1
Certified in Public Health (CPH) Exam Actual Exam
2026/2027: Questions and Verified Answers | Graded
A+ with Detailed Answers for Public Health Success –
Pass Guaranteed - A+ Graded
Section 1: Biostatistics (10 Questions)
Q1: In a study of 200 patients, 40 developed hospital-acquired infections. What is the cumulative
incidence of infection?
A. 20% [CORRECT]
B. 40%
C. 5%
D. 200 per 100,000
Correct Answer: A
Rationale: Cumulative incidence = (New cases / Total population at risk) × 100 = (40/200) × 100
= 20%. This measures the proportion of at-risk individuals who develop the outcome over a
specified time period. Option B uses the raw count without division. Option C is an incorrect
calculation. Option D converts to population rate unnecessarily (Rothman, Epidemiology: An
Introduction).
Q2: A dataset of household incomes is heavily right-skewed with extreme high values. Which
measure of central tendency is most appropriate to report?
A. Mean
B. Median [CORRECT]
C. Mode
D. Range
Correct Answer: B
Rationale: The median is resistant to outliers and skewed distributions, representing the 50th
percentile where half the data falls above and below. The mean (Option A) is pulled toward
extreme values in skewed distributions. Mode (Option C) may not exist or be meaningful. Range
(Option D) is a measure of dispersion, not central tendency (Gerstman, Basic Biostatistics).
Q3: In a clinical trial, the probability of a Type I error (α) is set at 0.05. What does this represent?
,2
A. The probability of failing to detect a true effect
B. The probability of rejecting the null hypothesis when it is actually true [CORRECT]
C. The probability of correctly accepting the null hypothesis
D. The study's statistical power
Correct Answer: B
Rationale: Type I error (alpha) is the false positive rate—rejecting a true null hypothesis
(concluding an effect exists when it doesn't). Type II error (Option A, beta) is failing to reject a
false null. Option C describes correct acceptance. Power (Option D) is 1-beta, the probability of
detecting a true effect (Fisher & Hebel, Biostatistics).
QQ4: A study reports an odds ratio of 2.5 with a 95% confidence interval of 1.8-3.4 for the
association between smoking and lung cancer. How should this be interpreted?
A. There is no association between smoking and lung cancer
B. Smokers have 2.5 times the odds of lung cancer compared to non-smokers, and this finding is
statistically significant [CORRECT]
C. 95% of smokers will develop lung cancer
D. The result is not statistically significant
Correct Answer: B
Rationale: OR=2.5 indicates smokers have 2.5× the odds of lung cancer. The 95% CI (1.8-3.4)
excludes 1.0 (null value), indicating statistical significance at α=0.05. Option A contradicts the
data. Option C confuses OR with risk. Option D is incorrect because the CI doesn't include 1.0
(Hennekens & Buring, Epidemiology in Medicine).
Q5: Which statistical test is most appropriate for comparing mean blood pressure between three
independent groups (control, treatment A, treatment B)?
A. Paired t-test
B. One-way ANOVA [CORRECT]
C. Chi-square test
D. Pearson correlation
Correct Answer: B
Rationale: One-way ANOVA compares means across three or more independent groups. Paired t-
test (Option A) compares dependent/paired samples. Chi-square (Option C) tests categorical
,3
associations. Pearson correlation (Option D) measures linear relationships between continuous
variables, not group differences (Rosner, Fundamentals of Biostatistics).
Q6: In a normal distribution with mean = 100 and standard deviation = 15, what percentage of
values fall between 85 and 115?
A. 50%
B. 68% [CORRECT]
C. 95%
D. 99.7%
Correct Answer: B
Rationale: In normal distributions, approximately 68% of values fall within ±1 SD of the mean
(100±15 = 85-115). 95% (Option C) falls within ±2 SD. 99.7% (Option D) falls within ±3 SD.
50% (Option A) would be the interquartile range approximately (Gerstman, Basic Biostatistics).
Q7: A screening test has sensitivity = 90% and specificity = 95%. What is the probability that a
person with a negative test result truly does not have the disease (negative predictive value,
NPV)?
A. Cannot be determined from information given
B. Depends on disease prevalence [CORRECT]
C. 95%
D. 90%
Correct Answer: B
Rationale: NPV = True Negatives / (True Negatives + False Negatives). While specificity
influences NPV, the actual NPV depends on disease prevalence—NPV increases as prevalence
decreases. Without prevalence information, exact NPV cannot be calculated. Sensitivity (Option
D) and specificity (Option C) alone don't determine predictive values (Fletcher & Fletcher,
Clinical Epidemiology).
