Study
Ariel and colleagues (2016) conducted a randomized controlled trial to examine the impact of a soft policing strategy for reducing crime in Peterborough, England. Located in East England, in the county of Cambridgeshire, Peterborough has a population of 200,000 residents. Residents are white (81 percent), Asian (12 percent), and Black (3 percent).
The study authors examined 72 hot spots, assigning 34 to the treatment group and 38 to the control group. In the treatment group, police community support officers (PCSOs) were deployed to hot spot areas. The PCSOs were instructed to employ proactive “soft” policing techniques. Specifically, PCSOs were unable to make arrests or carry weapons and were only required to be visible in their hot spot patrol areas. PCSOs were assigned to hot spots on a rotating basis to ensure that they patrolled different combinations of hots spots rather than remaining in certain few hot spots throughout the study. GPS locators were used to measure patrol time in hot spots. Because PCSOs were to spend only about 15 minutes per hot spot, the GPS tracker was set to ping every minute. By counting the number of pings, the GPS could detect how much time officers actually spent in each hot spot. The control group was assigned to use more traditional “hard” policing techniques, which consisted of regular armed police officers (i.e., regular officers) and PCSOs engaged in reactive policing by responding to emergency calls for service. Even though both groups included PCSOs and were tracked by GPS, the study compared the treatment group’s proactive and focused foot patrols (“soft” policing) with the control group’s primarily reactive (“hard” policing) approach.
Hot spots were defined as polygons with a radius of 150 meters that had had no fewer than 36 calls for service during the two-year period prior to the experiment. Calls for service were defined as calls made to the “999” emergency telephone number and included crimes such as antisocial behavior, suspicious circumstances, violence, burglary, and criminal damage. The crime outcome was defined as crime reports generated by victims as opposed to police-generated crime reports. Specifically, reported crimes such as theft, criminal damage, burglary, robbery, sexual offenses, and common assault were included as victim-generated crimes; however, crime reports that were generated by proactive police activity and investigation, such as drug offenses and stop-and-searches, were not included in the measurement of crime.
The study examined changes in the two outcome types (i.e., calls for service and reported crimes) from the pre-treatment period (the 2 years before the trial) to the post-treatment period (the 1 year period after the launch of the trial). The study reported standardized mean differences to measure treatment effects on calls for service and crime reported within the hot spot areas. No subgroup analyses were conducted.
