
Negative Liberty and Democracy at Risk provide detailed examinations of public support for civil liberties
and counterterrorism policies in the post-9/11 era. Anxious Politics explores the influence of anxiety on
public opinion across a variety of policy matters, including counterterrorism.
Barnum, David G. “Decision Making in a Constitutional Democracy: Policy Formation in the Skokie Free
Speech Controversy.” Journal of Politics 44 (1982): 480–508.
Gibson, James L., and Richard D. Bingham. “Skokie, Nazis, and the Elitist Theory of Democracy.”
Western Political Quarterly 37 (1984): 32–47.
These two articles discuss the Nazi-Skokie free speech controversy and its implications for the theory of
democratic elitism.
Claassen, Christopher, and James L. Gibson. “Does Intolerance Dampen Dissent? Macro-Tolerance
and Protest in American Metropolitan Areas.” Political Behavior 41 (2019): 165–185.
This article focuses on how tolerance influences the political context by demonstrating that areas with
higher levels of public intolerance witness fewer political protests.
Golebiowska, Ewa A. The Many Faces of Tolerance: Attitudes Toward Diversity in Poland. New York:
Routledge, 2014.
Sniderman, Paul M., Michael Bang Petersen, Rune Slothuus, and Rune Stubager. Paradoxes of Liberal
Democracy: Islam, Western Europe, and the Danish Cartoon Crisis. Princeton, NJ: Princeton University
Press, 2014.
Political tolerance in Europe is the focus of these two books. Golebiowska’s is a broad assessment of
tolerance in Poland, a nation known for public intolerance. Sniderman and co-authors examine a nation
(Denmark) in the midst of a freedom of expression controversy.
American Civil Liberties Union, www.aclu.org
One of the nation’s leading organizations supporting civil liberties, the ACLU aims, in a variety of ways,
to ensure the individual protections provided for in the U.S. Constitution are upheld. Its website details
the organization’s activities and presents a wealth of information about many civil liberties issues, such
as free speech (including on college campuses), religious liberty, mass incarceration, the death penalty,
privacy and surveillance, and immigrants’ rights.
Teaching Tolerance, www.tolerance.org
A program of the Southern Poverty Law Center, this website provides many free resources, including
teaching tools, for people who wish to fight bias and promote respect for diversity in their communities.
Bright Line Watch, https://brightlinewatch.org/
http://www.tolerance.org/
One of the greatest threats to democracy is the idea that it is unassailable.
“Trust, Facts and Democracy,” Pew Research Center, https://www.pewresearch.org/topics/trust-facts-
and-democracy/
In this chapter, we present results from both of these ongoing explorations of public support for
democracy, but we did not have enough room to discuss all their interesting results. Check out these
websites if you want to learn more.
https://www.pewresearch.org/topics/trust-facts-and-democracy/
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists percentage of importance and ranges from 0 to 100 in increments of 20. The
vertical axis lists various democratic principles. Approximate data from the graph are tabulated as
follows:
Democratic principles
Percentage consider
them to be very
important
Percentage who consider them to
be very or somewhat important
Rights and freedoms of all
people are respected
85 95
Judges are not influenced by
political parties
82 95
News organizations are
independent of government
77 92
People are free to peacefully
protest
77 95
Views of those not in the
majority on issues are
respected
62 95
Back to image
Data from the graph are tabulated as follows:
Country Percentage in supportof military rule
Percentage in support of rule
by a strong leader
Percentage in support of
rule by experts
Hungary 14 24 68
South
Korea
8 23 52
Poland 9 15 50
Spain 11 13 49
Country Percentage in supportof military rule
Percentage in support of rule
by a strong leader
Percentage in support of
rule by experts
Japan
15 31 49
Israel 10 25 48
Chile 14 27 47
France 12 17 44
Germany 4 6 44
UK 15 26 42
Canada 10 17 41
Australia 12 19 41
U.S. 17 22 40
Italy 17 29 40
Sweden 4 9 40
Netherlands 8 10 39
Greece 8 12 31
Back to Figure
The horizontal axis lists communist, socialist, and atheist. The vertical axis lists percentage and ranges
from 0 to 100 in increments of 20. Approximate data from the graph are tabulated as follows:
Political
minority
Percentage allowing
public speech
Percentage allowing book
in library
Percentage allowing
teaching
Political
minority
Percentage allowing
public speech
Percentage allowing book
in library
Percentage allowing
teaching
Communist 27 27 7
Socialist 55 52 35
Atheist 38 36 12
Back to Figure
The horizontal axis lists years and ranges from 1954 to 2018. The vertical axis lists percentage of
tolerance and ranges from 0 to 100 in increments of 20. Approximate data from the graph are tabulated
as follows:
Year
Percentage
tolerant to
communist
Percentage
tolerant to
atheist
Percentage
tolerant to
socialist
Percentage
tolerant to
racist
Percentage
tolerant to
militarist
Percentage
tolerant to
homosexual
man
Percentage
tolerant to
Muslim
extremist
1954 25 38 58 No data No data No data No data
1972 55 65 82 63 55 55 No data
1980 55 64 No data 61 56 68 No data
1990 65 74 No data 63 58 75 No data
2000 65 74 No data 59 62 80 No data
2010 66 74 No data 57 69 85 40
2018 65 78 No data 56 71 90 45
Back to Figure
The horizontal axis lists allow public speech and allow public rally. The vertical axis lists percentage and
ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated as follows:
Activity
Least liked
groups,
1978
Least liked
groups,
2005
Communists, speeches in
1978 and rallies in 2005
Atheists, speeches in
1978 and rallies in 2005Activity
Least liked
groups,
1978
Least liked
groups,
2005
Communists, speeches in
1978 and rallies in 2005
Atheists, speeches in
1978 and rallies in 2005
Allow
public
speech
50 35 63 65
Allow
public
rally
35 30 40 50
Back to Figure
The horizontal axis lists years and ranges from 1996 to 2016. The vertical axis lists percentage of
agreements and ranges from 0 to 80 in increments of 10. Approximate data from the graph are
tabulated as follows:
Year
Necessary for
people to give up
liberties
Government policies have not
gone far enough to protect
country
Government policies have
gone too far in restricting
liberties
1996 30 No data No data
1997 29 No data No data
Mid
September
2001
55 No data No data
2002 48 No data No data
2003 43 No data No data
2004 38 50 29
2005 40 52 31
2006 43 55 26
2007 40 No data 35
Year
Necessary for
people to give up
liberties
Government policies have not
gone far enough to protect
country
Government policies have
gone too far in restricting
liberties
2009
25 43 35
2010 No data 58 26
2013 No data 37 46
2014 No data 50 35
2015 No data 49 36
2016 No data 49 32
Back to Figure
The horizontal axis lists support for policy and ranges from 0 to 1 in increments of 0.1. The vertical axis
lists different policies. Approximate data from the graph are tabulated as follows:
Policy
Support
for
policy
NSA surveillance: Government should monitor telephone conversations between
American citizens and suspected terrorists living in other countries.
0.75
Airport security: Muslims, including U.S. citizens, should undergo more intensive security
checks before boarding airplanes in the United States.
0.68
Patriot Act: Support the Patriot Act, which makes it easier for the federal government to
access phone and email records.
0.67
Military commissions: Support the creation of a separate set of courts and prisons in
which individuals classified as enemy combatants can be held indefinitely.
0.65
Policy
Support
for
policy
Assassination: Support government targeting individuals suspected of being al-Qaeda or
Taliban leaders for assassination.
0.6
Detentions: Support detaining someone who is not a U.S. citizen indefinitely if that
person is suspected of belonging to a radical Muslim organization
0.55
Ethnic profiling: Law enforcement should be able to bring in for questioning people of
certain ethnic backgrounds if these groups are thought to be more likely to engage in
terrorism.
0.55
Waterboarding: Support the use of waterboarding on terrorist suspects in an effort to gain
information.
0.47
Torture: Government authorities should have the right to torture an American if they think
it will help prevent a terrorist attack
0.34
Back to Figure
The horizontal axis lists military commissions, NSA surveillance, and torture. The vertical axis lists
support for policy and ranges from 0 to 1 in increments of 0.1. Approximate data from the graph are
tabulated as follows:
Policy Support by foreignnationals
Support by
Americans
Support by Americans of Middle
East background
Military
commissions
0.79 0.6 0.64
NSA
surveillance
No data 0.16 0.32
Torture No data 0.38 0.35
CHAPTER 10 SUPPORT FOR CIVIL RIGHTS
THROUGHOUT 2019, Elizabeth Warren shared “pinky promises” with young girls, telling them that she
was running for president because that is “what girls do.”1 During her campaign stops, Kamala Harris
used female pronouns when mentioning the president, such as in this sentence: “The commander in
chief of the United States of America has as one of her most important responsibilities the responsibility
of keeping our nation secure.”2 Her goal with this language choice? To help citizens envision a female
president. Warren and Harris were two of the record-number six women running at the same time for a
major-party presidential nomination. Indeed, the field of candidates attempting to be the Democratic
Party’s 2020 nominee was the most diverse ever vying for the U.S. presidency. The field contained an
African American man, an American Samoan woman, an Asian American man, a biracial (African
American and Indian American) woman, a Latino, and a gay man (respectively, Cory Booker, Tulsi
Gabbard, Andrew Yang, Kamala Harris, Julián Castro, and Pete Buttigieg). This diversity is notable
given that nearly all U.S. presidents have been straight white men. Yet the United States has
experienced some historic firsts in recent years: in 2008 when Barack Obama became the first African
American president, in 2012 when Mitt Romney became the first member of The Church of Jesus Christ
of Latter-day Saints (Mormon) to win a major-party nomination, and in 2016 when Hillary Clinton
became the first woman to win a major-party nomination. Will 2020 be another first? What does the
public think of this turn of events? How supportive are citizens of candidates who break the mold?
One marker of a democratic society is the openness of its political system to citizens from many different
walks of life. If elected officials only come from a narrow stratum of society, it raises questions about
how representative the institutions of government truly are. In the United States, the institutions of
government do not accurately reflect the diversity of the citizenry. Take gender, for example. As of 2019,
the U.S. Congress is 24 percent female.3 Obviously that is low given the proportion of women in society.
It is also low compared with many other countries around the world. The United States ranks 78th out of
190 countries in terms of female representation in its legislature.4 And of course the presidency is the
ultimate bastion of male dominance. The closest women have made it to the presidency was in 2016
when Hillary Clinton lost the Electoral College vote to Donald Trump, although she did win more popular
votes than he did. Furthermore, the nation has never had a female vice president. Only two women
have run for that office on a major-party ticket, Representative Geraldine Ferraro (D-NY) in 1984 and
Alaska governor Sarah Palin in 2008. Certainly there are no laws on the books that stop women or
minorities from seeking political office, so why is there this inequality in officeholding? Are citizens’
attitudes toward women and minorities partially to blame for the lack of diversity in our institutions?
In this chapter, we discuss how supportive citizens are of political rights for women and minorities,
specifically in terms of holding the office of president. If even a small segment of society is biased
against women and minorities, it can have huge electoral implications. Over the past two decades,
presidential elections have been decided by differences in the popular vote of fewer than 10 percentage
points. The closest recent election was in 2000, when the difference between Al Gore and George W.
Bush was 0.5 percentage point. In 2008, the difference between John McCain and Barack Obama was
7.3 percentage points, the largest gap in the past 20 years. The difference between Hillary Clinton and
Donald Trump was 2.1 percentage points.5 Given such an evenly split electorate, hesitation on the part
of citizens to vote for someone due to a particular characteristic could spell disaster for that candidate.
Political rights are a specific type of civil right we discuss in this chapter, but we also examine support for
other civil rights. “Civil rights are government guarantees of equality for people in the United States
with regard to judicial proceedings, the exercise of political rights, treatment by public officials, and
access to and enjoyment of the benefits of government programs.”6 In particular, we examine public
opinion concerning civil rights policies geared toward four groups: African Americans, gays and
lesbians, transgender individuals, and people with disabilities.
Controversies over civil rights concern participatory democratic theorists a great deal. Participatory
democratic theorists worry about equality in the political sphere and, thus, are concerned when the path
to leadership is not open to all citizens. To these theorists, opposition to civil rights reflects the fact that
many citizens are not actively involved in politics. If all citizens participated in political decision making,
then women and minorities would have more power within the political system and thus would be able to
successfully push for civil rights. Furthermore, participatory democratic theorists believe that
participation in and of itself produces better citizens. That is, political interactions with many different
types of people allow citizens to see beyond their narrow worlds and to gain a better understanding of
what’s good for other citizens and for the nation as a whole. As a result, citizens would probably
demonstrate greater support for civil rights.
In contrast, elite democratic theorists are not particularly concerned with equality in society. Indeed,
some elite democratic theorists expect that certain inequities will exist in the political system. Thus, they
would not be surprised that our institutions of government do not reflect the diversity of society. Other
elite democratic theorists are not concerned about opposition to civil rights among the public because
these theorists emphasize that elites are the key political players. Because the elites appreciate and
support the rights of minorities, the public’s views on the subject don’t really matter.
PUBLIC OPINION AND PRESIDENTIAL CANDIDATES
There are very few formal qualifications to run for president of the United States. A person must be a
natural-born citizen, at least thirty-five years of age, and a resident of the United States for fourteen
years to be eligible for the presidency. That’s it. In practice, however, the story is much different. The
youngest president ever elected was John F. Kennedy at forty-three years old, and the oldest elected to
a first term was Donald Trump at seventy. Ronald Reagan was seventy-three when elected to his
second presidential term. The majority of presidents have been between the ages of fifty and fifty-nine
when elected. Only one president has been Catholic, and two have been Quakers. Just one has been
African American, and no woman has ever been elected to the highest office in the land. What explains
why presidents have come from such a limited pool of citizens? We turn to some fascinating public
opinion data to help answer this question.
