Measuring Crime Concentration in Place and Time: A Distance-Based Alternative

• In Event: Communities and Place: Crime Concentration, Rural Areas, and Police

Abstract

Crime concentration in place refers to the uneven distribution of crime activities across microgeographic units, which has been well documented in the existing literature on crime and place. However, some criticisms suggest that this non-uniform distribution may reflect largely a random process. Traditional measures of crime concentration, such as the generalized Lorenz curve and Gini coefficient, fail to address this concern adequately. Analyzing the spatiotemporal concentration of crime (i.e., near-repeat victimizations) typically relies on contingency tables and the Knox test. However, creating contingency tables involves converting continuous spatial and temporal distances (e.g., in meters and days) into categorical measures (e.g., 0-100 meters and 0-7 days). This transformation can decrease measurement accuracy, raising doubts about whether the observed concentration is merely an artifact of measurement limitations. This paper proposes a distance-based alternative measure designed to address these concerns. By comparing this approach with commonly accepted measures, the paper discusses its advantages and disadvantages. I argue that the proposed measure can complement existing measures, offering researchers a more comprehensive understanding of crime concentration in place and time.

Author

• Yichao Gao, University at Albany, SUNY