Bilingual Education and the Short and Longer-Term Outcomes of English Language Learners

We study the longer-term effects of early exposure to bilingual education programs for English language learning (ELL) students in Texas. In Texas and around the country, students whose native language is not English constitute a growing share of the student body. Much of the policy debate related to the education of ELL students revolves around the desirability of bilingual education as opposed to alternatives including English immersion. However, testing both the short and long-term efficacy of bilingual education programs is challenging, primarily because such programs are not randomly assigned. Texas law requires districts with 20 or more ELL students in a grade and language to offer bilingual education. Consequently, the probability that a district offers bilingual education is higher for districts at or above 20 ELL students than for districts below. The appeal is that districts on opposite sides of the 20-student threshold are likely to be very similar in the other factors that affect outcomes. A recent study by Chin, Daysal, and Imberman (2013) used school-level data to study how this law affected elementary test scores, finding no effect on Spanish speakers and positive effects on non-ELL students. Using longitudinal administrative data from Texas, our study builds on their work by focusing on early exposure to bilingual education and its relation to intermediate and longer-term student and teacher outcomes, including reclassification to non-ELL, grade retention, secondary-schooling persistence, behavioral outcomes, the probability of enrolling in college, college persistence, and labor market performance (i.e. early wages). Preliminary estimates suggest that students more likely to be exposed to bilingual education relative to students more likely to be exposed to English immersion experience higher rates of grade retention and lower rates of schooling persistence and graduation