ArticleOriginal scientific text
Title
Multivariate multiple comparisons with a control in elliptical populations
Authors 1, 2
Affiliations
- Department of Food Sciences, Tokyo Seiei College 1-4-6, Nishishinkoiwa, Katsushika-ku, Tokyo 124-8530 Japan
- Department of Mathematical Information Science, Tokyo University of Science 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan
Abstract
The approximate upper percentile of Hotelling's T²-type statistic is derived in order to construct simultaneous confidence intervals for comparisons with a control under elliptical populations with unequal sample sizes. Accuracy and conservativeness of Bonferroni approximations are evaluated via a Monte Carlo simulation study. Finally, we explain the real data analysis using procedures derived in this paper.
Keywords
comparisons with a control, Bonferroni approximation and Monte Carlo simulation
Bibliography
- F. Bretz, M. Posch, E. Glimm, F. Klinglmueller, W. Maurer and K. Rohmeyer, Graphical approaches for multiple comparison procedures using weighted Bonferroni, Simes, or parametric tests, Biometrical Journal 53 (2011) No. 6, 894-913. doi: 10.1002/bimj.201000239.
- T. Iwashita, Asymptotic null and nonnull distribution of Hotelling's T²-statistic under the elliptical distribution, Journal of Statistical Planning and Inference 61 (1997) 85-104. doi: 10.1016/S0378-3758(96)00153-X .
- R.J. Muirhead, Aspects of Multivariate Statistical Theory (New York, Wiley, 1982). doi: 10.1002/9780470316559.
- N. Okamoto, A modified second order Bonferroni approximation in elliptical populations with unequal sample sizes, SUT Journal of Mathematics 41 (2005) 205-225.
- N. Okamoto and T. Seo, Pairwise multiple comparisons of mean vectors under elliptical populations with unequal sample sizes, Journal of the Japanese Society of Computational Statistics 17 (2004) 49-66.
- T. Seo, The effects of nonnormality on the upper percentiles of
statistic in elliptical distributions, Journal of the Japan Statistical Society 32 (2002) 57-76. - T. Seo and M. Toyama, On the estimation of kurtosis parameter in elliptical distributions, Journal of the Japan Statistical Society 26 (1996) 59-68. doi: 10.14490/jjss1995.26.59.