Geographic reach considerations for intervention analysis and crime prevention

This study provides an evaluation of a park amenity, a fenced dog enclosure, on crime in its vicinity. This study also includes geographic variability and assesses the intervention’s impact at multiple distances from the intervention. This study uses an interrupted time series analysis focused on a before-and-after comparison. Monthly incidents of police calls-for-service (CFS) from 2010 to 2016 are analyzed (n = 84). We analyze two series: (1) within 50 feet of the intervention site; and (2) within 700 feet of the intervention. After the installation of the dog park, CFS within 50 feet of the intervention increased (b = 0.776, p < 0.05). However, the expanded geographic coverage of 700 feet shows that the intervention has no statistically significant effect on the series (b = 5.25, p > 0.05). This work found that the dog park intervention increased CFS immediately around the park, and that decisions regarding level of geographic aggregation can change the conclusions of an evaluation.