U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics Special Report Arrest-Related Deaths in the United States, 2003-2005 October 2007, NCJ 219534 ---------------------------------------------------------- This file is text only without graphics and many of the tables. A Zip archive of the tables in this report in spreadsheet format (.csv) and the full report including tables and graphics in .pdf format are available from: http://www.ojp.usdoj.gov/bjs/abstract/ardus05.htm ----------------------------------------------------------- By Christopher J. Mumola BJS Policy Analyst During the 3-year period from 2003 to 2005, 47 States and the District of Columbia reported 2,002 arrest-related deaths to the Bureau of Justice Statistics' (BJS) Deaths in Custody Reporting Program (DCRP). The DCRP counts represent the first national measurement of all types of deaths that occurred in the process of arrest. The collection of these counts was mandated by the federal Death in Custody Reporting Act (Public Law 106-297). All States were required under the act to submit a record of any death that occurred "in the process of arrest" as a condition of receiving federal correctional grants. The DCRP collection was the first time most States made a comprehensive count of these deaths. California and Texas were the only States that compiled data on all arrest-related deaths before the collection began. Three States -- Georgia, Maryland, and Montana -- never reported counts of arrest-related deaths to DCRP. Federal agencies were not required by the act to report these deaths. BJS worked with States to determine a contact to report information on arrest-related deaths and to develop a questionnaire to collect uniform data from all States. The governor's crime commission (22 States) was the most common reporting office, followed by the State attorney general and the State police (8 States each). State contacts relied on a wide range of data sources to complete the DCRP questionnaire. Forty-two of the 47 States that reported data used more than 1 source, and 30 States used at least 3 data sources. The most common source of information was the law enforcement agency involved in the death (43 States). Homicides by law enforcement officers made up 55% of all deaths during arrests by State and local agencies In each year between 2003 and 2005, homicides accounted for a majority of all reported arrest-related deaths. During this period, States reported 1,106 arrest-related homicides which represented 55% of arrest-related deaths from all causes. Homicides by law enforcement officers accounted for 1,095 arrest-related deaths, and 11 homicides were committed by other persons present at the scene. Homicide was reported four times more often than any other cause of death. Drug and alcohol intoxication accounted for 13% of all deaths, followed by suicides (12%), accidental injuries (7%), and illness or natural causes (6%). For 157 deaths (8%) a definitive cause was not reported. FBI and BJS counts of homicides by law enforcement were similar in most States Homicides by law enforcement are the only type of arrest-related deaths measured by another national statistical program--the Supplementary Homicide Reports (SHR) collected by the Federal Bureau of Investigation (FBI). The SHR and DCRP measures of homicides by law enforcement have one important difference. The SHR program includes only counts of homicides by law enforcement in which the use of force was ruled "justifiable." Deaths due to unjustified use of lethal force by officers are counted with other murders. The DCRP counts of homicides by law enforcement include all deaths that resulted from the use of lethal force. In most States, the SHR and DCRP counts of law enforcement homicides showed small differences. In 33 States, the 2 measures differed by fewer than 10 deaths over the 3 years. Nine States reported counts that differed by at least 20 deaths over this period, with 5 States reporting higher counts to DCRP. California had the largest measurement variation, reporting 354 homicides by law enforcement to SHR between 2003 and 2005, but less than half as many (160) to DCRP. Over the same period, Florida reported 98 law enforcement homicides to DCRP but did not report to SHR. In the aggregate, the two programs collected very similar counts between 2003 and 2005. A total of 1,095 law enforcement homicides were reported to DCRP, and 1,082 justifiable homicides by police were reported to SHR. Taking the higher count reported by each State for each year, there were a total of 1,489 reported law enforcement homicides. BJS and FBI law enforcement homicides had similar characteristics Analysis of the characteristics of the deaths in the two data collections also revealed similarities. Across both programs nearly all of the decedents were male, the average age was 33 years old, and over 80% were killed by a handgun. Arrests for violent crimes were involved in 3 of every 4 law enforcement homicides Three-quarters of the law enforcement homicides reported to DCRP involved arrests for a violent crime. Except for suicides (51%), violent offenders were involved in less than 30% of all other causes of death. Public-order offenders accounted for 8% of homicides, followed by property (4%) and drug offenders (2%). In 2% of cases, law enforcement personnel did not intend to charge the person, but took them into custody for reasons such as medical or mental health needs. Criminal charges were not reported for 8% of law enforcement homicides. Among all persons killed by law enforcement officers in the process of arrest, 9% would have been charged with the murder or attempted murder of a law enforcement officer, 17% would have been arrested for assaulting an officer, and 2% would have been charged with obstruction of police activity or resisting arrest. 