How the FBI Uses Big Data to Identify Crime Trends and Patterns

Evaluation of the Federal Bureau of Investigation’s Uniform Crime Reporting Program

Established in 1908, the current stated mission of the Federal Bureau of Investigation (FBI) is to “protect the American people and uphold the Constitution of the United States” (Our vision and mission, 2021, para. 2). To achieve this vital mission, the FBI investigates a wide range of domestic crimes, including white-collar, organized and violent crimes, among others. One of the key strategies that the FBI uses to fight these types of crimes is their Uniform Crime Reporting Program which can identify important trends to help the organization focus its efforts where they are most needed. To gain a better understanding of these tools, the purpose of this paper is to provide an evaluation of the FBI’s Uniform Crime Reporting Program website’s resources on crime and victimization, including a description of these online resources, the URL where they can be located and an analysis concerning how these selected resources differ in their respective approaches to information collection, including whether they report sampling information and what the reported statistics mean. In addition, a discussion concerning whether the FBI reports the sampling method used in the studies from which they obtained these statistics is followed by recommendations on improving these sampling methods in order to increase the representativeness and generalizability of these statistics. Finally, a summary of the research and key findings concerning the foregoing issues about the FBI and its Uniform Crime Reporting Program are presented in the conclusion.

Five selected resources from the FBI’s Uniform Crime Reporting Program Website

The five resources selected from the FBI’s Uniform Crime Report Program website at https://www.fbi.gov/services/cjis/ucr/ on crime or victimization and a brief description of these resources and their respective locations are provided below:

· Resource #1: National Incident-Based Reporting System (NIBRS) at https://www.fbi. gov/services/cjis/ucr/nibrs: This resource provides detailed information concerning every crime incident that is reported by nationwide law enforcement agencies together with separate offenses that occur during the same crime incident (i.e., information on the victims involved as well as known offenders, the relationships between offenders and victims, arrestees, and the property involved in crimes). At present, 57% of the nation’s 14,993 law enforcement agencies provide data for the NIBRS.

· Resource #2: Hate Crime Statistics (HCS) at https://www.fbi.gov/services/cjis/ucr/hate-crime: This information service collects data from data from city, county, college and university, state, tribal, and federal agencies concerning various types of violent crimes that were classified as being motivated by prejudice based on gender (or gender identity), race, disabilities, religious beliefs, sexual orientations or ethnicity; however, only federal agency participation is mandatory and all other law enforcement agencies participate on a voluntary basis.

· Resource #3: Crime Data Explorer (CDE) at https://crime-data-explorer.app.cloud.gov/ pages/ home: This webpage is the most comprehensive of those reviewed. The CDE features visualization tools that use the above-described NBRIS data (for 20 states only), NIBRS and the FBI Summary Reporting System (SRS) data (25 states only) and SRS data only (five states). In addition, the CDE also provides downloadable big data files concerning crimes in the U.S.

· Resource #4: Law Enforcement Officers Killed and Assaulted (LEOKA) at https://www.fbi.gov/ services/cjis/ucr/leoka: Based on specific inclusion and exclusion criteria, this webpage provides information collected from participating law enforcement agencies concerning law enforcement officers that have been killed or otherwise assaulted in the United States. In addition, this webpage also features resources for law enforcement agencies only that includes a training calendar, case studies of violent encounters, in line of fire incidents and a killed in the line of duty study. The webpage is silent, though, concerning the number of law enforcement agencies that contribute data.

· Resource #5: The National Use-of-Force Data Collection (NUFDC) at https://www.fbi.gov/services/cjis/ucr/use-of-force: This information service provides national-level statistics collected voluntarily from participating law enforcement agencies concerning law enforcement use-of-force incidents and basic information about the circumstances, subjects, and officers involved. with the first data collection from law enforcement agencies beginning on January 1, 2019. This webpage is also silent concerning how many federal, state, local, and tribal law enforcement agencies contribute data to this service.

How These Sources of Information Differ in Their Approaches to Collecting Information

Do they report the sampling information? In some cases (i.e., NIBRS, HCS, CDE) some indication concerning the sampling procedures that are used is provided; in the other cases (i.e., NUFDC and LEOKA), however, the sources of information listed above either do not provide any meaningful description concerning the sampling used by the respective information collection activities.

What statistics do they present? In each case, the statistics presented by these information services focus on law enforcement incidents that involve various types of violent and nonviolent crimes committed in the United States as well as their corresponding adverse effects they have on their victims.

Do they report the sampling method used in the studies from which they obtained these statistics? To its credit, the FBI has cast a very wide net in an effort to systematically collect as much reliable data as possible from nearly 15,000 law enforcement agencies across the country. This wholesale sampling method does not attempt to apply a sample method to these data; rather, the FBI systematically collects and reports as much data as possible from mandatory (federal agencies only) and voluntary participants (all other law enforcement agencies) in its crime reporting system.

Evaluation of Sampling Methods in Terms of Representativeness and Generalizability

Prior to the introduction of sophisticated computer-based data collection and statistical analytical tools, it would have been vitally important for the FBI to apply appropriate sampling methods in order to facilitate the identification of relevant trends and geographic patterns of violent and nonviolent crimes on a national basis. Indeed, it would have been extremely difficult or even impossible for the FBI to contact tens of thousands of city, county, college and university, state, tribal, and federal law enforcement agencies on a systematic basis.

Therefore, the representativeness and generalizability of the crime data that is currently collected by the FBI by the above-described information services only relates to local, state, regional and national differences that may exist but additional research using the ancillary tools provided by the FBI is required in order to identify these specific differences. This constraint together with the others identified above are discussed further below together with a salient recommendation for the FBI to improve the ease of use and robustness of the information is provide law enforcement agencies across the country.

Recommendations for Improvement

A recurrent constraint that was identified in all of the above-described information services was the voluntary nature of participation by the country’s city, county, college and university, state and tribal law enforcement agencies. Although participation by the federal government was mandatory for the Hate Crime Statistics Program, the FBI does not provide a report concerning the specific participants that contribute to each of its information services. Here again, this information may be available by reviewing other files provided by the FBI but at present no information concerning which or how many agencies participate in these services except for the HSC, but this webpage does not break down which federal agencies participate nor their respective participation rates.