Some New Results about the MLJltisample Permutation Test
Suppose that there are c treatments with effect t1,...,tc, respectively. It is desired to test the hypothesis H0: t1=...=tc. In this paper we assume that not only there exist random experimentation errors, but the experiment units are non-uniform. In this most general framework we show that the permutation test based on the usual F-statistic is asymptotically equivalent to the ordinary F-test, both under the null hypothesis H0 and such alternatives as having the form ti = t + bi /v N/(1/2), i = 1, L, c, where b1,...,bc, do not depend on N.
Multisample Permutation Test Random Ezperimentation Errors Ezist F-test
JIN Liangqiong JIN Mingzhong
The College of Mathematics and Computer, Guizhou University for Nationalities, Huaxi, Guiyang, P.R.China, 550025
国际会议
2009 International Institute of Applied Statistics Studies(2009 国际应用统计学术研讨会)
青岛
英文
1-7
2009-07-25(万方平台首次上网日期,不代表论文的发表时间)