PUBH 6012: Advanced Public Health Research Methods
Exam 2: Study Design, Biostatistics, Causal Inference, and Research Ethics
2026–2027 Academic Year Update – 100% Accuracy (Professor-Verified)
Graduate Program in Public Health
Department of Population Health Sciences
Date of Examination: June 2026
Total Questions: 30 | Points per Question: 3.33 | Total Points: 100
Abstract
This examination assesses advanced graduate-level competency in public health research
methods, biostatistics, epidemiological study design, and causal inference as covered in
PUBH 6012. The exam comprises 30 multiple-choice questions distributed across four
sections: (1) Advanced Epidemiological Study Designs and Causal Frameworks (8
questions), addressing DAGs, nested case-control, crossover trials, marginal structural
models, and AI-assisted systematic reviews; (2) Biostatistics, Measures of Association, and
Precision (8 questions), covering hazard ratios, attributable risk, confidence intervals, odds
ratio approximation, and meta-analysis heterogeneity; (3) Quantitative Data Analysis,
Hypothesis Testing, and Power (7 questions), encompassing ANOVA, power analysis,
intention-to-treat, sample size, and Kaplan-Meier survival analysis; and (4) Research
Ethics, Validity, Bias, and 2026 Methodological Updates (7 questions), integrating effect
modification, selection bias, propensity score matching, and NIH/ICMJE open science
standards. All statistical interpretations and methodological justifications are verified to
peer-reviewed accuracy.
Keywords: causal inference, DAGs, marginal structural models, hazard ratio, odds ratio, attributable
risk, confidence intervals, ANOVA, statistical power, intention-to-treat, Kaplan-Meier, effect
modification, propensity score matching, Berkson’s bias, healthy worker effect, ICMJE, open
science, 2026 research methods
,Examination Information
Course PUBH 6012 – Advanced Public Health Research Methods
Exam Exam 2: Study Design, Biostatistics, Causal Inference, and Research Ethics
Date June 16, 2026
Total Questions 30 (4 sections: 8, 8, 7, 7)
Points per Question 3.33 points
Total Points 100 points
Format Multiple-choice (A–D) with rationales
Accuracy 100% Professor-Verified
Section 1: Advanced Epidemiological Study Designs and Causal Frameworks (Q1–Q8)
Section 2: Biostatistics, Measures of Association, and Precision (Q9–Q16)
Section 3: Quantitative Data Analysis, Hypothesis Testing, and Power (Q17–Q23)
Section 4: Research Ethics, Validity, Bias, and 2026 Methodological Updates (Q24–Q30)
, Section 1: Advanced Epidemiological Study Designs and Causal
Frameworks
Q1: A Directed Acyclic Graph (DAG) shows that variable C is a common cause of both
exposure A and outcome B, with no direct arrow from A to B. In this causal structure, C
is classified as:
A. A mediator on the causal pathway from A to B
B. A confounder that must be controlled to block the non-causal (backdoor) association
between A and B
C. A collider that should not be controlled to avoid collider stratification bias
D. An instrumental variable that can be used for causal estimation
Q2: A nested case-control study is conducted within a large prospective cohort of 50,000
nurses. For each case of breast cancer that develops, four controls are selected from
cohort members who are still at risk at the time the case is diagnosed. The primary
advantage of this design over a traditional case-control study is:
A. It eliminates recall bias because exposure data were collected before disease onset
B. It eliminates all forms of selection bias in the control group
C. It provides a direct measure of disease incidence in the source population
D. It allows causal inference without any consideration of confounding
Q3: A crossover trial compares Drug X to placebo for chronic pain. Each participant
receives both treatments in random order with a 4-week washout period between them.
Which assumption is most critical for the validity of this design?
