CPH: CERTIFIED IN PUBLIC HEALTH
PRACTICE EXAM 2026–2027
(200 Original Practice Questions with Correct Answers & Rationales)
Public Health | NBPHE/ASPPH-Aligned Domain Practice
Epidemiology • Biostatistics • Environmental Health • Health Policy & Management • Social & Behavioral Sciences • Public Health Biology
IMPORTANT DISCLAIMER: This is an ORIGINAL educational practice resource for study purposes only. It is NOT an official CPH examination, does NOT
contain proprietary NBPHE items, and is NOT affiliated with NBPHE, ASPPH, or CEPH. Content aligns with commonly tested public health domains. Verify
against current NBPHE candidate materials. Certification requires meeting eligibility rules and passing the official exam.
Introduction
This 200-question set mirrors official CPH length. Correct answers appear in bold cyan with concise rationales. Use domain
scores to guide review.
• Simulate ~4 hours before checking answers.
• Score by domain; review where <75%.
• Build formula sheets for epi measures and screening 2×2 tables.
• Connect policy/behavior items to real programs.
EXAM QUESTIONS (1–200)
PART A — Epidemiology (Q1–40)
1. Epidemiology is best defined as the study of:
A. Only pharmaceutical manufacturing
B. The distribution and determinants of health-related states or events in specified populations, and application of
this study to control health problems
C. Only hospital billing practices
D. Only laboratory diagnosis of individual patients
Correct Answer: B. The distribution and determinants of health-related states or events in specified populations, and
application of this study to control health problems
Rationale: Classic definition emphasizes distribution, determinants, populations, and control. Clinical medicine focuses on individuals.
2. Prevalence is best described as:
A. Only deaths from a disease
B. The case fatality ratio exclusively
C. The proportion of a population that has a condition at a given time (or over a period)
D. Only new cases during a time period
Correct Answer: C. The proportion of a population that has a condition at a given time (or over a period)
Rationale: Prevalence = existing cases/population. Incidence = new cases over time. P ≈ I × D when stable.
3. Incidence rate measures:
A. Healthcare costs exclusively
B. The occurrence of new cases of a disease in a population over a specified period of time
C. All existing cases at one moment only
D. Only recovered cases
Correct Answer: B. The occurrence of new cases of a disease in a population over a specified period of time
Rationale: Incidence captures risk of developing disease; rates often use person-time denominators.
4. In a steady state, prevalence is approximately equal to:
A. Mortality only
B. Incidence ÷ duration
C. Attack rate squared
D. Incidence × average duration of disease
Correct Answer: D. Incidence × average duration of disease
Rationale: If incidence and duration are stable, P ≈ I × D.
5. A case-control study primarily compares:
A. People with a disease (cases) to people without (controls) regarding past exposures
B. Genetic sequences only
C. Only future disease without exposure history
D. Entire populations randomized to treatment
, Correct Answer: A. People with a disease (cases) to people without (controls) regarding past exposures
Rationale: Efficient for rare diseases; odds ratio estimates association.
6. A cohort study follows:
A. Cross-sectional snapshots exclusively
B. Only cases backward without controls
C. Randomized assignment only by definition
D. Groups defined by exposure status forward in time to observe disease outcomes
Correct Answer: D. Groups defined by exposure status forward in time to observe disease outcomes
Rationale: Relative risk is a natural measure; good for rare exposures.
7. The strongest observational design for temporal sequence is generally the:
A. Case report
B. Prospective cohort study
C. Ecologic study only
D. Case series without comparison
Correct Answer: B. Prospective cohort study
Rationale: Exposure precedes outcome by design. RCTs are experimental and stronger for causality.
8. Randomized controlled trials reduce confounding primarily through:
A. Selecting only healthy volunteers without comparison
B. Random assignment balancing known and unknown confounders
C. Avoiding any control group
D. Publishing only positive results
Correct Answer: B. Random assignment balancing known and unknown confounders
Rationale: Randomization supports internal validity; blinding reduces other biases.
9. Confounding occurs when:
A. The sample size is large
B. A third variable associated with exposure and outcome distorts the observed association
C. Randomization is perfect and complete
D. Measurement is perfectly accurate
Correct Answer: B. A third variable associated with exposure and outcome distorts the observed association
Rationale: Control via restriction, matching, stratification, or multivariable analysis.
