U.S. flag

An official website of the United States government, Department of Justice.

Predictive Policing Analytic and Evaluation Research Support

Award Information

Award #
2009-IJ-CX-K114
Funding Category
Competitive
Location
Congressional District
Status
Closed
Funding First Awarded
2009
Total funding (to date)
$1,084,846

Description of original award (Fiscal Year 2009, $1,084,846)

The purpose of this project is to provide technical assistance and analytic support to seven pilot projects on Predictive Policing, and to evaluate the implementation and effectiveness of three of the pilot projects. This project will be conducted in two phases. The first will involve the provision of technical assistance to seven predictive policing pilot sites. The second phase will involve three sites (selected from the initial seven) in an evaluation of the implementation and effectiveness of predictive policing interventions.
In phase one, technical and analytic/evaluation assistance will be provided to each of the seven grantees through site visits, cross-site meetings, and continued communications. Implementation plans, appropriate measurements, and research designs for each of the pilot projects will be developed through a series of focus groups, meetings, and semi-structured interviews. Pilot sites will also receive assistance in the development of predictive analytics to address specific local problems. Assistance will also be provided to pilot sites to ensure that predictive policing plans are feasible and that the burden of data collection during the evaluation will be minimal for the grantees and respective police departments.

In the second phase, process evaluations and with-in and cross-site evaluations will be conducted. Administrative data, implementation and fidelity measures will be collected as predictive policing interventions are deployed. Outcome evaluation will focus on determining the accuracy of the prediction element and the success of the prevention approach. These analyses may range from relatively simple to very complex analyses such as combinations of interrupted time series, propensity score matching, and double robust regression procedures.

ca/ncf

Date Created: September 28, 2009