Schizoaffective disorder is a relatively common illness with an uncertain relationship with bipolar disorder. The publication of DSM-III-R in 1987 operationalized the diagnosis of schizoaffective disorder, separating it from psychotic bipolar disorder by the presence of persistent psychosis in the absence of affective symptoms. Since that time, there have been few prospective outcome studies comparing schizoaffective and bipolar disorders. The authors recruited 27 hospitalized patients with schizoaffective disorder and compared their 12-month outcome to 27 sex-, age-, socioeconomic status- and race-matched hospitalized bipolar patients. The schizoaffective patients were significantly less likely to achieve syndromic recovery than the bipolar patients, although neither group achieved high rates of symptomatic or functional recovery. As expected, the schizoaffective patients were more likely to exhibit persistent psychosis, with or without affective symptoms, throughout the follow-up interval. These data provide clinical support of the predictive value of the DSM-III-R criteria for schizoaffective disorder in a naturalistic outcome setting.