A Meta-analysis of the Wisconsin Card Sort Task in Autism.

Oriane and colleagues pooled every Wisconsin Card Sort Task study that compared autistic and non-autistic groups. They looked at all kinds of WCST scores: how many rules you break, how long you stick to old rules, and how many cards you sort right.

They kept computer and paper versions separate to see if the screen helps. The final pile held 31 studies with people of all ages and IQ levels.

Across every study, autistic groups scored about half a standard deviation worse on every WCST index. That is a medium-sized gap and it stayed the same whether the test was on paper or on a screen.

Lower verbal IQ predicted more non-perseverative errors. Lower performance IQ predicted more perseverative errors. In plain words, language scores forecast random mistakes, while spatial scores forecast rule-stickiness.

Gilbert (2003) saw that pretend play deficits shrink when you add structure. Oriane’s team now shows that executive deficits do NOT shrink when you switch from manual to computer WCST. Same domain—autism—but structure helps play, not card sorting.

Kraijer (2000) found that autistic people with intellectual disability score far lower on social and communication scales than matched non-autistic peers, yet equal on self-help and motor skills. Oriane’s IQ split echoes this: verbal and non-verbal IQ each predict different kinds of WCST errors, so matching groups on self-help or motor skills alone (as D suggests) may still leave executive gaps hidden.

Parks (1983) warned that early autism scales had shaky validity. Oriane’s 2016 meta keeps that worry alive: WCST gives stable group differences, yet those differences still link tightly to IQ, raising the same old question—are we measuring autism-specific executive problems, or just IQ?

When you screen an autistic client, expect WCST scores to land below age norms even if the task is on a tablet. Do not blame the format—blindly adding computer prompts will not fix it. Instead, check both verbal and performance IQ before you interpret errors. A low verbal score flags random mistakes; a low performance score flags rule perseveration. Use these patterns to pick targets: teach flexible shifting for perseveration, and boost self-monitoring for random errors. Finally, if you match groups in research, balance both IQ domains, not just adaptive or motor scores, or you may miss real executive gaps.