Passing the Test Is Not the Same as Using the Skill

What the research found

Theory of mind is the everyday ability to grasp that other people have their own thoughts, beliefs and intentions, and that these can differ from your own. It is the mental step behind reading a raised eyebrow, catching sarcasm, or noticing that a friend has gone quiet because they are upset. Autistic children often find this harder than their peers. This study set out to ask a sharper question: do all autistic children struggle with it in the same way, and does doing well on the standard tests translate into doing well with real people?

The researchers studied 89 children aged 7 to 11. Fifty-two were autistic without intellectual disability, and 37 were typically developing children used for comparison. All had intelligence within the normal range, so the findings are not about IQ. The team measured theory of mind two ways. One was an explicit, test-like task, where a child reasons out what a character in a story knows or believes. The other, drawn from a parent questionnaire, rated how the child applies that reasoning in real situations.

When they grouped the 52 autistic children by ability, two distinct profiles appeared. A “Lower theory of mind” group made up 57.7% of the children (30 of them). A “Higher theory of mind” group made up the remaining 42.3% (22 children). The gap between them was large and statistically clear. On the verbal reasoning task the effect size was substantial (F = 50.39, p < 0.001, partial eta-squared = 0.50), and the same held for the applied parent-rated measure (F = 44.11, p < 0.001).

The two groups differed in ways parents will recognise. The Lower group had more severe communication symptoms (F = 14.61, p < 0.001) and weaker everyday language use. On the Children’s Communication Checklist, which captures pragmatic skills such as taking turns, staying on topic and using context, the Higher group scored 23.77 on average against 15.20 for the Lower group. Adaptive functioning followed the same pattern. On the Vineland scales, the Higher group scored higher for daily living skills (82.77 versus 74.37) and for socialisation (81.00 versus 71.43).

Here is the finding that matters most. The Higher group could match typically developing children on the explicit, sit-down reasoning tasks. Yet on the applied measures, all the autistic children, including this stronger group, still lagged behind their typically developing peers. The authors put it plainly: even the children whose formal theory-of-mind performance equalled that of typical peers “does not seem to successfully deal with social interactions in daily life.”

What it means for you and your child

The practical lesson is that passing a test of a skill is not the same as using the skill in the moment. A child can correctly explain what a story character is thinking, calmly, with time to reason it through, and still miss the same cue in a fast, noisy, unscripted playground exchange. Real interaction happens in real time, with feelings running and no pause button. This study suggests that gap is real and measurable, not just an impression parents have.

For families, this reframes what progress looks like. If your child has learned to answer social-reasoning questions well, that is genuine and worth having. But it should not be mistaken for the finish line. The children in the study who reasoned well on paper still needed support with the spontaneous, everyday business of getting along with others. Explicit understanding and fluent social participation are related, but they are not the same thing, and one does not automatically produce the other.

It also argues for looking closely at your own child rather than at a single label. The study found two quite different profiles within one diagnosis. A child in the Lower group, with more pronounced communication difficulties, may need direct work on the building blocks of language and interaction. A child in the Higher group may need something different: practice applying what they already understand, in the flow of real situations, rather than more instruction on the underlying concept.

A few cautions keep this honest. This is one study of 89 children, and it is cross-sectional, meaning it captured a single moment rather than tracking children over time. It cannot tell us whether these profiles are stable, or whether stronger reasoning eventually feeds through into smoother interaction as children mature. The sample was mostly boys, so the findings may not describe autistic girls as well. And it relied heavily on parent questionnaires rather than direct observation across many settings. These are the authors’ own stated limits, not afterthoughts.

What the study establishes is modest but useful. Among autistic children with typical intelligence, theory-of-mind ability varies enough to form distinct groups, and even the strongest performers on formal tests still find real-world social life demanding. It does not prove what causes this gap, nor what closes it. For a parent, the takeaway is a calmer, more targeted expectation: notice which profile fits your child, value formal gains without over-reading them, and give particular attention to the harder, quieter work of using social understanding when it actually counts.