volunteered for the program reported that escaping from gangs and drugs was their most important reason for participating in the program. Katz et al. (1999) used the experimental data from the Boston site to evaluate the effects of the program on, among other things, maternal reports of the health and behavior problems of children between ages 6 and 15 and on their own mental health. They found significantly fewer injuries, accidents, and asthma attacks for children in the experimental compared with the control groups. Furthermore, rates of behavior problems among boys in the experimental group were significantly lower than among boys in the control groups. Measures such as “is cruel or mean to others” showed a larger experimental effect than measures such as “is unhappy, sad or depressed.” In the case of behavior problems, there were no significant differences in program or control groups for girls. Ludwig et al. (in press) used the experimental data from the Baltimore site to evaluate the effects of the program on the frequency of criminal activity among adolescents, as reflected in the criminal offender records of the Maryland Department of Juvenile Justice. They found a sizable and statistically significant reduction in the proportion of youth who were arrested for violent offenses. However, property crime arrests appeared to be higher for the experimental group, particularly in the first year after the move, perhaps reflecting the greater opportunities for property theft in the new, more affluent locations.
How do we reconcile the large effects found in the Gautreaux and Moving to Opportunity studies with the more modest impacts found in the nonexperimental literature? One possibility is that the effects found in the experimental studies are less biased owing to their experimental designs. Another is that improving neighborhood conditions matters a lot for the development of children in the worst neighborhoods —a finding that could be masked in population-based studies that look at a wide range of neighborhoods. A third possibility is that the nature of the samples for these two interventions—which consisted of voluntary participants who were also, to some extent, screened by those administering the programs—produced larger effects than would be the case for a less selective sample of low-income, inner-city families. Whether beneficial effects extend from schoolage to younger children is another critical issue that needs to be addressed in research on neighborhoods.
Although most urban poor families still live outside high-poverty neighborhoods, the past quarter century has produced an alarming growth in the fraction of poor urban families who do live in such neighborhoods. The combination of family poverty and neighborhood poverty poses double risk to a substantial minority of black children and, to a lesser extent, to His-