Empathic accuracy in adults with a pervasive developmental disorder during an unstructured conversation with a typically developing stranger.

Ponnet et al. (2005) watched high-functioning adults with pervasive developmental disorder talk with a stranger. The chat had no script. Afterward the researchers asked each adult what the other person had thought or felt during the talk.

They asked the same questions to typical adults who had similar chats. Then they compared how often each group got the thoughts and feelings right.

The adults with developmental disorders scored just as well as the typical adults. Both groups guessed the partner's hidden thoughts and feelings with the same accuracy.

Ponnet et al. (2008) seems to disagree at first. In that later study young adults with ASD did worse when the talk was loose, but caught up when the talk followed clear rules. The key difference is structure: the 2005 chat was free-form, yet the sample was older and higher-functioning, so they handled the chaos.

Rum et al. (2025) adds a twist. They showed that simply telling a partner "I'm autistic" made the partner better at reading the autistic speaker's emotions. Together the three papers say: autistic adults can read others fine, and others can read them better once they know the label.

Granieri et al. (2020) widens the lens. Autistic adults still got rated more negatively, even though their accuracy was okay. The takeaway: skill and reputation are separate. You may need to work on both.

Stop assuming clients can't read faces or feelings. Instead, check if the setting is too loose or if negative bias is the real wall. Try adding clear turn-taking rules or letting clients disclose their diagnosis when they want. Measure social success by connection, not just correct answers.