80% of law enforcement homicides involved the use of a weapon by the arrest subject, 96% involved the use of a firearm by officers State respondents reported a number of circumstances surrounding each arrest-related death, including a variety of behaviors by the arrest subjects and officers. In 80% of the law enforcement homicides, the deceased reportedly used a weapon "to threaten or assault" the arresting officers. In 17% of the homicides, the arrest subject grabbed, hit or fought with the arresting officers. Few homicides by law enforcement involved persons who appeared intoxicated during the attempted arrest (18%). A third (36%) of persons killed by officers attempted to escape or flee from custody. About 4% of persons killed by police had been placed under physical restraints during the attempted arrest. Nearly all law enforcement homicides (96%) involved the use of firearms by officers. Handguns were used by officers in 84% of the homicides, and rifles and shotguns in 17%. Three deaths were caused by an officer's use of a nightstick, baton, or conducted-energy device such as a taser or stun gun. No weapons were reportedly used in 3% of law enforcement homicides. ------------------------------------------------------------- During 2003-2005, 380 law enforcement officers were killed, and nearly 175,000 assaulted According to the FBI's Law Enforcement Officers Killed and Assaulted (LEOKA) data, 380 law enforcement officers nationwide were killed in the line of duty between 2003 and 2005. Fewer than half (159) of these deaths were homicides. Accidental deaths during arrests (221) accounted for the majority of officer deaths in the line of duty. Analysis of DCRP records indicates that assaults on officers and the use of weapons by arrest subjects were common circumstances in homicides by law enforcement officers. However, FBI data suggest that assaults on law enforcement officers do not lead to arrest fatalities in most cases. The number of persons killed by officers in the process of arrest from 2003 to 2005 (1,095) was less than 1% of all reported assaults on law enforcement officers (174,760) over the same period. ------------------------------------------------------------ Half of all homicides by law enforcement involved large departments with at least 500 sworn officers According to the 2004 Census of State and Local Law Enforcement Agencies, 168 of the Nation's 17,876 law enforcement agencies employed at least 500 full-time sworn officers. These large law enforcement agencies employed 39% of all sworn officers, but were involved in half (50%) of all homicides by law enforcement officers reported to DCRP during 2003 through 2005. Most arrest-related homicides involved local police departments (74%), followed by county sheriffs' offices (19%) and State police or highway patrol agencies (6%). Special jurisdiction agencies, such as campus police, transit police authorities or park police departments, were involved in 1% of such homicides. Deaths from alcohol or drug intoxication were least likely to involve arrests for violent crimes After homicide, deaths from alcohol or drug intoxication were the second most common type of arrest-related deaths (13%).***Footnote 1: Excludes cases in which an arrest subject swallowed packets of drugs for concealment purposes (classified as "accidental injury").*** Intoxication deaths differed from other arrest-related deaths on several characteristics: *Arrest subjects under age 25 made up a smaller share of intoxication deaths (13%) than homicide (28%), accidental (25%), or suicide (20%) deaths. *Whites were involved in 33% of intoxication deaths, the lowest percentage of whites among all causes of death. *Intoxication deaths were less likely to involve violent offenders (28%) than any other cause of death. Despite the low percentage of violent offenders overall, assault (23%) was the most common criminal offense involved in intoxication deaths. Persons arrested for substance abuse-related offenses - such as drug law violations (19%), disorderly conduct including drunkenness (12%), and driving while intoxicated (1%) - combined for a third of all intoxication deaths. Persons with no criminal charges made up a larger share of intoxication deaths (9%) than any other cause. (See Appendix tables 4 and 6 for details on characteristics and offenses by cause of death.) In half of intoxication deaths at booking centers or police stations, arrestees were not visibly intoxicated Of the 252 arrest-related intoxication deaths, 198 occurred in the field, prior to booking the arrest subject. In 75% of the pre-booking intoxication deaths, the deceased appeared intoxicated to the arresting officers. Over 80% of the pre-booking intoxication deaths took place either at a medical facility or en route to a medical facility, indicating that some of the subjects who did not appear intoxicated were still provided medical care for other reasons. Among the 50 intoxication deaths that took place at a police station or booking center, half (51%) were visibly intoxicated at the time of booking. The percentages who were booked with visible medical (18%) or mental health problems (2%) were much smaller. Overall, in a third of cases where arrest subjects died from intoxication after booking, the deceased showed no apparent signs of intoxication, mental health problems, or medical problems at time of the booking. (See appendix table 8 for details.) In 4 cases, the location