样本数相同组方差齐性检验的探讨  

Research on Equdlity Examination of Variances of Serieses With Same Number Of Samples

作　　者：包和平[1] 包德喜[1] 牛英春[1] 

机构地区：[1]内蒙古民族大学数学与计算机科学学院,内蒙古通辽028043

出　　处：《内蒙古民族大学学报（自然科学版）》2006年第4期381-382,共2页Journal of Inner Mongolia Minzu University：Natural Sciences

摘　　要：方差分析的应用条件是各总体的方差相等.因此,在方差分析之前,需进行方差齐性检验,一般用Bartlett检验.但这一方法要求分布是正态分布且计算繁琐,而用Hartleg检验和Cochran检验计算方便,方法实用;尤其Cochran检验利用了更多的样本信息,因此显得更灵敏.对有轻微偏倚的分布,Cochran检验也适用.Applying Conditions of Variance analysis are that variance of any a totality is equal to another.So,the equality examinations(Bartlett examination is adopted usually) are needed before Variance analysis,But the eistribution must be normal distribution and calculation is complilated in the method.Hartleg examination and Cochran examination can give us convenience,Forthe more,Cochran examination is better because it makes use of more sample information.At the same time,Cochran examination is adapt to slight deviational distribution.

关 键 词：方差齐性检验 Hartleg检验 Cochran检验 

分 类 号：F212.1[经济管理—国民经济]