The Relationship Between Crime and Urban Location in Raleigh, North Carolina

by Elizabeth L. Davison
Appalachian State University

William R. Smith
North Carolina State University

    A Geographic Information System (GIS) analysis of 13,490 burglary, larceny (not including motor vehicle thefts) and robbery events that occurred in Raleigh during 1993 shows these events only happened in 6% of all Raleigh places (N=114,549).  This finding provides evidence that crime is not a random event, but happens because of social structures and processes.

    GIS represents a powerful research technology that can be used for mapping and studying the spatial distributions of various social phenomena, including events (crimes, illnesses, accidents), placements (social service, prison releases), and organizational services (police beats, school districts). One of the marvels of GIS is that it enables researchers to link seemingly unrelated data sets through a process called geocoding (matching addresses to geographical coordinates) which creates many types of relevant information for each spatial point. Five years ago we used GIS software to construct a data base containing tax assessor information, census data, official police records and 911 calls for service for the city of Raleigh.

    Another powerful asset of GIS software is that it enables the researcher to aggregate data to different levels.  For example, we aggregated the Raleigh data at the address level to face blocks (both sides of the street between two city blocks), city blocks, as well as larger geographic areas (census tracts).  In general we find that traditional ecological constructs are best in explaining smaller units of analysis such as the face block (both sides of a street between two intersections) rather than the more commonly used census tracts and block groups that are assailable because of spatial heterogeneity concerns (e.g., a large areas such as a census tract can be half African American and half white and still be racially segregated within the area).  Additionally, results from smaller units of analysis yield more specific policy recommendations.

Map Showing the Relationship Between Location and Robberies in Raleigh in 1993

    In addition to the statistical analysis, GIS can create powerful visuals to demonstrate findings. To illustrate, the Map of 1993 Robberies shows a cross-sectional view of the south western part of Raleigh. The map is not layered except for streets (black lines) and robbery incidents (a triangle represents one or more robbery incidents). The plotting of robberies among the streets allows for a better image of some of the city details since incidents of robberies are infrequent compared to burglaries or larcenies. Also, robbery is considered a more serious crime that is subsequently more likely to be reported and less likely to be influenced by policing bias. A striking observation of this map supports the assumption of the routine activity approach that crime is not a spatially random event. Emerging from the map are notable patterns in the distribution of robberies among face blocks. These trends can be explained by both social disorganization and routine activities factors (some of which we have not tested).

    True to the concentric zone models of the Chicago School, robberies are clustered around the center of the city (area 1) where businesses, government offices and motivated offenders are in close proximity of each other. In general, robbery incidents become less frequent away from the center of the city. Surrounding the central business district is the "transitional zone" (area 2) which consists of neighborhoods that are typical of social disorganization neighborhoods. These neighborhoods contain mixed parcels of residential and businesses, concentrations of public housing surrounded by substandard rental property, dense population, high unemployment, and many social ills that contribute to the motivation of offenders. A notable exception to the concentration of robberies in the downtown area, is the clustering of robberies on the "campus main drag" (area 3) which is located west of the capital and consists primarily of businesses that cater to the university community. The high rate of victimization in this area can be attributed to the routine activity concepts of "opportunity" and "awareness space." Heavy vehicular and pedestrian traffic brings a concentration of cars, people, products and money into the area creating more opportunities for robbery and "awareness space" for motivated offenders. Another surprising clustering of robberies that are away from the business and transitional zones is found in the lower left hand corner (area 4). This area consists primarily of student apartments that are characterized by a lot of transition and many anonymous interactions with neighbors. Both of these factors create opportunities for crime victimization. Other routine activities factors include that fact that no cul-de-sac street have a robbery event since these areas are not in the "awareness space" of most offenders. Conversely, the main thoroughfare has more crime due to increase in awareness space and opportunities. Area 5 (towards the upper left hand corner of the map) is virtually untouched by robbery incidents. This area consists of predominantly single family dwellings which caters to faculty and other professionals. A high concentration of home ownership acts as guardianship over properties in this area.

Here Are Some of the Major Findings:

  1. Automobile theft analysis reveals that proximity to areas where motivated offenders live increases the chances of automobile theft, as does proximity to locations where other automobile thefts occur.

  2. An analysis of street robbery reveals that
    • the potential of robbery occurring on a face block increases by 1% for every ten places,
    • decreases by 2% for every mile away from the downtown area,
    • increases by 9% for every motel or hotel and increases 5% for every gas station, bar or restaurant.
    • In addition, the potential of street robbery occurring on a face block results from a spatial diffusion effect.  In other words, places nearby existing robbery events are at a greater risk of robbery than places away from robbery incidents. 

  3. Uniform Crime Report data from the FBI are often touted as the most valid measures of crime, but findings suggest that 911 calls and UCR data perform similarly and provide equally valid measures of crimes for ecological theory testing.

  4. The existence of owner-occupied housing on a face block significantly decreases crime particularly in areas closer to downtown Raleigh.

    In conclusion, GIS software has aided us in compiling a database that has furthered the understanding of crime distributions for various geographical boundaries. Since individuals do not often experience the "real world" through administrative geographical boundaries such
as school districts, census tracts and voting precincts, researchers need to have the flexibility to study units of analysis that are determined by theoretical undergirding rather than available resources.

Note: findings include research by Elizabeth L. Davison, Sharon Glave Frazee, Kennon J. Rice, William R. Smith, and Claudia Squire.

For further information see Smith, William R., Sharon Frazee and Elizabeth Davison.  "Towards and Integration of  Routine Activity and Social Disorganization Theories: A Socio-ecological Analysis of Robbery." Criminology (forthcoming)

Source: This is a longer version of an article from April/May 2000 Sociation.

Want to see how urban design can influence crime rates? Oscar Newman's book Creating Defensible Space is now available online in *.pdf format.

 Return to Sociation Today Urban Sociology Reprint Series Outline