Moving Beyond Census Boundaries With Geographic Information Systems

George Rengert, Temple University

The arbitrary nature of Census boundaries has been recognized as a problem in urban analysis for some time. For example, old East Coast cities have not changed their political boundatires in many years although the contiguous built up area has continued to expand outward. The census bureau has recognized this problem by establishing Metropolitan Statistical Areas that include the politically bounded city plus the contiguously built up surrounding counties.

Within cities, census tracts, block groups and blocks incorporate ever decreasing areas that are as internally homogenous as possible. However, they are never completely homogenous with respect to a census variable. This creates an analytical problem. For example, suppose that a census tract is recorded as 50 percent black and 50 percent white. Does this mean that approximately every other household is of a different race or ore likely, that one side of the census tract is predominately black and the other side predominately white? Furthermore, we do not know the spatial arrangement of the racial groups from the census figures. They are assumed to be uniformly distributed since it is a characteristics of the bounded area.

In the present analysis, GIS is used to evaluate the importance of three theoretical explanations of the spatial arrangement of illegal drug sales arrests in Wilmongton, Delaware. The first explanation is taken from marketing geography and is a measure of the local demand for illegal drugs in residential communities taken from census variables. The second explanation is a measure of how accessible an area is to regional customers measured by access to expressway, highway and public transportation facilities using GIS methods. The third explanation is taken from routine activity theory that identifies facilities that are used on a routine basis by people with a high propensity for illegal drug use such as high schools and homeless shelters.

(Return to Program Resources)

Updated 05/20/2006