Religion and Presidential Candidates
For several decades, the Gallup organization has asked citizens whether they would vote for particular
types of candidates if nominated by their party.7 Let’s first examine support for candidates of different
religious faiths (see Figure 10.1). Support for Baptist presidential candidates has been high since the
question was first asked in 1958.8 In contrast, support for a Catholic was modest in 1937 but has
climbed steadily since then. Support for a Jewish candidate was under 50 percent in 1937 but has also
increased over time. In 2019, over 90 percent of citizens said they would vote for a well-qualified
Catholic or Jew if nominated by their party.
Support for a Mormon president has stayed remarkably steady over time. The question was first asked
in 1967 when George Romney (Mitt Romney’s father) ran for the Republican presidential nomination. At
that time, 75 percent of Americans said they would vote for a Mormon nominated by their party. Most
recently, Gallup has assessed public support for a Mormon candidate only in the years when Mitt
Romney was running for president (2007 and 2011) or considering a candidacy (2015). Eighty-one
percent of the public indicated they would vote for a Mormon in 2015, which is only a few percentage
points higher than public support was in 1967. Thus, there remains significant opposition to a Mormon
becoming president of the United States.
Description
Figure 10.1 Support for Presidential Candidates, 1937–2019: Religion
Source: Data from “The Presidency,” Gallup, https://news.gallup.com/poll/4729/presidency.aspx
Note: Gallup’s question wording has varied over time. The question wording from the 2019 Gallup poll:
“Between now and the 2020 political conventions, there will be discussion about the qualifications of
presidential candidates—their education, age, religion, race, and so on. If your party nominated a generally
well-qualified person for president who happened to be a __________, would you vote for that person?” The
2011 data for Muslim are from 2012.
A Mormon’s chances look good, however, compared with an atheist’s chances. In the late 1950s,
support for an atheist presidential candidate was very weak at 18 percent, and although it has tripled
since then, it is still fairly low. In April 2019, only 60 percent of Americans said they would vote for a well-
qualified atheist if nominated by their political party. Gallup only recently started measuring attitudes
toward Muslim and evangelical Christian candidates, so unfortunately we don’t know how those
attitudes have changed over time. Only 66 percent of Americans said they would support a well-
qualified Muslim nominated by their party in 2019, and 80 percent indicated they would vote for an
evangelical Christian. Perhaps not surprisingly, there are large partisan differences in attitudes toward
atheist, Muslim, and evangelical Christian candidates with Republicans much more supportive of
evangelical Christian candidates than Democrats, and Democrats much more supportive of atheists and
Muslims than Republicans.9
https://news.gallup.com/poll/4729/presidency.aspx
Another way to examine support for candidates with various religious affiliations is to ask citizens
whether they would be more or less likely to support a candidate with particular characteristics. The Pew
Research Center conducts polls doing just that (see Table 10.1). In 2016, Pew asked about support for
evangelical Christian, Mormon, Muslim, and atheist presidential candidates. For 55 percent of people,
whether a candidate is an evangelical Christian wouldn’t matter; when it does matter, being a Christian
is slightly more likely to help the person than hurt them. Sixty-nine percent of people said it wouldn’t
matter if a candidate were Mormon, but when it does matter, it works strongly against the candidate.
Twenty-three percent said they would be less likely to support a Mormon, whereas only 5 percent said
they would be more likely to support a Mormon. Just over one-half of Americans indicated that a
candidate being Muslim wouldn’t matter to them, whereas 42 percent said they would be less likely to
support such a candidate. Still, being a Muslim puts a candidate at an advantage over an atheist. Fifty-
one percent of Americans said they would be less likely to support a candidate who does not believe in
God. Given these current levels of public opposition, it seems unlikely that the United States will have
either a Muslim or an atheist president any time soon.
Table 10.1 Religion and Likelihood to Vote for a Presidential Candidate, 2016
“Regardless of the specific candidates who are running for president, we’d like to know how you
generally feel about some different traits. Would you be more likely or less likely to support a
candidate for president who is ___________, or wouldn’t this matter to you?”
Response Evangelical Christian Mormon Muslim Does not believe in God
More likely to support 22% 5% 3% 6%
Less likely to support 20 23 42 51
Wouldn’t matter 55 69 53 41
Don’t know/refused 2 3 2 1
Source: Data from “Faith and the 2016 Campaign,” Pew Research Center, January 27, 2016,
https://www.pewforum.org/2016/01/27/faith-and-the-2016-campaign/
https://www.pewforum.org/2016/01/27/faith-and-the-2016-campaign/
Race, Ethnicity, and Presidential Candidates
What about citizens’ attitudes toward presidential candidates who are racial or ethnic minorities? In
1958, Gallup first asked citizens whether they would support a well-qualified Black candidate nominated
by their party for the presidency. Less than 50 percent of Americans said they would do so (see Figure
10.2). This percentage increased gradually but steadily over time with 79 percent of Americans saying
they would support a Black candidate in 1987. The question was not asked again until 1997, at which
time support had jumped to 93 percent. Since then, support has remained in the low- to mid-90 percent
range. In recent years, the Gallup poll has also asked citizens whether they would vote for a well-
qualified Hispanic for president if nominated by their party. In 2007, 87 percent said they would do so.
By 2019, this percentage had increased to 95 percent.
In the lead-up to voting in the 2020 presidential contest, Pew did not repeat its national survey asking
Americans whether they would be more or less likely to vote for candidates who possess certain traits.
However, in the spring of 2019, they did poll Democratic citizens, asking whether they would be more
enthusiastic, less enthusiastic, or wouldn’t it matter if the Democratic presidential nominee were Black,
Hispanic, or Asian.10 For each category, approximately three-quarters of Democrats said the nominees’
race or ethnicity would not make a difference in their enthusiasm level. Of those who indicated their
enthusiasm level would be affected, Democrats were likely to be more enthusiastic than less
enthusiastic by margins of about 5-to-1 for a Black candidate (21 percent more vs. 4 percent less), 3.5-
to-1 for a Hispanic candidate (21 percent to 6 percent), but less than 2-to-1 for an Asian candidate (16
percent to 9 percent).
Description
Figure 10.2 Support for Presidential Candidates, 1937–2019: Gender, Race,
Ethnicity, and Sexual Orientation
Source: Data from “The Presidency,” Gallup, https://news.gallup.com/poll/4729/presidency.aspx
Note: See note for Figure 10.1.
Some political observers, including the authors of this book, were skeptical that an African American
could win the presidency prior to Barack Obama’s 2008 victory even with opposition to a Black
candidate in the single digits. Why the skepticism? First, these survey questions asking respondents to
directly report their attitudes toward a Black candidate may underestimate opposition because social
desirability pressures may lead some people not to answer them honestly. Many citizens recognize
that it is socially unacceptable to say they wouldn’t vote for a candidate just because of race. Thus,
prejudiced citizens, unwilling to reveal their true preferences to a survey interviewer, might simply refuse
to answer the question or say they would vote for a Black candidate even when they have no intention
of doing so in the privacy of the voting booth. Survey questions such as these measure explicit
prejudice, which is “consciously endorsed negative attitudes based on group membership.”11 Second,
https://news.gallup.com/poll/4729/presidency.aspx
other measures of explicit prejudice find significantly higher levels of bias against Blacks. In 2008,
survey questions measuring racial stereotypes showed that more than one-third of white citizens rated
Blacks as less intelligent than they rated whites, and around 50 percent rated Blacks as lazier than they
rated whites.12 If white citizens were to apply these negative stereotypes to Barack Obama, his chances
for winning the presidency would be hurt. Third, psychological research on implicit attitudes shows that
many white citizens hold unconscious prejudice toward Blacks.13 Specifically, implicit prejudice “refers
to associations that come to mind unintentionally, whose influence on thought and action may not be
consciously recognized and can be difficult to control.”14 Therefore, white citizens may honestly say
they would vote for a Black candidate when asked in a survey yet have unrecognized negative reactions
to Blacks that would lead them to oppose such a candidate in reality. In other words, automatic or
unintentional biases can shape citizens’ political judgments without them even knowing it’s happening.
Important research conducted since President Obama’s victory speaks to these concerns about public
opinion and support for a Black candidate. As Barack Obama led in the polls in the months and weeks
before the 2008 election, many political junkies debated whether or not social desirability was inflating
levels of support for Obama. In particular, Obama supporters worried that there would be a gap between
where Obama stood in the polls leading up to the election and the actual vote on Election Day. This
phenomenon is known as the “Bradley effect,” because Tom Bradley, a Black candidate for California
governor in 1982, led comfortably in preelection polls but ended up losing the election to a white
opponent.15 An analysis of polling and election outcomes showed that Black candidates experienced
the Bradley effect until the mid-1990s; however, in more recent elections, including the 2008 contest,
polling provided accurate estimates of voters’ support for Black candidates.16 Therefore, concerns about
the effects of social desirability on Obama’s poll results were unfounded.
On the other hand, there is evidence that explicit prejudice dampened support for Barack Obama in both
polls and on Election Day. For example, one study showed that white citizens who rated Blacks as less
intelligent and lazier than whites were less likely to vote for Obama than white citizens who did not
endorse those stereotypes.17 This effect was particularly strong among Independents and Democrats,
whereas Republican opposition to Obama was already so pervasive that prejudiced Republicans were
only slightly more likely to oppose Obama than nonprejudiced Republicans.18 Another study
demonstrated that white citizens who expressed racial resentment were more likely to oppose Obama
than white citizens who did not express those views. The pattern was different among Latinos, however.
Latinos held roughly the same level of racial resentment as non-Hispanic whites, but those sentiments
did not shape their attitudes toward Obama.19
Several scholars have examined whether implicit prejudice influenced American citizens’ attitudes and
vote choices in 2008 and 2012.20 Initial explorations found that citizens who were high in implicit
prejudice toward Blacks were less likely to vote for Obama.21 After further analyses of survey data from
2008 and 2012, however, scholars concluded that implicit prejudice had little if any systematic influence
on citizens’ voting behavior. Implicit prejudice exists among some citizens, but it appears not to creep
into political judgments, such as vote choice, that require cognitive effort. Scholars continued to find
strong evidence, however, of the influence of explicit prejudice on evaluations of Obama and vote
choice.22
Obviously, despite the impact of explicit prejudice on citizens’ vote choices, Barack Obama did win the
2008 presidential election and was reelected in 2012. His historic candidacy in 2008 benefitted from a
unique set of political and economic conditions, according to a group of scholars who specialize in
forecasting elections. These researchers concluded that Barack Obama should have won an
overwhelming victory in 2008 given the record-low popularity of the sitting president, George W. Bush,
and the biggest economic crisis since the Great Depression. They estimate that racial prejudice
depressed Obama’s vote by 5 percentage points, which was not enough to cost him the election but
enough to deny him a landslide.23 An analysis of 2012 data suggests that racial prejudice cost President
Obama votes in his reelection bid as well.24
Gender and Presidential Candidates
Since 1937, the Gallup organization has asked citizens whether they would vote for a well-qualified
woman nominated by their party for the presidency. Only one-third of Americans said they would do so
in 1937 (see Figure 10.2). During the 1950s and 1960s, support for a woman candidate hovered in the
50 percent range. Coinciding with the women’s movement, support for women candidates increased
significantly in the 1970s. By 1975, 73 percent of citizens said they would support a well-qualified
woman for the presidency. This percentage increased to 92 percent by 1999, only to fall back a few
percentage points in the 2000s.
Why did support for a woman candidate drop after 1999? It is difficult to answer that question by
examining the Gallup data because the organization did not ask the question between 1999 and 2003.
Another polling organization, however, did ask the question of a representative sample of citizens in
August and September 2002. In that poll, only 65 percent of respondents said they would vote for a
qualified woman candidate nominated by their party for the presidency, 7 percent said they would not do
so, and 28 percent said they were unsure.25 What led to all this uncertainty about a woman candidate?
What happened between 1999 and 2002 that might have led citizens to be hesitant about electing a
woman as president?
On September 11, 2001, terrorists struck the World Trade Center and the Pentagon, a tragic event
etched in the minds of most Americans. From that fateful day until the economic crisis of 2008, foreign
policy issues, especially the war on terrorism, dominated American politics. Why would this influence
whether citizens would be willing to support a woman candidate? Because many citizens believe male
politicians are more competent than female politicians when it comes to issues relating to terrorism and
national security. For example, in the 2002 poll just mentioned, 35 percent of citizens said that men
would be more competent to punish the people responsible for the September 11 attacks, and 40
percent indicated that men were more capable of protecting the homeland from future attacks. Nearly a
decade later, a 2010 survey found that 42 percent of the public think male elected officials are better
able than elected females to handle the issue of national security.26 In both years, virtually everyone
else said that women and men would be equally competent to deal with these issues, with only a tiny
percentage indicating women would be more competent. It turns out that the same citizens who doubted
the competency of women were the ones more likely to indicate they would not vote for a woman
candidate or were unsure whether they would do so.27
Thus, these data suggest that in an atmosphere focused on war and terrorism, women candidates might
be at a severe disadvantage because of the endorsement of gender stereotypes by a significant
minority of the American people. On the other hand, more recent research suggests that when voters
are faced with a choice between real-life candidates on Election Day, gender stereotyping is not very
likely, at least not in contests for the U.S. Senate, U.S. House, and state governor offices.28 The Gallup
question and the policy competence items we just mentioned ask about hypothetical candidates. For
actual candidates, voter determinations about whether a candidate is competent to handle certain
issues are not strongly driven by the sex of the candidate. In other words, in a contest between Jane
Williams and Richard Jones, Richard is not likely to be seen as better to handle national security simply
because he is a man. Furthermore, when deciding whether to vote for Jane or Richard, voters are more
influenced by factors such as the party and incumbency of the candidate or the competitiveness of the
race than gendered expectations about the candidates.
