Which correction reduces the alpha level by the number of tests to control the Type I error rate?

Which correction reduces the alpha level by the number of tests to control the Type I error rate?

• A. Bootstrap
• B. Box's test
• C. Bonferroni correction
• D. Brown-Forsythe F

Answer: (C)

When you run several statistical tests, the chance of getting a false positive somewhere rises. To keep the overall chance of a Type I error at a desired level, you adjust how strict each individual test must be. The Bonferroni correction does this by setting the significance level for each test to the overall alpha divided by the number of tests. For example, with an overall alpha of 0.05 and six tests, each test uses 0.05/6 ≈ 0.0083. This keeps the familywise error rate around 0.05 across all tests. Other options don’t perform this adjustment for multiple comparisons: Bootstrap is a resampling method, Box's test checks equality of covariance matrices, and Brown-Forsythe F tests equality of variances under non-normal conditions.

When you run several statistical tests, the chance of getting a false positive somewhere rises. To keep the overall chance of a Type I error at a desired level, you adjust how strict each individual test must be. The Bonferroni correction does this by setting the significance level for each test to the overall alpha divided by the number of tests. For example, with an overall alpha of 0.05 and six tests, each test uses 0.05/6 ≈ 0.0083. This keeps the familywise error rate around 0.05 across all tests.

Other options don’t perform this adjustment for multiple comparisons: Bootstrap is a resampling method, Box's test checks equality of covariance matrices, and Brown-Forsythe F tests equality of variances under non-normal conditions.