Statistical Assumption of Homogeneity of Variance: A Query on the Robustness of Analysis of Variance and Regression Statistics

By Dr Isaac Ofem Ubi, Anastecia William Osang & Aniyom Asuquo Obambon

Abstract

The study examined the robustness of analysis of variance (ANOVA) and regression statistics in the face of the homogeneity of variance assumption. The sample was made up of 640 students selected through stratified random sampling technique from secondary schools in Calabar Education zone of Cross River State, Nigeria. Two instruments – Students’ Attitude and Test Anxiety Questionnaire (SATAQ), and WAEC May/June SSCE Mathematics objective test for 2017/2018 academic session were used for the study. Results of the study showed that Multiple Regression Statistical tool was more robust than the Two–Way ANOVA when the homogeneity of variance assumption is violated. The study recommends strict adherence to the assumption of homoscedasticity when ANOVA and Regression Statistics are proposed for a study.

Read Full Article

Scroll to Top
Scroll to Top