Poster Session 10 Program Schedule
02/17/2024
09:00 am - 10:15 am
Room: Shubert Complex (Posters 1-60)
Poster Session 10: Neurodevelopmental | Congenital Conditions
Final Abstract #27
Can autistic and non-autistic viewers accurately identify autistic and non-autistic people in conversation?
Catherine Crompton, University of Edinburgh, Edinburgh, United Kingdom
Holly Sutherland, University of Edinburgh, Edinburgh, United Kingdom
Harriet Axbey, Durham University, Durham, United Kingdom
Danielle Ropar, University of Nottingham, Nottingham, United Kingdom
Sue Fletcher-Watson, University of Edinburgh, Edinburgh, United Kingdom
Category: Autism Spectrum Disorders/Developmental Disorders/Intellectual Disability
Keyword 1: autism spectrum disorder
Keyword 2: social cognition
Objective:
Autistic social communication difficulties are well described, and are a core component of diagnosis. Recent research has suggested that social communication between autistic people may be more comfortable and successful than communication between autistic and non-autistic people. This suggests that when two autistic people interact there may be no evidence or experience of communicative difficulties.
In this study, we examined (RQ1) whether it was possible for autistic and non-autistic viewers to correctly identify whether a person is autistic or not based on a brief video of them interacting; (RQ2) whether people are more accurate at identifying someone of their own neurotype – do autistic people more accurately identify other autistic people?; and (RQ3) whether autistic people were easier to detect when taking part in cross-neurotype interactions compared with autistic-autistic interactions.
Participants and Methods:
Autistic and non-autistic viewers (n = 78, 39 autistic; 28M, 2NB; age mean = 34, range 18-64) watched short video clips of autistic dyads, non-autistic dyads, and mixed (one autistic person and one non-autistic person) dyads in conversation. Viewers were then asked to guess whether each member of the dyad was autistic.
Results:
FDR-corrected two-tailed one-sample t-tests (with a reference value of 0.5) were used to establish whether accuracy ratings differed significantly from chance (50%). Both autistic and non-autistic viewers correctly identified diagnostic status with an accuracy significantly greater than chance (RQ1) (non-autistic viewers 58% (95% CI: 0.51 – 0.64; p< 0.05); autistic viewers 59% (95% CI: 0.53 – 0.65; p< 0.05)).
Non-autistic viewers were more accurate at identifying non-autistic people: autistic viewers performed poorly at this (RQ2). Autistic dyads were not well identified by either group of viewers: autistic viewers performed at chance, and non-autistic viewers performed well below chance (RQ2). Autistic people were easier to detect in the mixed dyad condition, where autistic viewers also outperformed.
Conclusions:
It is possible for autistic and non-autistic viewers to accurately identify whether someone is autistic or non-autistic to a greater likelihood than chance based on a short video interaction. However, there are factors that impact accuracy, including the diagnostic status of the person viewing the video and the social context of the interaction: whether it is between an autistic dyad, a non-autistic dyad or a mixed dyad. Mixed dyads appear to make autistic traits more ‘visible’ in comparison to autistic dyads, particularly for autistic viewers. More research is needed to determine whether this is an own-group guessing bias, or whether autistic and non-autistic people deploy different strategies in identifying whether someone is autistic or not.
| 
