We had previously reported that imipramine was superior to placebo for the treatment of chronic depression. As a part of that study, we subsequently investigated clinical and demographic variables which might be associated with favorable or poor outcome for treatment with imipramine or placebo. Results are reported herein. Eight-six patients were entered and 53 completed an 8-week protocol. Outcome was assessed based on a 6-week, double-blind treatment phase, which followed a 2-week, single-blind placebo phase. Outcome was not found to significantly relate to demographic variables, severity or course of depression, diagnostic subtype, symptom profile, or DST results. Some modest associations were found between 'neurotic' personality traits and poor outcome. Results are discussed and compared with prior studies of prediction of tricyclic antidepressant response in both acute and chronic depressions.