Making Americans Healthier: Social and Economic Policy as Health Policy

Chapter 12. Are Some Neighborhoods Better for Child Health than Others?

344 / Chapter 12 Are Some Neighborhoods Better for Child Health than Others? Rebecca C. Fauth and Jeanne Brooks-Gunn I t is believed that the neighborhoods in which children and youth live are associated with their health and well-being. The underlying premise of this belief is that living in more affluent and safe neighborhoods results in better health (as well as increases the likelihood of doing well in school and obtaining better jobs and housing). While this assumption, on its face, seems obvious, it is incredibly difficult to substantiate from a social-science perspective. The primary reason is that families have some choice as to where they live. Consequently, when neighborhood influences are reported, individual-level variables may be accounting, wholly or in part, for the neighborhood-health links. The problem of selection bias in estimating effects of neighborhood residence on well-being is so unsettling (and difficult to solve in traditional regression analyses of cross-sectional and shortterm longitudinal data) that social scientists are quite divided as to whether they accept this premise at all. To better understand the source of this division, we examine literature on neighborhood -health links with a focus on children and youth. Our interest is informed by several policy-relevant situations. The first has to do with increases in neighborhoods of concentrated poverty in urban areas during the 1970s and 1980s, as documented by William Julius Wilson (1987) and Douglas Massey (1990). Although neighborhood poverty rates have dropped slightly in recent years, in 2000 more than 15 million people in the United States lived in extremely impoverished neighborhoods. These neighborhoods are characterized by poverty rates of at least 30 percent and high rates of female-headed households, low educational attainment , and high unemployment (O’Hare and Mather 2003). The vast majority of these individuals were African American or Latino. In part, these trends are influenced by conditions such as the relocation of nonpoor families from central cities to ring communities and the suburbs, the lower demand for unskilled laborers in the workforce, exclusionary housing and zoning policies, the construction of massive public housing projects in cities during the 1940s and 1950s, and the deteriorating condition of the housing stock in older cities (Massey 1990; Wilson 1987, 1996). Second, even though the increase in poverty concentration attenuated somewhat in the 1990s, the divide between more- and less-advantaged families is growing (Sawhill and McLanahan 2006), as evidenced by rising work and income differences between single- and two-parent families as well as between parents with postsecondary education and those without it. If these family-level differences remain or accelerate, it is likely that neighborhood poverty concentration, or at least the differences between affluent and other neighborhoods, will increase in the coming decades. Third, racial- and social-class gaps in school achievement and health are, for the most part, not closing (Jencks and Phillips 1998; Rouse, BrooksGunn , and McLanahan 2005). If neighborhood conditions are part of the underlying cause for the persistence of these gaps, then alteration of such conditions may be one strategy for reducing them. There are multiple types of evidence for estimating neighborhood effects, and we focus on three research designs: neighborhood-cluster designs, “natural experiments ” (which allow researchers to explore variation in neighborhood conditions due to exogenous changes), and relocation experiments. For historical purposes, we also consider cross-sectional and longitudinal designs without a cluster feature . These various types of studies offer evidence about associations between neighborhood conditions and physical health outcomes (such as respiratory problems , low birth weight, lead poisoning, obesity, and ratings of poor health) and emotional distress (depressive and anxious symptoms). In addition, prior studies also allow for the examination of the relationship between neighborhood and health compromising behaviors (such as juvenile delinquency and crime, substance use, and early sexual behavior). Throughout these literatures, one can connect specific characteristics of neighborhoods to health outcomes. These characteristics include exposure to violence, social processes and normative control, availability of resources and institutions, quality of housing stock, and presence of environmental pollutants.1 STANDARDS OF EVIDENCE FOR ESTIMATING NEIGHBORHOOD EFFECTS The use of large-scale cross-sectional data sets to examine neighborhood effects on child and adolescent outcomes were the coin of the realm in earlier decades. Although these studies were not specifically designed to study neighborhood effects , they typically include a large range of neighborhood types—albeit with only a few children or families from each neighborhood—which allows for variation on measures of neighborhood dimensions (Duncan, Connell, and Klebanov...