Thread by @Noahsasson

A thread about our new open-access paper, just out today. We tested how well standardized measures of social cognition, social skill, and social motivation predict real-world social interaction outcomes for autistic and non-autistic (NA) adults. frontiersin.org/articles/10.33…

First some background: a deficit model of autism assumes that autistic adults often struggle in interactions w/ NA adults because they have poor or less normative social abilities. Surprisingly this assumption is rarely tested. Seems important to do!

Many psychosocial treatments of autism implicitly use a deficit framing, presuming that training autistic adults to mimic more “typical” social behavior will lead to better real-world social success & life outcomes but this seldom happens in practice.

These treatments do sometimes increase performance on standardized assessments of social cognition, social skill, and social motivation but these effects don’t often generalize to improvements in personal and professional interactions. Why not?

For one, the “deficit” being treated is an individual characteristic while social interaction isn’t. A deficit model assumes autistic people are solely to blame for poor interactions w/ NA people but this ignores that interaction is bidirectional.

In this study, we examined social cognition, social skill, & social motivation in *both* autistic & NA adults and assessed how well they predicted their social interaction outcomes. Importantly we examined effects relationally not just individually.

Compared to NA adults, autistic adults on average scored lower overall on social cognitive tasks (e.g., emotion recognition & theory of mind), were rated less normative on social skills, and endorsed fewer normative indicators of social motivation.

Autistic adults were also evaluated less positively by unfamiliar partners following a “get to know you” conversation, and NA (but not autistic) adults were less interested in future interaction with them. This is all consistent w/ previous research.

However, we found very few links between autistic adults’ measured social abilities & their social interaction outcomes. In some cases, the social abilities of NA adults were actually more predictive, particularly when interacting w/ autistic people.

In fact, most of our findings indicate relational, not individual, effects on social interaction outcomes for autistic adults. These challenge a “deficit” framework & argue for a broader understanding of social disability in autism. For instance…

NA and autistic adults w/ less normative social skill were rated as more awkward by their partners, but this only predicted lower social interest for NA adults. For autistic adults, rating someone as “awkward” was unrelated to social interest in them.

Ratings of social skill really didn’t predict much else. What constitutes good “social skill” may vary across groups & individuals and a single holistic rating may be inadequate for capturing social skill w/in a dynamic and emergent interaction.

We argue that conceptualizing social skill as some objective metric in which people can be rank ordered & a single standard applied to all is unhelpful for predicting complex social dynamics, particularly between neurologically diverse people.

Performance on standard social cognitive tasks also wasn’t very predictive of social interaction quality or outcomes for autistic adults. W/in autistic-NA conversations, it was actually the social cognition of NA adults that was more predictive.

NA adults w/ better social cognition rated interactions w/ autistic adults as higher in quality. It could be that these NA adults are better able to perceive cues & infer mental states from autistic partners, which many NA adults struggle to do.

Taken at face value, this would mean that social cognitive training programs for NA people might be more effective at improving autistic-NA interaction than programs for autistic people. After all we didn’t find effects for autistic adults.

Another interesting finding: autistic adults perceived NA adults w/ better social cognition as *more* awkward. This seems counterintuitive but NA social cognitive ability may manifest behaviorally in ways perceived as undesirable by autistic people.

This suggests (again) that what constitutes “good” social ability is not determined individually but relationally, and compatibility between partner abilities & expectations drives social interaction outcomes rather than individual characteristics.

Our findings align w/ double empathy (@milton_damian) & dialectical misattunement (@dimitrisbolis) theories of autistic-NA disconnection. Traditional notions of social ability may not always extend in anticipated ways to autistic-NA interactions.

In sum, autism research would benefit from developing and validating measures of real-world social abilities within an interactive context and continue to emphasize relational rather than individual predictors of social interaction success.

Many more findings & a much deeper dive can be found in the full paper. It’s open access so please read & feel free to share. Big thanks to Kerrianne Morrison, @kmdebrabander, @DesiRJones, & Rob Ackerman for working on this absolute beast of a study. We're very proud of it!