What about the highest seat in the land? The good news for women presidential candidates is that the
Gallup polls indicate support for a woman has recovered and stands at 94 percent in 2019. The bad
news is that some scholars believe that number is inflated due to social desirability effects. One study,
for example, demonstrated that as many as 26 percent of citizens were “angry or upset” by the notion of
a woman president, sentiments that are not revealed on surveys because citizens do not want to appear
sexist.29 Indeed, sexist attitudes, particularly hostile sexism, played a role in Hillary Clinton’s 2016
defeat. Hostile sexism is prejudice against women, rooted in beliefs that women are inferior to men and
are attempting to take power away from men.30 Statements measuring hostile sexism were created by
social psychologists; for examples, see Table 10.2. Note that agreeing with each statement is the hostile
sexist response. When it comes to the 2016 presidential election, citizens possessing the most hostile
sexist attitudes were much more likely to support Donald Trump’s candidacy than were citizens with the
least sexist attitudes.31 To be sure, there were many factors that explain the 2016 election outcome, yet
research demonstrates that sexism clearly helped Trump and hurt Clinton on Election Day.
Table 10.2 Measuring Hostile Sexism
“The statements … concern women, men, and their relationships in contemporary society. Please
indicate the degree to which you agree or disagree with each statement.”
“Many women are actually seeking special favors, such as hiring policies that favor them over men,
under the guise of asking for ‘equality.’”
“Most women interpret innocent remarks or acts as being sexist.”
“Most women fail to appreciate fully all that men do for them.”
“Women seek to gain power by getting control over men.”
“When women lose to men in a fair competition, they typically complain about being discriminated
against.”
Source: “The Ambivalent Sexism Inventory,” Understanding Prejudice, https://secure.understandingprejudice.org/asi/take
https://secure.understandingprejudice.org/asi/take
Sexual Orientation and Presidential Candidates
In 1978, the Gallup organization asked citizens for the first time about voting for a “homosexual”
presidential candidate. Only 26 percent of Americans said they would vote for such a candidate (refer
back to Figure 10.2). Over time, that percentage changed substantially as did the language used to
characterize such a candidate. In 2019, 76 percent indicated they would vote for a “gay or lesbian”
candidate nominated by their party for the presidency. When asked in 2016 if it would matter whether a
candidate was gay or lesbian, 69 percent of citizens said it wouldn’t matter, yet 26 percent said it would
make them less likely to support the candidate.32 Although a great deal of prejudice remains, this is a
significant improvement from 2007, when 46 percent said they would be less likely to support a
homosexual candidate. Overall, these polling data suggest that Americans are becoming more open to
a gay or lesbian presidential candidate; nevertheless, it seems unlikely that such a candidate, whether
that be Pete Buttigieg or someone else, could win the White House any time soon.
Socialism and Presidential Candidates
One of the personal characteristics that has more recently appeared on Gallup polls regarding support
for presidential candidates is whether the candidate is a socialist. The inclusion of this item is
undoubtedly related to the 2016 and 2020 presidential bids of Senator Bernie Sanders. Sanders is a
democratic socialist, a philosophy which he describes as prioritizing economic rights and justice for all
Americans, reducing economic inequality, and reforming the political system so that it responds to the
wishes of the public rather than to narrow special interests.33 As you might recall from Chapter 9,
socialism is an economic system whereby the means of production are communally owned, such as by
the government.
Overall, Americans seem very reluctant to vote for a well-qualified socialist nominated by their party.
Indeed, only 47 percent said they would in 2015 and 2019, whereas 45 percent indicated so in early
2020. In each of these years, a majority of Americans stated that they would not vote for a socialist
candidate, making socialism the most unpopular of all the candidate characteristics queried by Gallup in
recent years.34 Partisan differences emerge in support for a socialist presidential candidate. Three-
quarters of Democrats would support a well-qualified socialist for president compared to 45 percent of
Independents and only 17 percent of Republicans.35 Thus, to the degree that his opponents can attach
the label socialist to the democratic socialist Sanders, it seems Sanders would be able to clinch the
2020 Democratic presidential nomination but might not fare well in the general election contest.
SUPPORT FOR CIVIL RIGHTS POLICIES
“Equal Justice Under Law” are the words majestically incised above the entrance to the U.S. Supreme
Court building. This simple phrase captures the essence of civil rights, as suggested by the more
detailed definition of civil rights that we presented at the beginning of the chapter. In this section, we first
discuss public support for African American rights and how support for those rights has changed over
time. We focus on African Americans because they are an important group in the United States, both
historically and in contemporary society. African Americans have had to fight for their civil rights since
the founding of this nation, and the struggle has been a long and hard one. Today African Americans still
battle to preserve their civil rights and work to expand them further. In addition, we highlight opinion
toward policies dealing with African Americans for practical reasons. Social scientists have collected
extensive public opinion data on issues dealing with African Americans; unfortunately, other racial
minority groups have not received nearly as much attention.
We also examine public attitudes toward gay and lesbian rights, transgender rights, and disability rights.
In comparison with African Americans, the rights movements for these groups are more recent
phenomena. Having said that, over the past few decades, gay and lesbian issues have become quite
prominent. In the 1980s, fear of AIDS led to concerns about job discrimination against gays and
lesbians. Then, the (initially) divisive issues of the “Don’t Ask, Don’t Tell” policy (which allowed gays and
lesbians to serve in the military) and same-sex marriage emerged on the national agenda. As we will
see, public opinion has shifted on these issues across time.
Finally, our discussions of transgender rights and disability rights are notably shorter than the sections
exploring civil rights for Blacks and for gays and lesbians. The topic of transgender rights has reached
the national political agenda more recently, with issues such as public restroom access for transgender
individuals receiving attention. Civil rights for disabled individuals is not a new arena in U.S. politics, yet
polling of and scholarship regarding public views toward disability rights have been rare.
Support for Civil Rights for African Americans
Are citizens supportive of civil rights for Black Americans? In principle, the answer is yes. Both white and
Black Americans overwhelmingly endorse principles of equality. In practice, however, the answer is
much more complex. Support drops significantly among whites when it comes to the implementation of
civil rights policies, especially if the implementation means that Blacks will receive preferential treatment
or redress for past harms.36 Blacks also demonstrate less support for civil rights in practice than in
principle, but a majority still supports the enforcement of civil rights policies.
Support for School Integration
Let’s first discuss attitudes toward school integration to illustrate support for the principles of equality.
When white citizens were first asked in 1942 whether they thought Negro and white students should go
to the same or separate schools, only one-third supported integration (see Figure 10.3). White opinion
on this issue has shifted dramatically over time, with virtually all whites supporting integration in 1995.
Some scholars have characterized this change in attitudes as a “revolution.”37 Indeed, the GSS has
stopped asking the question because support for equality is so high. Regrettably, this question was not
asked of Blacks until 1972. Black support for integration was nearly universal at that time and has
remained so ever since.
In sharp contrast to their support for equality in principle, whites show much less support for putting this
principle into practice. Given high levels of racial segregation in residential patterns, it was rare that
white and Black students attended the same neighborhood schools. Thus, to achieve school integration,
several communities enacted busing plans during the 1970s. In other words, some cities bused white
children to schools in Black neighborhoods and Black children to schools in white neighborhoods as a
way to end segregation. Schools became more integrated as a result, but these policies were highly
unpopular with white citizens. When asked whether they favored or opposed “the busing of Negro and
white school children from one school district to another,” only 14 percent of white Americans supported
the policy in the 1970s (see Figure 10.4). Over time, white support for busing doubled, but it was still
quite low at 30 percent the last time the question was asked in 1996. Compare these low levels of
support for busing with the high levels of support white Americans voiced for whites and Blacks going to
the same schools (refer back to Figure 10.3). There is a huge gap between how whites think about
equality in theory versus what policies they are willing to support in practice.
Description
Figure 10.3 Support for Same Schools, by Race
Source: Data from Howard Schuman, Charlotte Steeh, Lawrence Bobo, and Maria Krysan, Racial Attitudes
in America (Cambridge, MA: Harvard University Press, 1997), 104–105, 240–241.
Note: Here is the General Social Survey question wording: “Do you think white students and (Negro/black)
students should go to the same schools or to separate schools?”
Blacks also demonstrate lower levels of support for the implementation of school integration policies
than they do for integration in principle. During the 1970s, 53 percent of Blacks supported school busing
(see Figure 10.4). By the 1990s, Black support for busing had increased to 61 percent. Although this
percentage is lower than the almost universal support among Blacks for the abstract notion of the two
groups attending school together (refer to Figure 10.3), Black support for busing is still double the level
of white support.
This lack of support for busing, especially among whites, at least partially explains why schools continue
to be highly segregated. In many cities, whites protested against busing, brought lawsuits to end the
policy, and moved to the suburbs in a practice called “white flight.” Further, in 2007, the Supreme Court
struck down voluntary school desegregation plans that took into account race in assigning students to
schools.38 As a result, not only has progress in integrating schools stalled, but resegregation is
occurring in school systems across the country.39 A recent examination, for example, concludes that
“nearly one in three Black students [in Tuscaloosa, Alabama] attends a school that looks as if Brown v.
Board of Education never happened” and “Black children across the South now attend majority-Black
schools at levels not seen in four decades.”40
Description
Figure 10.4 Support for School Busing, by Decade and Race
Source: Analysis of General Social Survey Cumulative Data File, 1972–2004.
Note: Here is the General Social Survey question wording: “In general, do you favor or oppose the busing of
(Negro/black/African-American) and white school children from one school district to another?”
Support for Affirmative Action
Affirmative action is an umbrella term for a variety of policies that ensure equal treatment of minorities
and whites in education and employment. Affirmative action policies range from a job advertisement that
includes a declaration that the company is an equal opportunity employer to a business taking race into
account when making hiring decisions to a college considering race as a factor in admissions. Some
affirmative action policies ensure equal opportunity, whereas other policies take special steps to turn the
principle of equal opportunity into practice by encouraging equal outcomes.41
Affirmative action policies have been justified primarily on two grounds. First, they are intended to make
up for past and continuing discrimination against racial minorities in our society. And second, they are
designed to ensure that the diversity of the United States is reflected in our colleges and universities,
government institutions, and workplaces. Policies to ensure equal treatment between men and women
are also referred to as affirmative action policies.
As with other civil rights policies, attitudes toward affirmative action are more positive in principle than
in practice. In 1995 and in 2007, the Pew Research Center asked citizens whether they supported
“affirmative action programs designed to help Blacks, women and other minorities get better jobs and
education.” This question focuses on the principle of enhancing opportunities rather than providing
special treatment. As a result, a majority of whites and almost all Blacks supported the policy.
Specifically, 53 percent of white citizens favored such programs in 1995.42 By 2007, the level of support
had increased to 65 percent. The fact that the question refers to both Blacks and women also boosts
white support for the policy.43 Among Blacks, support for this type of affirmative action policy was nearly
universal in both 1995 and 2007 (94 and 93 percent).
Similarly, 71 percent of citizens support the idea of affirmative action programs on college campuses.
According to a 2017 Pew Research Center survey, 66 percent of whites said that “affirmative action
programs designed to increase the number of Black and minority students on college campuses are a
good thing.” This is 11 percentage points higher than three years earlier. Eighty-two percent of Blacks
agreed with that sentiment as did 83 percent of Hispanics. A large majority of Democrats (84 percent)
compared to a slight majority of Republicans (52 percent) thought such programs were a good thing.44
Citizens’ support for affirmative action, however, falls dramatically when the policies refer to specific
steps to ensure the equal treatment of Blacks. In surveys conducted between 1986 and 2016, the
American National Election Study asked respondents whether “Blacks should be given preference in
hiring and promotion.” In 1986, only 14 percent of white Americans supported such policies. White levels
of support have varied little over time, with 18 percent favoring preferential treatment in 2016 (see
Figure 10.5). Among Blacks, we also see less support for affirmative action policies designed to put
equality into practice. Nevertheless, two-thirds of Blacks supported these policies in both 1986 and
2016.45
Description
Figure 10.5 Support for Preferences in Hiring and Promotion for Black
People
Sources: Data from Howard Schuman, Charlotte Steeh, Lawrence Bobo, and Maria Krysan, Racial Attitudes
in America (Cambridge, MA: Harvard University Press, 1997), 174–175, 268–269; Analysis of American
National Election Studies Cumulative Data File, 1948–2016.
Note: Here is the American National Election Study question wording: “Some people say that because of
past discrimination, Blacks should be given preference in hiring and promotion. Others say that such
preference in hiring and promotion of Blacks is wrong because it gives Blacks advantages they haven’t
earned. What about your opinion—are you for or against preferential hiring and promotion of Blacks?”
Thus, only a small percentage of white citizens favor concrete, affirmative steps to diversify the
workforce. Likewise, when asked a similar question about considering race in college admissions,
whites are not nearly as supportive as they are when evaluating college affirmative action programs in
principle. In 2016, only 14 percent of whites said they favored “allowing universities to increase the
number of Black students studying at their schools by considering race along with other factors when
choosing students.” Among Blacks, 42 percent favored such programs, considerably greater support
than among white citizens.46 But still, the bottom line is that white and Black support for affirmative
action in practice is below what it is for affirmative action in principle.
Support for Reparations
Beginning in 1989, Congressman John Conyers (D-MI) proposed the Commission to Study Reparation
Proposals for African Americans Act (HR 40) in every Congress until he retired in 2017. Since then,
Congresswoman Sheila Jackson Lee (D-TX) has introduced HR 40. The bill would establish a
commission to examine the effects of slavery on African Americans living today and make
recommendations regarding appropriate remedies to address slavery’s harmful effects. Such remedial
policies are referred to as reparations.
