Multiple Comparison Procedures under Heteroscedasticity

Shun-Yi  Chen; Su-Hao  Lee

doi:10.6180/jase.2011.14.4.02

Multiple Comparison Procedures under Heteroscedasticity

Shun-Yi Chen This email address is being protected from spambots. You need JavaScript enabled to view it.¹ and Su-Hao Lee¹

¹Department of Mathematics, Tamkang University, Tamsui, Taiwan 251, R.O.C.

Received: October 21, 2010
Accepted: February 22, 2011
Publication Date: December 1, 2011

Download Citation: ||https://doi.org/10.6180/jase.2011.14.4.02

ABSTRACT

When the homogeneity hypothesis of treatment means under heteroscedasticity is rejected, one may be interested in comparing each treatment with all other treatments. The objective of this paper is to derive two multiple range test procedures, the Newman-Keuls type and Duncan type, for multiple comparisons of treatment means with unknown and unequal variances. We investigate these pairwise comparison procedures with the Tukey type simultaneous confidence intervals approach with respect to comparisonwise and experimentwise error rates, as well as correct decision rates. The tables for applying the new multiple range test procedures are provided, along with a numerical example to demonstrate the methods.

Keywords: Pairwise Comparisons, One-Stage and Two-Stage Sampling Procedures, Significance Level, Protection Level, Correct Decision Rate

REFERENCES

[1] Carmer, S. G. and Swanson, M. R., “An Evaluation of Ten Pairwise Multiple Comparison Procedures by Monte Carlo Methods,” Journal of the American Statistical Association, Vol. 68, pp. 66 74 (1973).
[2] Bishop, T. A. and Dudewicz, E. J., “Exact Analysis of Variance with Unequal Variances: Test Procedures and Tables,” Technometrics, Vol. 20, pp. 419 430 (1978).
[3] Chen, S. Y. and Chen, H. J., “A Range Test for the Equality of Means when Variances are Unequal,” American Journal of Mathematical and Management Sciences, Vol. 20, pp. 145 170 (2000).
[4] Chen, S. Y., “One-Stage and Two-Stage Statistical Inference under Heteroscedasticity,” Communications in Statistics: Simulation and Computation, Vol. 30, pp. 991 1009 (2001).
[5] Bishop, T. A., “Some Results on Simultaneous Inference for Analysis of Variance with Unequal Variances,” Technometrics, Vol. 21, pp. 337 340 (1979).
[6] Hochberg, Y., “A Modification of the T-Method of Multiple Comparisons for a One-Way Layout with Unequal Variances,” Journal of the American Statistical Association, Vol. 71, pp. 200 203 (1976).
[7] Tamhane, A. C., “Multiple Comparisons in Model I One-Way ANOVA with Unequal Variances,” Communications in Statistics - Theory and Method, Vol. 6, pp. 15 32 (1977).
[8] Wilcox, R. R., “A Table of Percentage Points of the Range of Independent t Variables,” Technometrics, Vol. 25, pp. 201 204 (1983).
[9] Chen, S. Y. and Chen, H. J. “Single-Stage Analysis of Variance under Heteroscedasticity,” Communications in Statistics - Simulation and Computation, Vol. 27, pp. 641 666 (1998).
[10] Duncan, D. B., “Multiple Range and Multiple F Test,” Biometrics, Vol. 1, pp. 1 42 (1955).
[11] Chen, H. J. and Lam, K. “Single-Stage Interval Estimation of the Largest Normal Mean under Heteroscedasticity,” Communications in Statistics - Theory and Methods, Vol. 18, pp. 3703 3718 (1989).
[12] SAS Institute Inc., SAS Language: Reference, Version 8, Cary, NC. (1999).
[13] Chen, S. Y., Chen H. J. and Ding C. G., “An ANOVA Test for the Equivalency of Means under Unequal Variances,” Computational Statistics and Data Analysis, Vol. 33, pp. 151 175 (2000).