Errors in hypothesis testing are critical for B.Pharm students to understand because they affect conclusions in pharmaceutical research and clinical trials. Key concepts include Type I error (false positive), Type II error (false negative), alpha (significance level), beta (probability of Type II), and statistical power (1 − beta). These errors influence drug approval, sample size planning, and interpretation of results. Balancing alpha and beta, adjusting for multiple comparisons, and understanding effect size are essential for valid study design and regulatory compliance. Clear grasp of these topics reduces wrong decisions in safety and efficacy assessments.
Now let’s test your knowledge with 30 MCQs on this topic.
Q1. What is a Type I error in hypothesis testing?
- Failing to reject a false null hypothesis
- Rejecting a true null hypothesis
- Accepting the alternative when it is false
- Calculating the wrong p-value
Correct Answer: Rejecting a true null hypothesis
Q2. What is a Type II error?
- Rejecting a true alternative hypothesis
- Failing to reject a true null hypothesis
- Failing to reject a false null hypothesis
- Using the wrong statistical test
Correct Answer: Failing to reject a false null hypothesis
Q3. Which symbol typically denotes the significance level (probability of Type I error)?
- β
- γ
- α
- δ
Correct Answer: α
Q4. Which symbol represents the probability of committing a Type II error?
- α
- β
- 1 − α
- Power
Correct Answer: β
Q5. Statistical power is defined as:
- The probability of a Type I error
- The probability of failing to detect a true effect
- 1 − β, the probability of correctly rejecting a false null hypothesis
- α multiplied by β
Correct Answer: 1 − β, the probability of correctly rejecting a false null hypothesis
Q6. If α = 0.05 and β = 0.20, the power of the test is:
- 0.05
- 0.20
- 0.80
- 0.95
Correct Answer: 0.80
Q7. Increasing the sample size in a study generally has what effect on Type II error (β)?
- Increases β
- Decreases β
- No effect on β
- Converts β into α
Correct Answer: Decreases β
Q8. Which action will directly reduce the probability of a Type I error?
- Increasing sample size while keeping α unchanged
- Lowering the significance level α
- Increasing β
- Decreasing effect size
Correct Answer: Lowering the significance level α
Q9. Which action will directly increase the power of a hypothesis test?
- Reducing the sample size
- Decreasing the effect size
- Increasing α or increasing sample size
- Using a less precise measurement method
Correct Answer: Increasing α or increasing sample size
Q10. In a clinical trial context, a Type I error could result in:
- Failing to detect a harmful side effect
- Approving an ineffective drug as effective
- Dismissing a truly effective treatment
- Underpowered subgroup analysis
Correct Answer: Approving an ineffective drug as effective
Q11. In the same context, a Type II error could result in:
- Approving a harmful drug
- Approving an ineffective drug
- Failing to approve a truly effective drug
- Multiplying Type I errors
Correct Answer: Failing to approve a truly effective drug
Q12. What is the relationship between α and β when all other factors are fixed?
- They are independent and never affect each other
- Reducing α generally increases β, and vice versa
- Increasing α always decreases β without limit
- α plus β must equal 1
Correct Answer: Reducing α generally increases β, and vice versa
Q13. Which of the following helps control inflated Type I error when performing many comparisons?
- Ignoring multiple testing
- Using Bonferroni or other multiple comparison adjustments
- Increasing β for each test
- Decreasing sample size
Correct Answer: Using Bonferroni or other multiple comparison adjustments
Q14. A one-sided (one-tailed) test compared with a two-sided test of the same α will generally:
- Have the same power for effects in either direction
- Have increased power to detect an effect in the tested direction
- Double the Type II error in the tested direction
- Eliminate Type I error
Correct Answer: Have increased power to detect an effect in the tested direction
Q15. If a p-value is 0.03 and α = 0.05, the correct decision is to:
- Fail to reject H0
- Reject H0 and risk a Type II error
- Reject H0 and accept the alternative hypothesis
- Increase α to 0.10
Correct Answer: Reject H0 and accept the alternative hypothesis
Q16. Which factor does NOT directly affect statistical power?