In a 2014 article in The Atlantic, Ta-Nehisi Coates drew significant attention to the issue by presenting a
powerful case for reparations. Coates argued that America began “in black plunder and white
democracy, two features that are not contradictory but complementary,” and detailed the ways in which
Black lives and livelihoods continued to be plundered through the present. He also lamented that most
Americans are not interested in debating the issue.
Perhaps after a serious discussion and debate—the kind that HR 40 proposes—we may find
that the country can never fully repay African Americans. But we stand to discover much about
ourselves in such a discussion—and that is perhaps what scares us. The idea of reparations is
frightening not simply because we might lack the ability to pay. The idea of reparations
threatens something much deeper—America’s heritage, history, and standing in the world.47
The topic of reparations has been in the news a few times in the years after the publication of Coates’s
article, including these instances in 2019. In June, a U.S. House Subcommittee held a hearing on HR
40, the first time in over a decade that such a hearing was held in Congress.48 During a July nationally
televised debate among Democratic candidates for the presidency, Marianne Williamson defended her
plan to distribute payments to descendants of enslaved persons, stating that current income disparities
between whites and Blacks stem “from a great injustice that has never been dealt with.”49 Concrete
plans to provide reparations for past slave-holding were debated that fall throughout the nation, in
places such as statehouses, religious institutions, and universities.50 At Georgetown University, for
example, students asked the Board of Trustees to set up a fund to right a specific historical wrong. In
the 1830s, Georgetown leaders sold 272 enslaved people, using the money to pay off crippling
university debt. The proposed fund would be used to benefit the descendants of those slaves. Said one
of the student protestors, “there would be no Georgetown without [that sale]. It’s a debt we need to pay
back.”51
In the wake of such heightened national attention to the issue, polling firms have conducted surveys to
assess attitudes toward reparations. Respondents to a 2019 Pew poll were asked whether they
thought slavery and discrimination were factors affecting the position of Blacks in contemporary
American society. As you can see in Table 10.3, Blacks and whites had very different views on these
questions with Blacks much more likely to say that slavery and discrimination were relevant. Blacks and
whites also differed on whether the government should take steps to compensate Black Americans for
slavery. According to a 2019 Gallup survey, 73 percent of Black Americans supported cash payments to
descendants of slaves. On the other hand, whites were particularly resistant to the notion of cash
payments. Overall, whites and Blacks have strikingly different views on the subject of reparations for
slave descendants.
Table 10.3 White and Black Support for Reparations, 2019
Question
Percentage “a
great deal”
Whites Blacks
“How much, if at all, do you think the legacy of slavery affects the position of
Black people in American society today? A great deal, a fair amount, not much,
not at all?”
26 59
Percentage
“major reason”
Whites Blacks
“Do you think [racial discrimination] is a major reason, minor reason, or not a
reason why Black people in our country may have a harder time getting ahead
than white people?”
35 75
Percentage
“should”
Whites Blacks
“Do you think the government should — or should not — make cash payments to
Black Americans who are descendants of slaves?”
16 73
Sources: Data from Juliana Menasce Horowitz, Anna Brown, and Kiana Cox, “Race in America 2019,” Pew Research Center, April 9,
2019, https://www.pewsocialtrends.org/2019/04/09/race-in-america-2019/; Mohamed Younis, “As Redress for Slavery, Americans
Oppose Cash Reparations,” Gallup, July 29, 2019, https://news.gallup.com/poll/261722/redress-slavery-americans-oppose-cash-
reparations.aspx
Philip J. Mazzocco and his colleagues conducted a fascinating study to examine why whites oppose
reparations for Black Americans who are the descendants of slaves.52 One possible reason for people
to oppose reparations is that the crime of slavery happened long ago and neither the perpetrators of the
crime nor the victims of the crime are alive today. Thus, reparations are not fair because they would
https://www.pewsocialtrends.org/2019/04/09/race-in-america-2019/
https://news.gallup.com/poll/261722/redress-slavery-americans-oppose-cash-reparations.aspx
benefit people who were not directly harmed and penalize people who had nothing to do with slavery. If
this reasoning explains white opposition to reparations, then it makes sense that whites would oppose
reparations for any type of crime that happened long ago, not just reparations for slavery. Is this the
case?
To answer this question, Mazzocco et al. asked sixty-six white college students to read a scenario in
which someone had done significant harm to their great-great-grandfather, including gaining financial
advantage at the expense of their relatives. After reading the scenario, participants were asked whether
they would join in a class action suit to try to recover some of their great-great-grandfather’s assets.
Here is the scenario:
Imagine that about 150 years ago, in the mid 1800s, your great, great grandfather was
kidnapped by Fineus Jones. Jones demanded a million dollars from your great, great
grandfather’s shipping business. The family borrowed the money and paid the ransom and
your great, great grandfather was released. Jones escaped to Europe and was apprehended,
but none of the million dollars was found. Your great, great grandfather lost his business to pay
back the ransom loan and died in poverty. Recently it was proven that the lost money had been
transferred to one of Fineus Jones’s sons who started a successful banking company with a
successor firm now worth 100 million dollars. Your cousins have found a respected attorney
who will press a claim on the successor firm and will do the work on a contingency basis, that
is, the attorney will receive a portion of the amount awarded by the court. If all costs are
included in the claim, the amount awarded to each claimant will be about $5,000.00. Your
cousins have asked if you would wish your name to be included on the list of claimants.53
What would you do? Would you allow your name to be included in the class action suit? Of the
participants in the study, 61 percent said they would want their name to be listed in the lawsuit. Thus, it
seems that many whites are not opposed to the notion of reparations in principle; rather, they are
opposed to reparations for Black Americans who are the descendants of slaves.
Mazzocco et al. argue that this opposition to reparations for Blacks stems from white Americans
underestimating racial disparities in the United States. Across many areas of society—health care,
criminal justice, education, and the economy—blacks are at a disadvantage relative to whites. But
because many whites do not seem to understand or appreciate the cost of being Black in our society,
they do not support reparations.
The research by Mazzocco et al. was not conducted on a random sample of white citizens. Thus, we
should be cautious about generalizing these results to whites in general. Nevertheless, it is striking how
similar the level of white support for reparations based on the scenario is to the percentage of Blacks
who supported cash payments to the descendants of slaves in the 2019 poll (refer back to Table 10.3).
Ideally, future research will examine the reaction of a representative sample of whites to these types of
scenarios.
Support for Civil Rights for Gays and Lesbians
Do citizens support civil rights for gays and lesbians? It depends on which type of right. Citizens largely
oppose discrimination in employment but are less supportive of gay rights when it comes to the legality
of intimate relationships. Furthermore, attitudes toward gay and lesbian rights have changed a great
deal over time, with support for such rights increasing dramatically since citizens were first polled on
these issues in the 1970s.
Support for Equality in Employment
Let’s begin with a discussion of attitudes toward equality for gays and lesbians in the workplace. In
1977, the Gallup organization began asking citizens whether they think “homosexuals should or should
not have equal rights in terms of job opportunities.” A majority of Americans, 56 percent, supported such
rights in 1977 (see Figure 10.6). By the mid-1990s, support for equal job rights had jumped to over 80
percent, reaching 89 percent in 2008. The next time Gallup asked this question, in 2019, 93 percent of
Americans supported equal job opportunities for gays and lesbians. This is a large shift in public opinion
over time.54
It is both ironic and troubling that despite widespread public support for gay rights in employment, there
is no federal or constitutional ban against employment discrimination based on sexual orientation. In
other words, federal law and the U.S. Constitution allow employers to fire employees simply because
they are gay or lesbian. That situation might change, however, depending upon how the Supreme Court
rules on cases that were argued before it in October 2019. The justices were asked to decide whether
or not the 1964 Civil Rights Act prevents employers from firing people because they are gay or
lesbian.55 In the meantime, some states, localities, universities, and corporations have instituted bans
on discrimination based on sexual orientation. As a result, whether gay rights are protected depends on
where you live and where you work in the United States. We will use ourselves to illustrate this point.
Zoe Oxley lives in New York, a state that has a law that bans employment discrimination based on
sexual orientation. In contrast, Rosalee Clawson lives in Indiana, a state that does not ban such
employment discrimination, yet the city she lives in, Lafayette, Indiana, does have a gay rights
ordinance barring job discrimination. Both Clawson and Oxley work at universities where discrimination
based on sexual orientation is prohibited.
A particular type of employment received a great deal of attention in the 1990s and 2000s—gays and
lesbians serving in the U.S. military. Since 1977, the Gallup poll has asked citizens whether they thought
homosexuals should or should not be hired for the armed forces. Just over one-half of citizens believed
gays and lesbians should be allowed to serve in the military at that time (see Figure 10.6). Over the next
three decades, support increased substantially, reaching 80 percent in 2003. In 2019, 83 percent of
Americans said they would support the right of gays and lesbians to serve in the armed forces. Overall,
these data suggest many Americans have come to the conclusion that “you don’t have to be straight to
shoot straight.”56
Despite this level of public support for gays and lesbians joining the armed forces, the U.S. government
maintained a policy against gays and lesbians serving openly until President Obama signed legislation
repealing “Don’t Ask, Don’t Tell” in December 2010. The legislation did not go into effect immediately
because the Department of Defense and the president were required to take a number of steps to
ensure the military was prepared for the policy change. During summer 2011, President Obama certified
that the military was ready, and as a result, gays and lesbians have served openly since September 20,
2011.
Support for Equality in Intimate Relationships
Although most people are opposed to discrimination in employment, public opinion has historically been
much more divided when it comes to support for intimate relationships between gays and lesbians. In
1977, the Gallup poll asked citizens whether they think “homosexual relations between consenting
adults should or should not be legal.” Only 43 percent of citizens supported the rights of gays and
lesbians to have sexual relations at that time (see Figure 10.6). The level of support dropped even lower
in the mid-1980s when merely 33 percent of the public supported such rights. This drop in support might
have been a reaction to the Supreme Court ruling on gay sex in 1986. In Bowers v. Hardwick, the
Supreme Court ruled that the U.S. Constitution does not protect homosexual relations between
consenting adults even in the privacy of their own homes.57 Thus, the Court ruling might have given
“permission” to citizens who were hesitant about gay rights to decide it was acceptable to oppose the
legality of gay and lesbian sexual relationships. This was also a period during which people’s fears
about AIDS may have influenced their attitudes toward gay and lesbian rights.
Description
Figure 10.6 Support for Gay Rights, 1977–2019
Source: Data from “Gay and Lesbian Rights,” Gallup, Washington, DC,
http://news.gallup.com/poll/1651/gay-lesbian-rights.aspx
Note: The solid lines indicate employment policies, and the dashed lines indicate policies concerning
intimate relationships. There have been variations in question wording over time in response to changing
norms about how to refer to gays and lesbians. The following are the most recent question wordings:
Job opportunities: “As you may know, there has been considerable discussion in the news regarding the
rights of gay men and lesbian women. In general, do you think gays or lesbians should or should not have
equal rights as non-gays or non-lesbians in terms of job opportunities?”
Serve in armed forces: “Do you think gays or lesbians should or should not be hired for the following
occupations: The Armed Forces?”
Relations between consenting adults: “Do you think gay or lesbian relations between consenting adults
should or should not be legal?”
Marriage: “Do you think marriages between same-sex couples should or should not be recognized by the
law as valid, with the same rights as traditional marriages?”
By the late 1980s, support for homosexual relationships had gone back up and was at 47 percent. In
May 2003, support for the legality of gay and lesbian relationships reached 60 percent, but again public
opinion seemed to react to a Supreme Court decision on gay and lesbian rights, although this time the
Court was in support of such rights. On June 26, 2003, the Supreme Court reversed its earlier decision
on gay sex and ruled in Lawrence v. Texas that the U.S. Constitution protects consensual sexual
relationships between two people of the same sex.58 This ruling caused an immediate outcry from
conservative activists, especially because many argued this ruling was the logical precursor to the
legalization of same-sex marriage. Thus, some citizens thought the Court had overstepped its bounds,
and they rejected the Court’s position. As a result, public support for homosexual relations between
http://news.gallup.com/poll/1651/gay-lesbian-rights.aspx
consenting adults dropped to 50 percent in a July 2003 Gallup poll. By 2019, however, citizens’ support
for gay and lesbian relationships had climbed to 73 percent.59
In a prior (2013) edition of this book, we said: “That brings us to the most controversial gay rights issue
of our times—gay marriage.” In an incredibly short time period, same-sex marriage has gone from a
highly controversial issue to one that is accepted by a sizable majority of citizens. Gallup surveys show
that attitudes toward same-sex marriage have changed dramatically in the past twenty years. In 1996,
Gallup asked citizens for the first time, “Do you think marriages between same-sex couples should or
should not be recognized by the law as valid, with the same rights as traditional marriages?” At that time
only 27 percent said they supported same-sex marriage (see Figure 10.6). When the question was
asked again in 1999, 35 percent of Americans said they would support same-sex marriage. Between
2004 and 2010, support for same-sex marriage bounced around the 40 percent level, and then for the
first time in 2011 a majority of Americans came around to the marriage equality point of view. By 2018
support had reached an all-time high of 67 percent favoring same-sex marriage, then fell back slightly to
63 percent in 2019. Interestingly, analyses of the remarkable change in attitudes toward same-sex
marriage over time conclude that the transformation was overwhelmingly due to people changing their
minds on the issue (rather than simply generational replacement of older cohorts by younger and more
progressive citizens).60
It was not particularly surprising when the U.S. Supreme Court upheld same-sex marriage on June 26,
2015, in the landmark case Obergefell v. Hodges, because Court rulings are consistent with public
opinion roughly two-thirds of the time.61 To be sure, Democrats and Republicans continue to differ on
the issue, but with 83 percent of young adults under 30 in support of same-sex marriage, not to mention
same-sex couples across the country exercising their newly won right, marriage equality is here to
stay.62
Box 10.1 Support for Gay Rights across the World
Public Opinion in Comparative Perspective
In 2019, the European Union commissioned a study of attitudes toward gay, lesbian, and
bisexual political leaders. Respondents were asked to place themselves on a 1 to 10 scale with 1
meaning totally uncomfortable and 10 totally comfortable with having someone who is gay,
lesbian, or bisexual in the highest elected political position in their country. On average, 64
percent of a European Union nation’s residents would feel totally comfortable. Support varied a
great deal across countries within the European Union, ranging from a high of 93 percent in the
Netherlands to a low of 18 in Bulgaria. The average percentage of citizens for each country who
would be totally comfortable with a gay, lesbian, or bisexual leader is presented in the table.
