A Comparison of Efficiency for Homogeneity of Variance Tests under Log-normal Distribution
DOI:
https://doi.org/10.24203/ajas.v9i4.6692Keywords:
Homogeneity of variance test, Log-normal distribution, Type I error, Power of the testAbstract
In this paper, we compare five homogeneity of variance tests which are Bartlett’s test, Levene’s test, Cochran’s test, O’Brien’s test and Jackknife test. Considering their ability to control probability of type I error and the power of the test, when groups of population are distributed in log-normal distribution. The equal sample sizes are defined as 10, 15, 30 and 50 at the significance is 0.05. The results show that the Levene’s test become the best test for the high skewed distribution. For the lowed skew distribution, the Cochran’s test is the best test when a variance of group is different to the others. Moreover, Bartlett’s test provides the highest power when variances of all populations are different.
References
Foster, J., Bevis, M. & Raymond, W., “Precipitable water and the lognormal distribution”, Journal of Geophysical Research Atmospheres., vol.111, no.15, 2006
Levene, H. “Robust Tests for Equality of Variances”, In Contributions to Probability and Statistics. California: Stanford University Press. Pp.278-292, 1960
Miller, R.G. Jr. “Jackknifing variances”, Annals of Mathematical Statistics, vol.39, pp.567-582, 1968
O'Brien, R. G. “A simple test for variance effects in experimental designs”. Psychological Bulletin, vol.89, no.3, pp.570-574, 1981.
Sharma, M.A. & Singh, J.B. “Use of Probability Distribution in Rainfall Analysis”, New York Science Journal, vol 3, no.9, 2010
Snedecor, G.W. & Cochran, W.G. Statistical Methods ( 6th ed.). Oxford and IBH, New Delhi., 1983
Downloads
Published
Issue
Section
License
Copyright (c) 2021 Kunlanan Sritan, Bumrungsak Phuenaree
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Copyright © The Author(s). This article is published under the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
