Date created: | Last Updated:

Category: Project

Description: The shiny app is available at: https://aaroncaldwell.us/project/shinypowerapp/ and https://aaroncaldwell.us/project/shinyexactapp/ A complete manual using Superpower is available at: https://aaroncaldwell.us/SuperpowerBook/ Researchers often use an analysis of variance (ANOVA) when reporting results of study. In order for an ANOVA to be informative researchers need to ensure their study is adequately powered. Yet, power analyses for factorial ANOVA designs are often challenging. First, current software solutions do not enable power analyses for complex designs with many within-subject factors. Second, power analyses often need partial eta-squared or Cohen's f as input, but these effect sizes do not generalize to different experimental designs. We have created R functions and an online Shiny app that performs simulations for ANOVA designs. Our functions allow for up to three within- or between-subject factors, with an unlimited number of levels. Predicted effects are entered by specifying means, standard deviations, and correlations (for within-subject factors). The simulation provides power calculations for all ANOVA main effects and interactions. In addition, power calculations are provided for pairwise comparisons and there are range of options to correct for multiple comparisons. The simulation plots p-value distributions for all tests. This tutorial will demonstrate how to perform power analysis for ANOVA designs. The simulations illustrate important factors that determine the statistical power of factorial ANOVA designs. The code and Shiny app will enable researchers without extensive programming experience to perform power analyses for a wide range of ANOVA designs.