Comparing the Performance of Approaches for Testing the Homogeneity of Variance Assumption in One-Factor ANOVA Models

• Activity Number: 300
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Education
• Author(s): Aarti P. Bellara and Thanh Pham and Diep Thi Nguyen and Patricia Rodriguez de Gil and Yi-Hsin Chen Chen and Harold Holmes and Yan Wang*+ and Tyler Hicks and Isaac Li and Eun Sook Kim and Jeanine Romano and Jeffrey D. Kromrey
• Companies: University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida and University of South Florida
• Keywords: homogeneity of variance ; analysis of variance ; non-normality ; Type I error control ; statistical power

Abstract

The validity of the results of an ANOVA test is largely dependent on satisfying the homogeneity of variance, normality, and independence assumptions. Violations of these assumptions lead to distorted Type I error rates. Various tests to check the homogeneity of variance assumption for non-normal data have been proposed in the literature, yet there is no consensus as to which test is most appropriate. A simulation study was conducted to explore the Type I error rates and statistical power of fourteen approaches for testing the homogeneity of variance assumption in one-way ANOVA models. Seven factors were manipulated in the study: number of groups, average number of observations per group, pattern of sample sizes in groups, pattern of population variances, maximum variance ratio, population distribution shape, and nominal alpha level for the test of variances. Results from this study delineate the performance of the tests under a wide variety of conditions, providing researchers with information to guide the selection of a valid test for assessing the tenability of this critical assumption.