重复测量中两组均值是否相等的假设检验
Hypothesis Testing for Two Group Means with Repeated Measurements
-
摘要: 在重复测量中,存在着检验两个随机试验组的各次测量均值是否相等的问题,假定所有试验对象初始值无差异,那么只需对所有均值的对比进行检验就可以了;若这个假定不成立,则需要直接检验各次测量的均值是否均对应相等,在正态性、复合对称的假定下,前一种情形已有很好的结论,针对后面一种情况,本文采用似然比检验的方法进行了检验,构造出了检验统计量的分布函数表,并对计算的精确度作了较好的控制。Abstract: The autocorrelation of repeated measurements makes it different from independent measurements. In comparing two group effects, without the hypothesis of same initial values, we must compare the overall group means. With the assumption of normal distribution and compound symmetry, we use the Likelihood ratio test to solve this problem in this paper. A method of controlling the calculation precision is also given.