National Trends in Special Education and Academic Outcomes for English Learners With Disabilities

Context

Researchers have examined whether ELs are disproportionately identified for special education services. These studies are often guided by different data sources, including student-level records from districts (Umansky et al., 2017), nationally representative samples of students (Hibel & Jasper, 2012; Samson & Lesaux, 2009), or aggregate counts at the district or state level (Artiles et al., 2005; Sullivan, 2011). How ELs are identified in the data (e.g., time varying or time invariant) and the choice for comparison (e.g., White peers or non-ELs) also vary in studies (Artiles et al., 2005; Sullivan, 2011; Umansky et al., 2017). Measurement of time is another factor as studies may examine patterns by grade level (Artiles et al., 2005), merge cross-sectional data across years (Sullivan, 2011), or analyze longitudinal records of the same students (Umansky et al., 2017). The choice of data and methods has tradeoffs but, not surprisingly, the results differ: longitudinal records tend to show EL underrepresentation in special education, while aggregate data reveal overrepresentation.

National count data collected at the federal level for IDEA also have limitations but remain underutilized for examining the education of ELs with disabilities. Primarily, discussions and analyses of disproportionality in special education at the national level rarely include this population (U.S. Department of Education, 2022). An understanding of EL and non-EL representation in special education across multiple years at the national level would be helpful for monitoring the extent of disproportionality. Also relevant is how these trends may vary by disability type as previous research has focused mainly on specific learning disability (e.g., Hibel & Jasper, 2012) or overall totals (Artiles et al., 2005). A related issue is that the reporting of special education disparities rarely considers patterns in growth. For example, the proportional change in the number of ELs with disabilities compared with non-ELs with disabilities over time can be informative for how schools prepare for shifting demographics. This omission in special education research is surprising as EL research tends to emphasize the growing population. The degree to which change in the population of ELs with disabilities reflects patterns in the overall EL population is unclear but highlights the importance of examining school referral practices, especially if the former is increasing at a higher rate.

More recent research has focused on inequities in school placement for students with disabilities. The least restrictive environment mandate of IDEA requires schools to ensure that students with disabilities are taught in the same classroom as peers without disabilities to the extent possible. Yet, studies show that students of color with disabilities tend to have lower rates of inclusion in the general education classroom (Cooc, 2022; Grindal et al., 2019). General education placement for ELs with disabilities overall is less explored but some existing research at the state level indicates they are less likely to be placed in restrictive settings compared with White students (Sullivan, 2011). More recent research is needed on whether these trends still hold, especially since language supports in the general education setting may be affected by special education services (Cioè-Peña, 2021; Kangas, 2018).

Although inclusion is typically framed in terms of time in the general education classroom, the larger issue for students with disabilities is missed opportunities to learn. This is especially concerning for students with disabilities who also tend to experience higher rates of school suspension (Ryberg et al., 2021). Research on school discipline for ELs overall is scarce (Burke, 2015; Losen et al., 2015) but one theory is that challenges with school adjustment, including difficulties with learning English, may contribute to behavioral challenges (Patel et al., 2017). The one study that examined ELs with disabilities found in 2015-16 little evidence of disparities in suspension between ELs and non-ELs at the national level (Whitford et al., 2019). However, given the consequences of suspension on learning and its correlations without other adverse outcomes (Duxbury & Haynie, 2020), a focus on ELs with disabilities is warranted.

Beyond special education services and placement, research on the academic outcomes of ELs with disabilities is also limited. Annual IDEA data reports include exit data on high school graduation and drop rates for all students with disabilities but not ELs. Using data from the National Longitudinal Transition Study 2, Trainor et al. (2016) found similar rates of high school graduation between ELs and non-ELs with disabilities. Graduation rates are important, particularly for postsecondary education and employment, but may not fully capture academic achievement and learning, especially if diploma requirements change. Research shows ELs with disabilities are likely to struggle academically because of barriers to accessing dual services (Kangas, 2018) or placement into lower academic courses (Kangas & Cook, 2020). Unfortunately, although IDEA collects information on graduation, no study has pooled together information on how ELs with disabilities have performed on academic assessments and examined whether there has been progress. Greater national accountability for the learning of ELs with disabilities will require monitoring high school graduation and academic achievement.

