Exploring Multiple Comparison Procedures in Factorial ANOVA With Heterogeneous Variances

• In Event: AERA Poster Session 6
  In Poster Session: Best Practices in Applied Statistics

Abstract

When statistically significant main effects are present in a factorial ANOVA, researchers often wish to further explore the data in order to determine where differences might exist. When the homoscedasticity assumption is met with no apparent interaction, similar post hoc tests from one-way ANOVA are commonly used by researchers. However, researchers do not have as many options available in factorial ANOVA procedures for main effects post hoc tests with heterogeneous variances. The purpose of this paper is to investigate the characteristics of post hoc tests given statistically significant main effects with heterogeneous variances on some complex factorial ANOVA designs. Monte Carlo simulations were conducted to investigate Type I error and power for common post hoc approaches under various conditions.

Authors

• Michael Senteney, Ohio University

• Gordon P. Brooks, Ohio University