Percentage of Europeans Feeling Comfortable with a Gay, Lesbian, or Bisexual Political Leader
Country Percentage “totally comfortable”
Netherlands 93
Sweden 90
United Kingdom 86
Ireland 84
Luxembourg 84
Spain 81
Belgium 79
Denmark 78
France 74
Malta 73
European Union 64
Germany 64
Country Percentage “totally comfortable”
Finland 61
Portugal 57
Italy 55
Slovenia 55
Czechia 53
Austria 52
Poland 49
Croatia 45
Slovakia 45
Greece 44
Estonia 41
Cyprus 40
Hungary 33
Latvia 32
Lithuania 26
Romania 26
Bulgaria 18
Source: Data from “Special Eurobarometer 493: Discrimination in the European Union,” European Commission, October 2019,
https://ec.europa.eu/commfrontoffice/publicopinion/index.cfm/survey/getsurveydetail/instruments/special/surveyky/2251
https://ec.europa.eu/commfrontoffice/publicopinion/index.cfm/survey/getsurveydetail/instruments/special/surveyky/2251
Another way to examine the public’s attitudes toward gay rights is to consider whether people
believe homosexuality should be accepted by society. In 2013, the Pew Research Center
conducted a Global Attitudes survey in thirty-nine nations. The survey asked respondents
whether “homosexuality should be accepted by society.” People in Middle Eastern, African, and
some Asian countries were particularly opposed to homosexuality. Russians were also quite
opposed. In contrast, a majority of citizens in Canada, the United States, and most European
Union and Latin American countries deemed homosexuality acceptable.1 Perhaps attitudes
around the globe regarding the acceptance of homosexuality have changed in recent years. Pew
has not conducted another global survey on this topic so we cannot be sure. Pew did, however,
poll Europeans regarding their support for same-sex marriage between 2015 and 2017. One
clear pattern from their results is that support is lower in eastern Europe than central Europe and,
especially, western Europe. Even among young adults, opposition to gays and lesbians marrying
is especially high in eastern Europe, as the following map demonstrates.
Description
Source: “Eastern and Western Europeans Differ on Importance of Religion, Views of Minorities, and
Key Social Issues,” Pew Research Center, Washington, DC, October 29, 2018,
https://www.pewforum.org/2018/10/29/eastern-and-western-europeans-differ-on-importance-of-
religion-views-of-minorities-and-key-social-issues/
1 “The Global Divide on Homosexuality,” Pew Research Center, Washington, DC, June 4, 2013,
http://www.pewglobal.org/2013/06/04/the-global-divide-on-homosexuality
https://www.pewforum.org/2018/10/29/eastern-and-western-europeans-differ-on-importance-of-religion-views-of-minorities-and-key-social-issues/
http://www.pewglobal.org/2013/06/04/the-global-divide-on-homosexuality
Support for Civil Rights for Transgender People
Of the many civil rights issues that the transgender rights movement has focused on, fighting
discrimination in the workplace has been, and continues to be, one of the most important.63 There is no
national legal or constitutional protection against employment discrimination based on transgender
status. As with workplace protections for gays and lesbians, however, employment protection for
transgender individuals was on the Supreme Court’s docket in the fall of 2019. Before the Court was the
case of a transgender woman who was fired after she told her employer she would present as a woman
at work.64 The Justices are expected to issue a ruling on this case in 2020. Another issue related to
transgender employment centers on the U.S. military. During his first year in the White House, Donald
Trump announced a ban on all openly transgender service members. This ban, which was a reversal of
a Barack Obama–era policy, eventually was scaled back a bit to ban only future transgender people
from the military. Active-duty transgender troops could continue to serve. Finally, “bathroom bills” have
been debated in states and localities across the nation in recent years. One version of these bills, such
as that passed in North Carolina in 2016, would restrict access to single-gender public and school
restrooms to one’s sex assigned at birth. In contrast, other versions legally allow transgender individuals
to use public restrooms that align with their gender identity or require single-occupant restrooms to be
gender neutral.
When it comes to public attitudes toward transgender rights, support is highest for workplace rights.
Yet, fewer Americans support employment rights for transgender individuals compared to rights for gays
and lesbians.65 Take military service, for instance. In 2019, 71 percent of the public agreed that
transgender women and men should be able to serve in the armed forces (see Figure 10.7) versus 83
percent who supported gays and lesbians in the military (refer back to Figure 10.6). According to a 2019
survey conducted by the Economist magazine and YouGov, 59 percent of the public think that
employers should not be able to fire transgender employees who wear “appropriate work clothes that
match with their gender identity.”66 In that same survey, 73 percent of respondents said employers
should not be able to fire a worker “for being gay.”
As for attitudes regarding transgender rights in public accommodations, such as restrooms, and views
toward whether it is even possible for one’s gender identity to differ from their birth sex, public support is
lower than in the domain of transgender employment rights.67 Less than a majority (44 percent) of
Americans in 2019 agreed that transgender individuals should be able to use public bathrooms that
align with their gender identity, a figure that has remained quite stable since 2016.68 Finally, a 2017 Pew
survey queried respondents about whether people’s gender identities are determined by their sex
assignment at birth or can be different. A slight majority chose the former compared to 44 percent who
believed that gender identity can differ from one’s assigned sex at birth.
Description
Figure 10.7 Support for Transgender Rights, 2017 and 2019
Sources: Data for military service and public restrooms from Justin McCarthy, “In U.S., 71% Support
Transgender People Serving in Military,” Gallup, June 20, 2019,
https://news.gallup.com/poll/258521/support-transgender-people-serving-military.aspx. Data for employment
discrimination from Kathy Frankovic, “American Views of Transgender People: The Impact of Politics,
Personal Contact, and Religion,” YouGov, October 11, 2019, https://today.yougov.com/topics/politics/articles-
reports/2019/10/11/american-views-transgender-people-poll. Data for identity as a man or woman from
“Republicans, Democrats Have Starkly Different Views on Transgender Issues,” Pew Research Center,
November 8, 2017, https://www.pewresearch.org/fact-tank/2017/11/08/transgender-issues-divide-
republicans-and-democrats/
https://news.gallup.com/poll/258521/support-transgender-people-serving-military.aspx
https://today.yougov.com/topics/politics/articles-reports/2019/10/11/american-views-transgender-people-poll
https://www.pewresearch.org/fact-tank/2017/11/08/transgender-issues-divide-republicans-and-democrats/
Support for Civil Rights for People with Disabilities
Landmark disability rights legislation was passed by Congress and signed by President George H. W.
Bush in 1990. This civil rights law, the Americans with Disabilities Act (ADA), “guarantees equal
opportunity for individuals with disabilities in public accommodations, employment, transportation, state
and local government services, and telecommunications.”69 Attitudes toward disability rights are very
favorable, if we focus on public support for the ADA. For example, in 2015, around the twenty-fifth
anniversary of the law’s passage, 83% of Americans who had heard of the ADA supported it compared
to only 5% who opposed the act.70 In contrast, Americans were more divided when it came to
determining which disabilities qualify for protection under the ADA. Three-quarters or more of the public
believed that permanent vision or hearing loss, cerebral palsy, and multiple sclerosis should be covered
under the ADA but that the following conditions should not be covered: migraine headaches, morbid
obesity, anorexia or bulimia, drug addiction, and alcoholism. In contrast, there was a roughly 50-50 split
over whether learning disabilities (such as ADD or dyslexia), cancer, and schizophrenia merited ADA-
protected status.71
Much less is known regarding public opinion toward other civil rights for disabled individuals. Do
Americans support changes to the process of voting (location of polling places, availability of early
voting, etc.) to enable disabled Americans to cast their ballots more easily? Should governments and
private organizations do more to make their websites accessible to people who are visually impaired? Is
it acceptable that certain employers can legally pay workers with disabilities less than the minimum
wage? Do schools and universities provide sufficient accommodations and services to students with
disabilities? Hopefully more social science research and opinion polling will be conducted in the near
future that addresses questions such as these.
CONCLUSION
Are citizens hesitant about supporting women and minority candidates? Well, it depends. If you are a
Muslim or an atheist, your chances for winning the presidency seem quite slim. The public is not
predisposed to favor such a candidate. If you are a Mormon, an evangelical Christian, or gay, you also
have an uphill battle. Your prospects are better if you are a woman, a Jew, or a racial or ethnic minority,
but still there are some citizens who say they will not support someone like you. The good news is that
over time citizens have become increasingly more likely to say they would vote for a woman or minority
candidate, with one notable exception. Public support for a Mormon candidate has budged only slightly
over the past fifty years. But these survey questions ask about hypothetical candidates. Actual flesh-
and-blood people who happen to be nontraditional candidates might be able to win despite stereotypes
and prejudices or, then again, they might not be able to—just ask Barack Obama and Hillary Clinton.
What about citizens’ support for civil rights? When it comes to equality for Black Americans, citizens’
attitudes depend heavily on two factors. First, support varies based on whether the policy at hand refers
to equal rights in principle or equal rights in practice. Second, levels of support for civil rights differ a
great deal between Black and white citizens. In terms of public support for civil rights toward the other
groups we profiled in this chapter, citizens demonstrate strong support for equal job opportunities (in
and outside the military) for gays and lesbians and for the Americans with Disabilities Act. Citizens’
support for gay and lesbian sexual relationships and marriage rights have increased substantially over
time. The bottom line is that a sea change in attitudes toward rights for gays and lesbians has occurred
over the past forty years. On the other hand, support for transgender rights is notably lower.
Overall, these findings certainly provide some encouragement to participatory democratic theorists. That
citizens have become more supportive of women and minorities holding office and that citizen support
for several civil rights policies has increased over the years are heartening to theorists who worry about
inequality in society. Nevertheless, the news is not all good for participatory democratic theorists. A
meaningful minority (especially given the closeness of presidential races) still opposes women and
minorities running for the highest office in the land. Moreover, the influence of explicit racial prejudice on
vote choice during the 2008 and 2012 presidential elections and hostile sexism in 2016 is also
extremely troubling. Further, the hesitancy to support the implementation of civil rights policies on the
part of whites raises serious doubts about whether our society provides “equal justice under law,” as
proclaimed by the words on the Supreme Court building.
Because elite democratic theorists expect some inequality in society, they are not shocked by opposition
to women and minority candidates or by the lack of support among white Americans for implementing
civil rights policies or by opposition to same-sex marriage or transgender rights. As we discussed earlier,
elite democratic theorists emphasize that it is fine if the public doesn’t support civil rights because elites
hold the power. Because elites are more sophisticated, they appreciate the importance of civil rights in a
democratic society and are therefore better equipped to protect the rights of marginalized or unpopular
groups. Political theories, however, do not always accurately capture the practice of politics. Take the
issue of gay rights. For a long time elites in both Republican and Democratic parties opposed marriage
rights for gays and lesbians. Most Democratic leaders have finally changed their position, but many
Republican officials continued to speak out against same-sex marriage even after the Supreme Court
ruled it constitutional in 2015. Clearly these elites are no more protective of this particular civil right than
those citizens who persist in opposing same-sex marriage.
KEY CONCEPTS
attitudes toward affirmative action 322
attitudes toward disability rights 335
attitudes toward gay and lesbian rights 327
attitudes toward reparations 325
attitudes toward same-sex marriage 330
attitudes toward school integration 320
attitudes toward transgender rights 334
Bradley effect 314
civil rights 308
explicit prejudice 314
gender stereotypes 316
hostile sexism 317
implementation 319
implicit prejudice 314
principles of equality 319
social desirability 314
SUGGESTED SOURCES FOR FURTHER READING
Golebiowska, Ewa A. “Sexual Orientation, Gender Identity, and Political Decision Making.” In Oxford
Encyclopedia of Political Decision Making, ed. David P. Redlawsk. Oxford: Oxford University Press,
prepublished June 2019, DOI: 10.1093/acrefore/9780190228637.013.774.
Lee, Hye-Yon, and Diana C. Mutz. “Changing Attitudes toward Same-Sex Marriage: A Three-Wave
Panel Study.” Political Behavior 41 (2019): 701–722.
Golebiowska provides a comprehensive review of research pertaining to public support for the rights of
gay, lesbian, bisexual, and transgender individuals. Lee and Mutz explore why public support for same-
sex marriage has increased considerably.
Schneider, Monica C., and Angela L. Bos. “Measuring Stereotypes of Female Politicians.” Political
Psychology 35 (2014): 245–266.
Streb, Matthew J., Barbara Burrell, Brian Frederick, and Michael A. Genovese. “Social Desirability
Effects and Support for a Female American President.” Public Opinion Quarterly 72 (2008): 76–89.
These studies use innovative research methods to explore public attitudes toward women running for
political office. The first investigates stereotypes of female politicians compared to other relevant groups,
including politicians, male politicians, and female professionals. The second article uses a “list
experiment,” which is an unobtrusive method to assess prejudice toward a female presidential
candidate. The researchers find substantial evidence of prejudice.
Schuman, Howard, Charlotte Steeh, Lawrence Bobo, and Maria Krysan. Racial Attitudes in America.
Rev. ed. Cambridge, MA: Harvard University Press, 1997.