Conceptual Framework

The present study is grounded in understanding the outcomes for ELs with disabilities over time. Yet, the lack of attention in this area within IDEA reporting is also informative given that the data exists. This oversight, according to intersectionality theory (Crenshaw, 1989, 1991), is not surprising as the multiplicative effects of different identities, particularly minority statuses, are often ignored. Indeed, national reports may disaggregate data for ELs relative to non-ELs, or students with disabilities relative to peers without disabilities, but not both simultaneously. Intersectionality, however, suggests the learning challenges of English learning and a disability are likely compounding. Thus, one hypothesis is much lower academic outcomes for ELs with disabilities relative to non-ELs with disabilities or peers without disabilities, especially when considering that EL and disability status are each correlated with other dimensions of marginalization (e.g., household income, race, and immigration) that compound as well (Cioè-Peña, 2021; Yu, 2013).

Intersectionality is also useful for understanding policies and practices that may further exacerbate the effects of marginalization. More specifically, policies aimed at supporting one marginalized group may inadvertently harm another group. Kangas (2018) documents this effect for ELs with disabilities by showing how educators tend to weigh special education and EL laws differently when providing services. In upholding special education law, educators may overlook language services for ELs with disabilities. The prioritizing of disability over language is a manifestation of intersectionality where multiple identities are treated differently in practice, albeit strongly influenced by school accountability pressures. Such beliefs and practices, however, would likely contribute to differences in educational environments for ELs and non-ELs with disabilities and widen gaps in academic achievement.

The present study uses intersectionality to guide an analysis of national trends in special education services and outcomes for ELs with disabilities currently missing in IDEA reports. Using multiple data sources across more than 15 years, I ask the following research questions:

Method

Results

Service Outcomes

I begin with a display of EL representation among students with disabilities across the country in Figure 1. In 2006, among students with disabilities in each state, ELs are most concentrated in the southwest and west and the least in the Appalachia region. By 2020, this pattern remains but ELs among students with disabilities have also increased in the midwest and northeast. Separate analyses confirmed that much of the growth in the total number of ELs with disabilities has occurred within the Midwest (250%) and Northeast (150%). ELs with disabilities in the West and Southwest, though high in total numbers, have remained proportionally the same.

OPEN IN VIEWER

Figure 1.

Geographic Trends in English Learners among Students with Disabilities, 2006 to 2020.

Figure 2 summarizes the disparities in EL representation within special education from 2006 to 2020 at the national level. The top graph presents some evidence of underrepresentation in special education for ELs relative to non-ELs before 2012 and overrepresentation thereafter. However, as states tend to set thresholds of risk ratios above 2.0 for overrepresentation (e.g., California Department of Education, 2020), the results in this study would not reach the cutoff for significant EL disproportionality within special education at the national level. In contrast, the bottom graph displays changes in the total number of ELs and non-ELs with disabilities since 2006. Despite limited evidence of disproportionality in the risk ratios, ELs with disabilities have increased nearly 50%, compared with about 7% for non-ELs with disabilities.

OPEN IN VIEWER

Figure 2.

Risk Ratios of Special Education (Top) and Growth Rates for English Learners and Non-English Learners With Disabilities (Bottom).

I further disaggregate the risk ratios by disability type in Figure 3. Three main trends emerge. First, risk ratios have steadily increased from 2013 to 2020 for all but one disability (speech language impairment). Second, more than half of the disability categories have risk ratios above 1.0 that indicate some degree of overrepresentation of ELs relative to non-ELs but not significant disproportionality. Third, risk ratios above 1.0 were also detected for low-incidence physical disabilities, such as hearing and visual. Less surprising is the higher risk ratios for learning disability and speech language impairment, two categories where the challenge of disentangling learning difficulties related to English language development and disability is most prominent. ELs are consistently underrepresented relative to non-ELs for other health impairment and emotional disturbance.

OPEN IN VIEWER

Figure 3.

Risk Ratios of Special Education for English Learners Relative to Non-English Learners.

Figure 4 displays the proportional change in the total number of ELs and non-ELs by disability type. Across all disabilities, the change since 2013 is higher for ELs than non-ELs. Although the number of students for autism spectrum disorder and developmental delay has increased overall, ELs in those categories have increased more than 150%. In addition, while ELs may be underrepresented in other health impairment and multiple disabilities according to the risk ratios, they have increased by more than 50% in each category. ELs in specific learning disability and speech language impairment also increased more than non-ELs, but less than other disabilities. The results confirm that although ELs are most prevalent in these two areas, their recent growth is high in low-incidence disabilities.

