Shifting Landscapes of Professional Practices: English Learner Special Education Placement in English-Only States ch. 7
Introduction
Proposition 227 in California and Proposition 203 in Arizona led to a decrease in primary-language instruction.
The convergence of reforms like No Child Left Behind Act (NCLB) in 2001, Reading First in 2002, and Individuals with Disabilities Education Improvement Act (IDEA) reauthorization in 2004 may have negative consequences for low-income, ethnic, and English learner (EL) students.
The study examines the impact of reduced language support systems on ELs in English-only states.
The aim is to see if the focus shifts from language accommodation to concerns about ability level and if EL placement in special education increases.
The questions address changes in student identities and the outcomes of policy changes, considering the debate around special education placement of minority students.
The study aims to determine special education placement patterns after restrictive language policies, offering policy recommendations.
Emergent Research on the Effects of Restrictive Language Policies
Researchers have studied the impact of restrictive language policies in California, Arizona, and Massachusetts focusing on teacher practices, achievement gaps, school climate, and emotional impact on stakeholders.
A qualitative study by Mueller, Singer, and Grace (2004) found that special education teachers misinterpreted Proposition 227, lacked administrative support, and excluded parents from language of instruction decisions.
A study in California suggested some EL subgroups had a greater risk of disability identification, especially in high school (Artiles et al., 2005).
ELs in bilingual education had a lower risk of special education placement compared to those in English immersion programs.
Socio-economic status (SES) mediated EL placement risk, with low-SES ELs more likely to be identified with learning disabilities (LD) and high-SES ELs with speech and language impairment (SU).
There is limited research on academic outcomes for ELs with disabilities after restrictive language policies were implemented.
Investigation of English Learner Disability Identification in California and Arizona
The study addresses two questions:
EL special education placement trends before and after language support programs were drastically reduced.
The educational experiences of ELs in these states.
The study examines indicators such as placement segregation, the impact of school poverty level on disability identification risk, and opportunity to learn (e.g., access to language supports) for ELs.
EL is defined in Arizona Statute 15-751 and California Education Code §305.306 as a child who does not speak English or whose native language is not English and who is not currently able to perform ordinary classroom work in English, also known as a Limited English Proficiency or LEP child
Setting the Context: Evidence Sources and Procedures
Primary data was drawn from the Office of Civil Rights (OCR) Elementary and Secondary School Surveys for the 1998, 2000, 2002, 2004, and 2006 academic years.
Data sources included the Elementary/Secondary School Universe Data (U.S. Department of Education, 2008) for 1995-2006 and the Arizona State Department of Education (ADE) for the 1997-2006 academic years.
The analysis focused on placement patterns for English learners and English proficient students across and within the mild or "subjective" disability categories- learning disabled, mental retardation, emotional/behavioral disorders, and (whenever feasible) speech or language impairment.
These high-incidence categories comprise about 85% of students identified for special education in the nation (U.S. Department of Education, 2006).
Placement risks were calculated for each target group (i.e., ELs, English proficient students) in overall special education enrollment and in each target disability category.
A comparative measure called the relative risk ratio was used, which estimates the chance a target group of students (e.g., English learners) has to exhibit a particular outcome (such as placement in special education) compared to another group (e.g., English proficient students).
A value of 1 indicates that the target group (i.e., ELs) is equally likely to be identified as the comparison group in the category or placement of interest.
Following the research literature (Waitoller, Artiles, & Cheney, in press), 1.50 and 0.50 were used as the cutoff scores to determine over- and underrepresentation, respectively.
EL Special Education Placement Trends
Arizona
The number of districts reporting EL data increased from 1999-2006.
ELs were consistently less likely to be identified for special education overall.
From 2004 to 2005, EL overall enrollment decreased, EL special education enrollment increased, while English proficient student enrollment in special education remained unchanged.
English learners' risk for special education placement increased between 1999 and 2006, but risk indicators didn't reflect overrepresentation.
EL risks for learning disability and mild mental retardation identification were slightly lower before 2002, but risks increased substantially after 2003.
ELs were 63% more likely than their counterparts to be identified in the LD category after 2005.
English learners were not overrepresented in the mental retardation category between 1999 and 2006, though their placement risk approached overrepresentation levels in 2005 and 2006.
ELs had lower identification risks in speech and language impairments and emotional/behavioral disorders, with a marked underrepresentation trend in the latter category.
