Policing Predicted Crime Areas: An Operationally-Realistic Randomized, Controlled Field Experiment, Philadelphia, Pennsylvania, 2015-2016

The Philadelphia Predictive Policing Experiment was a place-based, randomized control trial to study the impact of different patrol strategies on violent and property crime in predicted crime areas. The experiment's goal was to learn whether different operationally-realistic police responses to crime forecasts, estimated by a predictive policing software program, would reduce crime. Specifically, the study tested whether greater awareness among general duties patrol officers of the predicted crime areas would be sufficient to deter crime, whether a dedicated uniform patrol attendance in predictive areas would increase visible police presence sufficiently in the local area to deter crime, or if dedicated plain-clothes units performing surveillance and unmarked patrol would increase interdiction and offender incapacitation sufficiently to reduce crime. With support of the Philadelphia Police Department, the study took place over two, three-month periods between 2015 and 2016.

费城预测警务实验（Philadelphia Predictive Policing Experiment）是一项基于地点的随机对照试验（randomized control trial），旨在研究不同巡逻策略对预测犯罪区域内暴力犯罪与财产犯罪的影响。本实验的核心目标是验证：由预测警务软件程序生成的犯罪预测结果所对应的、不同符合实战操作要求的警务响应措施，是否能够降低犯罪率。具体而言，本研究分别测试了三类方案：一是提升常规巡逻警员对预测犯罪区域的知悉程度，是否足以遏制犯罪；二是在预测犯罪区域安排专属着装巡逻勤务，是否能在当地大幅提升显性警力部署以威慑犯罪；三是部署专属便衣警力队伍开展监视与无标识巡逻，是否能有效提升执法拦截与违法人员羁押效能，进而降低犯罪率。在费城警察局（Philadelphia Police Department）的支持下，本研究于2015年至2016年间分两个为期三个月的阶段实施。