|POI Homicide Database Project|
ENTIRE ARCHIVE OF DATA SETS
THE DATABASES (Database name is linked. Downloads below.)
Youth Justice Coalition
This PDF document reports all cases – with name, age, race, location and where possible incident details – from January 1, 2007 – August 31, 2014
Supplementary Homicide Report
The SHR, Supplementary Homicide Report, is a database curated and publicly diffused by the Federal Bureau of Investigation (FBI) which contains information about particular kinds of offenses committed in 2012, defined by the FBI as: “Murders and Non-negligent Man Slaughters, including justifiable homicides”. Navigating its 152 variables and 13,063 entries, it is possible to narrow the research just to those justifiable homicides in which the victim was killed by a police officer.
The SHR data folder is comprised of 21,083,371 bytes (21.1 MB on disk) and includes 18 items. Data are presented in ten different formats, including a txt. In order for the reader to interpret the database, a codebook is provided. The codebook is 71 pages long and presents the directions to interpret the 152 columns of variables captured in the .txt file. Not all the 13,063 row possess 152 variables—on average entries are composed of approximately 35 columns. The most relevant variable for this study is the one that corresponds to characters 92-93 of the entries and refers to the “homicide circumstances,” with 81 as the code for “Felon Killed by Police Officer.” Other variables are associated with geographical and temporal information about the homicide, plus the age, sex, race, ethnicity of the victim, plus the weapon used and the relation of offender to the victim. All 50 US States are represented in the database, but not every agency in each state reports.
Once the database is imported in Excel and counting techniques for multiple criteria are applied, a cross search for justifiable homicides committed in the LA County by police officers can be generated. The result of this endeavor showed that there were 33 cases of officer-involved homicides in 2012. The data is of particular interest when compared to the same data provided by the LA Times Homicide Report, as will be discussed in the conclusion. The next step of our research will focus on the identification of the cases that are reported by the LA Times and not reported in the FBI database.
Wide-ranging Online Data for Epidemiologic Research
Sponsoring Organization—Center for Disease Control (CDC)
Source Type—Government CDC
The CDC provides information regarding causes of death in its Wide-ranging Online Data for Epidemiologic Research (WONDER) database. This data is collected by the National center for Health Statistics and provided to the WONDER database. Data are based on death certificates that identify the cause of death and demographic data for United States residents. National, state and county mortality data by underlying cause and up to twenty multiple causes of death are available for each death certificate. Data are available by age group, race, and sex. Cause of death is specified by ICD 10 codes. Data are updated annually.
On the CDC WONDER database search system, there are a number of fields to choose from in order to get the desired results. Its Multiple Cause of Death (MCD) database is made available through efforts of is based on a comprehensive collection of all death certificates issued nationwide. Because of the details in these certificates, the CDC WONDER database can often provide specific demographic information alongside cause of death.
One can limit data to California, then one can access any county within the state of California, including Los Angeles County. 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. The death records are typically reported by coroners and medical examiner's offices.
When one receives results in CDC WONDER, one may view them online and then, export the results onto personal computer so that you can load them into another program. One can export to a tab-separated file that can be loaded into a spreadsheet or a statistical analysis program, Microsoft Excel, for example. There are a number of formats to choose from when exporting results. One can select to get data for rates per capita at the 95% level of statistical significance.
Los Angeles Times Homicide Report
Abbreviation--LA Times Homicide Report
Sponsoring Organization--LA Times
This site documents and maps homicides reported by the LA Times in LA county. One can filter according to a number of parameters, including “only officer involved” to show those that were perpetrated by cops. The LA Times hosts The Homicide Report as a part of its interactive data journalism component, the LA Times Data Desk. This site documents and maps homicides reported biweekly to the LA Times by the coroner’s office of Los Angeles county. One can filter according to a number of parameters, including “only officer involved” to show those that were perpetrated by cops. An interactive viewer can see where these incidents occurred on a map for which the viewer can zoom in or out for clearer location information. More importantly, each officer-involved homicide is accompanied by a news story that gives pertinent information about events surrounding the homicide. Los Angeles is one of the only cities in the United States with such information available to the public, as a result of the efforts of the LA Times’ Data Desk. Thus, it serves as a good case study. If, in a city with the best documentation on officer-involved homicides, reports still show discrepancies and we can find reasons for these discrepancies, it is safe to assume that there are other cities for which this data would be crucially necessary. We are also working with the LA Times to get their data into a downloadable format that can be made available to the public.
