The Under-Identification of Autism in Females: A Review and Analysis of Sex-Based Scoring Differences Observed in Autism Diagnostic Observation Schedule (ADOS) Module 3.

Yu et al. (2026) built a computer model. The model tests if ADOS-3 items treat boys and girls the same way.

They tried a new way to pick anchor items. This rank-based method is meant to catch sex bias more cleanly.

The new anchor method gave more accurate DIF results. DIF means an item is harder for one sex even when ability is the same.

No real kids were tested; the study is a pure simulation.

Bottini et al. (2025) asked clinicians to score the same vignette. When the child was labeled female, clinicians saw more severe autism. This shows bias lives in people, not just test items.

Kopp et al. (2011) added girl-typical items like "avoids demands" to the ASSQ. Their tool caught girls earlier, proving the test itself can be fixed.

Backer van Ommeren et al. (2017) found girls with autism show better back-and-forth talk than boys. If ADOS still uses the same social cut-offs for both sexes, it will keep missing girls.

Kamp-Becker et al. (2013) updated the ADOS-3 algorithm for high-functioning youth. Yvonne’s method could now check if those new rules are also sex-fair.

Before you trust any ADOS score, run the rank-based DIF check in the spreadsheet the authors share. If an item shows sex bias, weight or skip it. Pair this with clinician self-audit tools like those Summer et al. suggest. You will cut the chance that quiet, socially masked girls walk away undiagnosed.