Special Education Representation Across the School Years by Language Status in Tennessee

• In Event: Equitably Supporting the Academic Achievement of Multilingual Learners in the United States

Abstract

Purpose
Mixed and contentious findings are reported on special education (SPED) representation of learners from non-English language backgrounds (NELB) across grades. We examine overall and disability-specific SPED trends in TN across a 10-year span, accounting for possible risk factors (e.g., economic disadvantage) in native English speakers (NES), English-proficient Bilinguals (EPB), and Current English learners (Current EL).

Theoretical Framework
SPED disproportionality constitutes part of the equity requirements (disproportionate representation, significant discrepancy, and significant disproportionality) under the 2016 Equity in Individuals with Disabilities Education Act (IDEA) Regulation, which aim to promote educational equity by identifying and preventing educational disparities. However, equity expectations in IDEA concern disproportionality between racial and ethnic groups, not linguistic diversity.

Methods
Following the recommendation from the U.S. Department of Education (2017), we calculated cross-sectional risk ratios for students who are NES, EPB (i.e., former ELs or never ELs), and Current ELs across ten years (2009-2019), addressing two questions:

RQ1. What are the cross-sectional unadjusted and adjusted risk and odds ratios by language status (NES, EPB, Current EL) for overall SPED placement and by disability type (i.e., SLD) over a 10-year span across grades 3-8 in TN?
RQ2: To what extent, if any, do income status (i.e., free-or-reduced price lunch eligibility), racial/ethnic minority status (i.e., Hispanic vs. non-Hispanic), and grade level (i.e., elementary vs. middle school) interact with language status to predict SPED placement accounting for covariates?

Data Sources
We utilized TN state-level data from 2009-2019 for students in grades 3-8, categorized by language (i.e., NES, EPB, Current EL) and income status (i.e., eligibility for free or reduced-price lunch). For calculating unadjusted and adjusted cross-sectional risk and odds ratios (i.e., likelihood of a particular group receiving a certain identification), our sample included 676,532 students, with a mean of 291,648 students (SD = 4,250) per academic year.

Results
Unadjusted risk ratios by language status were relatively stable across the focal 10-year span and did not reveal evidence of significant disproportionality in terms of over-identification by language status per the state threshold. However, EPB and Current EL had lower SPED risk ratios relative to NES. Further, language and income status interacted with and contributed to SPED risk ratios. Current ELs from low-income homes had a higher probability of being in SPED while EPB students were less likely to be in SPED, independent of income status. Additionally, adjusting for covariates lowered the likelihood of SPED identification for Current ELs. Current ELs who were Hispanic were more likely than their non-Hispanic NES peers to be identified for language-related disabilities (i.e., SLD and SLI) and, in recent years (2017-2019), Current ELs from lower SES were more likely than their higher-SES NES peers to be identified for SLD and SLI. See Figures 1a-1c.

Significance
SPED disproportionality is an issue of educational equity for all learners. Findings reveal that English proficiency and income are associated with SPED identification. We echo the national call to better understand NELB students’ representation in SPED, particularly potential under-identification, as there is little policy or research guidance in this critical area.

Authors

• Gigi Luk, McGill University

• Jeannette Mancilla-Martinez, Vanderbilt Peabody College

• Min Hyun Oh, Vanderbilt University

• Adam Rollins, Tennessee Department of Education