Modeling crime as pollution: Geographic variations in violent crime victimization risks

• In Event: Violent and Property Crime Victimization Social Context and Hot Spots

Abstract

Crime victimization has characteristics consistent with pollution and there may be benefits to thinking of these risks as pollutants (Eck, 2011). In environmental research GIS has been utilized to model the relative exposure of different demographic groups to pollutants (Hamer, Warner, Pierce & Huber, 2002). This paper proposes to modify this methodology to examine the how geography and race impact crime victimization risks in Portland, Oregon.

Using a comparative crime risk index, the paper explores the exposure to violent crime of White and Black residents, relative to the exposure of all other groups at the level of the police beat. Consistent with empirical literature on the geographic concentration of crime (Sherman et al., 1989; Weisburd et al., 2004), clustering and hot spot analysis reveal significant clustering of victimization hot spots: Both Anselin Local Moran's I and Getis Ord GI* indicate statistically significant clustering of both crime counts and hot spots for both groups, p > .001. Despite this clustering, variations in risk both within racial groups and between groups are identified. Ignoring these variations in risk may lead to suboptimal prevention efforts. In order to address this limitation, a weighted comparative crime risk index is proposed to help focus prevention.

Authors

• Greg Stewart, Portland Police Bureau

• Christian Peterson, Porltand Police Bureau

• Kimberly Barsamian Kahn, Portland State University