Evaluating Holland's and Gati's vocational-interest models: A structural meta-analysis

In a recent review, Gati (1991) criticized Holland's model of vocational interests and concluded that his own hierarchical model was superior. A major problem with Gati's comparison of his and Holland's models was that it was based on an incomplete specification of Holland's model. The relative validity of Holland's order and circumplex models and Gati's 3-group partition model as they fit 104 published (1965-1989) correlation matrices is evaluated. Using 3 separate structural meta-analytic techniques (Hubert & Arabic's, 1987, randomization test of hypothesized order relations, confirmatory factor analysis, and individual-differences cluster analysis), the fit of Holland's and Gati's models to the data were compared. Holland's order and circumplex models were found to be adequate representations of the structure of vocational personalities and work environments and were found to be superior to Gati's model.