Arrest Related Deaths
Sponsoring Organization—Bureau of Justice
Source Type—Government The Bureau of Justice’s Arrest Related Deaths (ARD) Statistical Tables
ARD provide state-level statistics from 2003-2009 on the circumstances of deaths that occur during, or shortly after, state or local law enforcement officers engage in an arrest or restraint process. In BJS’s interpretation, arrest related deaths include any time “a person is in the custody or under the restraint of law enforcement officers” – whether or not a formal arrest process was intended. Federal law enforcement were excluded from the tables.
State Reporting Coordinators (SRC) – often from a state justice commission or governor’s office – worked with BJS to collect this information, which they submit quarterly or annually. SRCs may use any methodology to determine arrest-related death occurrences – they can consult police departments, county medical examiners or coroners offices, or extract data from the UCR or CDC’s violent death reporting system. Media sources were not considered official information sources, though SRCs might consult them if official sources presented incomplete data. Data collected include date of death, manner and cause of death, the decedent's demographic characteristics, the decedent's behavior during the events leading up to the death, and the tactics and weapons used by law enforcement personnel.
The data fall short for understanding local patterns because the data remain in aggregate form, broken down by state by not county. Furthermore, SRCs report voluntarily – 47 states participated between 2003-06, and in CA and TX reporting was mandatory, but 16 states didn’t report at some point from 2003-2009, and three, Georgia, Maryland and Montana, never submitted.
National Archive of Criminal Justice Data
Sponsoring Organization—National Institute of Justice
The NIJ, or National Institute of Justice, is an organization dedicated to the leveraging and support of criminal justice and forensic science programs, technologies, and research. It is separated into six branches: the Office of Research and Evaluation, Office of Science and Technology, Office of Investigative and Forensic Science, Office of Research Partnerships, Office of Operations, and Office of Communications. Data from funded research projects are made available under the Data Resources Program through the National Archive of Criminal Justice Data. Registration is required for access to the materials, but is easily completed in a few minutes. After registering on the NACJD site, a user is given access to all available documents from past research online, and has the opportunity to personally download material. One’s ease of access may be affected by organizational or institutional affiliation.
The data provided in the National Archive of Criminal Justice Data is original data, datasets from various projects that can be chosen from a provided codebook and manipulated through statistical procedures and other methods. A selected portion of data is provided for these purposes in the Survey Documentation and Analysis portion of the site, while search functions for comparing variables across datasets is available in a separate Social Science Variables Dataset.
Searching ‘police homicides,’ as variables in the Social Science Variables Dataset, provided 920 results with most of the results on the first two pages being of significant relevance. Several datasets listed ‘Additional Justifiable Homicide Circumstances’ were available, with each listing relating to the given data in a year between 1991 and 2009. The information provided were numbers of incidents recorded involving a variety of variables:
Criminal attacked police officer, killed by that officer
Criminal attacked police officer, killed by other police officer
Criminal attempted flight from a crime
Criminal killed in commission of crime
Criminal resisted arrest
Unable to determine/not enough information
These variables were all taken from the National Incident-Based Reporting System of each year, as conducted by the FBI and United States Department of Justice. The National Incident-Based Reporting System is defined in the Codebook as an incident-based reporting system for crimes known to the police.
The information available for download includes a Codebook describing the Incident-Based Reporting System (description, general procedures, codes for recording incidents), a bibliographic document with related literature, and a spreadsheet including every crime incident reported that year, the city it occurred in, the incident number and categorization, and a few other number columns that do not clearly match up to a corresponding phrase in the codebook. The variables related to police homicides found in the Social Science Variables Dataset represent a very small percentage of the incidents listed in the entire FBI dataset. For example, the 1994 police homicides variable chart lists 14 police homicides of various categorization (categories as noted above) out of a total of 953,063 crime incidents recorded that year. As of now, a specific code for police homicides does not seem to be explicitly provided, though it was able to be teased out in the variables dataset.