Akademik Veri Yönetim Sistemi
International Journal of Assessment Tools in Education, cilt.12, sa.4, ss.1170-1185, 2025 (ESCI, TRDizin)
It is an important point to test the homogeneity of variances in statistical methods such as the t-test or F-test used to make comparisons between groups. An erroneous decision regarding the homogeneity of variances will affect the test to be selected and thus lead to different results. For this reason, there are many tests for homogeneity of variance in the literature. This study aims to examine the type I error and power ratios of Levene, Bartlett, Brown-Forsythe, and Fligner-Killeen tests under different conditions. In this study, conducted within the scope of basic research, analyses were performed using simulated data. The simulation conditions included variance ratio (1:1, 1:2, 1:3, 2:1, 3:1), distributions (normal, skewed-normal, gamma), sample sizes (60, 120, and 240), and ratio of group sizes (1/1, 1/2, 1/4, 1/9). According to the study results, when controlling for type I error is a primary concern, the Brown–Forsythe and Fligner–Killeen tests are recommended, particularly under non-normal distributions. If the power is a major concern for research, the Bartlett’s test and the Levene’s test should be used in general.