How Schools and Neighborhoods Shape Achievement Gaps

• In Event: Segregation in New York City Schools and Ongoing Challenges to the American Schooling Dream

Abstract

Racial gaps in student achievement have been a persistent feature of the U.S. educational landscape, although most scholarly work to date compares only White and Black students. This study, by contrast, focuses on Black and Hispanic students, tracking their educational outcomes as they progress through school. Recent literature highlights schools and neighborhoods as the two most important contextual influences on students (Owens, 2010; Carlson and Cowen, 2015) – a unique contribution of this paper is exploiting administrative and census data to explore how school and community contexts shape student experience and how they differ across Black and Hispanic students. Further, we explore differences in average outcomes and also across the range of the distribution, separately tracking high- and low-performing students. Understanding the evolution of student outcomes, and the external factors that shape them, is key to framing appropriate policies that address how achievement gaps persist and even grow.
This study makes use of detailed student-level longitudinal data and school-level administrative data obtained from the New York City Department of Education to explore the differences in achievement levels in 3rd grade between Black students and Hispanic students and how these vary by subject, gender, poverty and ELL status. New York City’s large and diverse student population provides a unique opportunity to explore racial achievement gaps, allowing one to further disaggregate students to explore potential heterogeneity. We also analyze how the achievement growth trajectory for Black students during elementary and middle school compares to that for their Hispanic counterparts. We then look at how racial gaps evolve based on initial performance – do they widen or narrow when one separately considers high-performing students and low-performing students? We also analyze how schools attended by Black students differ from those attended by Hispanic students. Finally, we incorporate neighborhood-level data from the U.S. Census Bureau’s American Community Survey using students’ residential census tract – using census tracts as the unit of analysis allows us to focus on differences across small geographic units while still ensuring adequate statistical precision. We explore how neighborhoods in which Black students reside differ from Hispanic students’ neighborhoods, and whether high-performing Black students come from more disadvantaged neighborhoods compared with high-performing Hispanic students.
Results show that Hispanic students lag behind Black students in 3rd grade English tests, but score on par in Mathematics. As they progress through school, however, Hispanic students outperform their peers - closing the gap in English and opening up a statistically significant lead in Mathematics. This is also true when we separately track high- and low-performing students, though there is important heterogeneity by subject, gender and initial performance. There are also disparities in terms of peers, schools and neighborhoods – among other things, Black students disproportionately hail from communities with a higher incidence of crime, and this is also true among high-performing students too.
Against the backdrop of recent concerns about resurgent segregation, our results highlight how school and neighborhood factors differ among disadvantaged minority students, and how they interact to shape student outcomes.

Author

• Joydeep Roy, Independent Budget Office & Columbia University