Individual and Relational Social Impairments in Adults With Autism Spectrum Disorder

Psychosocial interventions aimed at improving social interaction in adults with Autism Spectrum Disorder (ASD) have produced small effects and often fail to generalize to improvement outside the laboratory (Gates, Kang, & Lerner, 2017). These limitations may stem from an incomplete understanding of how adults with ASD interact with others in real-world settings, and how the dynamic interplay between social partners’ behaviors, perceptions, and characteristics influence social outcomes. 125 ASD (n=67) and TD (n=58) adults were assigned to three different dyad types (ASD-TD n=25; ASD-ASD n=22; TD-TD n=23) and completed a five minute unstructured interaction with an unfamiliar peer. Afterwards, partners evaluated their perceptions of their partner and the interaction, and reported their metaperceptions of how they believed their partners would evaluate them. Participants also completed measures of social cognition and social motivation, and their social skills during the interaction were coded by trained raters. Using dyadic analytic techniques to quantify effects of actors, partners, and the interaction between the two, ASD adults were rated as more awkward, and less attractive and warm relative to TD adults by both TD and ASD and partners, with TD but not ASD partners reporting a lower desire to interact with them. Although ASD adults performed worse on measures of social cognition, social motivation, and social skill, these abilities in TD partners, not in ASD partners, were most predictive of interaction outcomes. TD adults with better social cognitive skills in ASD-TD pairs rated the interaction as higher in quality, but these TD adults were rated as more awkward by their ASD peers. Results also revealed that, contrary to prediction, ASD adults were more accurate than TD adults at predicting others’ desire to interact with them in the future, suggesting an awareness by autistic adults that TD individuals view them poorly. These findings suggest that while ASD adults are evaluated more unfavorably across many traits, few of their measured abilities in social cognition and social motivation predicted their social interaction outcomes, at least in the context assessed here, which raises questions about the ecological validity of some of these measures within ASD research In contrast, however, objectively measured social skill was predictive of ratings of awkwardness for both ASD and TD adults. In sum, findings replicate prior first impression work demonstrating less favorable evaluations of ASD adults relative to TD controls, and extend these findings to real-world interactions and to impressions formed by ASD raters. These results therefore support social interaction impairments in autistic adults under more ecologically valid conditions but attempts by this study to specify the mechanisms of these impairments demonstrated, at best, mixed success.