Micro Places – Sorg (2015)

Study Reference:

Sorg, E. (2015). An ex post facto evaluation of the Philadelphia GunStat model. Doctoral Dissertation, Temple University.


Location in the Matrix; Methodological Rigor; Outcome:

Micro places; Focused; Proactive; Rigorous; Backfire effect

*This intervention also focused on individuals as described by the author


What police practice or strategy was examined?

Philadelphia’s GunStat Program is both place-based and offender-focused and involves multiple stages. First, the Philadelphia Police Department’s (PPD) research analyst identified gun crime hot spot locations using gun crime incident data. Then, offenders who offended in the hot spots were identified and placed on a violent offender priority list to be targeted during the intervention. Offenders were selected based on arrest records, criminal intelligence, community input, and additional police intelligence. The third stage involved monitoring the individuals on the list by instructing officers to pay close attention to them, and when appropriate, to take law enforcement actions, such as making arrests or issuing citations. Finally, GunStat also involved enhanced sanctions for the target individuals if they were arrested. This aspect was handled by the district attorney’s office.


How was the intervention evaluated?

The intervention was evaluated by comparing those hot spots with offenders who received the intervention against similar hot spots with serious offenders who did not. Comparison hot spots were identified using propensity score matching that matched hot spots on levels of serious and minor crimes and disorders and police activity data. Violent street felonies during and post-treatment periods within those treatment and comparison areas were then collected and compared.


What were the key findings?

GunStat was not effective in reducing violent street felonies during the treatment period in the targeted hot spots. A backfire effect was detected during the post-treatment period (one year after GunStat), as violent crime levels were higher in treatment areas than in comparison areas.


What were the implications for law enforcement?

Although GunStat was ineffective in reducing violent crime, the authors argue that it was effective in building networks and relationships between law enforcement and district attorney offices.


Where can I find more information about this intervention, similar types of intervention, or related studies?