Comparing means under heteroscedasticity and nonnormality: Further exploring robust means modeling

Researchers are commonly interested in comparing the means of independent groups when distributions are nonnormal and variances are unequal. Robust means modeling (RMM) has been proposed as an alternative to ANOVA-type procedures when the assumptions of normality and variance homogeneity are violated. This paper extends work comparing the Type I error and power rates of RMM to those for the trimmed Welch procedure. A Monte Carlo study was used to investigate RMM and the trimmed Welch procedure under several conditions of nonnormality and variance heterogeneity. Our results suggest that the trimmed Welch provides a better balance of Type I error control and power than RMM.