When is the Bonferroni correction typically applied?

• A. Before any test to choose sample size
• B. After an ANOVA shows significant differences among groups
• C. Only for t-tests with a single comparison
• D. During data collection to randomize order

Answer: (B)

When you perform several statistical comparisons, the chance of a false positive accumulates. The Bonferroni correction tackles this by making the significance threshold more stringent for each test—usually by dividing the overall alpha level by the number of comparisons. It’s typically applied after an overall test, such as ANOVA, shows that there are differences among groups. If the ANOVA is significant, you proceed to post hoc pairwise comparisons, each using the adjusted alpha to control the family-wise error rate. For example, with four groups there are six pairwise tests, so each test would use alpha = 0.05/6 ≈ 0.0083. This approach isn’t about deciding sample size beforehand, and it’s specifically used when you’re making multiple comparisons after an overall significant finding.