• In the National Institute of Mental Health Collaborative Study of the Psychobiology of Depression, six-month followup evaluations are available for 24 patients with schizoaffective disorder (depressed type), 56 with psychotic depression, and 274 with nonpsychotic major depression. Outcome for patients with schizoaffective depression was significantly worse than for patients with nonpsychotic depression. The psychotic depression group held an intermediate position on most outcome measures and on psychosocial measures had outcomes significantly worse than those of the nonpsychotic group. Recovery rates assumed a very similar pattern in another cohort admitted more than 40 years ago and followed up without somatic treatment. Follow-ups of 12, 18, and 24 months are available for proportions of each diagnostic group. Survival curves suggest similar outcomes in psychotic depression and nonpsychotic depression, whereas outcomes in schizoaffective depression remain disparate. These trends together with family history studies suggest that a small proportion of patients with schizoaffective disorder, depressed type, will have a long-term course consistent with schizophrenia. Moreover, these data show that outcome studies of schizoaffective disorder must control for follow-up length and the effects of psychosis per se.