|
Organizers |
Robustness and power of some statistics in one-way classification
by
Kunja Deka
Department of Statistics, Handique Girls' College, Guwahati 781001, Assam, INDIA
Coauthors: Bipin Gogoi, Department of Statistics Dibrugarh University, Dibrugarh 786004, Assam, INDIA
This paper deals with the study of robustness in terms of type I and type II errors of a few parametric and nonparametric test statistics under different distributions in one-way layout. The test statistics considered here are ANOVA F test, Kruskal-Wallis test, Studentized Brown and Mood test, Brown and Forsythe test, Welch multisample extension test and Trimmed test. The levels and the powers of the test statistics are calculated by using Monte Carlo Simulation technique. The performance of the test statistics are compared on the basis of the calculated values of their levels and powers respectively. It turns out that in case of equal sample size and equal variances in symmetric distributions Kruskal-Wallis test is the most powerful one whereas for unequal sample size and unequal variances Trimmed test and Welch test are to be preferred.
Date received: July 24, 2005
Copyright © 2005 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # caqt-33.