Research on social skills interventions has been mixed. This study evaluates a group-administered, manualized social skills intervention program. Twenty-three boys and 9 girls between the ages of 7 and 13 participated. Participants were included in the groups based on peer relationship difficulties rather than diagnostic status, resulting in a diagnostically heterogeneous sample. As a group, children who were symptomatic at pretest made statistically significant and, in some cases, clinically significant improvements on socially relevant constructs. The results provide modest support for this intervention. Implications for future research, clinical work, and younger children are discussed. (Contains 4 tables and 3 figures.)