Since autism was first described, major difficulties in social interaction have been a defining feature of individuals with autism spectrum disorders (ASD). Social skills groups are a common intervention for individuals with ASD. Although a frequently recommended practice, the few studies that have addressed the efficacy of social skills groups have shown mixed results. To determine the effectiveness of social skills groups for improving social competence, social communication, and quality of life for people with ASD who are six to 21 years of age. We searched the following databases in December 2011: CENTRAL (2011 Issue 4), MEDLINE (1948 to November Week 3, 2011), EMBASE (1980 to Week 50, 2011), PsycINFO (1887 to December Week 2, 2011), CINAHL (1937 to current), ERIC (1966 to current), Sociological Abstracts (1952 to current), OCLC WorldCat (12 December 2011), Social Science Citation Index (1970 to 16 December 2011), and the metaRegister of Controlled Trials (20 December 2011). We also searched the reference lists of published papers. Randomized control trials (RCTs) comparing treatment (social skills groups) with a control group who were not receiving the treatment for participants aged six to 21 years with ASD. The control group could be no intervention, wait list, or treatment as usual. Outcomes sought were standardized measures of social competence, social communication, quality of life, emotion recognition, and any other specific behaviors. Two review authors independently selected and appraised studies for inclusion and assessed the risk of bias in each included study. All outcome data were continuous and standardized mean difference effect sizes (ES) with small sample correction were calculated. We conducted random-effects meta-analysis where possible. We included five RCTs evaluating the effects of social skills groups in 196 participants with ASD aged 6 to 21 years old. The results show there is some evidence that social skills groups improve overall social competence (ES = 0.47, 95% confidence interval (CI) 0.16 to 0.78, P = 0.003) and friendship quality (ES = 0.41, 95% CI 0.02 to 0.81, P = 0.04) for this population. No differences were found between treatment and control groups in relation to emotional recognition (ES = 0.34, 95% CI -0.20 to 0.88, P = 0.21) assessed in two studies or social communication as related to the understanding of idioms (ES = 0.05, 95% CI -0.63 to 0.72, P = 0.89), which was assessed in only one study. Two additional quality of life outcomes were evaluated, with results of single studies suggesting decreases in loneliness (ES = -0.66, 95% CI -1.15 to -0.17) but no effect on child or parental depression. No adverse events were reported. Given the nature of the intervention and the selected outcome measures, the risk of performance and detection bias are high. There is limited generalizability from the studies as they were all conducted in the US; they focused mainly on children aged 7 to 12, and the participants were all of average or above average intelligence. There is some evidence that social skills groups can improve social competence for some children and adolescents with ASD. More research is needed to draw more robust conclusions, especially with respect to improvements in quality of life.