Overall special education placement risks didn't reach overrepresentation levels between 1999 and 2006 (mean risk indices ranged between 0.95 and 1.43), but between 16% and 30% of all districts in Arizona had overrepresentation levels during this time period.
ELs' risk for LD identification fluctuated, with data reflecting overrepresentation after 2002.
In 1999, about one-fourth of all Arizona districts showed EL overrepresentation in the LD category. By 2006, this figure had doubled to about 50%.
EL risks for speech and language impairments (SLI) and mental retardation (MR) identification tended to be about the same or lower than those of their English proficient peers.
Between 17% and 26% of the districts had EL overrepresentation in the MR category, and between 9% and 24% of districts exhibited EL overrepresentation in SLI.
English learners' risk for an emotional/behavioral disorder (E/BD) diagnosis was the lowest, reflecting substantial underrepresentation.
Between 3% and 9% of the state's districts reached EL overrepresentation levels in the E/BD category.
California
Overall school enrollment increased from 3.8 million to 4.7 million students between 1998 and 2006.
The proportion of students identified in special education remained relatively stable at around 10% of the total enrollment.
Enrollment of ELs decreased slightly in 2000 from 26% to 24%; nevertheless, this enrollment was at about 27% between 2002 and 2006.
English proficient student special education enrollment decreased from 11 % in 1998 to 9% in 2006, EL special education enrollment almost doubled from 6% in 1998 to 11 % in 2006.
ELs had a lower special education placement risk than English proficient learners between 1998 and 2002. This pattern changed in 2004 when both groups of students had the same probability of being placed in these programs. The same pattern was observed in 2006.
English learners had either an equal or lower chance of being diagnosed with the MR label in 1998, 2000, and 2006. In 2002 and 2004, ELs were 59% and 30% more likely, respectively, to be placed in this category.
ELs were less likely to be diagnosed as having LD in 1998 and 2000. After 2002, however, English learners were more likely to be identified as LD, though the risk index never reached overrepresentation levels.
ELs' risk level in 2006 approached a significant level of overrepresentation (34% more likely than their English proficient peers).
The risk indicators for E/BD placement reflected an upward trend, these data were at a significant underrepresentation level across all years.
Poverty and Placement Risk
In Arizona, ELs were underrepresented in all high-incidence disability categories in high-poverty schools in 2004.
ELs in low-poverty schools had a higher placement risk in overall special education (27% more likely, relative risk ratio of 1.27) and in the LD category (24% more likely).
EL placement risk increased as school poverty level decreased across the three disability categories.
Overrepresentation was observed only in the MR category in low-poverty schools (51 % more likely than English proficient students), a considerable proportion of low-poverty districts in Arizona had overrepresentation risk ratios (i.e., 20% to 24% of districts), most notably in the LD and MR categories.
In California, EL overrepresentation was observed at the state level only in high-poverty schools in the MR category (52% more likely).
The EL overall special education placement risk at the state level was higher in low-poverty schools. English learners had a greater risk for identification in each high-incidence disability category in high-poverty schools.
20% to 26% of local education agencies had EL overrepresentation in the LD category.
EL placement trends must be conducted at multiple levels to obtain a comprehensive understanding of the differences in placement risks.
Special education placement risks at the state level often did not reflect the presence of disproportionate EL representation in special education at the district level.
School poverty level had complex associations (both positive and negative) with EL placement risk.
Both underrepresentation and overrepresentation can constitute serious inequities. In the case of the former, students may not be receiving critically important services; in the case of the latter, they may be unfairly labeled and inappropriately assigned to classes and programs that are more harmful than helpful to their academic growth.
Educational Experiences of ELs With High-Incidence Disabilities
The second goal of this study was to examine the educational experiences of ELs in terms of opportunity to learn.
Most low-income, ethnic minority, and linguistic minority students attend schools with students who have similar demographic profiles.
These schools tend to be underfunded, staffed by less qualified teachers, and characterized by less demanding curricula (Anyon, 2005).
Indicators of opportunity to learn for ELs were examined, including school suspensions, expulsions, placement in programs for gifted students, and segregation level of special education placements.
Arizona reported a stable proportion of ELs who qualified for but did not receive language supports between 2000 and 2004 (about 10% ), whereas California's percentage increased from 10% to 14%.
