Essays on immigrant students in public schools---evidence from North Carolina

Recent immigration has played a large role in changing the demographic make-up of public schools across the United States. Although some research has focused on the academic performance of immigrant students once they arrive to the United States, less is known about the interaction between these students with others attending the same local education systems. Further, even less is known about these interactions in new destination areas where recent immigration is a relatively new phenomenon. This study consists of three parts that focus on the influx of immigrant students into North Carolina, the state with the highest growth rate of immigrants over the 1990s. Along with data from multiple public sources, confidential student level data were obtained through an agreement with the North Carolina Education Research Data Center at Duke University. These data are used to describe the changing population, study how others react in their schooling decisions, and test for impacts on the academic performance of those that attend schools with immigrant students.;Part I of this study has two goals. The first is to detail changes in population and household characteristics using population surveys. The second is to detail the public schools attended by, and academic performance of, immigrant students in grades three through eight. This is done by focusing on those who have been identified as English language learners. I find the demographic changes are mostly driven by Hispanic immigrants, but there is also significant growth of Asian immigrant families as well. However, there are large differences between these two immigrant groups that is materialized in parents' education, household finances, and academic performance. Overall, Asian immigrant families are more educated than the native population, have higher family incomes, and better English speaking ability, where the reverse is true for Hispanic immigrants. In public school grades three through eight, Asians attend better schools than Hispanics, regardless of English learner status. When controlling for student and school characteristics, after exiting Limited English Proficiency status, Asians perform .15 standard deviations better at math than white students, while Hispanics perform .06 standard deviations worse. However, both Asian and Hispanic students struggle at reading, scoring 0.13 standard deviations below whites even after testing out of Limited English Proficiency status. Based on the lower socioeconomic standing of most Hispanic families, these students are at particular risk for lower educational outcomes.;The second part of this study is concerned with the schooling decisions of other students as a response to the changing demographics caused by Hispanic immigrants. I focus on the sorting response away from immigrant students as measured by the avoidance of public schools that attract large shares of Hispanic immigrants. I first document a strong relationship between measures of sorting and the share of immigrant students attending public elementary and middle schools. To test for a causal relationship, I then focus on the academic years ending in 2003 and 2004 to estimate demand for public elementary schools in a differentiated goods framework. The endogenous relationship between immigrants moving to North Carolina and unobservable factors influencing schooling decisions is addressed using instruments from lagged data on industries that employ large shares of low-skilled Hispanic immigrant workers. Although I find that students are avoiding immigrants in schools, this does not vary by race. Rather, it is a function of household income and education levels. After controlling for school characteristics, including educational quality, students whose parents do not have a BA, but are wealthier, are avoiding schools with higher shares of immigrants the most.;The final part of this study focuses on the academic performance of students that remain in schools with an increasing number of Limited English Proficient (LEP) students. LEP students use additional education resources and change classroom dynamics. Indeed, this has caused some to claim these students are a burden on public schools. Using the administrative data, I test for peer effects of English learners on others in public elementary schools, where the outcomes of interest are standardized Math and Reading exams. To control for adverse selection, I create forecasts of LEP shares and limit comparisons to students attending schools with similar forecasts, where variation in actual shares still exists. The large dataset allows for use of covariate matching methods that highlight a fine disaggregation of heterogeneous effects. Contrary to fears of negative effects, both positive and negative effects exist, which are small in magnitude and reveal few generalizable trends across demographic groups.