Examining the Role of Race, Ethnicity, and Gender on Social and Behavioral Ratings Within the Autism Diagnostic Observation Schedule.

Taylor et al. (2017) looked for hidden bias inside ADOS scores. They checked if race, ethnicity, or gender changes how items rate social and behavior skills.

They ran a statistical scan of ADOS records. The goal was to spot items that score kids differently even when skill levels are the same.

Three ADOS items showed small but real racial/ethnic bias. Gender items came out clean.

In plain words, the test gives minority children slightly lower scores on a few social points, even when their true skills match peers.

Magaña et al. (2013) saw a similar story with the ADI-R. Social and communication scores held steady across Latino and white adolescents, but restrictive behavior scores ran lower for Latino youth.

Tafolla et al. (2025) pushed further and found ADOS-2 severity scores stayed the same whether the test was given in English or Spanish to 94 bilingual clients. This seems to clash with J et al.'s bias finding, yet the two studies looked at different questions: language form versus racial/ethnic item wording.

Hedley et al. (2010) and Lugo-Marín et al. (2019) add more Spanish-language validations. Their positive results for the ADEC-SP and AQ-Short show tools can work well once translated and re-normed, backing the idea that cultural tweaks, not wholesale replacement, fix the bias.

You already use the ADOS to help decide eligibility. This paper tells you to pause if a child is from a racial or ethnic minority group. A small score dip on a few items might reflect test bias, not less skill. When the total score is close to the cutoff, gather extra parent input, school records, or a second observation before you rule in or out. The same three items may also shift slightly if you test in Spanish, yet Tafolla et al. (2025) show the overall severity score holds across languages, so keep using your bilingual protocol with confidence.