A Group-Administered social Skills Training for 8- to 12- Year-Old, high-Functioning Children With Autism Spectrum Disorders: An Evaluation of its Effectiveness in a Naturalistic Outpatient Treatment Setting.

The groups were not truly randomised; allocation depended on waiting-list availability and was described as quasi-random. Most children also received other treatment before or during the study, so the effect of the social skills training alone cannot be separated. Parents knew whether their child had started training, so parent ratings were not blinded. The sample was mostly boys, with 47 boys and 5 girls, so findings may not apply equally to girls. 29 of 81 eligible children did not participate, and outcome data were missing at some time points. The diagnoses used DSM-IV-TR categories, including PDD-NOS, and no standardised autism diagnostic instruments were used. The paper-and-pencil Social Skills Observation had not been explicitly psychometrically tested in this version. Follow-up was only 3 months, so longer-term effects are unknown. Loneliness was a secondary outcome and did not improve more than waiting list.

Do not conclude that the training alone caused the social skills gains or that it reduces loneliness. Allocation was quasi-random, 79% of children also received other treatment before or during the study, and loneliness did not improve more than waiting list.