OPEN IN VIEWER

Figure 4.

Growth Rates Among English Learners and Non-English Learners With Disabilities.

I next focus on missed instructional time due to potential disparities in general educational placement and discipline in Figure 5. General education placement rates have been consistently lower for ELs with disabilities (top graph), even as overall rates have increased for all students. In 2020, about 66% of non-ELs with disabilities were in general education classrooms, compared with about 62% of ELs with disabilities. In contrast, any in-school suspension (middle graph) tends to be consistently lower for ELs than non-ELs with disabilities, but the gap is less than 2 percentage points. The same pattern is evident for any out-of-school suspension or expulsion (bottom graph). The overall decline in suspension in the last decade for all students with disabilities corroborates trends in studies of the overall student population (Ryberg et al., 2021). Although students with disabilities tend to have higher suspension rates, the rates for ELs with disabilities are similar to students without disabilities (Ryberg et al., 2021). Overall, gaps in general education placement and in discipline tend to be small.

OPEN IN VIEWER

Figure 5.

General Education Inclusion (Top) and School Discipline Rates for English Learners and Non-English Learners With Disabilities by in School (Middle) and Out of School Suspension (Bottom).

School Outcomes

The final figures focus on academic school outcomes. Figure 6 shows that the share of ELs and non-ELs with disabilities who obtain regular high school diplomas have increased since 2006 (top graph). Although ELs with disabilities tend to have lower rates of diplomas than non-ELs with disabilities, the gap began closing in 2014. In 2019, about 70% of ELs with disabilities exited with a regular high school diploma, compared with about 76% of non-ELs with disabilities. Dropout rates have been declining overall (bottom graph) and are only slightly higher for ELs with disabilities. The academic disparities between both groups are more evident when examining NAEP long-term trends in Figure 7. The results show consistently lower scores for ELs with disabilities than their counterpart, a gap that has widened over time in math and reading for 9-year-olds. In 2020, the gap between ELs and non-ELs with disabilities in math was about 0.65 standard deviations. As a comparison, the gap between ELs and non-ELs without disabilities is about 0.63 standard deviations. These trends are similar when examining math scores at age 13. Reading scores at age 13 tend to be lower than math but the gap between ELs and non-ELs with disabilities also show little evidence of closing. Last, Figure 8 displays NAEP main results. Recall that NAEP main updates test items to reflect current curricula and standards, while the NAEP long-term assesses students in the same content area each time to better measure progress. The results in Figure 8 continue to show ELs with disabilities scoring lower than non-ELs with disabilities. However, the gaps are closing somewhat, partly from improvement among ELs with disabilities and slight declines among non-ELs with disabilities.

OPEN IN VIEWER

Figure 6.

Graduation (Top) and Dropout (Bottom) Rates for English Learners and Non-English Learners With Disabilities.

OPEN IN VIEWER

Figure 7.

Comparison of NAEP Long-Term Trends.

OPEN IN VIEWER

Figure 8.

Comparison of NAEP Main Scores.

Discussion

This study pooled together annual data from multiple sources over a 15-year period to examine trends in special education services and outcomes for ELs with disabilities. I make several contributions to the literature. First, I analyzed disproportionality in special education representation and growth rates for all ELs with disabilities and by disability type. Second, I assessed disparities in school placement for ELs and non-ELs with disabilities in terms of suspension and general education classroom participation. Third, I investigated trends in high school graduation and standardized math and reading test scores for both student populations.

Studies of EL disproportionality in special education show that results may depend on the grade level and how EL is defined (Hibel & Jasper, 2012; Samson & Lesaux, 2009; Umansky et al., 2017). Although data limitations in the present study only allowed for aggregate trends across grade level for current ELs, the results show that the risk of special education for ELs has been increasing relative to non-ELs since 2012. The pattern is similar in research at the state level (Sullivan, 2011). The evidence of overrepresentation increasing, though limited depending on the cutoff used, is an area of concern for existing identification procedures. These patterns are magnified when disaggregated by disability type. The higher rates for ELs in specific learning disability, speech language impairment, intellectual disability, and developmental delay were consistent with prior research (Sullivan, 2011; Umansky et al., 2017), but the surprise was EL overrepresentation in the hearing and visual disability category. This new finding may be related to the challenges of learning English and potentially American Sign Language at the same time for ELs (Cannon & Guardino, 2012).