of the intoxication death was not reported. In 62% of intoxication deaths, the deceased resisted arrest, without force or weapons in most cases Among intoxication deaths that took place in the field prior to booking, 62% of the persons resisted arrest. According to several measures, these actions were usually less violent and involved weapons less often than in deaths due to homicides by law enforcement. ----------------------------------------------------------- Increasing number of arrest-related deaths involved the use of tasers or other conducted-energy devices Conducted-energy devices (CEDs), such as stun guns or tasers, were involved in 36 arrest-related deaths reported to DCRP during 2003 through 2005. In about half of these deaths (17), the CED was reported as the cause of death. In the remaining 19 deaths, the use of a CED was indicated, but it was not reported as the cause of the death. *** Among medical and law enforcement experts, the ability of CEDs to cause a death is a subject of debate. Due to reporting gaps these 36 cases do not represent a complete count of all deaths in which the use of a CED was involved.*** The involvement of CEDs in arrest-related deaths increased from 3 deaths in 2003 to 24 in 2005. Every type of arrest-related death was reported among the 36 deaths involving the use of CEDs. The most common cause of death was intoxication (10), followed by accidents (8), and homicides by law enforcement (7). One death was attributed to illness and another death was a suicide. In 9 cases where CED-use was reported, a cause of death could not be determined. Arrests for violent crimes were involved in 16 of the CED- involved deaths, and 8 deaths involved property crimes. In two cases, the deceased was detained for mental health care, and in five cases no information on criminal offense was reported. ----------------------------------------------------------- Higher percentages of violent offenders, whites involved in arrest-related suicides Suicide accounted for nearly as many arrest-related deaths (234) as intoxication (252). Reports of suicides also increased 63% from 2003 (56) to 2005 (91). Other than homicides (74%), suicides were the only type of arrest-related death to involve arrests for violent crimes in most cases (51%). Suicide was the only type of death in which a majority (57%) of the decedents were white. This finding is consistent with the higher suicide rates observed for white prison and jail inmates in other DCRP collections (see Suicide and Homicide in State Prisons and Local Jails ). Two-thirds (67%) of arrest-related suicides occurred at the scene of the attempted arrest, while the remainder took place at a police station or booking facility. In terms of their personal characteristics, the arrest subjects who committed suicide in these two settings showed many similarities. Arrest-related suicides in the field differed sharply from those in police stations or booking centers The circumstances of the suicides showed sharp differences based on the setting of the events. Suicides that took place at the scene of the attempted arrest were more likely to involve large law enforcement agencies and typically involved firearms. Suicides that took place at a police station or in a booking center were more likely to involve small law enforcement agencies and almost exclusively involved hanging. Suicides that took place after booking also involved a higher percentage of arrest subjects who were visibly intoxicated. Methodology The Death in Custody Reporting Act of 2000 (Public Law 106-297) required the U.S. Department of Justice to begin a quarterly collection of individual death records for all persons incarcerated in State or local correctional facilities, as well as "any person who is in the process of arrest." The act linked eligibility for funds under the Violent Offender Incarceration and Truth In Sentencing (VOI/TIS) grant programs to reporting death records to the Attorney General. At the time the law was enacted, all 50 States and the District of Columbia participated in the VOI/TIS grant programs. The act did not apply to Federal law enforcement agencies or the Federal Bureau of Prisons. BJS was tasked with implementing the act and phased in data collection activity over a 4-year period. Collection of death records from local jail facilities began in 2000, followed by a separate collection from State prison authorities in 2001. In 2002 BJS began collecting death records from all State juvenile correctional systems. The collection of arrest-related records began in 2003. These four data collections make up the Deaths in Custody Reporting Program (DCRP). Identifying State-level data providers In developing the collection of arrest-related death records, BJS staff contacted multiple offices in each State to determine the appropriate reporting contact. While each State was required to report these death records as a condition of eligibility for VOI/TIS grant funds, the grants were awarded to State departments of correction. In developing the collection, BJS did not find any State correctional authorities that collected information on the operations of law enforcement agencies in their State. At the time the Death in Custody Reporting Act was passed, only two States (California and Texas) collected information on all types of arrest-related deaths. For the remaining 48 States and the District of Columbia, the new DCRP collection was the first attempt to perform a comprehensive count of all arrest-related deaths. In California and Texas, State statutes required the reporting of all arrest-related deaths to the State Attorney General's