In one sentence
Among 40 autistic and 39 non-autistic children aged 9–15 in Japan, parent questionnaires clearly marked the autism group, while a clinician observation (ADOS-2) showed a sex difference that parents did not — and the two kinds of score were only weakly linked.
What the researchers did
The team at Nara Medical University Hospital compared everyday parent reports with a structured clinic observation. Parents completed the Social Responsiveness Scale, Second Edition (SRS-2) and the Japanese Autism-Spectrum Quotient (AQ-J). Clinicians scored the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2), modules 3–4, which are for verbally fluent children. The autism group had a clinical DSM-5 diagnosis plus an ADOS-2 autism classification. Children with a Das–Naglieri Cognitive Assessment System total below 70 were excluded. The autism sample was 20 boys and 20 girls; the comparison group was 23 boys and 16 girls, recruited from the community and not formally matched beyond age and sex.
What they found
- Parent SRS-2 and AQ-J totals separated the autism group from the comparison group strongly. Sex effects on those parent totals were mostly absent.
- On the ADOS-2 total, autistic boys scored higher than autistic girls (mean 12.50 vs 10.75, p = 0.034). Most ADOS-2 domain scores did not differ by sex.
- Parent subscales were only weakly related to clinician-rated ADOS-2 severity. Sex still predicted the ADOS-2 total after parent scores were in the model.
- Among girls, the SRS-2 versus ADOS-2 gap shrank with age — older girls showed closer parent–clinician agreement. That age pattern was not the same for boys.
What this means for families and therapists
A mismatch between “what we see at home” and “what happened in the assessment room” is common in this study, not proof that a parent is exaggerating or a clinician is missing everything. Parents see months of context; ADOS-2 is a timed, structured sample. For girls, the authors themselves warn that ADOS-2 was largely developed on male samples and can miss some female presentations — so a lower clinic severity score is not a reason to dismiss a parent’s report. Therapists can name both sources in the plan: home load and clinic snapshot. Families can ask how the session compared with expected home behaviour, and what supports follow from the difference, not only from one number.
Limitations and what we don't know yet
This is one Japanese clinic sample of 79 children, all verbally fluent and without low DN-CAS scores, so it does not represent minimally speaking children or intellectual disability. The comparison group was convenience-sampled. Cross-sectional data cannot show how gaps change year to year. The authors treat informant discrepancy as meaningful context, not as measurement error — that is a framing, not a proof that every gap is “real extra information.”
This is a plain-language summary of Discrepancies between parent-reported and clinician-rated autism severity: roles of sex and age by Ishida R., Yamamuro K., Kashida N. et al., Frontiers in Psychiatry (2026). Source license: CC-BY-4.0. It is not medical advice — talk to a qualified clinician before changing therapy.