Explaining Dosage Diffusion During Hot Spot Patrols: An Application of Optimal Foraging Theory to Police Officer Behavior

Abstract

Drawing from the concepts of optimal foraging theory, this paper presents and tests the assumptions of a foraging theory of police behavior during hot spots patrols. The theory explains why, over time, officers involved in hot spots policing interventions would leave the hot spots they are assigned to police and begin working within other locations. We test what factors influence the amounts of activity that officers undertake outside of their assigned hot spots and at nearby streets using data gathered as part of the Philadelphia Foot Patrol Experiment. Officers performed more activity outside of their beats as the experiment progressed. Several theoretically relevant variables predict the level of activity that officers perform outside their beats, including the size of the target area and the amounts of crime occurring within and outside of the target area. “Dosage diffusion” might be one reason why hot spot interventions have diminishing effects over time. From an optimal foraging theory perspective, hot spots requiring police officers to constrain their actions to pre-defined areas can be perceived as counter-intuitive by the officers, especially over extended periods of time. The results of this study support the suggestion that hot spots patrols should be short-term and randomly rotated across hot spots.

Keywords:

• dosage diffusion
• hot spots policing
• initial deterrence decay
• crackdown decay
• Philadelphia Foot Patrol Experiment

Acknowledgements

We would like to thank reviewer number three for raising these points. Future research might address whether the acceleration of officer conducted activity outside of their beats linked directly to the decreases in crime within targeted beats over time. Another important question for theory and practice would be whether officer movement outside of their assigned beats linked to a decay in the crime reduction benefits that our previous work had uncovered (Sorg et al., 2013). Although we attempted to model these relationships in our analysis, the HLM model reliability estimates dropped below the .05 threshold suggested by Raudenbush and Bryk (2002). Notwithstanding, the support for our foraging theory of police behavior has implications for how hot spots policing in particular is designed, organized and evaluated.

Notes

¹ Although there were 60 hot spots, the close proximity of the hot spots required some buffer zones to be combined. In total, 55 buffer zones were drawn, as 10 beats shared a buffer zone with another beat.

² Disorder stops include being stopped for prostitution, public drunkenness, loitering, and the violation of city ordinances such as carrying an open container or public urination. We excluded activities and arrests related to other offenses because these activities are likely to be a combination of reactive response to 911 calls and proactive policing. Since we are only interested in the extent to which officers were foraging for activity, we only include activities that are proactive and police initiated. As Ratcliffe et al. (2011:813) note, the activities we include in our outcome measure “are largely left to police to initiate, especially in higher crime areas.” This was also observed during field observations. Since the PPD generates a separate record if a pedestrian stop results in an arrest (one for the pedestrian stop and one for the arrest), there was no need to include arrests in our figure because this would have in effect double counted the incident if an arrest was made. Although the officers were on foot, our field observations suggested that officers were performing vehicle stops by standing on the corner and waving cars over if vehicle infractions was observed.

³ We recognize that by excluding activities undertaken outside of the buffer zones we are missing some activities that the officers conducted. However, in order to calculate variables for patch attributes, a closed geographic boundary is required. We are therefore modeling only one patch that officers could have chosen to forage in (the buffer zone). Drawing standardized boundaries for other patches outside of the buffer zone was not possible.

Additional information

Notes on contributors

Evan T. Sorg

Evan T. Sorg is an Assistant Professor in the Department of Law and Justice Studies at Rowan University, an instructor and affiliated researcher in the Center for Security and Crime Science, Temple University, and a former New York City Police Officer. His research involves police innovation, evaluation and geographic criminology.

Jennifer D. Wood

Jennifer Wood is an Associate Professor in the Department of Criminal Justice at Temple University. She conducts research in the areas of policing and regulation. Her current work centers on the public health dimensions of police work, including pre-booking strategies for addressing the health needs of justice-involved people. She is the North American editor for Policing and Society: An International Journal of Research and Policy.

Elizabeth R. Groff

Elizabeth R. Groff is an Associate Professor in the Department of Criminal Justice at Temple University where she is also associated faculty in the Center for Security and Crime Science and has a secondary appointment in the Department of Geography and Urban Studies. Her research interests are in the areas of geographic criminology, agent-based modeling, police practices and the use of technology in policing.

Jerry H. Ratcliffe

Jerry Ratcliffe is a Professor of Criminal Justice at Temple University. His research examines police effectiveness in crime reduction, intelligence-led policing and spatial analysis of urban crime patterns.