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

ABSTRACT:

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.

Notes

1 Several theoretical perspectives have been posited to elaborate why and how neighborhoods have their effects (see Gephart, 1997; Jencks & Mayer 1990). After initially finding limited evidence for neighborhood-level effects (see Jencks & Mayer’s 1990 review), research over the last two decades has demonstrated more promising results (see Sampson, Morenoff, & Gannon-Rowley, 2002).

2 In addition to phenomenological approaches, other alternative methods include examination of pedestrian accessibility and use patterns (Demerath & Levinger, 2003; Grannis, 1998) and systematic researcher observation of physical and social features of blocks or other geographic units (e.g., the “windshield surveys” used by Spencer, McDermott, Burton, & Kochman, 1997; see also Sampson & Raudenbush, 1999, 2004).

3 This qualitative study was a part of the Denver Neighborhood Study, which was itself part of the MacArthur Research Program on Successful Adolescent Development, a larger multi-site qualitative and quantitative study of neighborhood effects on adolescent development (see Elliott et al., 1996).

4 It is also worth noting that block groups are relatively small areas (averaging about 27 contiguous square blocks) that include a mix of resident ages and a mix of residential, commercial, and city properties. While we do not claim to have reached saturation in any block group, our relatively small sample size for each block group (14–19), should be viewed in light of the fact that the block groups themselves cover a small area, and our sampling criteria required us to identify households which included adolescents (census data indicate this would include between 16% and 33% of households in each of the block groups).

5 To preserve the confidentiality of study participants, the actual names of people, neighborhoods, streets, and parks have been replaced with pseudonyms, and some neighborhood features have been modified, including in some cases directionality. These efforts to maintain confidentiality are a requirement of the study, as outlined in the approved Institutional Review Board DNS protocol.

6 The race/ethnicity by block group is the following: Northside (non-Allenspark)—eight white, three Latino respondents; Allenspark—five Latino, one Asian American, one African American respondent; Martin Park—10 white, three Latino, one respondent ethnicity not identified; Westside—19 white respondents; Broadmore—17 white, two mixed (white and African American) heritage respondents.

7 The overall sample had 39/70 (55.71%) respondents with a relative in the sample, the breakdown for each of the four study neighborhoods was 50% (7/14) in Martin Park, 42% (8/19) in Broadmore, 68% (13/19) in Westside, and 61% 11/18) in Northside. Differences among the block groups in terms of numbers of related individuals did not appear to systematically relate to findings on shared agreement about neighborhood boundaries.

8 To ease readability of the figures, boundaries are not demarcated on the three block group maps that show significant boundary variability (Figures 2–4). The three shared boundaries of Broadmore participants are included on Figure 1.