Identifying changes in spatial patterns from police interventions: the importance of multiple methods of analysis

Abstract

The policing initiative of foot patrol was implemented to reduce crime and disorder as well as benefit the relationship between the community and the police in the community of Lower Lonsdale, North Vancouver, British Columbia. In this paper, police incident data are analyzed to evaluate the impact of police foot patrol on the hot spots of crime in this community. Specifically, two spatial analysis techniques (kernel density estimation and local Moran’s I) are used to show how the nuances of changes in spatial crime patterns can emerge when multiple methods of analysis are used.

Keywords:

• policing
• crime hotspots
• kernel density
• LISA

Notes

1. Kelling, Pate, Dieckman, and Brown (1974) is often cited as the known study of police patrol, but this study was in reference to a motorized patrol.

2. Mischief, in this particular context, is the term used to describe disorder.

3. First-order Queen’s contiguity is used for two reasons. First, it is commonly employed in the literature that uses LISA; and second, in a spatial regression (not reported here) first-order Queen’s contiguity successfully filters out spatial autocorrelation from the error terms.