To assess for possible racial or ethnic bias in disability identification, federal regulations being implemented require U.S. school districts to monitor for significant disproportionality in special education. Districts do so by reporting unadjusted risk ratios. For example, school districts in Maryland reporting significant disproportionality based on race or ethnicity above 2.0 over two consecutive years will be required to take corrective action. A hypothetical Maryland school district where 25% of Black or Hispanic students are in special education relative to 10% of other student groups would report a risk ratio of 2.5 (= .25/.10). This would exceed the state’s threshold. If the district’s risk ratio exceeded the state’s threshold for two consecutive years, the district would be viewed as overidentifying students as having disabilities based on race or ethnicity and so be required to take corrective actions. Corrective actions include reallocating up to 15% of federal special education funding.
Because the risk ratios are unadjusted, they do not account for other explanatory factors that may help explain the observed racial and ethnic disparities in special education. That is, and although designed to monitor for biased identification practices, the new federal regulations do not require that the risk ratios contrast similarly situated students who differ in their race or ethnicity. Yet differential treatment of similarly situated students—particularly those who are similarly situated across the most materially relevant factors—is how the U.S. Department of Education assesses for biased practices including in disability identification . By failing to contrast similarly situated students, the regulations may be misdirecting efforts to ensure that students with disabilities are appropriately identified and provided services to which they may be entitled.
Why might this be? Students in the U.S. receive special education services if their disabilities adversely affect their educational performance. Academic achievement is often the most materially relevant factor for assessing whether students with disabilities should receive special education services. Prior work has described academic achievement as “the key variable in special education eligibility for most students.” Recent work finds that students who are academically proficient are very unlikely to receive special education services in U.S. schools.
Yet, because of historical and ongoing racial segregation, students of color are more likely to be exposed to the risk factors that contribute to disabilities including those resulting in academic difficulties. Examples include a greater likelihood of being exposed to lead , low birthweight , environmental pollutants , and living in low income and unsafe neighborhoods . Currently, among U.S. fourth graders, students who are Black or Hispanic are about twice as likely to display below basic levels of reading achievement as students who are White. Academic difficulties are often evident by school entry . By failing to account for disparities in the population of students experiencing academic difficulties, the new federal regulations may mistake reasonable efforts by U.S. school districts to provide special education services to those students whose underlying disabilities result in academic difficulties as instead use by school districts of biased and discriminatory disability identification practices.
How might U.S. school districts be more appropriately monitored for potential bias in their disability identification practices? We recently examined this by merging and then analyzing three national datasets. Specifically, we merged district-level data collected by (a) the U.S. Department of Education’s Office of Civil Rights on racial and ethnic disparities in special education with (b) the Stanford Education Data Archive on achievement gaps and (c) the Common Core of Data’s socio-demographic information. Our two analytical samples included 1,952 and 2,571 U.S. school districts. We estimated the size of a district’s Black- and Hispanic-White risk ratios while adjusting for the district’s Black- and Hispanic-White achievement gaps. We also controlled for the district’s racial and economic composition and enrollment size as well as state fixed effects.
Adjusting for the district’s achievement gaps indicated that students who are Black or Hispanic were less likely on average to be enrolled in special education than students who are White. Districts with larger achievement gaps reported larger risk ratios. Properly accounting for the overrepresentation of students in color in the population of students with academic difficulties—which itself occurs due to historical and ongoing societal inequities including during early childhood— resulted in consistent evidence that Black and Hispanic students are relatively less and not more likely to be identified as having disabilities while attending U.S. schools. This finding of under-identification is consistent with repeated studies analyzing student-level datasets . Our study provides further evidence that the overrepresentation of students of color in special education is largely explained by these students being more likely to experience academic difficulties due to being more likely to be exposed to the risk factors for disabilities as well as the other correlates including poverty.
We did observe that some districts reported larger risk ratios than might be expected after adjusting for achievement gaps and other potential explanatory factors. These districts might be potential candidates for civil rights monitoring by the U.S. Department of Education, especially if larger-than-expected risk ratios indicating over-representation were repeatedly observed across several years. Doing so would be consistent with the federal regulation’s emphasis on disproportionality that is “ systemic and otherwise indicative of persistent problems.”
Our analyses suggest a method that could be used by state and federal auditors evaluating for significant disproportionality in special education based on race or ethnicity. This is to adjust the district’s Black- and Hispanic-White risk ratios for the district’s Black- and Hispanic-White achievement gaps as well as other explanatory factors. Doing so would provide a more methodologically and substantively justifiable method for identifying U.S. school districts where significant disproportionality in special education based on race or ethnicity—and not other factors—may be occurring.