Key points are not available for this paper at this time.
There has been a sharp increase during the past decade in the use of multivariate analysis in research in the behavioral sciences and education. This is a healthy trend because educational research inherently involves many variables, both as predictors and criteria. We seldom seek to predict educational outcomes from a single variable, and we seldom assess them with single criterion variables.
Maurice M. Tatsuoka (1973) studied this question.