Co-Principal Investigator: Julian Betts (UC San Diego)

Purpose: The purpose of this study is to evaluate the impact of four different reclassification policies in two different districts on a variety of academic outcomes. English Learners (ELs) are an academically vulnerable population to whom schools provide language support until they reach a certain level of mastery in spoken and written English. Once ELs reach their state or district's level of mastery, they are then reclassified, and no longer receive English language support services. A key policy question is whether some districts are reclassifying EL students too soon or too late. In this study, researchers will estimate and compare the effects of transitioning out of EL services due to reclassification, versus remaining an EL for one additional year. Their findings will contribute to a growing body of research evidence about the effects of reclassification, generally, and the contextual factors (at the district, school, and student levels) that moderate these effects across different settings.

Project Activities: The research team will work with large-scale retrospective datasets from two large California districts to evaluate the impact of different reclassification policies the districts have implemented since the 2001–02 school year. They will use regression discontinuity to compare the academic outcomes of students immediately and above and below the cut-score for reclassification, to see whether one condition (reclassified versus remain an EL) leads to better academic outcomes in the near or long-term. Due to the large number of policies and outcomes, the researchers will ultimately complete 40 different analyses with each district's data.

Products: Researchers will produce evidence of the impact of reclassification under a variety of different policy regimes, and moderated by a variety of different district-, school-, and student-level variables. The team will also produce peer-reviewed publications, white papers, and conference presentations to share with policymakers and other stakeholders.

Structured Abstract

Setting: The data for this study comes from the two largest districts in California. These districts account for 15 percent of the ELs in California and 5 percent of the EL population nationwide.

Sample: The student samples for this study consist of all elementary and secondary ELs in the two districts who were either reclassified or earned scores within a certain threshold below the reclassification cutoff between 2003 and 2012. Exact sample sizes vary depending on the district, grade-level, and reclassification policy, but they range from approximately 1000 to 150,000 students, depending on the specific analysis.

Intervention: In this regression discontinuity design, the intervention is defined as the additional year of EL services that students receive who are not reclassified. The exact nature of these services will vary depending on factors such as the student's district or grade-level. Due to the retrospective nature of this study, the researchers will not be able to observe directly what the intervention looks like for individual students. They will, however, attempt to measure fidelity of implementation to the intervention by testing 1) whether students' course assignments are appropriate for their English language development level, 2) whether students' course assignments align with each district's Master Plan for English Language Development, and 3) whether the ELD levels of classroom peers change as a result of reclassification.

Research Design and Methods: This team's research will address the following questions: 1) Do EL students who are reclassified subsequently perform differently academically than those who are not? 2) What moderating factors contribute to heterogeneity in the effects of being reclassified? The research will test for outcome differences along seven dimensions: stringency of reclassification requirements, home language, demographics of the student body, including concentrations of students by language and diversity of languages spoken at the school, neighborhood characteristics, the degree of mismatch between the EL student's achievement level and average achievement at the school, average teacher qualifications at the school, grade level at reclassification, and fidelity to English language development course assignment prior to and post reclassification.

Control Condition: The control condition for this study is removal of English language support due to reclassification.

Key Measures: The researchers will examine several outcome measures including annual standardized test scores, on-time grade progression, passing California's high school exit exam, enrollment in high school courses required for California's college system, persistence, and postsecondary enrollment and degrees obtained at two and four-year colleges. Scores on California's annual standardized assessment of English language proficiency will also be used to identify eligible students for the study.

Data Analytic Strategy: The researchers will use a fuzzy regression discontinuity design to estimate the local average treatment effect of reclassification on a variety of outcome measures of interest. The research team will build separate models for each outcome, district, and grade-level. The researchers will also explore whether the effects of reclassification dissipate, intensify, or stay constant over time.