非均衡数据两因素方差分析中多种原假设形式的一致性

The Consistency in the Variety Form of the Null Hypothesis in Two-way ANOVA for Unbalanced Data

基于自变量之间的非正交性或相关性,非均衡数据两因素方差分析方法区别于经典的均衡数据方差分析。文章给出了非均衡数据两因素方差分析的模型退化理论与线性推断理论的等价性证明。给出了非均衡数据两因素方差分析的多种原假设表现形式及它们在三种不同处理方法中的不同应用方便性。

Based on the non-orthogonality or correlation between independent variables, two-way ANOVA for unbalanced data is different from the classical ANOVAfor unbalanced data. This paper gives the equivalence between the model reduction the- ory and the linear inference theory in two- way ANOVA for unbalanced data. This paper gives the variety form of the null hypothe- sis in two-way ANOVA for unbalanced data and their different convenient application in three different processing methods.