Some Tests and Comparisons for Homogeneity of Variances

Creative Commons License

Öztürk N. N., Gamgam H., Altunkaynak B.

5 th INTERNATIONAL RESEARCHERS, STATISTICIANS AND YOUNG STATISTICIANS CONGRESS, Aydın, Turkey, 18 - 20 October 2019, pp.381-388

  • Publication Type: Conference Paper / Full Text
  • City: Aydın
  • Country: Turkey
  • Page Numbers: pp.381-388
  • Erzincan Binali Yildirim University Affiliated: No

Abstract

The equality of variances is one of the basic assumptions of classical variance analysis. There are many test statistics known in the literature to determine whether this assumption is met or not. The aim of this study is to introduce the Bartlett (B), Hartley (H), Levene (L), Fligner-Killeen (FK) and Bootstrap (FRMD) tests for testing the homogeneity hypothesis of variances and compare these tests. Using Monte Carlo simulation method, these tests were compared in terms of type I error rate and power under various scenarios. With these comparisons, it was aimed to determine the test suitable for testing the homogeneity hypothesis of variances in both normal distribution and some nonnormal distributions.