Moberg, Sarah Patton, Maria Krysan, and Deanna Christianson. “The Polls–Trends: Racial Attitudes in
America.” Public Opinion Quarterly 83 (2019): 450–471.
The classic book by Schuman and coauthors examines trends in racial attitudes, of both Blacks and
whites, since the early 1940s. Moberg et al. update and extend this analysis.
Human Rights Campaign, www.hrc.org
The Human Rights Campaign is the largest civil rights organization working for lesbian, gay, bisexual,
and transgender equality. This website contains information about LGBTQ issues, such as marriage,
adoption, and workplace discrimination. You can look up LGBTQ policies in your state and community
on this site.
National Association for the Advancement of Colored People (NAACP), www.naacp.org
Founded in 1909, the NAACP is the nation’s oldest civil rights organization. Its website contains both
historical and contemporary information about civil rights issues. The site includes a “Youth & College”
section that students will find of particular interest.
UnidosUS, https://www.unidosus.org/
UnidosUS, formerly the National Council of La Raza, is the largest national Latino civil rights
organization in the United States. This website contains information about civil rights, including racial
profiling, criminal justice issues, education, and voting rights.
Campaign Zero, www.joincampaignzero.org
http://www.hrc.org/
http://www.naacp.org/
http://www.joincampaignzero.org/
This grassroots organization focuses on ending police violence, a civil rights issue that has recently
gained significant national attention. This website provides information on political leaders’ positions on
a range of policies, including body cameras for police, community oversight, and demilitarization of
police departments.
Descriptions of Images and Figures
Back to Figure
The horizontal axis lists years and ranges from 1937 to 2019. The vertical axis lists percentage of
support and ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year
Percentage
support
from
Baptists
Percentage
support
from
Catholics
Percentage
support
from Jews
Percentage
support
from
Mormons
Percentage
support
from
Atheists
Percentage
support
from
Evangelical
Christians
Percentage
support
from
Muslims
1937 No data 60 46 No data No data No data No data
1940 No data 61 No data No data No data No data No data
1955 No data 69 No data No data No data No data No data
1956 No data 71 No data No data No data No data No data
1958 92 67 61 No data 18 No data No data
1959 93 68 70 No data 20 No data No data
1960 No data 69 No data No data No data No data No data
1961 No data 83 67 No data No data No data No data
1963 No data 85 77 No data No data No data No data
1965 No data 88 79 No data No data No data No data
1967 93 90 80 75 No data No data No data
1969 No data 88 87 No data No data No data No data
1978 No data 91 83 No data 39 No data No data
Year
Percentage
support
from
Baptists
Percentage
support
from
Catholics
Percentage
support
from Jews
Percentage
support
from
Mormons
Percentage
support
from
Atheists
Percentage
support
from
Evangelical
Christians
Percentage
support
from
Muslims
1983 No data 91 88 No data 40 No data No data
1987 No data No data 90 No data 41 No data No data
1999 No data 92 91 77 46 No data No data
2003 No data 91 90 No data No data No data No data
2007 No data 92 91 71 44 No data No data
2011 No data 90 90 76 46 No data 61
2015 No data 91 90 78 61 77 62
2019 No data 92 91 68 63 78 68
Back to Figure
The horizontal axis lists years and ranges from 1937 to 2019. The vertical axis lists percentage of
support and ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year
Percentage
support from
women
Percentage
support from
Blacks
Percentage support
from Hispanics
Percentage support
from gay or lesbian
1937 32 No data No data No data
1945 32 No data No data No data
1949 48 No data No data No data
1955 51 No data No data No data
Year
Percentage
support from
women
Percentage
support from
Blacks
Percentage support
from Hispanics
Percentage support
from gay or lesbian
1958 53 37 No data No data
1959 57 49 No data No data
1961 No data 49 No data No data
1963 53 48 No data No data
1965 No data 57 No data No data
1967 55 51 No data No data
1969 50 66 No data No data
1971 65 68 No data No data
1975 71 71 No data No data
1978 74 74 No data 27
1983 77 76 No data 28
1987 80 78 No data No data
1997 No data 94 No data No data
1999 91 95 No data 56
2003 88 93 No data No data
2007 88 93 87 52
2011 91 93 88 65
Year
Percentage
support from
women
Percentage
support from
Blacks
Percentage support
from Hispanics
Percentage support
from gay or lesbian
2015 90 90 90 71
2019 91 93 92 74
Back to Figure
The horizontal axis lists years and ranges from 1942 to 1995. The vertical axis lists percentage of
support and ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year Percentage support from Whites Percentage support from Blacks
1942 31 No data
1956 49 No data
1965 70 No data
1972 85 97
1980 87 97
1985 91 97
1995 95 98
Back to Figure
The horizontal axis lists the decade. The vertical axis lists percentage of support and ranges from 0 to
100 in increments of 20. Approximate data from the graph are tabulated as follows:
Decade Percentage support of Whites Percentage support of Blacks
1970s 15 54
Decade Percentage support of Whites Percentage support of Blacks
1980s 22 57
1990s 30 61
Back to Figure
The horizontal axis lists years and ranges from 1986 to 2016. The vertical axis lists percentage of
support and ranges from 0 to 100 in increments of 20. Approximate data from the graph are tabulated
as follows:
Year Percentage support of Whites Percentage support of Blacks
1986 13 67
1988 13 63
1990 15 75
1992 12 59
1994 10 52
1996 11 55
1998 11 55
2000 10 61
2004 11 54
2008 11 54
2012 11 57
2016 20 63
Back to Figure
The horizontal axis lists years and ranges from 1977 to 2019. The vertical axis lists percentage of
support and ranges from 0 to 100 in increments of 10. Approximate data from the graph are tabulated
as follows:
Year
Percentage support for
equal job opportunities
for homosexuals
Percentage support
for homosexuals
serving in armed
forces
Percentage support
for legality of gay or
lesbian relations
Percentage in
favor of same-
sex marriage
1977 57 51 43 No data
1982 59 52 44 No data
1985 No data 54 44 No data
1986 No data No data 33 No data
1987 No data 54 33 No data
1989 70 58 48 No data
1992 72 57 48 No data
1996 81 No data 43 27
1999 81 69 50 32
2001 83 70 54 No data
May
2003
85 78 59 No data
July
2003
85 No data 50 No data
2004 86 No data 51 39
2005 85 76 50 32
Year
Percentage support for
equal job opportunities
for homosexuals
Percentage support
for homosexuals
serving in armed
forces
Percentage support
for legality of gay or
lesbian relations
Percentage in
favor of same-
sex marriage
2006
86 No data 54 34
2007 86 No data 57 40
2008 86 No data 54 35
2009 No data No data 54 35
2011 No data No data 61 50
2013 No data No data 61 50
2015 No data No data 63 58
2017 No data No data 71 60
2019 91 81 73 59
Back to image
Text reads, Young adults in Central and Eastern Europe largely oppose gay marriage. Percentage of
those ages 18 to 34 who say they oppose or strongly oppose allowing gays and lesbians to marry
legally. Data from the map are tabulated as follows:
Country Percentage opposing gay marriage
Portugal 14
Spain 7
Country Percentage opposing gay marriage
Ireland 20
UK 13
France 17
Belgium 9
Switzerland 16
Italy 27
Denmark 6
Netherlands 10
Norway 14
Sweden 5
Finland 12
Germany 15
Czech Republic 18
Austria 14
Croatia 61
Bosnia 79
Estonia 61
Country Percentage opposing gay marriage
Latvia 70
Lithuania 74
Poland 50
Slovakia 42
Hungary 52
Serbia 75
Belarus 75
Ukraine 82
Moldova 89
Romania 66
Bulgaria 68
Greece 51
Russia 86
Georgia 94
Armenia 95
Back to Figure
The horizontal axis lists percentage of agreements and ranges from 0 to 100 in increments of 10. The
vertical axis lists four rights. Approximate data from the graph are tabulated as follows:
Statement PercentageagreementStatement
Percentage
agreement
Openly transgender men and women should be able to serve in the military. 71
Employers should not be able to fire transgender employees who wear work
clothes that match their gender identity.
59
Public restroom policies should allow transgender individuals to use the restroom
that corresponds with their gender identity.
44
Whether a person is a man or a woman can be different from their sex assigned at
birth.
44
PART V WHAT IS THE
RELATIONSHIP BETWEEN
CITIZENS AND THEIR
GOVERNMENT?
D O U.S. CITIZENS trust the government? Do they support the
institutions of government and the people who occupy key political
offices? Does the government respond to the policy preferences of
the public? If so, do all citizens’ opinions weigh equally, or do the
views of some matter more? Alternatively, is policy not related to
public opinion but, rather, influenced by the wishes of others, such as
interest groups or elected officials’ own goals?
All of these questions point, in one way or another, to the relationship
between citizens and their government, the focus of Part V. In
addition to these empirically oriented questions, we also consider
many normative ones. How important to democratic governance is
citizens’ trust in their government? Can a democracy survive without
it? What if citizens display little support for the institutions of
government? Are there worrisome implications that arise under this
situation? Finally, why does it matter whether public opinion
influences public policy? Are there some circumstances where the
public should have a large influence? A small one? No influence?
CHAPTER 11 TRUST IN GOVERNMENT, SUPPORT
FOR INSTITUTIONS, AND SOCIAL CAPITAL
LET’S RECAP MAJOR NATIONAL EVENTS of the past decade or so. The financial and subprime
mortgage crises of 2008 kick-started the worst economic downturn since the Great Depression. For
weeks during the summer of 2011 President Barack Obama and congressional leaders engaged in a
long and, let’s face it, not very pretty debate over raising the federal government’s debt ceiling. An
agreement was finally reached, but only hours before the government would have been prevented from
borrowing money to pay its bills. Two years later, the debt ceiling was on the agenda again. Both sides
hunkered down yet again. Leaders failed to reach an agreement on this or the nation’s annual budget,
leading to the federal government shutting down for sixteen days. In the summer of 2014, police officers
in Ferguson, Missouri, shot and killed Michael Brown, an unarmed Black man. The deaths of African
Americans in police custody, including Eric Garner in New York City, Freddie Gray in Baltimore, and
Sandra Bland in Texas, received widespread national attention in 2014 and 2015. The 2016 presidential
election featured not only nastiness, but also two candidates—Donald Trump and Hillary Clinton—who
were especially unpopular among large swaths of the public. Hurricane Maria hit a number of Caribbean
islands in 2017, including Puerto Rico. This massive and deadly storm caused widespread damage, yet
relief efforts from the U.S. federal government were slow in coming. January 2017 brought a new
occupant to the White House, but not lessened governing tensions between Congress and the
president. Due to a failure to pass legislation to pay the government’s bills, the federal government shut
down twice in 2018, for a brief two days in January and then for thirty-five days from late December into
January 2019. The latter was the longest government shutdown in U.S. history.
This is admittedly only a partial list of recent events, and, by design, they are all negative. Even so, is it
any wonder the American public does not view the government and public officials very favorably these
days? One important feature of democratic public opinion is citizen assessment of government.
Furthermore, democratic citizens are expected to not only evaluate their government and their political
leaders but also have the means to enact change if they are dissatisfied. Citizens can hold elected
officials accountable by voting for their opponents on Election Day. Severe dissatisfaction with the
government could lead to calls for changing governmental procedures or even for replacing the
structure of government with a new one. This type of citizen control is one key characteristic of
democracies, as so clearly stated by Thomas Jefferson in the Declaration of Independence. “Whenever
any Form of Government becomes destructive of these ends [securing individual rights],” wrote
Jefferson, “it is the Right of the People to alter or abolish it, and to institute new Government.”1
Furthermore, a belief that the government is legitimate is related to citizen obedience to authorities and
to laws, whereas alienation from government may suppress involvement in political activities.2 Given its
importance to democratic functioning, therefore, public opinion scholars have long been interested in
whether citizens demonstrate support for their government.
There are, of course, many aspects of government toward which the public holds attitudes, including
constitutional principles, day-to-day functioning, governmental institutions, governmental decisions, and
the performance of elected and appointed officials. One useful way to categorize these diverse attitude
objects was presented by political scientist David Easton in the 1960s.3 Easton suggests there are two
types of public support for political systems: diffuse support and specific support. Diffuse support refers
to public opinions about the political system, such as contentment with the form of government and
attachment to the norms and structure of the regime. In contrast to this broad attitude, the public also
holds attitudes toward the performance of incumbent political leaders and governmental outputs, such
as public policies. Easton calls this specific support. As for the relationship between the two, Easton
has this to say: “One major characteristic [of diffuse support] is that since it is an attachment to a political
object for its own sake, it constitutes a store of political good will. As such, it taps deep political
sentiments and is not easily depleted through disappointment with outputs.”4 In other words, if citizens
are unhappy with governmental policy decisions, specific support for the government will be low, but
diffuse support can remain high.
In this chapter, we discuss attitudes tapping both diffuse and specific support for government. We first
examine public trust in government, which encompasses characteristics of both. We demonstrate that
public trust in government has declined over time, present explanations to account for changing levels
of trust, explore trust in the criminal justice system, and discuss the implications of lower trust levels.
Second, we examine confidence in particular governmental institutions: the executive branch, Congress,
and the Supreme Court. Public faith in those institutions is also a function of both diffuse and specific
support. We demonstrate that attitudes toward the members of the institutions are distinct from attitudes
toward the institutions themselves. In the final section, we move away from an assessment of public
evaluations of government to explore citizen interaction with other citizens. Social capital, or the degree
to which people connect with and trust other citizens and engage in civic activities, is related to both
trust in government and support for government institutions. As we will see, however, social capital has
other important consequences for the public and for democratic governments. Also, as with trust in
government, the stock of social capital in America has declined of late, a trend we examine. Finally, to
help us think through the importance and implications of public trust in government, support for national
institutions, and social capital, we turn to relevant democratic theories throughout the chapter.