10. Effect modification (interaction) means:
A. A third variable fully explains away an association as pure bias only
B. The effect of an exposure on an outcome differs across levels of a third variable
C. Sample size is infinite
D. Measurement error is absent
Correct Answer: B. The effect of an exposure on an outcome differs across levels of a third variable
Rationale: Report stratum-specific estimates; modification is a finding, not merely bias to remove.
11. Selection bias refers to:
A. Systematic error from how participants are selected or retained, making groups non-comparable
B. Perfect representative sampling
C. Only interviewer mood
D. Random variation only
Correct Answer: A. Systematic error from how participants are selected or retained, making groups non-comparable
Rationale: Examples: differential loss to follow-up, volunteer bias, Berkson's bias.
12. Information bias includes:
A. Misclassification of exposure or outcome (e.g., recall bias, interviewer bias)
B. Only random sampling error
C. Only ecological fallacy
D. Only confounding by age
Correct Answer: A. Misclassification of exposure or outcome (e.g., recall bias, interviewer bias)
Rationale: Non-differential misclassification often biases toward the null; differential can go either way.
13. Relative risk (risk ratio) is:
A. NNT exclusively
B. Prevalence minus incidence
C. Incidence in exposed divided by incidence in unexposed
, D. Odds of exposure in cases only without controls
Correct Answer: C. Incidence in exposed divided by incidence in unexposed
Rationale: RR=1 null; >1 increased risk; <1 decreased risk.
14. Odds ratio approximates the risk ratio when:
A. Confounding is maximal
B. The disease is extremely common always
C. The study is always cross-sectional only
D. The disease is rare (rare disease assumption) in the source population
Correct Answer: D. The disease is rare (rare disease assumption) in the source population
Rationale: In case-control studies OR is usual; when disease is common OR may overestimate RR.
15. Attributable risk (risk difference) is useful because it:
A. Quantifies excess absolute risk in the exposed attributable to the exposure
B. Ignores baseline risk
C. Only provides a ratio without absolute meaning
D. Is identical to OR always
Correct Answer: A. Quantifies excess absolute risk in the exposed attributable to the exposure
Rationale: Absolute measures inform prevention impact better than ratios alone.
16. Number needed to treat (NNT) is calculated as:
A. Odds ratio × 100
B. 1 / absolute risk reduction
C. Relative risk reduction only
D. 1 / relative risk
Correct Answer: B. 1 / absolute risk reduction
Rationale: Lower NNT suggests greater absolute benefit; specify time frame.
17. Sensitivity of a screening test is:
A. The probability that a person with disease tests positive (true positive rate)
B. Prevalence of disease
C. The probability a person without disease tests negative
D. Only positive predictive value
Correct Answer: A. The probability that a person with disease tests positive (true positive rate)
Rationale: Sensitivity=TP/(TP+FN). High sensitivity helps rule out disease when negative.
18. Specificity of a screening test is:
A. The probability that a person without disease tests negative (true negative rate)
B. Only negative predictive value
C. The probability a diseased person tests positive
D. Incidence rate
Correct Answer: A. The probability that a person without disease tests negative (true negative rate)
Rationale: Specificity=TN/(TN+FP). High specificity helps rule in disease when positive.
19. Positive predictive value depends heavily on:
A. Only sensitivity, never prevalence
B. Only the manufacturer logo
C. Sample size alone
D. Disease prevalence (pretest probability) as well as sensitivity and specificity
Correct Answer: D. Disease prevalence (pretest probability) as well as sensitivity and specificity
Rationale: When prevalence falls, PPV falls—more false positives in low-prevalence screening.
20. Lead time bias in screening occurs when:
A. Treatment truly cures everyone
B. Tests have 100% specificity only
C. No one is screened
D. Earlier diagnosis makes survival seem longer without changing the disease course
Correct Answer: D. Earlier diagnosis makes survival seem longer without changing the disease course
Rationale: Lead time, length-time, and overdiagnosis can inflate apparent screening benefit.
