A foundational assumption shaping modern autism discourse — that autistic people communicate better with each other than with neurotypical peers — has come under direct empirical scrutiny, with results that complicate the prevailing narrative and raise fresh questions about what "empathic accuracy" actually measures in neurodivergent populations.
In a controlled empathic accuracy paradigm involving 106 adults (55 with autism, 51 neurotypical), participants watched video clips of a dyadic interaction between one autistic and one non-autistic individual, then estimated what each person was thinking at specific time points. Contrary to the hypothesis derived from the double empathy problem (DEP) framework — which predicts superior thought-inference among autistic dyads — adults with autism were not more accurate at reading the mental states of other autistic individuals. In fact, they scored generally lower on empathic accuracy across conditions. Additionally, autistic individuals' inner states were rated as harder to infer overall, suggesting a directional asymmetry in legibility rather than a simple bidirectional mismatch.
The double empathy problem, proposed by Damian Milton in 2012, has rapidly gained traction in both clinical and self-advocacy communities as a reframe shifting blame for social friction away from autistic deficits toward neurotype incompatibility. This study does not disprove that premise entirely — communication friction may still be mutual — but it does challenge the specific prediction that same-neurotype matching confers measurable cognitive empathy advantages in controlled settings. An important limitation is ecological validity: laboratory video-clip paradigms strip out the dynamic, reciprocal nature of live interaction, where autistic individuals may leverage shared context or behavioral cues unavailable in passive observation. The sample size, while reasonable, may also lack the power to detect subtle within-group effects. This is a single study, and the field needs larger, longitudinal, and interaction-based designs before the DEP framework is substantially revised. For now, the finding is best read as a caution against overclaiming the theory's empirical foundations.