“Small is Beautiful,” But Maybe Not so Much in Micro-geographic Randomized Policing Experiments

• In Event: Theoretical and Methodological Advances in Policing I

Abstract

Numerous forces have been pushing randomized experiments on police crime prevention tactics, for all the right reasons, to ever smaller spatial and temporal scales. These scaling shifts, however, create policy, theoretical, and evaluation challenges. To investigate these challenges, the current work uses data from a randomized control trial gauging impacts of micro-geographic-targeted police patrol interventions on the weekly occurrence of daytime property crime. Both temporal and spatial scaling get considered. An effective intervention, assignment of a marked patrol car, associated with a halving of the property crime odds, is examined. Epidemiological tables show that at least a tripling of the duration of the 14-week experiment, or at least a doubling of the size of the mission areas, would be needed to demonstrate conventional significance. We detail why such expansions may not be feasible or advisable and consider implications.

Authors

• Ralph B. Taylor, Temple University

• Jerry Ratcliffe, Temple University