Predictors of long-term outcome were identified among a sample (n = 163) of largely chronic DSM-III schizophrenic patients (80% ill for more than two years) from the Chestnut Lodge follow-up study using three methods of analysis (correlation, multiple regression, and discriminant function). Predictors accounted for approximately one third of the outcome variance across six outcome dimensions. The following variables regularly (ie, individually and as components in multivariate equations) predicted better global outcome: less family history of schizophrenia, better premorbid instrumental functioning (interests and skills), more affective signs and symptoms (especially depression) in the manifest psychopathology, and absence of psychotic assaultiveness.