Analysis of growth rates provides a different picture of the future. Not only has the number of ELs with disabilities increased at a rate seven times that of non-ELs with disabilities since 2006 but this growth has occurred primarily in regions (Midwest and Northeast) without large EL populations in the past. Similarly, the greatest growth by disability type for ELs is not in specific learning disability or speech language impairment but developmental delay, autism spectrum disorder, and other health impairment. The growth in the number of ELs with disabilities in schools is perhaps not surprising given that the total number of ELs has also increased during that time by about 14% (NCES, 2021). However, this explanation is incomplete as the number of ELs with disabilities increased nearly 50%. These trends together would suggest that procedures within schools are more likely to identify ELs for special education than before. Although it is unclear whether the increase is due to unequal learning opportunities or greater learning challenges among ELs that lead to special education, the results underscore the need to monitor and assess current practices that identify ELs for special education. More attention is also needed for disabilities that have been increasing among ELs even though they may still comprise a smaller share of all cases, such as autism spectrum disorder and developmental delay.

The findings on school placement for ELs extend recent studies that show inequities for students of color with disabilities (Cooc, 2022; Grindal et al., 2019). On one hand, results show that instructional time in the general education classroom is lower for ELs with disabilities. That is, the intersecting effects of EL and disability classification on placement appear to compound. If removal from the general education classroom is related to receiving more intensive outside supports, potentially for English language learning and disability, then the results may be less concerning. Other research suggests that dual social identities and classifications, despite good intentions, may lead schools to prioritize special education services over language needs (Kangas, 2018). Without dual services, EL with disabilities may be unable to develop English language proficiency, which would further compound learning challenges and may lead to more intensive services outside of the general education classroom. The lower rates of general education placement for ELs with disabilities, however, even for additional supports, may limit the social benefits of interacting and learning with peers without disabilities (Cooc, 2019; Ruijs & Peetsma, 2009; Stiefel et al., 2018).

On the other hand, school removal via suspension or expulsion is lower for ELs with disabilities than their non-EL counterparts. As the discipline rates for ELs with disabilities are also comparable to students without disabilities, the results suggest that the intersecting effects of language and disability may be less prominent in this outcome. This is consistent with the much lower rates for emotional disturbance among ELs with disabilities in this study, an area where suspension is typically the highest. Another possible explanation is that procedures for suspension may operate differently when students are dually identified for EL and special education. Nonetheless, given that instructional time lost due to suspension and discipline is more frequent for students with disabilities overall, this finding is encouraging for ELs.

The final analyses provide a mixed picture of academic achievement for ELs with disabilities. Graduating high school with a regular diploma has improved for all students with disabilities but a small gap persists between ELs and non-ELs. Dropout rates, though declining, are also slightly higher for ELs with disabilities. The pattern in graduation rates for ELs and non-ELs is similar to findings in previous studies using nationally representative samples of students with disabilities (Trainor et al., 2016) but the smaller gap in graduation is surprising given the multiplicative effects of language learning and disability. The increasing high school graduation rates for ELs and non-ELs with disabilities, though encouraging, should be viewed within the context of higher graduation rates overall in the 21st century (Murnane, 2013).

In contrast to high school completion, inequality in test scores between ELs and non-ELs with disabilities are clearer. Gaps in NAEP long-term trends have either widened or stagnated over time, a pattern shared by students overall since 2004. NAEP main results, which in contrast account for changes in curricular content in schools, also reveal large disparities but ELs with disabilities show more improvement over time. These results suggest that curricular content may be more aligned with the learning of ELs with disabilities or schools are more adept at teaching these standards for this student population. Although the large test score gaps clearly highlight the role of EL status among students with disabilities, EL status is correlated with other disadvantages, including lower parent income and education level and high-poverty schools, that would affect academic achievement (Trainor et al., 2016). The lower test score trends may also reflect the tracking of ELs with disabilities into lower academic classes (Kangas & Cook, 2020).