Application and power of criteria for testing the homogeneity of variances. Part I. Parametric criteria
A comparative analysis is made of the power of classical (Fisher, Bartlett, Cochran, Hartley, and Levene) tests of variance homogeneity. The distributions of the statistics of the tests are studied when the assumption that the sample obeys a normal law breaks down.
Key wordstests of homogeneity of variances Fisher Bartlett Cochran Hartley and Levene tests power of tests
Unable to display preview. Download preview PDF.
- 1.B. Yu. Lemeshko and S. B. Lemeshko, “Statistical distribution convergence and homogeneity test power for Smirnov and Lehmann–Rosenblatt tests,” Izmer. Tekhn., No. 12, 9 (2005); Measur. Techn., 48, No. 12, 1159 (2005).Google Scholar
- 2.B. Yu. Lemeshko and S. B. Lemeshko, “Power and robustness of criteria used to verify the homogeneity of means,” Izmer. Tekhn., No. 9, 23 (2008); Measur. Techn., 51, No. 9, 950 (2008).Google Scholar
- 3.B. Yu. Lemeshko and E. P. Mirkin, “Bartlett and Cochran tests in measurements with probability laws different from normal,” Izmer. Tekhn., No. 10, 10 (2004); Measur. Techn., 47, No. 10, 960 (2004).Google Scholar
- 4.B. Yu. Lemeshko and V. M. Ponomarenko, “Distributions of the statistics used for testing hypotheses of equal variances for non-normal law error,” Nauch. Vestnik NGTU, No. 2 (23), 21 (2006).Google Scholar
- 8.H. Levene, “Robust tests for equality of variances,” in: Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling (1960), p. 278.Google Scholar
- 14.B. Yu. Lemeshko and S. N. Postovalov, Computer Technology for Data Analysis and Investigation of Statistical Behavior: A Textbook [in Russian], Izd. NGTU, Novosibirsk (2004).Google Scholar
- 15.B. Yu. Lemeshko, Statistical Analysis of Uniform Observations of Random Quantities: A Program System [in Russian], Izd. NGTU, Novosibirsk (1995).Google Scholar
- 16.L. N. Bolshev and N. V. Smirnov, Tables for Mathematical Statistics [in Russian], Nauka, Moscow (1983).Google Scholar
- 17.“Levene test for equality of variances,” in: Handbook of Statistical Methods, http://www.itl.nist.gov/div898/handbook/eda/section3/eda35a.htm (checked January 25, 2010).
- 18.J. H. Neel and W. M. Stallings, “A Monte Carlo study of Levene’s test of homogeneity of variance: empirical frequencies of type I error in normal distributions,” Paper Presented at the Annual Meeting of the American Educational Research Association Convention, Chicago, Illinois (April 1974).Google Scholar