Spatiotemporal modelling of correlated small-area outcomes: Analyzing the shared and typespecific patterns of crime and disorder

Quick, Matthew, Li, Guangquan and Law, Jane (2019) Spatiotemporal modelling of correlated small-area outcomes: Analyzing the shared and typespecific patterns of crime and disorder. Geographical Analysis, 51 (2). pp. 221-248. ISSN 0016-7363

Preview

Text ga-crimedisorder-spacetime-final.pdf - Accepted Version Download (2MB) | Preview

Abstract

This research applies a Bayesian spatio-temporal modeling approach to jointly analyze physical disorder, social disorder, property crime, and violent crime over five years at the small-area scale. Despite crime and disorder outcomes exhibiting similar spatio-temporal patterns, as hypothesized by broken windows and collective efficacy theories, past research often analyzes one outcome and overlooks correlations between related crimes as well as underlying spatial and/or temporal patterns common to multiple crime and disorder types. In this article, the best fitting model partitions the area-specific risk of each type of crime and disorder into one spatial shared component and four type-specific spatial, temporal, and space-time components. The shared component captures the spatial pattern common to physical disorder, social disorder, property crime, and violent crime. Results show that the spatial shared component explains the largest amounts of variability for all types of crime and disorder and that temporal components explained the least. Space-time interaction hotspots are identified via posterior probabilities and are examined to contextualize the broken windows theory. The applications of joint spatio-temporal modeling to ecological crime theories, policing, and urban policy are discussed.

Actions (login required)

View Item

Downloads