Brief Report: Post Hoc Power, Observed Power, A Priori Power, Retrospective Power, Prospective Power, Achieved Power: Sorting Out Appropriate Uses of Statistical Power Analyses

Abstract

This brief note hopes to dispel some apparent confusions concerning statistical power, especially ones stemming from power computations based on treating an observed sample effect size as the population effect of interest ("post hoc," "observed," or "retrospective" power). Some after-the-fact power analyses can be useful for presentation purposes but only when based on population effect sizes of independent interest. It is recommended that labels such as "post hoc power," "observed power," "retrospective power," and "a priori power" be avoided, that reported power figures be accompanied by specification and justification of the values used to compute power, and that results characteristically be described with effect sizes, confidence intervals, and p values.