Brief Report: Neuropsychological Testing and Informant-Ratings of Children with Autism Spectrum Disorder, Attention-Deficit/Hyperactivity Disorder, or Comorbid Diagnosis.

Ng et al. (2019) gave kids with autism, ADHD, or both a short battery of neuropsych tests. They also asked parents and teachers to fill out rating scales.

The team wanted to see if test scores and informant ratings told the same story about social skills, attention, and executive function.

The lab tests only picked up social differences between groups. They showed no clear gaps in attention or executive function.

Parent ratings flipped the picture: they flagged attention and executive problems, but were quieter on social issues.

Bottom line—tests and ratings disagreed more than they agreed.

Harkness et al. (2025) later asked the same question with a huge national sample. Parent-report symptom scales again beat direct tests at telling autism or ADHD from typical development. Their newer data supersede the small Rowena sample and confirm the main warning.

Hanson et al. (2013) saw a similar rift earlier. When both parent and teacher had to agree, the rate of ADHD in kids with autism dropped from 16% to 2%. That drop previews the Rowena finding—single sources can over- or under-call symptoms.

Reus et al. (2013) add a twist. They showed that adding ADHD to autism pumps up parent scores on autism severity interviews, even though direct behavior codes stay flat. This earlier study helps explain why Rowena’s parent ratings looked more impaired than the lab scores.

You can’t trust one score. A clean neuropsych test does not cancel a messy classroom report, and vice versa. Collect both, then look for patterns. If test and rating clash, schedule an observation or bring in a second informant before you write goals.