Accuracy and confidence in competency to stand trial evaluations: A comparison of clinical and legal decision making

This study used a three group (forensic psychologists, clinical psychology graduate students, and lawyers) lens-model design. Protocols containing demographic, psychological, and criminological data taken from actual defendants referred for competency evaluations at an inpatient psychiatric facility were presented to participants. Participants were asked to make decisions of competency to stand trial and to rate their confidence in the decision for each protocol. Participant decisions were compared to decisions of competency to stand trial rendered by the legal system. Results suggest courts relied on a defendant's performance on measures of competency to stand trial, hospitalization history, and criminal offense while the majority of raters primarily relied on a defendant's performance on measures of competency to stand trial when making decisions of competency. There were no substantial differences in accuracy between the different disciplines. As predicted, confidence was unrelated to accuracy. For most raters there were no differences between human models of judgments and linear models of judgments on accuracy rates. When differences were observed, human models tended to outperform linear models of judgments.