Examining predictors of disproportionate representation and differential access in special education using the National Educational Longitudinal Study of 1988: Do race, ethnicity or English language proficiency matter

This dissertation utilizes the National Educational Longitudinal Study of 1988 (NELS) to consider whether race, ethnicity or English language proficiency influence disability classification or service receipt after controlling for student, parent, and school factors that proxy the need for special education services. The NELS provides a nationally representative dataset of 8th graders and their experiences in secondary school. However some student groups are disproportionately excluded from the NELS, including students with mental retardation and students served in restrictive special education settings.;This dissertation finds no evidence of disproportionate representation by race and ethnicity among the majority of the disability classification categories considered. When controlling for determinants of student need, students of color generally appear less likely than their white peers to be classified as having a disability. Moreover, this dissertation provides no evidence of differential access by ethnicity to special education services in the secondary grades among students classified as having a disability.;English language learners, however, appear more likely to have both mild cognitive and physical disabilities and to receive special education services once they are identified as having a disability. Whether these patterns reflect differences in student need for services or bias remains unclear. The greater level of access to special education services prior to 8th grade found among whites with disabilities in the NELS, however, suggests that students of color with disabilities may flounder in general education for a longer duration before receiving early intervention services or being referred to special education.;Future research on differences in classification and service delivery by diverse student groups will benefit from more recent and nationally representative student level data that are aligned to IDEA---definitions of disability and offer more detailed information regarding the receipt of specific special education services. This dissertation recommends that federal policymakers require a multivariate approach to determining whether disproportionate representation and differential access exist in ways suggestive of bias and include systemic data collections on student disability status by English language proficiency.