An Analysis of Correlation Matrices: Variables Cross-Classified by Two Factors
提出一种类似方差分析的方法，用于检验按两个因素交叉分类的相关矩阵（如多特质多方法矩阵）的结构，检验等相关假设并指示备选模型。
A procedure similar to an analysis of variance is presented for examining the structure in correlation matrices whose variables are cross-classified by two factors, such as multitrait-multimethod matrices. The method tests an equal correlation hypothesis against natural alternative hypotheses suggested by group representations. When there is significant departure from the hypothesized pattern, alternative models to be pursued are indicated. The method is illustrated using an example.