Self-categorization and Autism: Exploring the Relationship Between Autistic Traits and Ingroup Favouritism in the Minimal Group Paradigm.

Bertschy et al. (2020) asked adults to play a simple team game. The teams were fake—just random colors.

People with more autistic traits took the test online. The goal was to see if they still favored their new "team."

The study used the minimal-group trick: no history, no contact, just a label.

Higher autistic traits meant less favor for the new team.

The authors say this shows weaker self-categorization, fitting the ISCA model of autism.

Van Overwalle et al. (2025) extend the idea. They found autistic adults also learn picture categories more slowly and show odd brain responses. Together, the two studies map a broad category problem, not just a social quirk.

Jellema et al. (2009) and Jolliffe et al. (1999) set the stage. They showed autistic people miss automatic social cues and fail strange-story questions. Kristen’s result plugs neatly into that line: if you don’t auto-code "us versus them," ingroup bias never kicks off.

Beaurenaut et al. (2024) seem to clash. Their online game found autistic people could learn from social rewards as well as anyone. The key difference is task type: Beaurenaut used simple win/lose feedback, while Kristen used identity labels. Spared reward learning but weak self-tagging shows not all social learning is broken.

If clients don’t auto-slot themselves into groups, team-based social skills groups may fall flat. Start with explicit roles and clear badges. Check that the learner feels "inside" the group before you bank on peer pressure or modeling.