Despite the recent proliferation of social scientific literature focusing on neighborhood effects on an array of child outcomes, little is known about the magnitude of such effects or the mechanisms by which these effects transpire across developmental stages, gender and ethnicity. In this paper, we contribute to this literature by using a natural experiment in Denver to quantify the relationships between various measures of neighborhood context and Latino and African American educational outcomes. We address the question: For Latino and African American children who spent significant periods during childhood residing in public housing, are there significant differences in their elementary, middle and high school educational outcomes (grades, timely promotions, dropping out) that can be attributed to differences in their concurrent, lagged, and/or cumulative neighborhood environments, all else equal? What do caregivers and their children perceive as risk and protective factors?
Methods:
In this study, we utilize data from a natural experiment in Denver. Since 1969, the Denver Housing Authority has randomly assigned public housing residents to neighborhoods, providing a unique opportunity to disentangle neighborhood effects by controlling for selection bias. Our analysis is based on a sample of Latino and African American children who resided in Denver public housing for the majority of one or more developmental stages (N=1692). Data sources include (1) survey data from parent/caregivers; (2) administrative data from the U.S.Census Bureau and the Piton Foundation; and (3) 82 in-depth interviews with caregivers and their young adult children. Data gathered from parent/caregivers were geocoded for each year of their child(ren)'s life thereby providing a rare opportunity to comprehensively examine neighborhood exposure.
Results:
Using logistic regression with a clustered robust error adjustment to account for clustering at the family level, we find that neighborhood poverty rate is a weak and incomplete measure of neighborhood context affecting school performance compared to neighborhood: violent crime rates, social capital, child abuse rates, homeownership rates, and indoor recreational facilities, all of which show important (and often nonlinear) associations with grades and on-time promotion. At the high school level, we found that prior grades and on-time promotion substantially mediate the effects of neighborhood contexts on dropping out. Neighborhood effects appear much stronger for males and when measured concurrently with developmental stage, although patterns differ substantially by ethnicity. Caregivers and their children underscore the importance of neighborhood recreational facilities as protective factors enhancing educational outcomes across childhood.
Conclusions and Implications:
Our findings suggest that education results from the MTO demonstration are not general. Further, they underscore significant differences in the timing, magnitude and mechanisms by which neighborhood contexts influence the educational outcomes of low-income, minority children.
