To determine whether supportive-expressive psychotherapy (SET), a form of dynamic psychotherapy, and pharmacotherapy + clinical management (MED) for major depressive disorder (MDD) are more effective than pill-placebo + clinical management (PBO). This National Institute of Mental Health (NIMH)-sponsored randomized controlled trial was conducted (from November 2001 through June 2007) at the University of Pennsylvania Medical School. The sample included 156 patients diagnosed with MDD (DSM-IV) and having a 17-item Hamilton Rating Scale for Depression (HRSD(17)) score ≥ 14 for at least 2 consecutive weeks. This was an underserved sample in which 41% were male, 52% were self-designated minorities, and 76% had an annual income under $30,000. Treatment lasted 16 weeks. Medication patients not responsive by week 8 (maximum dose 200 mg/d of sertraline) were switched to venlafaxine (maximum dose 375 mg/d). Nonresponsive placebo patients at week 8 were switched to a different placebo. Patients' depression improved over the 16 weeks (P < .0001), with no between-group differences (P = .95), even among severely (HRSD(17) score ≥ 20) depressed patients (P = .45). Response rates did not differ between groups (P = .73). Gender and minority status moderated outcome (P = .014), with psychotherapy more efficacious for minority men than MED (P = .027, Cohen d = 1.02) and PBO (P = .019, d = 1.09). PBO was more efficacious for white men than MED (P = .03, d = 0.62) and SET (P = .003, d = 1.07). For white women, MED (P = .005, d = 0.77) and SET (P = .033, d = 0.71) were more efficacious than placebo. No differences among treatments were found for minority women. This trial of urban MDD patients failed to confirm that either active treatment was better than placebo. Minority status and gender had significant and differential effects on outcome that warrant replication in future studies. clinicaltrials.gov Identifier: NCT00043550.