Statistical power

Power is usually the flip side of a Type II error: a study with 80 percent power has a 20 percent chance of failing to detect a real effect of the assumed size. Convention aims for 80 to 90 percent, and reaching it requires an a priori power analysis that specifies the smallest effect worth detecting before data collection begins.

Underpowered research does more than waste effort. When a small, noisy study does cross the significance threshold, the estimated effect is almost always inflated, a phenomenon known as the winner's curse or Type M (magnitude) error. This is a major engine of the replication crisis, because splashy findings from tiny samples rarely hold up.

Reviewers should check whether a power calculation exists, whether its assumed effect size is plausible or conveniently large, and whether post hoc power (calculated from the observed effect) is being misused to excuse a null result. A credible sample-size justification tied to a meaningful minimum effect is a hallmark of careful design.