A student claims that because a 95% confidence interval for μ is (45.2, 52.8), the null hypothesis H₀: μ = 50 should not be rejected at α = 0.05 in a two-tailed test. Is this reasoning correct?
- ANo, because confidence intervals and hypothesis tests use different assumptions.
- BYes, because 50 is within the confidence interval, so it is a plausible value for μ.Correct
- CYes, but only if the sample size exceeds 30.
- DNo, because the confidence interval only applies to one-sided tests.
Why B is right
A 95% confidence interval and a two-sided hypothesis test at α = 0.05 are mathematically equivalent for the same data. If the null value μ₀ falls within the 95% CI, then the p-value > 0.05 and we fail to reject H₀. Since 50 is within (45.2, 52.8), we correctly fail to reject H₀: μ = 50.
Why the other 3 choices miss
- A
- Incorrect because for the same test conditions (two-sided, α = 0.05, same data), the confidence interval and hypothesis test are dual procedures and yield identical decisions.
- C
- Incorrect because the equivalence between confidence intervals and hypothesis tests holds regardless of sample size, as long as both use the same distribution (t or z) and the same α.
- D
- Incorrect because a two-sided confidence interval corresponds exactly to a two-sided hypothesis test; one-sided tests require one-sided intervals for duality.