Advances and Applications in Statistics
Volume 7, Issue 3, Pages 417 - 434
(December 2007)
|
|
EXACT PERMUTATION PARADIGM IN MULTISAMPLE EXPERIMENTS
J. I. Odiase (Nigeria) and S. M. Ogbonmwan (Nigeria)
|
Abstract: For the commonly used F-test in analysis of variance, the permutation of the observations in an experiment is the only possible way of truly constructing exact tests of significance. The permutation paradigm requires no distributional assumptions for a test statistic andtherefore no requirement that it should agree with a mathematically definable frequency distribution. The exact permutation algorithms of a multisample experiment to actually generate all the distinct permutations of the observations of an experiment is presented and implemented for analysis of variance problems. The one-way analysis of variance problem is examined empirically by using the algorithms developed in this paper to generate the exact distribution of the F statistic. |
Keywords and phrases: algorithm, ANOVA, complexity, permutation test, p-value. |
|
Number of Downloads: 380 | Number of Views: 1160 |
|