Intersectional Disparities in School-Based Autism Identification Are Occurring in U.S. Elementary Schools

We investigated whether intersections of race or ethnicity, gender, family income, and language use are associated with school-based autism identification (SBAI) in U.S. elementary schools. We estimated logistic and multilevel logistic regression models with interaction terms, adjusting for reading achievement and school contextual factors. We used school record of autism as a student's primary disability for which special education services were being provided to indicate SBAI. We analyzed a very large, aggregated sample of U.S. fourth graders (typically 9-10 years old; N = 1,419,440) participating in the National Assessment of Educational Progress across 2005 to 2022. The sample included stratified subsamples of Black (n = 246,040), Hispanic (n = 273,620), Asian/Pacific Islander (n = 68,800), and White (n = 759,800) students. Within the racial and ethnic groups, girls (adjusted odds ratio [aOR] range = 0.10-0.27) and those from low-income families (aOR = 0.21-0.60) were less likely to be identified. Among Hispanic and White students, MLs were less likely to be identified (aOR 0.57 and aOR = 0.40, respectively). Attending schools with greater percentages of low-income students was associated with lower odds of SBAI (aOR range = 0.73-0.81). Among Black, Hispanic, and White students, those attending schools with greater percentages of special education students were more likely to be identified (aOR = 1.16-1.23). Overall, Black and Hispanic girls, low-income Black and other race or ethnicity students, and White and Hispanic multilingual learners (MLs) had especially low odds of SBAI. Intersectional disparities are occurring in SBAI in U.S. elementary schools.Lay AbstractWe examined how multiple social positions across race and ethnicity, gender, family income, and language use are associated with whether students were identified with autism in U.S. elementary schools. We analyzed a very large, aggregated sample of U.S. fourth graders (typically 9-10 years old; N = 1,419,440) participating in the National Assessment of Educational Progress across 2005 to 2022. The sample included stratified subsamples of Black (n = 246,040), Hispanic (n = 273,620), Asian/Pacific Islander (n = 68,800), and White (n = 759,800) students. We used school record of autism as a student's primary disability for which special education services were being provided to indicate school-based autism identification (SBAI). The analyses adjusted for student academic achievement, school contextual factors, and other alternative explanatory factors. We repeatedly observed interactions between race or ethnicity and family income. Girls and those from low-income families were consistently less likely to be identified with autism including within samples of Black, Hispanic, Asian/Pacific Islander (PI), and White students. Black and Hispanic girls, low-income Black and other race or ethnicity students, and White and Hispanic multilingual learners (MLs) had especially low odds of being identified. Students across the racial and ethnic groups were less likely to be identified with autism when attending schools attended by higher percentages of low-income students. Students also were less likely to be identified in schools when their race or ethnicity was the same as those of other attending students. Our findings repeatedly indicate that intersectional disparities in SBAI are occurring in U.S. elementary schools. Referral practices, evaluation and eligibility procedures, and resource allocation and outreach efforts may need to address intersectional disparities to ensure that multiply marginalized student populations have equal access to autism services and supports in U.S. elementary schools.