21. Length-time bias favors detection of:
A. Injuries only
B. Only rapidly fatal disease exclusively
PRACTICE EXAM 2026–2027
(200 Original Practice Questions with Correct Answers & Rationales)
Public Health | NBPHE/ASPPH-Aligned Domain Practice
Epidemiology • Biostatistics • Environmental Health • Health Policy & Management • Social & Behavioral Sciences • Public Health Biology
IMPORTANT DISCLAIMER: This is an ORIGINAL educational practice resource for study purposes only. It is NOT an official CPH examination, does NOT
contain proprietary NBPHE items, and is NOT affiliated with NBPHE, ASPPH, or CEPH. Content aligns with commonly tested public health domains. Verify
against current NBPHE candidate materials. Certification requires meeting eligibility rules and passing the official exam.
Introduction
This 200-question set mirrors official CPH length. Correct answers appear in bold cyan with concise rationales. Use domain
scores to guide review.
• Simulate ~4 hours before checking answers.
• Score by domain; review where <75%.
• Build formula sheets for epi measures and screening 2×2 tables.
• Connect policy/behavior items to real programs.
EXAM QUESTIONS (1–200)
PART A — Epidemiology (Q1–40)
1. Epidemiology is best defined as the study of:
A. Only pharmaceutical manufacturing
B. The distribution and determinants of health-related states or events in specified populations, and application of
this study to control health problems
C. Only hospital billing practices
D. Only laboratory diagnosis of individual patients
Correct Answer: B. The distribution and determinants of health-related states or events in specified populations, and
application of this study to control health problems
Rationale: Classic definition emphasizes distribution, determinants, populations, and control. Clinical medicine focuses on individuals.
2. Prevalence is best described as:
A. Only deaths from a disease
B. The case fatality ratio exclusively
C. The proportion of a population that has a condition at a given time (or over a period)
D. Only new cases during a time period
Correct Answer: C. The proportion of a population that has a condition at a given time (or over a period)
Rationale: Prevalence = existing cases/population. Incidence = new cases over time. P ≈ I × D when stable.
3. Incidence rate measures:
A. Healthcare costs exclusively
B. The occurrence of new cases of a disease in a population over a specified period of time
C. All existing cases at one moment only
D. Only recovered cases
Correct Answer: B. The occurrence of new cases of a disease in a population over a specified period of time
Rationale: Incidence captures risk of developing disease; rates often use person-time denominators.
4. In a steady state, prevalence is approximately equal to:
A. Mortality only
B. Incidence ÷ duration
C. Attack rate squared
D. Incidence × average duration of disease
Correct Answer: D. Incidence × average duration of disease
Rationale: If incidence and duration are stable, P ≈ I × D.
5. A case-control study primarily compares:
A. People with a disease (cases) to people without (controls) regarding past exposures
B. Genetic sequences only
C. Only future disease without exposure history
D. Entire populations randomized to treatment
, Correct Answer: A. People with a disease (cases) to people without (controls) regarding past exposures
Rationale: Efficient for rare diseases; odds ratio estimates association.
6. A cohort study follows:
A. Cross-sectional snapshots exclusively
B. Only cases backward without controls
C. Randomized assignment only by definition
D. Groups defined by exposure status forward in time to observe disease outcomes
Correct Answer: D. Groups defined by exposure status forward in time to observe disease outcomes
Rationale: Relative risk is a natural measure; good for rare exposures.
7. The strongest observational design for temporal sequence is generally the:
A. Case report
B. Prospective cohort study
C. Ecologic study only
D. Case series without comparison
Correct Answer: B. Prospective cohort study
Rationale: Exposure precedes outcome by design. RCTs are experimental and stronger for causality.
8. Randomized controlled trials reduce confounding primarily through:
A. Selecting only healthy volunteers without comparison
B. Random assignment balancing known and unknown confounders
C. Avoiding any control group
D. Publishing only positive results
Correct Answer: B. Random assignment balancing known and unknown confounders
Rationale: Randomization supports internal validity; blinding reduces other biases.
9. Confounding occurs when:
A. The sample size is large
B. A third variable associated with exposure and outcome distorts the observed association
C. Randomization is perfect and complete
D. Measurement is perfectly accurate
Correct Answer: B. A third variable associated with exposure and outcome distorts the observed association
Rationale: Control via restriction, matching, stratification, or multivariable analysis.
