The Effect of Immigrant Composition on Student Achievement: Evidence from New York City
There has been a large body of recent literature focused on the effects of school composition on student outcomes. These studies have focused on peer group characteristics such as achievement, gender composition, ethnic and racial composition, and socioeconomic composition. This area of research has been commonly called "peer effects." A relatively unexplored area of peer effects research involves the effect of immigrant children on their schoolmates. Because of the heterogeneity between immigrant groups, this study focuses on East Asian and Dominican immigrant children. As these two groups are on opposite sides of the socioeconomic spectrum, comparing results of the two analyses should provide a reasonably complete picture of immigrant composition effects.
The data for this study come from New York City. New York City is arguably the ideal place to study immigration. Immigrants from though out the world attend New York City schools. While New York remains an outlier, it is quickly becoming the norm. In recent decades, various parts of the country that have not experienced large waves of immigration are doing so now. The experience of New York has potential to inform the larger debate on the cost of providing public services to immigrants. If immigrant children have negative effects on their schoolmates, they will increase the cost of education. On the other hand, if they have positive effects, they can serve as a positive externality and reduce the cost of public education.
The estimation of peer effects is a daunting challenge. One of the most challenging of these problems is called the selection problem. The selection problem occurs because immigrant children are not randomly assigned to classes, schools, or neighborhoods. To overcome this problem, this study uses credibly exogenous variation that occurs as a student progresses with a cohort within a school.
The results suggest that both East Asian and Dominican immigrant composition has a negative and significant effect on student achievement. This effect occurs for all subgroups and for both English-Language Arts and mathematics. Surprisingly, this immigrant composition effect is not driven by ELL status.
This coefficient can be considered something of a reduced-form measure of immigrant composition effect. Regressions that control for other country variables suggest that schools with growth in East Asian and Dominican immigrant composition also have growth in other forms of immigrant composition. When including these other variables, the results suggest a cultural effect. East Asian immigrants have positive effects in mathematics while Dominican immigrants continue to have negative effects, though at smaller magnitudes. These results suggest that culture matters.
As a matter of policy though, given that immigrants move together, it is not practical to separate specific ethnic immigrant effects. Rather policy recommendation should look at the "reduced form" effects. Potential policy recommendations include additional resources for immigrant education such as English as a second language and civics classes or newcomer schools. Ethnographic research on how immigrant children interact with their classmates and schools could also be valuable in deciphering the exact mechanism behind this negative effect.