Behavioral Analysis of Clinical Judgment

Abstract

Research on clinical judgment has produced conflicting information about effects of training on predictive accuracy, effects of the information available to the clinician, the correlates of accurate judgment, as well as the stability of such judgments. Such conflicting findings are largely related to lack of an adequate methodology. The present study examined the accuracy of judges making longitudinal clinical judgments with systematic increments in available information. Six judges, two non-experienced first-year graduate students in psychology, two medium-experienced practicum students, and two advanced interns, were asked to make discriminations between test protocols belonging to men convicted of first and second degree murder and men convicted of crimes against property. A discriminant function analysis classified 83% of these protocols accurately. Levels of information were increased in four phases. Most of the judges discriminated at a 50% accuracy level. No substantial increment in accuracy across phases was found. Judges of low accuracy discriminated best with the addition of the statistical findings. Daily judgments remained stable within the guessing range.