Subjective Constructions of Neighborhood Boundaries: Lessons from a Qualitative Study of Four Neighborhoods

This article explores the boundaries of neighborhoods as subjectively constructed by 37 adolescents and 33 parents across four census-defined block groups in a Western city. We examine the degree of consensus among participants on the spatial boundaries of their neighborhoods, the stability of participants’ subjectively constructed neighborhood definitions, and the overlap between subjectively constructed definitions and census block group and tract definitions. Through an analysis of qualitative interviews, we isolate four factors that appear to influence how participants define their neighborhood boundaries: physical and institutional characteristics of the neighborhood, its class, race, and ethnic composition, perceived criminal threats from within and outside the neighborhood, and symbolic neighborhood identities. These factors can operate to facilitate or compromise consensus and stability about neighborhood boundaries and identity. The study findings are exploratory but suggest several avenues for further investigation into how parents and adolescents construct neighborhood boundaries and the possible influences that subjective neighborhood definitions have on families. Neighborhood has long been recognized as a defining social context of American life (e.g., Coleman, 1988; Massey & Denton, 1993; Park, 1936; Sampson & Morenoff, 1997; Sampson, Morenoff, & Earls, 1999; Suttles, 1972; Wilson, 1987). Chaskin (1997) refers to the intuitive appeal of the neighborhood construct, suggesting the term’s power comes from “its nuanced Direct Correspondence to: Julia R. Henly, School of Social Service Administration, University of Chicago, 969 East 60th Street, Chicago, IL 60637. E-mail: [email protected]. JOURNAL OF URBAN AFFAIRS, Volume 31, Number 4, pages 461–490. Copyright C © 2009 Urban Affairs Association All rights of reproduction in any form reserved. ISSN: 0735-2166. DOI: 10.1111/j.1467-9906.2009.00450.x 462 II JOURNAL OF URBAN AFFAIRS II Vol. 31/No. 4/2009 complexity as a vernacular term” (p. 524) more than from its precise scientific meaning. Indeed, despite a long tradition of neighborhood research, scholars continue to disagree about how to define neighborhood (Bursik & Grasmick, 1993; Gephart, 1997; Hallman, 1984; Keller, 1968; Lee & Campbell, 1997). Arguments range from what the appropriate physical borders of a given neighborhood are to whether place-based definitions of neighborhood still matter in the modern information age. Some network analytic approaches refute the a priori importance of place in favor of person-centered models of interpersonal ties, some of which may be geographically bound and others not (Elliott & Sims, 2001; Fischer, 1982; Wellman & Leighton, 1979; White & Guest, 2003). Still, the hypothesis that place matters continues to attract considerable scholarly attention. The “neighborhood effects” literature, for example, posits that physically bounded spatial units independently affect a range of outcomes from teen pregnancy and school drop-out, to employment, marriage, parenting, and perceptions of social disorder and crime (e.g., Brooks-Gunn, Duncan, Klebanov, & Sealand, 1993; Crane, 1991; Furstenberg, Cook, Eccles, Elder, & Sameroff, 1999; Hipp, 2007; Hogan & Kitagawa, 1985; Sampson, Morenoff, & Gannon-Rowley, 2002; Wilson, 1987, 1996). Furthermore, research links neighborhood characteristics to various health outcomes from regular preventive care to mortality (Cummins, Curtis, Diez-Roux, & Macintyre, 2007; Kirby & Kaneda, 2005; Roos, Magoon, Gupta, Chateau, & Veugelers, 2004). From such a perspective, a neighborhood is not good or bad for family functioning and child development solely because of the individual characteristics of its residents, but also because the structural characteristics of the neighborhood, its institutions, and the social and cultural forces operating within it independently promote or undermine positive developmental outcomes of residents.1 Among the many challenges of studying neighborhood effects is determining where the boundaries of a neighborhood should be drawn in the first place. For example, a researcher’s characterization of the social, organizational, and cultural capacity of a neighborhood, its level of disorganization, or its cohesiveness, depends on where the neighborhood’s boundaries are drawn a priori. If these initial boundaries miss the mark, the study findings may be misleading. For example, if census tracts are chosen to proxy neighborhood boundaries, but the concept under study is sensitive to relatively proximate influences such as those operating at the block level, true neighborhood effects will get masked when analyzed at the census tract level. In his recent American Sociological Review article, Hipp (2007) convincingly demonstrates the sensitivity of neighborhood effects research to boundary definition by providing empirical evidence that neighborhood characteristics differ in their effects on perceptions of crime and disorder depending on whether they are measured at the tract or block level. Echoing concerns identified in previous work (e.g., Openshaw & Taylor, 1979, 1981; Armhein, 1995), Hipp argues that selecting a geographic unit of analysis that is appropriate for the particular phenomenon under investigation is critical to accurately identifying neighborhood effects. Despite its importance, however, he points to a relative dearth of attention paid by researchers to questions of “appropriate aggregation” (662). Further complicating neighborhood boundary placement is its sensitivity to resident interpretation. Coulton, Korbin, Chan, and Su (2001) suggest that the effect a neighborhood has on individuals partially depends on residents’ subjective definitions of neighborhood boundaries, such that “variation among residents in how they see their neighborhoods may, in part, be responsible for the weak neighborhood effects found in many studies” (p. 382). Building on Coulton and her colleagues’ research emphasizing resident-defined neighborhood boundaries, this article examines the subjective representations of neighborhood boundaries in a sample of residents from four neighborhoods chosen for their variation in social disadvantage and adolescent developmental outcomes. The intent of the current analysis is to consider a set of questions about neighborhood boundary definition that may have implications for the conduct of neighborhood research and the interpretation of findings from neighborhood studies. Our intent is not to evaluate neighborhood effects, but II Subjective Constructions of Neighborhood Boundaries II 463 rather to elaborate our understanding of neighborhood boundary definition: How much consensus is there across residents’ subjective boundary definitions (i.e., their “cognitive neighborhood maps”)? How similar are cognitive neighborhood maps to administratively defined boundaries such as census tracts or block groups? And over the course of an interview, how stable are residents’ subjective definitions of neighborhood boundaries? After examining these questions, we explore four factors that we find influence participants’ subjective neighborhood boundaries: physical and institutional; characteristics; race, ethnic and class composition of the neighborhood; proximate criminal threats; and symbolic neighborhood identities. These factors emerged from the interviews as critical to informing subjective neighborhood definitions, and their importance is also indicated in previous research on neighborhood definition (e.g., Haeberle, 1988). Before addressing the key research questions, we review the literature on neighborhood meaning and measurement, and introduce our study design and methods. The results are presented in two sections. First, we examine neighborhood boundaries constructed by respondents, and compare these with the most commonly used administrative boundaries, the block group, and the census tract. We then present qualitative interview data to elucidate the factors that we found to most influence neighborhood boundary selection and definition. Finally, we provide an extended discussion of the findings and implications for neighborhood effects research generally.