A Comparison of Robust Tests for Homogeneity of Variance in Factorial ANOVA

• In Event: Marriott Roundtable Session 52

Abstract

The factorial analysis of variance (ANOVA) design is widely used to compare means across two or more independent variables. Homogeneity of variance (HOV), which refers to equal population variances across cells, is a critical assumption and should be checked before applying factorial ANOVA. A simulation study is conducted to examine the efficacy of nine frequently-used HOV tests (z-variance test, Bartlett test, Pseudo jackknife test, Levene with squared residuals, Levene with absolute, Box-Scheffé test, Brown and Forsythe, Bootstrap Brown and Forsythe, and O’Brien) under comprehensive simulation conditions with and without the normality being held. This study is intended for enhancing the knowledge of those HOV tests and providing guidelines for HOV test selection.

Authors

• ZHIYAO YI, University of South Florida

• Yue Yin, University of South Florida

• Ke Cheng, University of South Florida - Tampa

• Thanh Vinh Pham, University of South Florida

• Diep Thi Nguyen, University of South Florida

• Yan Wang, University of South Florida

• Eun Sook Kim, University of South Florida

• Yi-Hsin Chen, University of South Florida