Which methods are commonly used to control for false positives when conducting multiple statistical tests?

• A. Bonferroni correction or false discovery rate control.
• B. Increasing the alpha level.
• C. Ignoring p-values.
• D. Ignore multiple testing altogether.

Answer: (B)

When many hypotheses are tested, the chance of a false positive somewhere among them increases. To keep that risk under control, researchers use methods that account for multiple comparisons. The Bonferroni correction is a straightforward approach: you divide the overall significance level by the number of tests so each individual test has a stricter threshold to meet. This helps keep the overall probability of making any false claim at or below the desired level. More nuanced and often more powerful are procedures that control the false discovery rate, which is the expected proportion of false positives among the tests you declare significant. Methods like Benjamini-Hochberg adjust thresholds in a way that preserves more true findings while still limiting false discoveries.

Raising the alpha level would do the opposite, increasing the likelihood of false positives. Ignoring p-values or ignoring the issue of multiple testing treats the problem as if there were only one test, which also leads to inflated type I error.