Non-autistic observers both detect and demonstrate the double empathy problem when evaluating interactions between autistic and non-autistic adults

Consistent with a "double empathy" framework, autistic adults often experience better interaction with autistic compared with non-autistic partners. Here, we examined whether non-autistic observers detect differences in autistic interactions relative to non-autistic and mixed ones. Non-autistic adults (N = 102) rated the interaction quality and traits of 42 autistic and 44 non-autistic male participants interacting in same or mixed-neurotype dyads. Non-autistic interactions and participants were evaluated most positively, with participants rated more favorably when interacting with non-autistic partners and rated as less intelligent and awkward when interacting with autistic partners. Observers perceived mixed interactions as the least successful overall. Whereas non-autistic interactions were rated as smoother and more enjoyable than mixed interactions, they were not rated differently from autistic interactions on any measure of interaction quality. Observers also perceived that non-autistic participants but not autistic participants disclosed more to non-autistic partners. However, they evaluated autistic participants more negatively than their partners in the interaction evaluated them; they disproportionately underestimated trust and intelligence ratings made by autistic participants; and they and reported lower social interest in participants than did the autistic and non-autistic people in the interactions. Collectively, these findings indicate that non-autistic adults both detect and demonstrate the double empathy problem when observing social interactions involving autistic people. Lay Abstract The "double empathy problem" refers to breakdowns in communication and understanding that frequently occur between autistic and non-autistic people. Previous studies have shown that autistic people often establish better rapport and connection when interacting with other autistic people compared to when interacting with non-autistic people, but it is unclear whether this is noticeable to non-autistic observers. In this study, 102 non-autistic undergraduate students viewed and rated video recordings of "get to know you" conversations between pairs of autistic and non-autistic adults. Sometimes the pairs were two autistic people, sometimes they were two non-autistic people, and sometimes they were "mixed" interactions of one autistic and one non-autistic person. Observers tended to rate non-autistic participants and their interactions the most favorably, but-consistent with the "double empathy problem"-they rated mixed interactions between autistic and non-autistic people as the least successful. They also perceived that only non-autistic people disclosed more when interacting with a non-autistic conversation partner. Autistic participants' partners in the conversations tended to evaluate them more favorably than did outside observers, suggesting that personal contact may facilitate more positive evaluations of autistic people. Furthermore, observers expressed less social interest in participants than did the autistic and non-autistic participants in the interactions. Together, these findings suggest that non-autistic observers both detect and demonstrate some aspects of the double empathy problem.