This research note explores complications with standard methods to evaluate place-based policing interventions. It identifies and explains issues of boundary misspecification during evaluation as a result of boundary adjustment by police during an intervention.
Using geographic data gathered during post-experiment focus groups with officers involved in the Philadelphia Foot Patrol Experiment, the authors highlight the practice of boundary adjustment on the part of officers and we explain why such adjustments occurred. They find that officers involved in the focus groups who identified the active boundaries of their hot spot assignments all reported policing outside of their delineated beats. On average, their active beats were 0.13 square miles larger than the originally delineated treatment beats. Some active beats overlapped catchment and control locations.
Boundary misspecification could cause researchers to incorrectly label a direct benefit of receiving treatment as a diffusion of crime control benefits; underestimate immediate spatial crime displacement; and underestimate treatment effects.