Fifty-one out-patients with social skills deficits (two-thirds men) completed ten weekly sessions of 75-minute group treatment; 44 were followed up for a mean of 16 months. Random assignment was to one of three conditions: (1) Cohesive group discussion; (2) Modelling and role-rehearsal; or (3) Modelling and role-rehearsal+daily social homework.
All three treatment conditions produced significant but incomplete improvement at the end of treatment and follow-up. The two role-rehearsal conditions were significantly superior to group discussion on several measures. Patients who completed daily social homework assignments did significantly better than patients who completed control homework. Alcohol and drug abuse patients usually dropped out. Schizophrenic patients in remission had lost their improvement at follow-up. Patients with other diagnoses retained their gains to 16-month follow-up.