Variability in first impressions of autistic adults made by neurotypical raters is driven more by characteristics of the rater than by characteristics of autistic adults.

Barton et al. (2019) asked 159 neurotypical adults to watch short videos of 20 autistic adults. Each rater said how much they wanted to talk with the person on screen.

Half the raters were told the person had autism. Half were told nothing. The team then looked at what best predicted the rater’s choice: the autistic traits shown in the video or the rater’s own stigma and autism knowledge.

The rater’s stigma level, not the autistic traits seen in the video, most strongly shaped their willingness to interact. When raters had low stigma, saying “this person has autism” helped a little. When raters already had high stigma, the label made things worse.

Higgins et al. (2021) conceptually replicate this finding with college students. They showed that special interests do not add extra stigma to an autism label—again, the label itself drives rejection.

Someki et al. (2018) extend the line by showing that a 30-minute online training can lower stigma in US and Japanese students. The training works because it targets the rater, not the autistic person.

KAgiovlasitis et al. (2025) look at the same population from a different angle. They find that higher autistic traits predict less eye gaze to social pictures in Indian adults. This supports the idea that autistic traits are real and measurable, yet E et al. show that stigma still comes mainly from the observer.

If you run social-skills groups or workplace inclusion programs, focus on the environment, not just the learner. Brief stigma-reduction lessons for peers, teachers, or co-workers can open doors faster than teaching autistic adults to mask their traits. Try adding a short autism facts sheet or 10-minute discussion before peer-mentoring starts.