office. These offices agreed to complete statewide reports of arrest-related deaths for submission to BJS. In all other jurisdictions, BJS worked with State officials to determine which agency would collect arrest-related death reports. A State criminal justice commission, commonly administered by the governor's office, was the most common data reporting contact (22 States), followed by the State attorney general and State police department (8 States each). In five States, the department of corrections took a lead role in compiling records because of the VOI/TIS grant funding. In over 30 States, the reporting office also served as a State criminal justice Statistical Analysis Center (SAC). While every State identified a data provider, three States -- Georgia, Maryland, and Montana - never submitted records. Arkansas, Louisiana, Mississippi, and Tennessee began participating in 2004, and Nevada and Wyoming stopped participating in 2004. A total of 43 States participated in 2003, and 45 States participated in both 2004 and 2005. The District of Columbia submitted records in all 3 years. Many of the arrest-related deaths undergo lengthy investigations by prosecutors, police departments, and coroner's offices to determine a cause of death or decide a legal disposition of the case. It took up to 15 months to finalize cause of death information for some cases. Multiple contacts were made to the State data providers to complete missing items and receive a final cause of death. Data sources for State reporters Regardless of which State office ultimately took the lead in compiling the death records, most States utilized multiple data sources in this effort. Of the 47 States that participated at some point, 42 used multiple sources of information on arrest-related deaths, while 30 states reported at least 3 different sources of data. (See Appendix table 1 for a detailed listing of reporting methods used by each State.) State and local law enforcement agencies were the most common source of data used by State reporters. Law enforcement agencies voluntarily reported arrest-related deaths to 40 data providers. In California and Texas, law enforcement agencies were required by State law to submit arrest-related death records. The Metropolitan Police Department was the only source of arrest-related data for the District of Columbia. Thirty States used media searches to identify arrest-related deaths and followed up with a data request to the law enforcement agency involved in the incident. In some cases when the agency did not respond to this information request, the State contact used the media accounts of the death to complete the DCRP questionnaire. Twenty-three States involved county coroners or the State medical examiner's office in compiling these records, and 19 States involved the State police. Nine States used their Uniform Crime Reporting office, and 6 States collected information from prosecutors' offices. Seven States used other resources or agencies in compiling these reports. Defining deaths "in the process of arrest" BJS had to define the term "in the process of arrest," specified in the Death in Custody Reporting Act (PL 106-297). BJS staff consulted with the International Association of Chiefs of Police (IACP), the National Sheriffs' Association (NSA), and criminal justice researchers to identify which circumstances involved an "arrest process." All deaths of persons in the physical custody or under the physical restraint of law enforcement officers were included. This resulted in the reporting of 75 deaths over three years in which no criminal charges were involved. Law enforcement responses to people exhibiting mental health problems accounted for 44 of these cases, while another 9 cases involved persons who had to be restrained by police for medical transportation. In another 22 cases, the reason for law enforcement involvement was not specified, but the record indicated that no criminal charges were involved. The deaths of any other persons not subject to an attempted arrest were excluded, including bystanders and law enforcement officers killed during an attempted arrest. State contacts were instructed to include all deaths resulting from use of force by law enforcement officers. Arrest-related suicides were also included in this collection, provided that law enforcement officers were in some type of contact with the arrest subject prior to the suicide. For example, if an armed suspect was surrounded by officers and chose to take his own life rather than surrender, the death would be included. However, if an offender was actively sought by police but committed suicide before the police located him, the death would be excluded. The reason for the exclusion is that no officers were present at the time of death to attempt an arrest. Vehicular accident deaths that were not specifically related to arrest activities were excluded from the collection. States were instructed to include vehicular accident deaths only when law enforcement officers actively took some role in causing the accident during an arrest attempt. This included shooting at the vehicle or driver or forcing the vehicle off the road with a police vehicle or other obstructions (such as a spike strip to blow out tires or a roadblock). All other vehicular deaths were excluded. States were also instructed to disregard whether an arrest warrant had been issued. Because officers frequently make arrests in response to unexpected events, requiring an arrest warrant would leave many arrest-related deaths unreported. Likewise, States were told to exclude the deaths of persons who had arrest warrants issued