When small effects are impressive.

Abstract

Effect size is becoming an increasingly popular measure of the importance of an effect, both in individual studies and in meta-analyses. However, a large effect size is not the only way to demonstrate that an effect is important. This article describes 2 alternative methodological strategies, in which importance is a function of how minimal a manipulation of the independent variable or how difficult-to-influence a dependent variable will still produce an effect. These methodologies demonstrate the importance of an independent variable or psychological process, even though they often yield effects that are small in statistical terms