Date of Award

Summer 8-1-2020

Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Tim Sass

Second Advisor

Daniel Kreisman

Third Advisor

Thomas Mroz

Fourth Advisor

Ross Rubenstein


Using longitudinal administrative data and quasi-experimental methods, this dissertation evaluates the impact of changing demographics in schools and education policies designed to enhance the academic achievement of English Learners.

Chapter 1 contributes to an emerging literature on the externalities of refugee integration by providing evidence on how this population affects the academic performance and behavior of incumbent students. Leveraging variation in the share of refugees within schools and across grades, I find that increasing the share of grade-level refugees by 1 percentage point leads to a 0.01 standard deviation increase in average math test scores. I find suggestive evidence that the positive spillovers in math achievement are driven by changes in classroom resources and access to academic support programs. While I find no effect on average English Language Arts (ELA) test scores, using nonlinear-in-means specifications I find evidence of negative spillovers in ELA performance among low-achieving students and positive spillovers among high-achieving students.

Chapter 2 estimates the impact of a temporary intensive English program aimed at English Learners with very low English proficiency. Access to the program is based on a maximum score on a standardized English proficiency screening assessment and grade level at screening, which I leverage to employ difference-in-differences and regression discontinuity approaches in my analysis. I estimate the impact of program access and participation on ELA and math test scores in the short term, relative to receiving traditional English as a Second Language support. On average, students who are eligible for the program have lower ELA test scores one year after program eligibility. However, the impact is large and positive among the subsample of refugee students. Results on the impact on program enrollment also show that students who participate have lower ELA achievement. I also find lower math test scores among program participants, relative to English Learners who receive traditional English as a Second Language support.

Chapter 3 presents new evidence on the impact of Dual Language Immersion programs on student academic outcomes. Leveraging enrollment lotteries from five oversubscribed DLI schools, I estimate intent-to-treat and local-average-treatment effects of bilingual education on English Language Arts, reading, and math test scores. On average, I find no difference in reading and ELA achievement between students with access to DLIs and those enrolled in traditional public schools. However, I find weak evidence of lower math achievement among lottery winners. Results vary by students’ initial EL status. Specifically, I find that native English speakers with access to DLI programs have higher reading and ELA achievement, relative to DLI lottery losers.