Exploratory Cluster Analysis and Nonmetric Scaling of Multitrait-Multimethod Matrices

A review of existing methods for exploratory analysis of the multitrait-multimethod matrix finds them lacking because of inferential biases, overly stringent data input requirements, and linear assumptions. The advantages of a cluster analysis and nonmetric scaling approach to the exploratory analysis of multimethod matrix is illustrated in comparative analyses of “successful” and “weak developmental” matrices.