- Sample size
- Effect size
- Significance level α
- Researcher’s age
Correct Answer: Researcher’s age
Q17. In power calculations, a larger effect size will generally:
- Decrease power
- Increase power
- Have no impact on power
- Increase α but not power
Correct Answer: Increase power
Q18. Which statement best connects confidence intervals and hypothesis testing?
- If a 95% CI for a mean difference excludes 0, a two-sided test at α = 0.05 would reject H0
- A confidence interval cannot inform hypothesis tests
- If a 95% CI includes 0, H0 is always false
- Confidence intervals and p-values are unrelated
Correct Answer: If a 95% CI for a mean difference excludes 0, a two-sided test at α = 0.05 would reject H0
Q19. Multiple interim analyses during a trial without adjustment most likely will:
- Reduce sample size without consequences
- Inflate the overall Type I error rate
- Have no effect on Type I error
- Eliminate Type II error
Correct Answer: Inflate the overall Type I error rate
Q20. Which of the following is an appropriate method to reduce Type II error?
- Decrease α to 0.01
- Reduce measurement precision
- Increase sample size or improve measurement accuracy
- Ignore variability estimates
Correct Answer: Increase sample size or improve measurement accuracy
Q21. In non-inferiority trials, a Type I error corresponds to:
- Concluding non-inferiority when the new treatment is actually inferior
- Concluding superiority when treatments are equal
- Failing to detect superiority
- Detecting a side effect
Correct Answer: Concluding non-inferiority when the new treatment is actually inferior
Q22. Which practice increases the risk of Type I error via selective reporting?
- Pre-registering study methods
- P-hacking and selective outcome reporting
- Blinded analysis
- Using appropriate correction for multiple testing
Correct Answer: P-hacking and selective outcome reporting
Q23. Sensitivity and specificity in diagnostic testing are conceptually analogous to which hypothesis testing errors?
- Sensitivity ~ Type I, Specificity ~ Type II
- Sensitivity ~ Type II, Specificity ~ Type I
- Sensitivity ~ Power, Specificity ~ 1 − α
- Sensitivity ~ 1 − β, Specificity ~ 1 − α
Correct Answer: Sensitivity ~ 1 − β, Specificity ~ 1 − α
Q24. Which choice best describes the trade-off when lowering α to make the testing criterion more stringent?
- It decreases β without other changes
- It may increase β unless sample size or effect size is increased
- It always increases power
- It eliminates the need for confidence intervals
Correct Answer: It may increase β unless sample size or effect size is increased
Q25. Which is a regulatory consideration for setting α in pivotal drug trials?
- Regulators prefer very large α to speed approval
- Regulators balance patient safety and evidence, often accepting α = 0.05 or adjusted levels
- α is irrelevant for regulators
- Regulators set β, not α
Correct Answer: Regulators balance patient safety and evidence, often accepting α = 0.05 or adjusted levels
Q26. If a study is underpowered, which outcome is most likely?
- High chance of Type I error only
- High chance of Type II error and missing true effects
- Guaranteed significant results
- Reduced measurement error
Correct Answer: High chance of Type II error and missing true effects
Q27. What happens to the width of a confidence interval if you increase sample size?
- Width increases
- Width decreases
- Width remains the same
- CI becomes invalid
Correct Answer: Width decreases
Q28. Which element is NOT part of a power/sample size calculation?
- Desired α (significance level)
- Desired power (1 − β)
- Expected effect size
- Investigator’s preference for publication
Correct Answer: Investigator’s preference for publication
Q29. In hypothesis testing, the critical region is best described as:
- The set of sample outcomes where H0 is not considered
- The set of sample outcomes that lead to rejection of H0
- The region where β is computed only
- The range of plausible parameter values
Correct Answer: The set of sample outcomes that lead to rejection of H0
Q30. Which strategy is ethically important in drug studies when balancing Type I and Type II errors?
- Prioritize rapid approval at any cost
- Choose α arbitrarily without justification
- Set α and power based on clinical consequences, patient safety, and regulatory guidance
- Ignore sample size and rely on post-hoc analyses
Correct Answer: Set α and power based on clinical consequences, patient safety, and regulatory guidance