TRUST IN GOVERNMENT
Although a number of different definitions of trust in government have been proposed, scholars
generally agree that trust refers to “the public’s basic evaluative orientation toward the government in
Washington.”5 As such, trust seems to be a measure of diffuse support for government. Consider,
however, the more focused definition provided in the 1970s by Arthur Miller, a leading researcher of
public trust: “The belief that the government is operating according to one’s normative expectations of
how government should function.”6 More recently, trust has been described “as a pragmatic running tally
of how people think the government is doing at a given point in time.”7 These definitions suggest that
public trust involves assessing the performance of government. The attitude of trust therefore can also
be classified as a measure of specific support because it involves some evaluation of governmental
outputs. The lack of trust is commonly referred to as cynicism or distrust.8 Following from Miller’s
definition of trust, cynicism “reflects the belief that the government is not functioning in accordance with
individual expectations of efficiency, honesty, competence and equity.”9
Many democratic theorists agree that public trust in government is important for democratic societies.
Citizens place governing duties in their elected representatives and appointed officials. Given their
distance from government and their lack of knowledge regarding the many complex policy matters that
leaders must address, citizen trust in leaders and governing institutions is a salient feature of democratic
decision making.10 If citizens trust their government, they will accept and comply with its decisions,
leading to a stable democracy. If, on the other hand, citizens do not trust their government, it might be
difficult for government to enforce the law, leading to political and social disruption.11
Proponents of participatory democracy and related variants, such as deliberative democracy, further
emphasize that citizen trust in government reacts to the political environment. Trust can increase the
more that government procedures are transparent and the more that citizens become involved in
debating issues. “Deliberative arenas … provide opportunities to explain oneself, one’s group, one’s
problems,” leading to greater understanding of the views of others and perhaps engendering trust in
others’ motives.12 An obvious extension of this view is that declines in public trust, whether of the diffuse
or specific support variety, could be indicative of too little involvement of the public in decision making.
Elite democratic theorists have a more nuanced view of trust in government. One key feature of liberal
democracy, a theoretical precursor to contemporary elite democracy, is a presumption that citizens
should distrust the people in government.13 After all, the Federalists designed the complex checks and
balances of the U.S. political system with the presumption that political leaders would not pursue the
public good but would, rather, look out solely for their own interests. Yet the founders wanted citizens to
trust the representative institutions of government so that people would not feel it necessary to pursue
direct democracy, which was anathema to the founders.14 Thus, elite democrats would expect the public
to place less trust in government leaders than in the institutions and procedures of government that
were established to hold leaders in check. In other words, low specific support would be acceptable, but
diffuse support should remain high.
Measuring Public Trust
To measure public trust in government, survey researchers working on the American National Election
Studies (ANES) developed four specific questions in the early 1960s.15 These questions are the ones
most commonly used to examine trust, so it is worth considering the content of the survey items in
detail. The complete wording of these questions appears in Table 11.1. The first question asks
respondents directly how often they trust the national government. The next two items query people
about the behavior of politicians, specifically whether they waste taxpayer money and whether they work
for “the benefit of all the people” or only a “few big interests.” The final question originally required
respondents to assess whether politicians were crooked. In 2012, this question was changed slightly.
Corrupt replaced crooked, perhaps because the latter is an outdated term. As noted in the table, the
question wording for the first trust item also changed recently. Respondents in 2016 were provided with
five possible response options instead of only three. We thus need to be careful when comparing
responses to these questions over time, as we do in the next section.
What specific criteria were these four items intended to assess? Donald Stokes, one of the creators of
the questions, points to “the honesty and other ethical qualities of public officials” as well as “the ability
and efficiency of government officials and the correctness of their policy decisions.”16 To be sure, these
are varied criteria. Despite this, Stokes’s analysis indicated that public responses to the individual
questions correlated strongly with each other to form a general evaluation of government, usually
referred to as the trust in government scale.
Table 11.1 Assessing Public Trust: Survey Questions from the American National Election Studies
“How much of the time do you think you can trust the government in Washington to do what is right
—just about always, most of the time, or only some of the time?” (before 2016); “How often can you
trust the federal government in Washington to do what is right—always, most of the time, about half
the time, some of the time, or never?” (in 2016)
“Do you think that people in government waste a lot of the money we pay in taxes, waste some of it,
or don’t waste very much of it?”
“Would you say the government is pretty much run by a few big interests looking out for themselves
or that it is run for the benefit of all the people?”
“Do you think that quite a few of the people running the government are crooked, not very many are,
or do you think hardly any of them are crooked?” (before 2012); “How many of the people running
the government are corrupt? All, most, about half, a few, or none?” (2012 and later)
Sources: American National Election Studies Cumulative Data File, 1948–2012 and American National Election Studies 2016 Data File,
https://electionstudies.org/data-center/
Although the ANES questions are often used by scholars, many acknowledge they are not perfect
measures of trust.17 The first item, with its explicit focus on trusting “the government in Washington,”
most closely resembles the notion of trust as a characteristic of the political regime. The other items,
however, appear to tap specific attitudes toward politicians’ actions and motives. Because of this,
argued Jack Citrin, an early critic of the ANES measures, the survey questions register “mere
disapproval of incumbent political leaders” rather than “alienation from the political regime.”18 This
seems especially likely because the attitude objects that are the focus of these questions alternate
between the general government and politicians (“the government in Washington” vs. “the people
running the government,” for example). In place of these survey items, Citrin preferred questions asking
respondents whether they favored changing the form of government or whether they were proud about
the form of government. For Citrin, such questions better tap individuals’ views of the political regime
than do the ANES items. In contrast, because the ANES questions specifically mention politicians, Citrin
viewed these as measures solely of specific support. Most public opinion researchers, however, accept
https://electionstudies.org/data-center/
that the ANES items tap a combination of diffuse and specific support.19 Finally, this debate is about
more than simply how to measure trust in government on a public opinion survey. As we see later in this
chapter, disagreement also exists over how to interpret public attitudes toward government when these
survey questions are used.
Decline in Public Trust
One benefit of the ANES trust questions is that they have been asked of the American public since
1964, allowing us to examine levels of public trust over many decades. The graph in Figure 11.1
displays public responses to the four ANES items in presidential election years since 1964. The most
obvious conclusion to be drawn from this figure is that public trust has declined considerably. Nearly 80
percent of the public felt that government could be trusted to do the right thing most of the time or just
about always in 1964 (refer to the solid line). This percentage steadily declined to 26 percent in 1980.
Public trust increased quite a bit by 1984, with 45 percent of the public trusting the government, but then
declined in 1988 and 1992. Between 1992 and 2004 slightly and steadily increasing percentages of the
public demonstrated trust in the government. Trust has decreased since then. In 2012 only 22 percent of
Americans trusted the government just about always or most of the time, far lower than in 1964. In 2016
this percentage dropped further to 11.5, the lowest level ever recorded for the ANES. Some of this
decline is undoubtedly due to the question wording changes noted earlier.20 Specifically, the 2016
survey offered a new option of trusting the government “about half the time,” which was nestled between
the previous options “most of the time” and “some of the time.” Some respondents who might have
chosen “most of the time” in previous years likely selected “about half the time” in 2016 when explicitly
given that option. Having said that, the drop in trust recorded by the 2016 ANES is probably not only an
artifact of the questions asked that year. Other polling agencies, which had not changed their questions,
recorded near historic low levels of public trust in government between 2016 and 2019.21 Finally, turning
to the other three trust questions (the dashed lines in Figure 11.1), the overall trends are very similar.
Significant increases in negative evaluations occurred between 1964 and 1980, followed by somewhat
more positive assessments in the mid-1980s and between 1992 and 2004. Evaluations have been more
negative since then.
Description
Figure 11.1 Public Trust in Government, 1964–2016
Sources: Analysis of American National Election Studies Cumulative (1948–2004), American National
Election Studies 2008, American National Election Studies 2012, and American National Election Studies
2016 data files.
Note: For 2016, the result for “Trust government most of the time/just about always” corresponds to “Trust
government most of the time/always.” For 2012 and 2016, the result for “Not many/hardly any government
officials are crooked” corresponds to “A few/no people running the government are corrupt.”
What explains this variation in public trust over time? The large and steady decline in trust that ended in
1980 began, in part, with the tumultuous events that occurred during the 1960s and 1970s. U.S.
involvement in and ultimate withdrawal from Vietnam demonstrated the inability of the American military
to succeed in this conflict. The urban uprisings of the late 1960s were visible manifestations that the
government was incapable of preventing social unrest. The Watergate scandal further quickened the
pace of decreasing public trust. The burglary of the Democratic National Committee headquarters in the
Watergate complex occurred in 1972. This was followed two years later by President Richard Nixon
resigning his office among allegations of abuse of power and trying to cover up his involvement in the
burglary. Thus, it is perhaps not surprising that the largest four-year decline in the trust-in-government
item occurred between 1972 and 1976, when the percentage of the public who believed government
could be trusted always or most of the time fell from 54 to 35 percent.
Presidential scandals are also cited to explain later trust declines, such as those witnessed between
1984 and 1988 and again at the beginning of Bill Clinton’s first term in office.22 In 1986, the public
became aware that members of President Ronald Reagan’s administration had sold arms to Iran to help
fund a group (the Contras) fighting the communist government in Nicaragua. Details relating to these
events, known as the Iran-Contra scandal, dogged Reagan for the rest of his time in office. Clinton was
embroiled in a number of scandals shortly after assuming office, including allegations that he and his
wife Hillary were involved in a shady real estate deal (the Whitewater development), the firing of the
entire White House travel office, and continued rumors of his marital infidelities. These scandals
contributed to a decline in trust between 1992 and 1994. By 1996, however, the news was dominated by
other topics, and trust in government had increased. Notably, public trust in government did not drop
during President Clinton’s biggest scandal—an affair with White House intern Monica Lewinsky, which
eventually led to an unsuccessful effort to remove him from office. Why might that be the case?
According to one political scientist, President Clinton’s approval ratings withstood the scandal because
the economy was strong, the country was at peace, and Clinton pursued moderate policies.23 Perhaps
these factors also kept trust in government from declining. As we discuss shortly, good economic times
and moderate policies can increase trust.
Other explanations for the long-term decline in trust include changes in public expectations of
government. In particular, Jane Mansbridge focuses on broader societal and cultural changes to
explain why government increasingly does not meet public expectations.24 Societal changes force new
problems onto the government’s agenda, problems that the public expects the government to address
but that the government might not necessarily be equipped to solve. When these problems persist over
time, as Mansbridge argues has been the case with crime and childhood poverty, government is
perceived as being unable to solve problems. Public distrust results.
Improved economic conditions increased trust in the 1980s. The high inflation and high unemployment
rates of the 1970s turned around in the 1980s. Public perceptions of the government’s handling of the
economy became more favorable, leading to increased trust in government.25 Economic performance
and the public’s evaluations of the government’s economic policies have been correlated with public
trust at other times as well. Declines in trust often accompany a poorly performing economy, as was the
case in the 1970s and during President George H. W. Bush’s term in office (1989–1993), whereas trust
rebounds when the economy improves (such as between 1992 and 2000).26 The impact of the economy
on trust is not symmetrical, however. Trust does indeed increase when the economy improves, yet trust
declines more significantly when the economy is poor. Why? “When the economy is performing badly,
more people care about it, which increases its (negative) effect on trust. Since fewer people care about
the economy when times are good, trust increases less in periods of prosperity than it decreases in
periods of weakness.”27
Furthermore, the state of the national economy and public trust in government do not always move in
tandem. Economic conditions as well as public perceptions of the economy worsened in 2000, for
instance, yet the increase in trust that had begun in the mid-1990s continued. This was largely because
of the September 11, 2001, attacks and the attention placed on terrorism, foreign policy, and domestic
security that resulted. Upsurges in trust follow events that place public attention squarely on terrorism
and foreign affairs, such as the 9/11 attacks themselves or the 2002 congressional resolution giving
President George W. Bush the authority to begin the war in Iraq.28 It is not only during the post-9/11 era
when we see that trust is influenced by foreign affairs. Over the past few decades, whenever the public
has been especially concerned about global issues and national security, trust in government has
increased.29 This is largely due to the propensity of the public to rally around the president during times
of crisis or threat and to the public’s positive evaluation of the U.S. military.
How can we account for the recent drop in trust? Analyzing this downward trend in 2010, the Pew
Research Center identified “a perfect storm of conditions associated with distrust of government—a
dismal economy, an unhappy public, bitter partisan-based backlash, and epic discontent with Congress
and elected officials.”30 Clearly, elements of this perfect storm have persisted, given that public trust has
declined further since 2010, even in the face of improved economic conditions. As we indicated in the
opening of this chapter, there has been much to dislike about the activities of national politicians of late.
In fact, the lowest levels of trust in government that Pew has recorded came about during the debt
ceiling debate of 2011, the 2013 government shutdown, and shortly after the 2019 shutdown.31
Contemporary distrust was especially high among supporters of the Tea Party movement and those
who voted for Donald Trump in 2016. For instance, in 2013, 80 percent of the overall public trusted the
federal government only some of the time or never, whereas fully 96 percent of Tea Party supporters
expressed this view.32 The Tea Party movement emerged in 2009, shortly after President Obama’s
inauguration. For the purposes of public opinion, this movement is best described as a set of not always
well-connected local groups, the members of which possess conservative views toward government.33
Demographically, citizen Tea Partiers are overwhelmingly white, majority male, and typically over the
age of fifty.34 As for why their trust in government was lower than that of the broader public during the
Obama administration, one analysis pointed to two key factors: grave unhappiness with the direction of
the country and the belief that public policies were aiding the poor rather than middle or upper
classes.35 As for Trump voters, one political scientist summed up the situation as follows: “years of
mounting antigovernment politics came to fruition in 2016 when a candidate with no experience in
government or politics secured the [Republican] nomination … [then] soared to victory in the Electoral
College, his rise fueled by those who had grown especially hostile toward government.”36
Sources of Trust
Citizens’ trust in government arises from a variety of sources. We encountered some of these in the
preceding discussion of variation in public trust over time, such as real-world events and government
activities, the state of the economy, and assessments of governmental economic policy performance.
