Research That Matters (January 17 - 20, 2008)


Governor's Room (Omni Shoreham)

Defining Neighborhood Boundaries for Social Measurement

Kirk A. Foster, MDiv, MSW, Washington University in Saint Louis.

Background The effects of place on social outcomes have long been an interest of social work research. Much of the current neighborhood-based research uses variables aggregated on administrative boundaries such as census tracts. However, such units are not necessarily useful in exploring how the spatial concentration of neighborhood factors affects individuals and groups. As collective attributes of neighborhoods are important to measure because of their potential to elucidate how ecological factors moderate the effects of other variables, our methods of defining neighborhood boundaries must better respect the natural logic and flow of spatial relationships. This paper examines the relationship between demographic characteristics and street boundaries as a first step in exploring a conceptually new method of defining neighborhood boundaries for social measurement. Using ArcGIS software allows for the visual examination of aggregate data as they relate to specific geographical boundaries governing the natural logic of neighbor interactions.

Methods This conceptual approach uses a GIS (geospatial information system) to examine the relationship between aggregate characteristics and geography. In this GIS, two contiguous zip codes in a Midwestern city are examined and were chosen due to the variance in demographic characteristics and the prominence of geographic barriers limiting the ease of foot travel across street boundaries. The GIS combines data from the 2000 Census at the tract and block group level with street grids to explore the spatial relationship of census data to primary (busy streets that are difficult to cross and serve as psychological barriers) and tertiary (walker-friendly) streets. Data were layered for spatial analysis.

Results Analysis of the data reveals that geographic barriers such as primary streets, parks, and gated streets impact the natural settlement patterns of people into neighborhoods. The GIS shows that primary streets bisect census tracts and block groups and that demographic characteristic vary between block groups within areas that can be defined as aggregate neighborhoods. The GIS also reveals that land use patterns have a downward bias on neighborhood characteristics when aggregated solely along administrative boundaries.

Conclusions To improve the reliability of measures and the veracity of conclusions from neighborhood studies, neighborhoods must be treated as ecological collectives and data aggregated accordingly. The effort here proposes a method to define and measure this concept in a way that minimizes within group variance, maximizes between group variance, and reduces measurement error. The GIS in this exploratory study demonstrated that researchers may make incorrect assumptions about the differences between residents of contiguous block groups. Analysis indicated that while administrative boundaries are an easy and inexpensive way to examine neighborhood effects, they may yield misleading results by aggregating too widely or too narrowly potentially creating false dichotomies within neighborhoods. Utilizing a GIS has the potential to provide better specification in neighborhood research and increase the homogeneity among neighborhoods as a unit of measure.