Study 1
Groff and colleagues (2015) evaluated offender-focused (OF) policing as part of the Philadelphia Policing Tactics Experiment. The study involved evaluations of three tactical approaches employed in the experiment: foot patrol, problem-oriented policing, and OF policing. As part of the experiment, intelligence analysts identified 81 violent-crime hot spots within Philadelphia, and police commanders determined which one of the three tactical approaches would be most appropriate to employ at each hot spot. Commanders chose 27 hot spots for each of the three interventions. Hot spots in each of these groups were randomly assigned to treatment and control categories. With regard to the OF policing approach, 20 hot spots received OF policing for 12 weeks and seven hot spots served as control areas. These control areas received no additional police attention and officers patrolled as usual. The treatment and control areas were equivalent at baseline due to the randomization procedure.
The primary outcome of interest was change in violent crime during the intervention period, which was examined using repeated-measures multilevel models, controlling for seasonality and the size of the hot spot. Contrast coding was used to examine crime rates during the intervention period rather than before or after the tactics were instituted. Violent crime was measured in two ways: violent felonies and all violent crimes. Violent crimes included all homicides, robberies, and aggravated assaults (simple assaults were excluded). The Philadelphia Police Department provided official data on the number of violent crimes biweekly during the study period. Spatial displacement of violent crimes in areas surrounding the treatment hot spots was also measured.
Study 2
Ratcliffe and colleagues (2015) examined the extent to which the Philadelphia Policing Tactics Experiment (which included OF policing) influenced citizens’ perceptions of the police. The researchers administered mail surveys 90 days before the tactics were implemented and 90 days afterwards. The surveys gauged how OF policing, foot patrol, and problem-oriented policing altered perceptions of crime, safety, disorder, and satisfaction with the police in the 81 hot spots.
Any taxable residence within the hot spot was eligible to receive the survey. With regard to OF policing specifically, 1,830 surveys were sent to addresses in the 20 OF hot spots, as well as 1,855 surveys to residents in the control areas. A total of 152 surveys from OF hot spots and 159 from control areas were returned before the policing tactics were implemented. Post-intervention, 160 surveys were returned from the OF hot spots and 177 from control areas, amounting to a response rate of 9 percent. Due to high non-response bias, citizens who returned the surveys were likely not representative of all citizens in the intervention and control areas. Ordinary least squares (OLS) regressions were used to estimate the impact of the policing tactics on changes in community perceptions of crime, disorder, safety, and satisfaction with the police.
A limitation of this study was the inability to match survey responses to specific households before and after the policing tactics were implemented. The study authors suggested that it was unlikely that the responses were attained from the same residents at both times, making it difficult to know whether the dependent variables captured changes in perceptions or reflected the perceptions of two different groups at two different times.