Micro-level spatio-temporal relationships between firearm arrests and shootings in Philadelphia: Implications for understanding of crime, time, place, and policing

Abstract

The current study examines the relationship between shootings and police firearm arrests at a more detailed spatial and temporal level than has previous work. Using data from Philadelphia during the years 2004 to 2007 two dynamics are investigated: the relationship between a shooting and subsequent police firearm arrests nearby in space and time; and the relationship between a police firearm arrest and subsequent shootings nearby in space and time. In order to simultaneously consider spatial and temporal variation at a more micro-level, the current study uses a modified version of the Knox (1964) close pair method, a spatio-temporal clustering technique first used to study contagious diseases, and later used to analyze near-repeat patterns in the study of crime. The first question explored the relationship between a shooting and subsequent police firearm arrests. Results showed elevated patterns of firearm arrests were approximately two and a half times greater than would be expected levels of firearm arrests than if shootings and subsequent firearm arrests lacked a spatio-temporal association. Greater than expected elevated patterns persisted for up to about a fifth of a mile away and about one week but the strength of these associations waned. The observed patterns suggest an immediate and geographically targeted police response to a shooting and a somewhat sustained effort. Turning attention to the next question, an initial slightly elevated level of shootings followed a firearm arrest but for only a couple of days and about one block; shooting swiftly dropped below expected levels as one moves away in time and space. The waning and eventual significant drop in shootings may arise from ecological deterrence, but any suppression of shootings was short-lived. Overall, the current work highlights the close associations in space and time between police and offenders and suggests that police and offender activity is not simultaneous as the police response to a shooting immediately whereas potential offender's response to police actions is moderately delayed. Potential implications for theory and policy regarding both police behavior/police organizational responsiveness and ecological deterrence are discussed.