TESTING THE EQUALITY OF CENTRAL TENDENCY MEASURES USING VARIOUS TRIMMING STRATEGIES
Keywords:
Heterogeneity, asymmetric, trimmed mean, MOM, robustnessAbstract
Ft statistic test is a non-classical method of comparing two or more groups. This statistical procedure is able to handle problems of sample locations when non-normality occurs but the homogeneity of variances assumption still applies. This method is not robust under the existence of variance heterogeneity. To make this test less sensitive when either one or both of the common assumptions are violated, in this study the test is modified and improved by replacing the test’s original central tendency measure that is, the fixed symmetric trimmed mean with a predetermined asymmetric trimmed mean and a modified one-step M-estimator (MOM) trimmed mean. The finding shows that when the data is suspected to be extremely skewed, then, it will be advantageous to adopt MOM procedures for homogeneous variance cases. On the other hand, for heterogeneous variances, a trimmed mean which uses predetermined asymmetric trimmed mean should be considered as an alternative, particularly for testing the equality of four groups
Published
How to Cite
Issue
Section
Copyright (c) 2024 International Journal of Interdisciplinary Research in Statistics, Mathematics and Engineering (IJIRSME)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
References
Babu JG, Padmanabhan AR, Puri MP (1999). Robust one-way ANOVA under possibly non-regular conditions. Biometrical J., 41(3), 321 – 339.
Bradley JV (1978). Robustness? British J. Math. Stat. Psychol., 31: 144 - 152.
Huber PJ (1981). Robust statistics. New York: Wiley.
Keselman HJ, Othman AR, Wilcox RR, Fradette K (2004). The new and improved two-sample t test. Psychol. Sci., 15: 47-51.
Kulinskaya E, Staudte RG, Gao H (2003). Power approximations in testing for unequal means in a one-way ANOVA weighted for unequal variances. Commun. Stat. – Theory Methods, 32: 2353 – 2371.
Lee H, Fung KY (1985). Behaviour of trimmed F and sine-wave F statistics in one-way ANOVA. Sankhya: The Indian J. Stat., 47 (Series B): 186 - 201.
Lix LM, Keselman HJ (1998). To trim or not to trim: Tests of location equality under heteroscedasticity and nonnormality. Educ. Psychol. Measure., 58(3): 409 – 429.
Luh WM, Guo JH (1999). A powerful transformation trimmed mean method for one-way fixed effects ANOVA models under non- normality and inequality of variances. British J. Math. Stat. Psychol., 52: 303-320.
Othman AR, Keselman HJ, Padmanabhan AR, Wilcox RR, Fredette K (2004). Comparing measures of the 'typical' score across treatment groups. British J. Math. Stat. Psychol., 57: 215-234.
Reed III JF, Stark DB (1996). Hinge estimators of location: Robust to asymmetry. Comput. Methods Programs Biomed., 49: 11 – 17.
Reed III JF (1998). Contributions to adaptive estimation.
: 651 – 669.
SAS Institute Inc. (1999). SAS/IML User's Guide version 8. Cary, NC: SAS Institute Inc.
Staudte RG, Sheather, SJ (1990). Robust estimation and testing. New York: Wiley.
Wilcox RR (1997). Introduction to robust estimation and hypothesis testing. San Diego, CA: Academic Press.
Wilcox RR, Keselman HJ, Kowalchuk, RK (1998). Can tests for treatment group equality be improved? The bootstrap and trimmed means conjecture. British J. Math. Stat. Psychol., 51: 123 – 134.
Wilcox RR, Keselman HJ (2002). Power analysis when comparing trimmed means. J. Modern Appl. Stat. Methods, 1(1): 24-31.
Wilcox RR, Keselman HJ (2003). Modern robust data analysis methods: Measures of central tendency. Psychol. Methods, 8: 254-274.
Yuen KK (1974). The two-sample trimmed t for unequal population variances. Biometrika, 61: 165 – 170.