A. Participants must have equal baseline pain severity
B. There must be no carryover effect of Drug X into the placebo period
C. The sample size must be at least 100 to achieve adequate power
D. Drug X must be administered before placebo in all participants
Q4: A retrospective cohort study examines the association between occupational asbestos
exposure and lung cancer using employment records from 1970 to 2020. The most
significant limitation of this design compared to a prospective cohort is:
A. It cannot calculate a relative risk because incidence cannot be determined
B. It is vulnerable to information bias because exposure and outcome data may be incomplete
or inaccurately recorded in historical records
C. It cannot assess multiple outcomes from a single exposure
D. It is faster and less expensive than prospective cohorts, which reduces validity
Q5: In a DAG, a variable that has two arrows pointing into it from two other variables is
called a:
A. Confounder
B. Mediator
C. Collider
D. Effect modifier
Q6: Marginal structural models (MSMs) are used in epidemiology primarily to:
A. Eliminate all measurement error from exposure data
Exam 2: Study Design, Biostatistics, Causal Inference, and Research Ethics
2026–2027 Academic Year Update – 100% Accuracy (Professor-Verified)
Graduate Program in Public Health
Department of Population Health Sciences
Date of Examination: June 2026
Total Questions: 30 | Points per Question: 3.33 | Total Points: 100
Abstract
This examination assesses advanced graduate-level competency in public health research
methods, biostatistics, epidemiological study design, and causal inference as covered in
PUBH 6012. The exam comprises 30 multiple-choice questions distributed across four
sections: (1) Advanced Epidemiological Study Designs and Causal Frameworks (8
questions), addressing DAGs, nested case-control, crossover trials, marginal structural
models, and AI-assisted systematic reviews; (2) Biostatistics, Measures of Association, and
Precision (8 questions), covering hazard ratios, attributable risk, confidence intervals, odds
ratio approximation, and meta-analysis heterogeneity; (3) Quantitative Data Analysis,
Hypothesis Testing, and Power (7 questions), encompassing ANOVA, power analysis,
intention-to-treat, sample size, and Kaplan-Meier survival analysis; and (4) Research
Ethics, Validity, Bias, and 2026 Methodological Updates (7 questions), integrating effect
modification, selection bias, propensity score matching, and NIH/ICMJE open science
standards. All statistical interpretations and methodological justifications are verified to
peer-reviewed accuracy.
Keywords: causal inference, DAGs, marginal structural models, hazard ratio, odds ratio, attributable
risk, confidence intervals, ANOVA, statistical power, intention-to-treat, Kaplan-Meier, effect
modification, propensity score matching, Berkson’s bias, healthy worker effect, ICMJE, open
science, 2026 research methods
,Examination Information
Course PUBH 6012 – Advanced Public Health Research Methods
Exam Exam 2: Study Design, Biostatistics, Causal Inference, and Research Ethics
Date June 16, 2026
Total Questions 30 (4 sections: 8, 8, 7, 7)
Points per Question 3.33 points
Total Points 100 points
Format Multiple-choice (A–D) with rationales
Accuracy 100% Professor-Verified
Section 1: Advanced Epidemiological Study Designs and Causal Frameworks (Q1–Q8)
Section 2: Biostatistics, Measures of Association, and Precision (Q9–Q16)
Section 3: Quantitative Data Analysis, Hypothesis Testing, and Power (Q17–Q23)
Section 4: Research Ethics, Validity, Bias, and 2026 Methodological Updates (Q24–Q30)
, Section 1: Advanced Epidemiological Study Designs and Causal
Frameworks
Q1: A Directed Acyclic Graph (DAG) shows that variable C is a common cause of both
exposure A and outcome B, with no direct arrow from A to B. In this causal structure, C
is classified as:
A. A mediator on the causal pathway from A to B
B. A confounder that must be controlled to block the non-causal (backdoor) association
between A and B
C. A collider that should not be controlled to avoid collider stratification bias
D. An instrumental variable that can be used for causal estimation
Q2: A nested case-control study is conducted within a large prospective cohort of 50,000
nurses. For each case of breast cancer that develops, four controls are selected from
cohort members who are still at risk at the time the case is diagnosed. The primary
advantage of this design over a traditional case-control study is:
A. It eliminates recall bias because exposure data were collected before disease onset
B. It eliminates all forms of selection bias in the control group
C. It provides a direct measure of disease incidence in the source population
D. It allows causal inference without any consideration of confounding
Q3: A crossover trial compares Drug X to placebo for chronic pain. Each participant
receives both treatments in random order with a 4-week washout period between them.
Which assumption is most critical for the validity of this design?
A. Participants must have equal baseline pain severity
B. There must be no carryover effect of Drug X into the placebo period
C. The sample size must be at least 100 to achieve adequate power
D. Drug X must be administered before placebo in all participants
Q4: A retrospective cohort study examines the association between occupational asbestos
exposure and lung cancer using employment records from 1970 to 2020. The most
significant limitation of this design compared to a prospective cohort is:
A. It cannot calculate a relative risk because incidence cannot be determined
B. It is vulnerable to information bias because exposure and outcome data may be incomplete
or inaccurately recorded in historical records
C. It cannot assess multiple outcomes from a single exposure
D. It is faster and less expensive than prospective cohorts, which reduces validity
Q5: In a DAG, a variable that has two arrows pointing into it from two other variables is
called a:
A. Confounder
B. Mediator
C. Collider
D. Effect modifier
Q6: Marginal structural models (MSMs) are used in epidemiology primarily to:
A. Eliminate all measurement error from exposure data
