100%(1)1 out of 1 people found this document helpful
This preview shows page 41 - 44 out of 58 pages.
specifications in columns B and G, it appears that relative exposure to low income neighbors has as important an effect as does relative exposure to minority neighbors. VII.SUMMARY AND CONCLUSIONSIn this paper we present new evidence on the effects of racial segregation on the relative achievement of black students. Building from a model in which the racial composition of school and neighborhood peer groups exerts a causal effect on student achievement, we show that the black-white achievement gapin a city will vary with the relative segregation of schools and neighborhoods in the city. The model also suggests that
41in measuring the effects of racial segregation it is important to control for relative exposure of black and white students to other characteristics of school and neighborhood peer groups, such as family income. Otherwise, these differences will tend to lead to an overstatement of the effects of race per se. Our main empirical evidence is based on SAT outcomes for all the black test takers and one-quarter of the white test takers in the 1998-2001 test cohorts. We match test-takers to information on the racial composition of their high schools and to an extensive set of family background characteristics of black and white students in different cities. We use data from the summary files of the 2000 Census to construct estimates of the relative exposure of white and black students in a city to a variety of neighborhood characteristics, including racial/ethnic composition, income, and family structure. To address concerns about potential selectivity biases in the SAT outcomes, we also use 2000 Census micro data to construct measures of the relative achievement of black and white youth in different metropolitan areas. Without controlling for neighborhood segregation, we observe that school segregation has a negative effect on black relative test scores and on achievement measures from the Census. In models that include both school and neighborhood segregation, however, the effects of relative exposure to black and Hispanic schoolmates are uniformly small and statistically insignificant, whereas the effects of relative exposure to black and Hispanic neighbors are negative. Probes into possible explanations for the absence of school segregation effects, including instrumental variables estimates and assessments of correlated differences in unobserved school or peer quality, give no indication that our estimates are biased in a way that would obscure negative effects of school segregation. Finally, in models that include school segregation, neighborhood segregation, and measures
42of the relative exposure of blacks to other characteristics of their neighbors, even the neighborhood segregation effects are diminished.