Abstract
It has been suggested that the procedures of construct validation include at least two distinct processes: one is the validation of the test for measuring a hypothesized construct and the other is the validation of the construct, i.e., its usefulness and significance as a part of a theoretical system. Factor analysis has been re-evaluated as a method for studying construct validity of new instruments in face of both theoretical and computational advances in the field of factor analysis. A numerical index of construct validity has been suggested to be the multiple correlation of the observed variables with the factor (hypothesized construct). The notion of universe of content, the limiting value of the multiple correlation when n → ∞, etc. have been shown to have important implications for the concept of construct validity. The usefulness of such a procedure for the study of content and predictive validities, item analysis, interpretation of scores and validation of psychological theories has been outlined.
Get full access to this article
View all access options for this article.