PROBLEM
Clinical judgment consists of accuracy (outcome) and the contributing ratiocinations (process). In quasi-clinical contexts, accuracy is unrelated to experience('* but may be increased by formal training'l', 12). Process designs have suggested that cognitive set(ls), sex of judge(17* l a ) , and kind of feedback''') are related to accuracy.
Relatively simple, linear models as well as more complex, configural models may be used to study clinical judgment. Goldberg'?) prefers linear models and finds that clinical judgment is inaccurate and unreliable, with confidence, experience and amount of information, all unrelated to accuracy. Nonetheless, configural models merit examination because clinicians believe their interpretations stem from configural analysis. Brunswik(*) was first to describe the two-sided, multiple-cue, probability model (Lens), a correlational method for abstracting the relative importance of cues in a response series'", lo).
The present series of studies deals with prediction of Rorschach human movement, or M%, within a Lens Model design, based on data from tasks which were found to be M-related in the empirical literature"). An initial study(s) considered the predictions of graduate student judges. This report compares M judgments by this original sample with other samples of judges varying in clinical experience and task-relevant training.