The subjects were children in the second, third, and fourth years at an Oxford middle school (age 10-13 years), who were screened to identify those who were observed to have behavioral problems and/or experience difficulties in social interactions, by teacher or peer report. The selected subjects were allocated at random to eight sessions of social skills training with four booster sessions later or to a no treatment control group. Change was assessed by parent, teacher, and self report measures. Results suggested significant improvement in social activity, parental report of social behavior, and self-esteem in the treatment group as compared to the control group. The same results were noted at a six-month follow-up. The article discusses the implications for school-based social skills programs. 8 tables, 24 references. (Author abstract modified)