Public trust is also related to evaluations of governmental policymaking across many other domains. In
other words, people’s assessments of the products of government (especially public policies) influence
whether they trust the government. Generally speaking, distrust rises for citizens whose policy
preferences are furthest from the policy decisions enacted by leaders or the platforms of the parties.
Between 1964 and 1970, this meant that people with strongly liberal or strongly conservative views were
the most distrusting because government policy was increasingly centrist, or moderate, rather than
either liberal or conservative.37 By the 1990s, however, both political parties had become more
ideologically extreme. Because the ideological and policy preferences of the citizens did not change
substantially, as the parties drifted more to the extremes, centrist citizens became the ones who felt
removed from the parties’ goals and who trusted government less.38 This dynamic also has partisan
links. Democratic citizens are most trusting of government when the president is Democratic and vice
versa among Republicans, likely because citizens are more pleased with national policy goals when
their party is in power. The partisan gap in trust has widened recently, driven by polarization among
party elites and by the public’s growing animosity toward members of the opposition party.39
Public attitudes toward government are also shaped by messages from the political environment. When
news media coverage becomes more cynical toward leaders and more questioning of politicians’
motives, as has been the case over the past few decades,40 public trust declines. Not all news outlets
present the same level of cynicism toward politics, of course, but people who read more critical
newspapers are less trusting of government.41 Other changes to the media landscape also have
consequences for political trust. For instance, political television is more likely today than in the past to
contain uncivil political discussions. As it turns out, “watching politicians and pundits hurl insults at one
another on television” results in lower trust in both government and politicians.42
The media are not the only source of information about government. Political leaders themselves
communicate their views about the government to the public. It is not uncommon for candidates,
especially newcomers to national politics, to run for office by criticizing the government and portraying
themselves as outsiders. This was the successful strategy pursued by Jimmy Carter, former governor of
Georgia, when he ran for president in 1976 and by Donald Trump in 2016. Sitting presidents have also
been known to speak harshly about government. Ronald Reagan did during his first inaugural address
when he stated, “Government is not the solution to our problems; government is the problem.”43 Trump
often does, such as by criticizing judges and courts, alleging that civil servants and law enforcement
officials are part of an entrenched “Deep State,” or warning against “the steady creep of government
bureaucracy that drains the vitality and wealth of the people.”44
With such attacks on government delivered by politicians, it is perhaps not surprising that public trust is
low. Could a more positive tone toward government increase trust? Two experiments address this
question. In one, respondents who had just answered questions about the federal government’s ability
to run specific programs (the military, Social Security, and Medicare) displayed greater overall
confidence in government than did those who had not yet answered the specific program questions.45
The other experiment demonstrates that trust can increase, albeit slightly, after people are exposed to
statements placing government in a favorable light, such as by emphasizing that government
regulations protect citizens. On the other hand, statements highlighting government inefficiency or
incompetency decrease public trust, often more so than the favorable statements increase trust.46
These studies suggest that if elite and media communication more often contained details of well-
regarded government policies and outcomes, levels of trust among the public could rise, provided that
antigovernment rhetoric is not also present.
Does actual experience with government programs produce more trust in government? Not so much,
concludes Suzanne Mettler in her book The Government-Citizen Disconnect. She comprehensively
tracked Americans’ use of a variety of federal social programs in the aggregate, noting that usage
increased in recent decades at the same time that overall public trust declined. To determine whether
being a policy beneficiary is related to individuals’ views toward government, she conducted a survey in
2008 with follow-up interviews in 2015. Mettler focused on twenty-one different federal policies that vary
along two characteristics.47 Government provision of benefits is visible for some programs (e.g., Social
Security, Pell Grants for college education). Government provision is more hidden for others, typically
because the benefit is received via a reduction in federal income taxes rather than as a direct payment
(e.g., home mortgage interest tax deduction, earned income tax credit). The other characteristic relates
to whether the program applies to everyone or whether it is means-tested, meaning that income
eligibility requirements exist to qualify for benefits. For each pair of examples provided earlier in
parentheses, the first policy is non-means-tested while the second is means-tested.
Of these four types of policies, Mettler concludes that personal experience with federal programs is
associated with more favorable attitudes toward government only among beneficiaries of visible
programs that are means-tested.48 The greater number of such programs that people use, the more
favorable their attitudes are. Assessments of the government are actually more negative the more
means-tested hidden programs people access. Thus, policy design matters, in that citizens are more
likely to credit the government for helping them if the benefits they receive are obviously being provided
by the government. In contrast, recipients of means-tested hidden programs, most notably the EITC,
describe these programs as “a service not affiliated with the government” or as an incentive that gives
“people a reward for using their money the right way.”49 Furthermore, some beneficiaries of the EITC
and other programs embedded in the tax code contrast these programs with those that are visible and
means-tested, characterizing the latter as “giving money to those who don’t want to work” or “creating
an entitlement generation, not a generation of workers.”50 Mettler concludes that such attitudes
significantly contribute to negative evaluations of government among those using means-tested hidden
programs.
Another factor that is not strongly related to people’s trust in government is their demographic
characteristics. Beginning with the initial analyses of the ANES trust items and continuing on to more
recent times, many scholars have found that levels of trust do not differ substantially across subgroups
of people.51 As you can see from Figure 11.2, in 2016 levels of trust in government were not
significantly different for women and men. Trust levels vary a bit across age groups as well as across
levels of education and income, although the differences are not that large. Younger adults, those who
have not graduated from high school, and people with annual incomes less than $40,000 were the most
trusting.
Racial and ethnic differences in trust are larger than for other demographic groups. Whites were the
least trusting of government, followed by Blacks and then Hispanics. One explanation for Hispanics’
higher level of trust hinges on the more recent arrival in the United States of some Hispanics. Newly
arrived immigrants generally have more positive views toward the government than do individuals who
have been in the country longer. Because a larger proportion of Hispanics are foreign-born than are
whites and Blacks, trust levels among Hispanics have been, on average, higher than among the other
two groups.52 We see this same trend in 2016, but only when Hispanics are compared to whites. Unlike
in the past, there are not large differences in trust between Hispanics and African Americans, primarily
because trust in government increased among Blacks when Barack Obama was president. Historically,
Blacks tended to be less trusting of government than either whites or Hispanics.53 Most explained this
difference using the political reality model.54 With less political power than whites in the United States
and a government less responsive to their needs, Blacks were more distrustful. As the face of
government literally changes, however, so can levels of trust.55 More broadly, and not surprisingly given
that Blacks are much more likely to be Democratic than Republican, Blacks display higher trust in the
government when the president is Democratic. Since Donald Trump’s inauguration, Blacks’ trust in
government has plummeted.56
Trust and the Criminal Justice System
Citizen experiences with the criminal justice system also influence trust in government, typically by
lowering it. The more serious the contact with the justice system, the lower the trust. For example, a
national survey of adults in their twenties demonstrates that, for those who had not had any encounters
with the criminal justice system, only 18 percent distrusted the federal government.57 But nearly 30
percent of the young adults who had been questioned by police or arrested were distrustful. Being
convicted of a crime or serving jail time further increased distrust (conviction, 31 percent distrustful;
incarceration, 43 percent; and more than one year in jail or prison, 55 percent). The authors of this study
argue that “punitive encounters with the state foster mistrust of political institutions,” and their results
demonstrate that the mistrusting attitudes extend to national political institutions and actors beyond the
police officers, local courts, and correctional facilities of the criminal justice system.58
Description
Figure 11.2 Trust in Government for Specific Demographic Groups, 2016
Source: Analysis of American National Election Studies 2016 data file.
Note: For these comparisons, we constructed an index of trust by adding up people’s responses to the four
ANES trust items. The possible range of the index was from 4 to 15, with higher numbers indicating more
trust. The average for all respondents was 7.83.
What about trust in the criminal justice system? One notable pattern is that trust and confidence in law
enforcement are lower among Blacks than among other racial groups. These racial differences are not
new,59 although they have received renewed attention in the years since the 2014 shooting death of
Michael Brown in Ferguson, Missouri. A Pew survey fielded in 2016 found that, on average, only 14
percent of African Americans had a lot of confidence in their local police, compared to 31 percent of
Hispanics and 42 percent of whites (see Figure 11.3). The 2016 Pew poll also assessed confidence in
two specific police activities: treating racial and ethnic groups equally and using the right amount of
force. For both, Blacks were least and whites most likely to view police performance as excellent or
good in these domains. The gap between these racial groups was substantial, with nearly three-quarters
of whites assessing the police favorably compared to only one-third of Blacks.
Description
Figure 11.3 Racial Differences in Views Regarding the Police, 2016
Source: Data from Rich Morin and Renee Stepler, “The Racial Confidence Gap in Police Performance,” Pew
Research Center, Washington, DC, September 29, 2016, https://www.pewsocialtrends.org/2016/09/29/the-
racial-confidence-gap-in-police-performance/
Note: Bars represent the percentage having “a lot” of confidence for the first item and viewing the police as
doing an “excellent” or “good” job for the other two items.
https://www.pewsocialtrends.org/2016/09/29/the-racial-confidence-gap-in-police-performance/
Implications of Declining Public Trust
What are we to make of the decline in public trust since the 1960s? Answers to this question depend, in
part, on what you think the trust survey questions are measuring. Thus, we return to the debate between
Arthur Miller and Jack Citrin. Recall that Miller believes the ANES questions tap into system-level
attitudes toward government. For him, then, a sustained decline in public trust suggests an unhealthy
democracy and the potential for citizens to demand radical change to the democratic political system.60
In contrast, according to Citrin such consequences are unlikely to result from declines in trust as
measured by the ANES because these measures primarily register assessments of incumbent
politicians rather than evaluations of the broader political system. To Citrin, expressions of cynicism are
“ritualistic rather than genuine … to agree verbally that many people ‘running the government’ are
corrupt, incompetent, or untrustworthy is like shouting ‘Kill the umpire!’ at a baseball game.”61
Even with sustained low levels of public trust in government, there is little evidence that the serious
consequences predicted by Miller and others have resulted. Public levels of patriotism remain high, and
pleas for radical changes to the governmental system are uncommon. Despite registering very low
levels of public trust, a 2019 Pew survey also recorded that Americans are not so much angry with the
federal government as frustrated.62 This, however, does not mean that cynical attitudes are largely
ritualistic and without any effects. Research conducted by Marc Hetherington, in two different projects,
demonstrates a variety of repercussions.
One of Hetherington’s arguments is that declining public trust has contributed to the increase in
conservative public policies adopted by the national government.63 Liberal solutions to public problems
generally involve government, such as the provision of federal money and subsidies to fight poverty.
Conservatives, in contrast, tend to prefer minimal government involvement in the economic sector, so
they generally favor policies that pursue this goal. Such conservative policies have been more common
in the United States since the 1960s, especially in contrast to President Lyndon Johnson’s Great Society
legislation, which established a variety of government programs to fight racism and poverty.
Hetherington demonstrates that this conservative turn in public policy did not occur because citizens’
attitudes became more conservative. Instead, he finds that public opinion has fluctuated toward social
welfare policies that require most people to sacrifice so that others (such as the poor and racial
minorities) will benefit. And these fluctuations have been largely driven by levels of public trust. When
trust is higher, the public supports liberal antipoverty and racial policies. As trust declines, public support
for liberal policies is much lower. In the end, Hetherington shows that decades of increasing public
distrust in government have “undermined public support for federal programs like welfare, food stamps,
and foreign aid, not to mention the entire range of race-targeted programs designed to make equality
between the races a reality.”64
Hetherington’s other research project, which he conducted with Thomas Rudolph, concludes that low
levels of public trust make governing difficult.65 This is especially the case when trust is polarized along
partisan lines, as it is today. Republican citizens are particularly likely to withhold trust from Democratic
leaders. Democrats repay the favor when Republicans are in charge. Through analyses of the 2009
economic stimulus legislation and the 2010 Affordable Care Act, Hetherington and Rudolph demonstrate
that low and polarized trust constrain the policy choices of the president and Congress. For example,
although the stimulus plan was passed to bring the economy out of recession, it did not contain the tools
that many policy experts would recommend to meet this goal. The plan consisted of government
spending and tax cuts, although the former is more likely to actually stimulate the economy than the
latter. Yet, a plan consisting of only spending was impossible, given low public trust (especially among
Republican citizens) and the accompanying reluctance to support liberal policies.66 More broadly,
Hetherington and Rudolph conclude that compromise becomes difficult during times of polarized trust
because members of the minority party in Congress have no electoral incentive to work with the majority
party to craft legislation. Their constituents do not trust the party in power, after all, and thus do not want
their legislators to work with members of the majority party. Gridlock and stalemate result.
Finally, some have argued that public distrust is a good thing, so we should not worry about any
negative implications. As some democratic theorists posit, suspicion of those in power can be healthy
for democracy, and thus declining trust might be a reasonable reaction to government actions.67
Further, public distrust in the past has contributed to the inclusion of new groups into democratic
decision making. Political movements such as the women’s movement, the civil rights movement, and
the environmental movement rose up to challenge existing power arrangements, in part because
members of these groups lost trust in the government’s willingness to address the groups’ concerns.68
Increasing public distrust has also been mentioned as an indicator of the level of development of a
nation. When a society’s basic needs must be met by the government, trust in the government is
expected to be high. Yet, “as material well-being increases, trust in political institutions and elites is likely
to decline as publics begin to evaluate their leaders and institutions by more demanding standards.”69
Box 11.1 Levels of Public Trust in Other Nations
Public Opinion in Comparative P
