Empiryczna ocena mocy testów dla wielu wariancji
A Comparative Study of Tests Power for Homogeneity of Variances

Summary/Abstract: Some statistical tests, for example the analysis of variance, assume that variances are equal across groups or samples. Tests for homogeneity of variances can be used to verify that assumption and for pooling of data from different sources to yield an improved estimated variance. Tests for homogeneity of variances can be used usually under assumption of normal distributions or nearly normal distributions. In this paper some tests for homogeneity of variances are examined under the null hypothesis and under the alternative, for various sample sizes, for various symmetric and asymmetric distributions. Monte Carlo simulations has been used for this. In this paper the following procedures have been analyzed: Levene’a test, Brown-Forsythe test, Fligner-Killeen test and O’Brientest.

• Details
• Contents