Quick Power Estimates Incorporating the Joint Effects of Measurement Error and a Covariate.

AbstractThe statistical power of a test of significance for the difference among means is dependent on a number of factors, among them the presence of a covariate, which often serves to increase power, as well as measurement error, which reduces power. Formulas are provided for estimating power under a variety of conditions. This paper includes a table, which incorporates these adjustments, for quick power estimates that require no computation for comparing two means in ANOVA and ANCOVA. Power estimates are provided for four effect sizes, three alpha levels, and eight sample sizes.