Social Adaptation to Change in Peer Behavior Before and After a Pilot Randomized Clinical Trial of a Theatre-Based Intervention for Adults With Autism.

Small numbers: 47 people were analysed after the intervention and 61 before it, so only a large effect would have been detectable. The waiting-list comparison is a pilot, not a definitive trial. The participants were mostly white, non-Hispanic and of average or above-average intelligence, with an IQ above 70 required for entry. The authors state the sample was not representative and that it is unclear how far the results apply to other autistic adults, including those with an intellectual disability. Depression was measured only by self-report questionnaire, with no clinical assessment of whether anyone actually met criteria for depression and no information on other co-occurring conditions. The trained conversation partners were instructed to act interested or bored regardless of what the participant did, and their behaviour was never independently checked, so the test is the participant responding to a fixed script rather than a two-way conversation. Several participants worked out at the second testing that they would meet one interested and one bored partner, so the test may not resemble an unpredictable real conversation. The encouraging result in the 10 people who started with the least adaptive behaviour came from a post-hoc analysis, was not statistically significant, and is too small to act on.

One author, Blythe Corbett, founded SENSE Theatre, the non-profit organisation behind the programme, but the paper states she receives no financial compensation from it. The other authors declared no conflicts of interest. No funding statement appeared in the text supplied.

Do not read the exploratory result in the 10 participants who started with the least adaptive behaviour as evidence the programme works. That finding favoured the theatre group with a large effect size but did not reach significance (p = .163), involved only 10 people, and came from a post-hoc analysis; the main whole-group analysis found no difference between the groups.