The participants included 47 inpatients at the Western State Hospital in Virginia, 32 outpatients at regional mental health centers in Virginia, and 35 siblings or parents of the patients. Each participant completed diagnostic and vocabulary tests and took part in a gambling procedure that involved 28 paired alternatives; monetary rewards were based on the gambling performance. Results revealed that the inpatients were significantly less able to weigh risks, benefits, and probabilities in a consistent manner than the outpatients. Similarly, the inpatients performed significantly worse on this measure of decisionmaking than their nonpatient family members. However, when the vocabulary subtest score of the Wechsler Adult Intelligence Scale-Revised was statistically controlled, no significant differences between any of the groups remained. Further analysis revealed that the vocabulary subtest predicted decisionmaking behavior for the outpatients and the controls, but not for the inpatients. The severity of psychiatric symptoms and the number of prior hospitalizations predicted the decisionmaking for the inpatients. Results suggested that competence assessments that rely mainly on verbal abilities may be inadequate to assess competence in acutely ill psychiatric patients. Figures, table, and 30 references