Methods and Problems in the Analysis of Multivariate Data

The purpose of this paper is to review some meth ods for analyzing multivariate data in education and psychology and to discuss issues and problems which may often be associated with such analyses. In particular, the concern is with studying complexes of dependent variables with respect to multiple independent variables. Although this review is written chiefly in the language of experimental design, with special focus on analysis of variance methods, the discussion is also generally pertinent to situations where nonexperimental independent variables are under study. The ultimate aim of the review is to facilitate scientific understanding of subject matter questions through applications of methodology. For this reason, attention is drawn away from canons of mathematical statistics and associated statistical inferential procedures; description of variation in data is viewed here as more important than formal inferential procedures. While methodological theories are indispensable to scientific progress, the formal