Evaluating the Use of Automated Vehicle Locator Technologies in Policing

Dallas AVL Experiment: Evaluating the Use of Automated Vehicle Locator Technologies in Policing


Police Foundation & National Institute of Justice


David Weisburd, Elizabeth Groff, Greg Jones, Karen L. Amendola, & Breanne Cave

Law enforcement agencies lack specific information describing where police officers patrol when not responding to calls for service. Instead, they have snapshots of events that are handled by police, such as the locations of crime reports, arrests, traffic citations, and pedestrian stops. And, despite a move toward more scientific “smart policing,” agencies still have little ability to assess the effectiveness of their deployment strategies in relationship to their goals.<>The authors sought to examine these two key gaps in the advancement of recent police innovations and answer several important questions:

  1. If the police have knowledge about where patrol resources are concentrated in a police agency, can commanders more successfully manage broad patrol resources?
  2. Within the context of a Compstat model, can they ensure that crime hot spots gain increased levels of patrol?
  3. If such knowledge were available to the police, will that help them to prevent crime?

The authors used trajectory analysis to identify four groups of beats with similar crime patterns. Commanders received information on the measured deployment levels—the amount of hours of vehicle presence as measured by an Automated Vehicle Locator (AVL) system. In addition, they received AVL-measured deployment information about Compstat hot spots (those identified for specific deployment strategies) in the treatment areas.

At the beat level, access to AVL-measured deployment information led commanders to request significantly higher amounts of patrol presence but did not result in an increase in actual patrol levels. At the hot-spot level, AVL did not lead commanders to request higher levels of patrol, but it did lead to higher actual levels of patrol at those places. Also, in contrast to the beat-level findings, the authors found treatment hot spots have about a 20 percent relative “decline” in crime.

The Dallas (Texas) AVL Experiment provides important information to improve understanding of how AVL technologies can be used to maximize patrol in police agencies. The data suggest that, at least in cities like Dallas with large geographies, AVL information will not aid patrol allocations because patrol coverage in beats is largely a function of cross-district dispatch rather than commander-specified deployment. However, it is effective in achieving higher levels of patrol in hot spots and significant reductions in crime.

You can access this resource online here.