Children and Neighborhoods: A Randomized Study of Mobility

National Bureau Of Economic Research Inc, Cambridge MA

Investigators

Abstract

During the past several decades, poverty in the U.S. has become more urban and more spatially concentrated, particularly among racial and ethnic minorities. In this proposal, we outline a strategy designed to assess the effects of residential location in high-poverty neighborhoods on child learning and development, and to illuminate the basic mechanisms through which these effects occur. To execute this strategy, we propose to utilize the research platform provided by the Moving To Opportunity (MTO) demonstration program, developed by the U.S. Department of Housing and Urban Development (HUD). MTO used a random lottery among housing-project residents in five of the nation's largest cities to offer rent subsidies that could be used to move to lower poverty neighborhoods. Data collected from children and their parents will be used to examine the effects of neighborhoods on child outcomes, and on the role of family and neighborhood factors that may lead to these outcomes. The randomization of the subsidy will be used to identify the causal effect of residential location on child and family outcomes using a more scientifically credible research design than has been used in previous research. This experimental design will be used to solve the fundamental selection problem in non-experimental research, where comparisons between individuals living in different neighborhoods may be biased by the fact that families choose their residential location. Furthermore, we will be able to minimize potential survey attrition bias, because substantial tracking has been done to keep sample location information current. Preliminary research has been conducted at several MTO sites. In Boston, a survey designed by Kling and his colleagues was completed approximately two years after random assignment with 520 of 540 sample households, for a 96% response rate. Moves to lower poverty neighborhoods induced by MTO subsidy offers appear to have produced significant improvements in the problem behaviors of boys, the physical health of boys and girls, and the overall and mental health of household heads, without substantially worsening social isolation of participating families. Analyses by Brooks-Gunn and Leventhal of initial data from the New York site suggest similar results. Research by Duncan and his colleagues using data from the Baltimore site suggests reductions in criminal behavior among adolescent boys. Additional research is needed, however, to directly assess educational achievement of children, to understand the mechanisms through which neighborhoods affect learning, and to see if initial program benefits persist. Larger sample sizes will also facilitate analyses of particular subgroups, by age, race, gender, or time since random assignment. This proposal provides a plan to expand MTO research to include an in-home interview with up to two children ages 5 to 11 per household (and a parent), for an approximate sampling universe of 3900 children and with a survey response rate target of 90%. Data will be collected on children's educational achieve-ment; children's behavior problems and delinquency; social networks, peer deviance, and exposure to violence; future expectations and efficacy in school, community, and family; home environ-ment, family routines, primary caregiver characteristics, and child care arrangements; and the economic and demographic situation of the family. This research is designed to complement HUD's concurrent policy evaluation centered on MTO household heads, and will enable a much deeper exploration of the relationship between children and their neighborhoods. The analysis will provide critical evidence on basic environmental factors that influence key developmental outcomes such as knowledge and skill acquisition and social behavior.