MMPI-A and Rorschach indices related to depression and conduct disorder: an evaluation of the incremental validity hypothesis.

This investigation extends the earlier research by Archer and Gordon (1988) by examining the extent to which combining indices from the newly released MMPI-A and the revised Rorschach Comprehensive System furnishes incremental validity in terms of improved diagnostic prediction. The predictive accuracy of selected MMPI-A and Rorschach variables conceptually related to diagnoses of depression and conduct disorder were compared in a clinical sample of 152 adolescents. Results of multivariate analyses of variance revealed some significant differences between diagnostic groups on several MMPI-A scales, and 1 significant difference on the Rorschach involving the Vista variable. Stepwise discriminant function analyses resulted in 2 MMPI-A scales and 2 Rorschach variables that collectively accounted for a small proportion of variance in the diagnosis of depression, and 3 MMPI-A scales that accounted for a significant component of variance in the conduct disorder diagnosis. Classification accuracy results indicated that the hit rate for depression diagnosis did not improve using an optimal linear combination of the 4 variables over the rates produced by the single use of either the MMPI-A Depression content scale (A-DEP) or Scale 2. For the conduct disorder diagnosis, the optimal linear combination of MMPI-A Conduct Problems (A-CON), Cynicism (A-CYN), and Immaturity (IMM) scales served as the best predictor, and no Rorschach variables contributed significantly to classification accuracy. Our results replicated the findings of Archer and Gordon (1988) in indicating that the combined use of MMPI-A and Rorschach variables does not appear to produce incremental increases in accuracy of diagnostic classification.