Law enforcement officers are second only to taxi cab drivers in terms of the rates at which they are murdered on the job. Their rate of non-fatal violent victimizations exceeds that of taxi drivers and, indeed, exceeds the rates of all other occupational groups. Despite the seriousness and importance of the problem of violence against the police and despite considerable changes within agencies over recent decades geared toward improving officer safety, surprisingly little research has been conducted to study the impact of various law enforcement agency initiatives on the level of violence against their personnel. The purpose of this project was to identify the factors, both internal and external to law enforcement agencies, which impact the rate at which police officers are assaulted and murdered. The dependent variable is the rounded average count of officer killings and assaults in each subject jurisdiction for 2000 through 2002 as reported in the National Incident-Based Reporting System (NIBRS). This represents the first time that data from NIBRS are used to study violence against police. Unlike other sources of data on police violence, NIBRS provides valid and reliable information on both homicides of and assaults against police. Two sets of independent variables represent (1) the factors internal to the agency that might impact on officer safety (e.g., training, policies, practices, equipment), and (2) factors external to the agency that might impact on the rate at which officers are assaulted/killed (e.g.,violent crime rate, poverty level). Independent variables are also classified according to our theoretical constructs of "exposure" and "guardianship." Based on routine activity theory, exposure refers to the proximity of potential crime victims to motivated offenders (in our study, proximity of police to subjects who are motivated to harm them). Guardianship refers to Violence Against Police protective factors, such as the policies, practices, tools and training that agencies use to promote officer safety. Information to measure the independent variables comes from the U.S. Census, Uniform Crime Reports, and an agency survey. The subjects are the 121 law enforcement agencies that submitted NIBRS data for 2001, serve populations of 50,000 or more, and returned the project survey. (The response rate for the survey was 77 percent.) Descriptive data from the survey provide important information regarding the extent to which agencies across the nation use various policies, procedures, tools and training programs to enhance officer safety. Negative binomial regression was used to assess the impact of those policies, practices and tools on violence against police. Only four variables produced statistical significance: number of Part I arrests per officer, percentage of female headed households in the jurisdiction, number of officers in the jurisdiction per 100,000 people, and agency policies/practices to promote the wearing of body armor. These results are discussed in the context of police practices, the methods used to produce the results, prior research and future research.
Police Executive Research Forum, 1120 Connecticut Ave., NW, Suite 930, Washington, DC 20036