Integration of information in a clinical judgment task, an empirical comparison of six models

Six models were compared for their effectiveness in reproducing six clinical psychologists' judgments of 38 patients on intelligence, ability to establish contact, and control of affect and impulses. In two of the models, subjective weights were used in the prediction of a judge's ratings. The judges based their judgments solely on verbal protocols from the Rorschach, a sentence completion test, and the Thematic Apperception test. The stability of the linear aspect of the judgment process was very high but decreased as the depth of interpretation of the rating variable increased. The nonlinear aspect of the judgment process had considerably low stability. In general, a model based on subjective weights was most effective in reproducing the judges' ratings.