Posts tagged: p-hacking

The Multiple Comparisons Problem

Run enough tests at a 0.05 threshold and something will look significant by chance. What the family-wise error rate means, why it matters, and what to do about it.

May 7, 2026