Heterogeneity of variance in clinical data

Traditional parametric (t, F) and nonparametric (Mann-Whitney-Wilcoxon U, K ruskal-Wallis H) statistics are sensitive to heterogeneity of variance (het eroscedasticity). Moreover, there are theoretical reasons to expect, and em pirical results to document, the existence of heteroscedasticity in clinica l data. Transformations to reduce heteroscedasticity are problematic. This article reviews the literature on robust methods that are available and tha t should be widely used to control rate of Type I error and maintain power. No one robust method is ideal for all situations, but such methods are sup erior to the traditional tests. Specific recommendations are made for appli cation under various conditions of heteroscedasticity.