Q8: What is the primary purpose of randomization in a randomized controlled trial?
A. To ensure equal sample sizes in all groups
B. To minimize selection bias and balance known and unknown confounders [CORRECT]
C. To eliminate the need for statistical analysis
D. To guarantee that the treatment will be effective
, 4
Correct Answer: B
Rationale: Randomization creates comparable groups by distributing confounders (known and
unknown) evenly across treatment arms, minimizing selection bias. It doesn't guarantee equal
sizes (Option A, though usually similar), eliminate analysis needs (Option C), or ensure
effectiveness (Option D) (Rothman & Greenland, Modern Epidemiology).
Q9: A researcher calculates the sample size needed to detect a 10 mmHg difference in systolic
blood pressure with 80% power and α=0.05. What happens to required sample size if power is
increased to 90%?
A. Sample size decreases
B. Sample size increases [CORRECT]
C. Sample size stays the same
D. Alpha increases automatically
Correct Answer: B
Rationale: Higher power (probability of detecting a true effect) requires larger sample sizes to
reduce Type II error probability. Power and sample size are directly related for fixed effect size
and alpha. Alpha (Option D) is set independently (Cohen, Statistical Power Analysis).
Q10: In a logistic regression analysis, what does an odds ratio of 1.0 for a predictor variable
indicate?
A. The predictor is perfectly associated with the outcome
B. The predictor has no association with the outcome [CORRECT]
C. The outcome is certain to occur
D. The model is invalid
Correct Answer: B
Rationale: OR=1.0 is the null value indicating no association—the outcome odds are identical
regardless of predictor status. OR>1 indicates increased odds; OR<1 indicates decreased odds.
Option A describes very large OR. Option C describes probability, not OR. Option D is incorrect
(Hosmer & Lemeshow, Applied Logistic Regression).
Section 2: Epidemiology (15 Questions)
Q11: In a cohort study of 1,000 smokers and 1,000 non-smokers followed for 10 years, 100
smokers and 20 non-smokers developed lung cancer. What is the relative risk?
A. 5.0 [CORRECT]
Certified in Public Health (CPH) Exam Actual Exam
2026/2027: Questions and Verified Answers | Graded
A+ with Detailed Answers for Public Health Success –
Pass Guaranteed - A+ Graded
Section 1: Biostatistics (10 Questions)
Q1: In a study of 200 patients, 40 developed hospital-acquired infections. What is the cumulative
incidence of infection?
A. 20% [CORRECT]
B. 40%
C. 5%
D. 200 per 100,000
Correct Answer: A
Rationale: Cumulative incidence = (New cases / Total population at risk) × 100 = (40/200) × 100
= 20%. This measures the proportion of at-risk individuals who develop the outcome over a
specified time period. Option B uses the raw count without division. Option C is an incorrect
calculation. Option D converts to population rate unnecessarily (Rothman, Epidemiology: An
Introduction).
Q2: A dataset of household incomes is heavily right-skewed with extreme high values. Which
measure of central tendency is most appropriate to report?
A. Mean
B. Median [CORRECT]
C. Mode
D. Range
Correct Answer: B
Rationale: The median is resistant to outliers and skewed distributions, representing the 50th
percentile where half the data falls above and below. The mean (Option A) is pulled toward
extreme values in skewed distributions. Mode (Option C) may not exist or be meaningful. Range
(Option D) is a measure of dispersion, not central tendency (Gerstman, Basic Biostatistics).
Q3: In a clinical trial, the probability of a Type I error (α) is set at 0.05. What does this represent?
,2
A. The probability of failing to detect a true effect
B. The probability of rejecting the null hypothesis when it is actually true [CORRECT]
C. The probability of correctly accepting the null hypothesis
D. The study's statistical power
Correct Answer: B
Rationale: Type I error (alpha) is the false positive rate—rejecting a true null hypothesis
(concluding an effect exists when it doesn't). Type II error (Option A, beta) is failing to reject a
false null. Option C describes correct acceptance. Power (Option D) is 1-beta, the probability of
detecting a true effect (Fisher & Hebel, Biostatistics).
QQ4: A study reports an odds ratio of 2.5 with a 95% confidence interval of 1.8-3.4 for the
association between smoking and lung cancer. How should this be interpreted?
