Power Analysis of Moderation Effects in Multisite Randomized Trials

• In Event: 31.038 - Power Analysis in Multilevel Modeling

Abstract

Although advance was made for power analysis of a binary site-level moderator effect in two-level multisite randomized trials (MRTs), in which the individuals are randomly assigned to the treatment group within each site, there is no comprehensive statistical tool for power analyses of moderator effects for different types of moderators (e.g., a continuous site-level moderator, a binary or continuous individual level moderator) and different types of models (e.g., random or non-randomly varying slope models). We derive formulas for the statistical power and the minimum detectable effect size difference with confidence intervals for moderator effects in two-level MRTs, and executes the formulas in the free-available software (PowerUp!-Moderator). We demonstrate how to use PowerUp!-Moderator, and provide suggestions in designing multisite moderation studies.

Authors

• Nianbo Dong, University of North Carolina at Chapel Hill

• Benjamin Kelcey, University of Cincinnati

• Jessaca K. Spybrook, Western Michigan University