Abstract:
Empirical research in criminology has traditionally focused on individuals and communities as primary units of analysis, but there has been growing interest in micro units of geography to study crime. The current trend toward smaller geographic units of analysis has much to offer, but this transition also requires greater attention to the selection of methods and analytical techniques. This dissertation discusses the untapped potential of spatiotemporal methods to improve evaluation and development of criminal justice policies and programs. This research relies on two innovate spatiotemporal methods that have been neglected in the field of criminology to demonstrate the utility of taking advantage of methodological advances that analyze data across both space and time simultaneously. The first technique demonstrated is a bivariate spatiotemporal Ripley’s k-function, which is employed to assess the deterrent effect of stop-question-frisk practices on crime across space at a daily level. The second technique demonstrated, agent-based modeling, is used to explore whether significant reductions in crime can be achieved if police use unallocated patrol time to engage in focused-deterrence strategies at hot spots rather than randomly patrolling a large geographic area. This dissertation focuses on how such spatiotemporal techniques can be useful tools to examine place-based criminal justice interventions and discusses ways the methods are advantageous over traditional, non-spatial evaluation methods.