A. There is no association between smoking and lung cancer
B. Smokers have 2.5 times the odds of lung cancer compared to non-smokers, and this finding is
statistically significant [CORRECT]
C. 95% of smokers will develop lung cancer
D. The result is not statistically significant
Correct Answer: B
Rationale: OR=2.5 indicates smokers have 2.5× the odds of lung cancer. The 95% CI (1.8-3.4)
excludes 1.0 (null value), indicating statistical significance at α=0.05. Option A contradicts the
data. Option C confuses OR with risk. Option D is incorrect because the CI doesn't include 1.0
(Hennekens & Buring, Epidemiology in Medicine).
Q5: Which statistical test is most appropriate for comparing mean blood pressure between three
independent groups (control, treatment A, treatment B)?
A. Paired t-test
B. One-way ANOVA [CORRECT]
C. Chi-square test
D. Pearson correlation
Correct Answer: B
Rationale: One-way ANOVA compares means across three or more independent groups. Paired t-
test (Option A) compares dependent/paired samples. Chi-square (Option C) tests categorical
,3
associations. Pearson correlation (Option D) measures linear relationships between continuous
variables, not group differences (Rosner, Fundamentals of Biostatistics).
Q6: In a normal distribution with mean = 100 and standard deviation = 15, what percentage of
values fall between 85 and 115?
A. 50%
B. 68% [CORRECT]
C. 95%
D. 99.7%
Correct Answer: B
Rationale: In normal distributions, approximately 68% of values fall within ±1 SD of the mean
(100±15 = 85-115). 95% (Option C) falls within ±2 SD. 99.7% (Option D) falls within ±3 SD.
50% (Option A) would be the interquartile range approximately (Gerstman, Basic Biostatistics).
Q7: A screening test has sensitivity = 90% and specificity = 95%. What is the probability that a
person with a negative test result truly does not have the disease (negative predictive value,
NPV)?
A. Cannot be determined from information given
B. Depends on disease prevalence [CORRECT]
C. 95%
D. 90%
Correct Answer: B
Rationale: NPV = True Negatives / (True Negatives + False Negatives). While specificity
influences NPV, the actual NPV depends on disease prevalence—NPV increases as prevalence
decreases. Without prevalence information, exact NPV cannot be calculated. Sensitivity (Option
D) and specificity (Option C) alone don't determine predictive values (Fletcher & Fletcher,
Clinical Epidemiology).
Q8: What is the primary purpose of randomization in a randomized controlled trial?
A. To ensure equal sample sizes in all groups
B. To minimize selection bias and balance known and unknown confounders [CORRECT]
C. To eliminate the need for statistical analysis
D. To guarantee that the treatment will be effective
, 4
Correct Answer: B
Rationale: Randomization creates comparable groups by distributing confounders (known and
unknown) evenly across treatment arms, minimizing selection bias. It doesn't guarantee equal
sizes (Option A, though usually similar), eliminate analysis needs (Option C), or ensure
effectiveness (Option D) (Rothman & Greenland, Modern Epidemiology).
Q9: A researcher calculates the sample size needed to detect a 10 mmHg difference in systolic
blood pressure with 80% power and α=0.05. What happens to required sample size if power is
increased to 90%?
A. Sample size decreases
B. Sample size increases [CORRECT]
C. Sample size stays the same
D. Alpha increases automatically
Correct Answer: B
Rationale: Higher power (probability of detecting a true effect) requires larger sample sizes to
reduce Type II error probability. Power and sample size are directly related for fixed effect size
and alpha. Alpha (Option D) is set independently (Cohen, Statistical Power Analysis).
Q10: In a logistic regression analysis, what does an odds ratio of 1.0 for a predictor variable
indicate?
A. The predictor is perfectly associated with the outcome
B. The predictor has no association with the outcome [CORRECT]
C. The outcome is certain to occur
D. The model is invalid
Correct Answer: B
Rationale: OR=1.0 is the null value indicating no association—the outcome odds are identical
regardless of predictor status. OR>1 indicates increased odds; OR<1 indicates decreased odds.
Option A describes very large OR. Option C describes probability, not OR. Option D is incorrect
(Hosmer & Lemeshow, Applied Logistic Regression).
Section 2: Epidemiology (15 Questions)
Q11: In a cohort study of 1,000 smokers and 1,000 non-smokers followed for 10 years, 100
smokers and 20 non-smokers developed lung cancer. What is the relative risk?
A. 5.0 [CORRECT]