against them that went unenforced. For example, if an offender had a bench warrant issued for their arrest, but later died before any officers attempted to enforce this arrest warrant, the State was told to exclude that record. In such cases, the arrest warrant indicated an administrative criminal justice status and not an attempt to bring the subject into custody. Deaths of arrestees were subject to the data collection from the time police encountered them in the field until the time they were booked into a local jail facility. This included deaths of arrest subjects who died at medical facilities due to injuries or medical problems, as well as any persons who died in transit from an arrest scene in a police vehicle or ambulance. All deaths in jails are reported to BJS under a separate DCRP collection with different questionnaires. Once records of arrest-related deaths were submitted to BJS, the forms were reviewed to ensure that each case met the established guidelines. Deaths were checked against the DCRP database of jail facility deaths for the same year to avoid double-counting. BJS staff and the State contacts routinely discussed and resolved cases that were ambiguous or appeared to involve circumstances that would exclude them from the collection. Arrest-related deaths from the Supplementary Homicide Reports While DCRP is the only national statistical program that measures all types of arrest-related deaths, two other national programs measure law enforcement homicides. Law enforcement agencies can submit Supplementary Homicide Reports (SHR) as part of the FBI's Uniform Crime Reporting (UCR) program. The Centers for Disease Control and Prevention's National Center for Health Statistics (NCHS) also compiles mortality statistics, including a category for law enforcement homicides. The SHR records include basic data on the type of homicide, the relationship between the deceased and assailant, and demographic characteristics of the deceased. Law enforcement agencies describe the event as a "justifiable homicide by police." During the 3 years covered by this report, Florida and the District of Columbia did not report SHR data to the FBI. In other States, an SHR record should have been filed by police for all justifiable homicides. It is unknown what percentage of justifiable homicides by police were actually submitted to SHR. For a detailed discussion of SHR reporting procedures regarding justifiable homicides by police, see the Methodology section of Policing and Homicide, 1976-98: Justifiable Homicide by Police, Police Officers Murdered by Felons . A major difference between the SHR and DCRP counts of law enforcement homicides is the source of the data. SHR records are reported by the law enforcement agencies involved in the deaths while DCRP records are compiled by State data reporters. State law enforcement agencies were involved in 6% of law enforcement homicide cases reported to DCRP. In the remaining 94% of law enforcement homicides, State data reporters provided data on local law enforcement agencies (see appendix table 7). A comparison of the SHR and DCRP counts of law enforcement homicides between 2003 and 2005 is discussed on page 2. Data for each State by type of data collection are presented in appendix table 3. Arrest-related deaths from the National Center for Health Statistics A national measure of law enforcement homicides is also contained in mortality statistics collected by the Centers for Disease Control and Prevention's National Center for Health Statistics (NCHS). In collecting death certificates for all deaths in the United States, NCHS classifies causes of death according to the International Classification of Disease, 10th revision (ICD-10). Under the ICD-10 system, deaths can be classified as "deaths by legal intervention" (or code Y35). These records include only deaths by police use of force and mandated executions in prisons. None of the other forms of arrest-related deaths are included. The NCHS death records are typically reported by coroners and medical examiner's offices. These data providers may not always use the specific ICD-10 code to designate law enforcement homicides separately from other homicide cases. The respondents may also not know if the deceased was involved in an attempted arrest at the time of death. Final NCHS' mortality data on legal intervention deaths were not available for 2005. For 2003 and 2004, NCHS counts of legal intervention deaths by police (679 nationwide) were lower than the counts of law enforcement homicides reported to DCRP (731) and justifiable homicides by police reported to SHR (739). NCHS counts of legal intervention deaths are available at . ------------------------------------------------------------ The Bureau of Justice Statistics is the statistical agency of the U.S. Department of Justice. Jeffrey L. Sedgwick is director. This Special Report was written by Christopher J. Mumola. Margaret E. Noonan verified the report and prepared data tables. Data collection and processing of death records were carried out by Lara E. Allen and Margaret E. Noonan. Erica L. Smith analyzed data from the FBI's Supplementary Homicide Reports. Carolyn C. Williams of BJS produced and edited the report. Jayne Robinson prepared the report for final printing, under the supervision of Doris J. James. October 2007, NCJ 219534 ------------------------------------------------------------ This report in portable document format and in ASCII and its related statistical data and tables are available at the BJS World Wide Web Internet site: . ------------------------------------------------------------ End of file 10/3/2007 JR