10. Effect modification (interaction) means:
A. A third variable fully explains away an association as pure bias only
B. The effect of an exposure on an outcome differs across levels of a third variable
C. Sample size is infinite
D. Measurement error is absent
Correct Answer: B. The effect of an exposure on an outcome differs across levels of a third variable
Rationale: Report stratum-specific estimates; modification is a finding, not merely bias to remove.
11. Selection bias refers to:
A. Systematic error from how participants are selected or retained, making groups non-comparable
B. Perfect representative sampling
C. Only interviewer mood
D. Random variation only
Correct Answer: A. Systematic error from how participants are selected or retained, making groups non-comparable
Rationale: Examples: differential loss to follow-up, volunteer bias, Berkson's bias.
12. Information bias includes:
A. Misclassification of exposure or outcome (e.g., recall bias, interviewer bias)
B. Only random sampling error
C. Only ecological fallacy
D. Only confounding by age
Correct Answer: A. Misclassification of exposure or outcome (e.g., recall bias, interviewer bias)
Rationale: Non-differential misclassification often biases toward the null; differential can go either way.
13. Relative risk (risk ratio) is:
A. NNT exclusively
B. Prevalence minus incidence
C. Incidence in exposed divided by incidence in unexposed
, D. Odds of exposure in cases only without controls
Correct Answer: C. Incidence in exposed divided by incidence in unexposed
Rationale: RR=1 null; >1 increased risk; <1 decreased risk.
14. Odds ratio approximates the risk ratio when:
A. Confounding is maximal
B. The disease is extremely common always
C. The study is always cross-sectional only
D. The disease is rare (rare disease assumption) in the source population
Correct Answer: D. The disease is rare (rare disease assumption) in the source population
Rationale: In case-control studies OR is usual; when disease is common OR may overestimate RR.
15. Attributable risk (risk difference) is useful because it:
A. Quantifies excess absolute risk in the exposed attributable to the exposure
B. Ignores baseline risk
C. Only provides a ratio without absolute meaning
D. Is identical to OR always
Correct Answer: A. Quantifies excess absolute risk in the exposed attributable to the exposure
Rationale: Absolute measures inform prevention impact better than ratios alone.
16. Number needed to treat (NNT) is calculated as:
A. Odds ratio × 100
B. 1 / absolute risk reduction
C. Relative risk reduction only
D. 1 / relative risk
Correct Answer: B. 1 / absolute risk reduction
Rationale: Lower NNT suggests greater absolute benefit; specify time frame.
17. Sensitivity of a screening test is:
A. The probability that a person with disease tests positive (true positive rate)
B. Prevalence of disease
C. The probability a person without disease tests negative
D. Only positive predictive value
Correct Answer: A. The probability that a person with disease tests positive (true positive rate)
Rationale: Sensitivity=TP/(TP+FN). High sensitivity helps rule out disease when negative.
18. Specificity of a screening test is:
A. The probability that a person without disease tests negative (true negative rate)
B. Only negative predictive value
C. The probability a diseased person tests positive
D. Incidence rate
Correct Answer: A. The probability that a person without disease tests negative (true negative rate)
Rationale: Specificity=TN/(TN+FP). High specificity helps rule in disease when positive.
19. Positive predictive value depends heavily on:
A. Only sensitivity, never prevalence
B. Only the manufacturer logo
C. Sample size alone
D. Disease prevalence (pretest probability) as well as sensitivity and specificity
Correct Answer: D. Disease prevalence (pretest probability) as well as sensitivity and specificity
Rationale: When prevalence falls, PPV falls—more false positives in low-prevalence screening.
20. Lead time bias in screening occurs when:
A. Treatment truly cures everyone
B. Tests have 100% specificity only
C. No one is screened
D. Earlier diagnosis makes survival seem longer without changing the disease course
Correct Answer: D. Earlier diagnosis makes survival seem longer without changing the disease course
Rationale: Lead time, length-time, and overdiagnosis can inflate apparent screening benefit.
21. Length-time bias favors detection of:
A. Injuries only
B. Only rapidly fatal disease exclusively
