PM 510 Homework 7 (Week 8)Problems not marked with [SPSS] should be done by hand, as problems similar to that could be on the
Midterm and/or Final. (Of course, feel free to check your work using SPSS!)
Problems marked with [SPSS] are intended to be done with SPSS. For these problems, please
provide the edited output file (e.g., Word).
1. For each of the scenarios described below, identify the study design and explain why you think it is
the design you selected. In your explanation, be sure to identify the exposure and outcome for each
study, if that is applicable.
(a) A group of 100 women were identified from hospital records for post-partum hemorrhage
(PPH), and their level of physical activity during pregnancy was then assessed. The same
information was obtained on the mothers of a sample of 100 women admitted for childbirth in
the same hospital during the same time period without any history of PPH. No clear
associations were found between prenatal physical activity level and PPH.
(b) To study the effect of postmenopausal hormone therapy on cardiovascular disease risk, 16,608
postmenopausal women aged 50–79 years were randomly assigned to receive either
postmenopausal hormone therapy or placebo. While a major reduction in coronary heart
disease incidence in the hormone therapy arm had been hypothesized on the basis of a substantial
body of observational research, this study found an elevation in coronary heart disease risk
over an average of 5.6 years of follow up.
(c) The measures of per capita alcohol consumption from 15 countries in the European Union were
tested for correlation with the percentage of liver disease among men in those same countries.
Alcohol consumption was positively correlated with the percentage of men reporting liver
disease in these data.
(d) According to World Health Organization (WHO), poliomyelitis is considered to be eradicated
from South Asia. On June 10, 2015, a physician in Nepal encountered a young boy with history
of fever followed by paralysis of limbs. His initial diagnosis was that it was a case of
poliomyelitis. The details of this observation were published in an article.
(e) A group of 57,000 Swedish women completed a self-administered questionnaire that included
demographic characteristics, tobacco and alcohol use, coffee consumption, participation in
sports or exercise, female reproductive history, hormone use, and dietary habits. They were
followed for breast cancer incidence for 10 years, and at the end of the follow up, the
association of alcohol consumption and breast cancer was investigated.
(f) A study on eye disease was conducted among Latinos of Los Angeles in 2003. A report from
the study describes a prevalence of 23% for diabetic retinopathy.
(g) A report from a study looked at the association of glycosylated hemoglobin and diabetic
retinopathy. Data on diabetic retinopathy and blood level of glycosylated hemoglobin was
collected roughly during the same time.
(h) In 1996, a study was conducted in Texas to study the association between diet and renal cell
carcinoma (RCC). A total of 203 cases of adults with RCC, aged between 18–50 years, were
included in that study. Additionally, 203 healthy adults without RCC living within one mile
of the patients were recruited to participate in the study. In a phone interview, investigators
obtained information on diet and physical activity from the respondents. No association was
found between any dietary factors and renal cell carcinoma risk.
2. A city is worried about the water quality in several buildings in the downtown area. They believe
that something in the water in those buildings might be causing irritation in the esophagus leading
to excessive cases of heartburn. They perform a survey of 200 people who work in the downtown
area and discover the following data:
Have heartburn
Work in
relevant
buildings
Yes
No
Total
Yes
15
5
20
No
97
83
180
Total
112
88
200
(a) What kind of study is this?
(b) What are the exposure and outcome in this study?
(c) What is the prevalence of working in the relevant buildings? What is the prevalence of
heartburn?
(d) Perform a statistical test for association between exposure and outcome and write your
conclusion. What is the odds ratio?
(e) The city does a case-control study with 200 subjects, taking an equal number of cases and
controls. They discovered 14% of the cases were exposed and 6% of the controls were
exposed. Construct the 2×2 contingency table, perform a statistical test for association between
exposure and outcome, and write your conclusion. What is the odds ratio?
(f) The city does a cohort study with 200 subjects, taking an equal number of exposed and
unexposed individuals who do not yet have heartburn. They discover 75% of the exposed
subjects develop heartburn over the next year, compared to 54% of the unexposed subjects.
Construct the 2×2 contingency table, perform a statistical test for association between exposure
and outcome, and write your conclusion. What is the odds ratio?
(Assume that OR is a good estimate of the relative risk in this study. You will learn other
measures of the relative risk in cohort studies in other classes. Also see question (h). )
(g) Compare the odds ratios and p-values you obtained for all three studies.
(h) For a cohort study, the odds ratio is not typically used to describe the effect size; measures
that are usually used instead are the rate ratios and the risk ratios. The risk ratio is defined
as the proportion of exposed subjects who have the outcome of interest divided by the
proportion of unexposed subjects who have the outcome. Calculate the risk ratio for the
cohort study from part (f) and compare it to the odds ratio.
3. Describe the differences between simple randomization, adaptive randomization, and block
randomization.
PM 510 Homework 8 (Week 9)
Problems not marked with [SPSS] should be done by hand, as problems similar to that could be
on the Midterm and/or Final. (Of course, feel free to check your work using SPSS!)
Problems marked with [SPSS] are intended to be done with SPSS. For these problems, please
provide the edited output file (e.g., Word).
1. [SPSS] In a survey conducted in Italy, physicians with different specialties (internal medicine,
surgery, radiotherapy, oncology, and gynecology) were questioned regarding the surgical
treatment of early breast cancer. In particular, they were asked whether they would recommend
radical surgery regardless of patient’s age (R), conservative surgery only for younger patients
(CR), or conservative surgery regardless of age (C). The results are saved in the file
HW_8_data_1.sav. At the 0.05 level of significance, test the null hypothesis that there is no
association between physician specialty (variable name: physician_specialty) and recommended
treatment (variable name: surgery).
(a) Write the null and alternative hypotheses.
(b) Fill in the table below with the expected value for each cell. (Show the detailed work
from manual calculation for at least 2 of the cells; you can get all other expected counts
from SPSS.)
Internal medicine
Surgery
Radiotherapy
Oncology
Gynecology
Total
R
Surgery
CR
C
33
110
297
Total
70
211
59
56
44
440
(c) What are the degrees of freedom?
(d) What is the value of the χ2-test statistic?
(e) What is your conclusion?
(f) From the SPSS output, does Fisher’s Exact Test need to be used in this scenario? Why or
why not?
(g) What is the Fisher’s Exact Test p-value? Does SPSS provide the requested Fisher’s Exact
Test p-value?
2. [SPSS] Use the Lecture8_Peritonitis_v2.sav data set to answer the following questions.
(a) Is there an association between sex and occurrence of peritonitis?
(b) Is there an association between sex and occurrence of peritonitis, after controlling for age?
For the purpose of this analysis use the age_group variable where 1 = young (≤ 50) and 2
= old (> 50).
(c) Is there any evidence of a statistically significant difference between the stratum-specific
odds ratios?
3. The calculations for the Mantel-Haenszel test statistic for the association between severe vertex
balding and myocardial infarction are presented in the paper by Lesko et al. (Table 7). Perform
the calculations necessary to fill the summary table given below. Calculate the appropriate test
statistic.
Summary Table:
Age Group
< 45
45–49
50+
ai
ei
vi
