Do Children with Autism Spectrum Disorder Benefit from Structural Alignment When Constructing Categories?

Snape et al. (2018) asked if kids with autism learn categories better when they see many examples or just one.

They ran a lab task with children with ASD and neurotypical peers. The kids had to group new objects after seeing either a single item or several similar items.

Extra examples gave zero boost to the autistic group. Their category scores stayed flat no matter how many exemplars they saw.

The authors say this shows children with ASD do not use 'structural alignment'—the mental trick of lining up matching parts across examples.

The finding clashes head-on with Schnell et al. (2018). Schnell’s team taught tacts with serial multiple-exemplar training and saw clear gains for every child. The difference is the goal: Schnell taught object names; Simon taught abstract categories. Names stick, abstract patterns don’t.

MacDonald et al. (2015) also used multiple exemplars to teach observational learning to kids with ASD and it worked well. Again, the skill was concrete—copying actions—not forming a new category rule.

Ropar et al. (2007) helps explain why. They showed that autistic children default to sorting by color or size, not by hidden category rules. Put together, the papers say: give many examples when you want rote learning or labeling; stick to one clear exemplar when you need flexible categorization.

Stop wasting time showing five red apples, six red balls, and seven red cars just to teach 'red.' Pick one clear item, teach the label or rule, then test with novel objects. Save the exemplar flood for tact programs and imitation chains where previous studies prove it helps.