ELs had a lower risk of being suspended from school than their English proficient peers.
Arizona showed an unstable trend from one year to the next on school expulsions. ELs were almost twice as likely as their non-EL counterparts to be expelled from school in Arizona in 2000 (relative risk ratio of 1.98) but had a lower expulsion risk in 2002 and 2006. The California data show lower expulsion risks for English learners, though there was a clear upward trend.
ELs were consistently less likely to be placed in gifted and talented programs over time in both states.
The LRE offers a continuum of options that range from most integrated (e.g., general education classroom) to least integrated (e.g., residential facility).
In Arizona, ELs' chances to receive services in general education classrooms oscillated between having an equal chance to being underrepresented in general education classrooms between 1999 and 2006.
After 2002, ELs had a greater risk than their English proficient peers of receiving special education services in resource programs that removed them from the general education classroom for a significant portion of the school day.
ELs were consistently underrepresented in the most segregated placement options, which include separate school and residential facilities or hospitals.
EL expulsion data varied across states. Arizona had substantial shifts that included lower and greater risk patterns. Access to gifted-and-talented programs remained elusive. Finally, ELs tended to be overrepresented in programs that kept them outside of the general education classroom between 20% and 60% of the schoolday.
Implications of the Findings
The study investigated whether ELs were increasingly placed in special education at a time when general education was given greater responsibility to educate these learners.
Challenging the Invisibility of the Problem: The Need for Multilevel Analysis
Disproportionality patterns do not always show up at aggregate levels. The relative risk ratios in both states do not reflect disproportionate representation levels.
Increasing numbers of ELs have been placed in special education since restrictive language policies were passed in California and Arizona. The placement analysis by disability category suggests that ELs tend to be overrepresented in LD in California and in MR in Arizona.
Although disproportionality levels were not problematic at the state level, a sizable proportion of school districts in Arizona in which ELs were overrepresented in the high-incidence categories.
The case of LDs was the most acute in Arizona, since over half of the districts in the state had EL overrepresentation in 2005 and 2006. California had a similar pattern in the LD category, though the proportion of districts was lower.
ELs in both states had a greater risk ratio in high-poverty schools when the data were examined across all disabilities. The California data in these categories showed the opposite pattern-that is, lower risk ratios in low-poverty schools.
It is possible California's history of litigation regarding testing bias with minority students, particularly in the MR category mediated placement trends in California. Arizona, however, does not share this legacy.
The EL special education enrollment doubled in California, as did the EL enrollment in LD in Arizona. Are the general education systems in these states increasingly relying on special education as a way of coping with the unrealistic requirements of these restrictive language policies or the lack of preparation of teachers to meet these students' needs?
Beyond Disability Placement Patterns: Tracking Opportunity to Learn
Disproportionate representation has been studied historically through the analysis of placement patterns in special education programs or disability categories. Antecedents to these practices, such as opportunity-to-learn gaps, have been neglected.
The trend in Arizona to remove ELs from their general education classrooms for special education services at increasingly greater rates than their English proficient peers is cause for concern.
Few special education teachers have received adequate preparation in working with ELs and are ill equipped to provide the linguistic support they need (Baca & Cervantes, 2004).
Zehler and her colleagues (2003) found that most ELs in special education tend to be educated in English immersion models by teachers with little training in serving EL students.
A substantial proportion of ELs lost access to language supports in both states, and this trend is increasing in California. Restricted access to programs that can significantly advance a student's career (e.g., gifted-and-talented programs) is also a reason for concern.
Policy Recommendations
First, monitor the extent to which particularly vulnerable groups are affected by restrictive language policies.
Second, improve information infrastructures to gauge policy impact. For example, we did not find databases that included specific information about ELs' language proficiency levels and disability status, and existing data sets do not provide sufficient information about the local practices that affect the opportunities to learn provided to ELs or that lead to their placement in special education.
Greater articulation is needed across institutions that specialize in language minority students, students with special needs, and educational equity in order to develop a comprehensive system of data collection that will allow researchers and policymakers to answer these urgent questions.
Finally, we have used relative risk ratios to analyze disproportionality of special education placement because we argue that simply looking at placement data can obscure inequities among groups.
There is a need to refine guidelines to measure and track the problem. There is no agreement in the field about what level of disproportionality constitutes a problem.