Methods

Overview

The data presented in this application come from the Federal Bureau of Investigation's (FBI) National Incident-Based Reporting System (NIBRS). NIBRS contains detailed information on crimes reported to law enforcement, including the nature and types of specific offenses in the incident, characteristics of the victim(s) and offender(s), and characteristics of persons arrested in connection with the crime incident. For a crime to be included the incident need only be reported; it does not have to be cleared or result in an arrest. In 2016, data from a sample of law enforcement agencies in 38 states and the District of Columbia were reported to the FBI through NIBRS.

In 2016, there were at least some law enforcement agencies in 38 states and the District of Columbia that did not report their crime and arrest data to the NIBRS effort. While it is difficult to quantify the actual level of reporting, it is fair to say that in some States the level of reporting was over 75% while in other states it was far less. This places several obvious limitations on any interpretation of NIBRS analysis. First, any state-specific analysis must not be considered an analysis of all incidents reported to law enforcement in the State, given that an unknown number of agencies did not report their data to NIBRS. Second, it cannot be assumed that crime data from agencies that reported to NIBRS have the same characteristics as those that did not report - implying that the patterns found in the reported data are not necessarily representative of the entire state. And finally, given that the data from each State do not reflect all crimes reported to law enforcement and that those that are reported to NIBRS might not have the same attributes as the universe of crimes reported to law enforcement in the State, state-to-state comparisons are not recommended.

Data source

The Inter-University Consortium of Political and Social Research (ICPSR) at the University of Michigan prepared the data underlying this application. ICPSR restructured the original NIBRS data files provided by the FBI into a series of "extracts" intended to facilitate access to these data. Working with the "raw" NIBRS data collected by the FBI requires considerable data manipulation skills that are distinct from data analysis. This is particularly true when working with multiple NIBRS segments (record types). As such, ICPSR restructured the various segments of the original hierarchical NIBRS data into set of research extracts to simplify analysis.

The following is a description of how ICPSR created the research extracts. [This material is adapted from the "National Incident Based Reporting System, 2013: Extract Files" codebook prepared by ICPSR, available from the National Archive of Criminal Justice Data website.]

NIBRS data as formatted by the FBI are stored in a single file, organized by various segment levels (record types). There are six main segment levels: administrative, offense, property, victim, offender, and arrestee. Each segment level has a different length and layout. There are other segment levels that occur with less frequency than the six main levels. Significant computing resources are necessary to work with the data in its single-file format. In addition, the user must be sophisticated in working with data in complex file types. For these reasons and the desire to facilitate the use of NIBRS data, ICPSR created the extract files.

ICPSR merged, restructured, and subsetted the original FBI data as follows. Four data files were created whose units of analysis are: incident, victim, offender, and arrestee. For each of the four files, all of the related NIBRS segment levels (record types) were merged onto a single record. The incident, victim, and offender files were subsetted based on the incident date occurring in 2016. The arrestee file was subsetted based on the arrest date occurring in 2016, regardless of incident date.

For a given crime incident, NIBRS records up to 10 offenses, 999 victims, 99 offenders, and 99 arrestees. In creating the extract files, the number of records from each of the various segment levels that were merged onto a record was limited to three. For example, in creating the victim-level file up to three records from each of the offense, property, offender, and arrestee segments were merged with the victim record. When there were four or more related records, the fourth through N records were excluded. However, all the victim records with crime incidents in 2016 were included. The exclusion of records does not apply to records that correspond to the unit of analysis. Approximately 99% of NIBRS incidents have three or fewer offense, victim, offender, and arrestee records.

Citation for the 2016 Extracts:
National Archive of Criminal Justice Data. National Incident-Based Reporting System, 2016: Extract Files. ICPSR36851-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2018-08-17. https://doi.org/10.3886/ICPSR37066.v1

While the National Center for Juvenile Justice (NCJJ) has benefited greatly from the work of others, any errors in this analysis and data presentation package are the responsibility of NCJJ. This package has been made available to the public so that interested users can explore the characteristics of victims of violence. NCJJ assumes no liability for any alleged or actual damages arising from use of this application.

Creating the sample

The 2016 Victim-level extract prepared by ICPSR includes information on nearly 6 million victims reported by law enforcement agencies from 38 states and the District of Columbia with an incident date occurring in year 2016. With this as the starting point, NCJJ created the data files used in this application as follows.

In developing this data analysis tool, NCJJ prepared two separate data extracts. The first extract represents all victims of violence reported by participating law enforcement agencies, while the second extract contains only victims of domestic violence. The primary distinction between these two data files is how we identified victims of domestic violence. For both extracts, the data were limited to only records in which the age of the offender was at least 7 years or "missing". A detailed explanation of the procedures used by NCJJ to create the data files used in this application can be found below.

Victims of Violence

NCJJ modified the contents of the 2015 Victim-level extract prepared by ICPSR by keeping only those records that involved a victim of violence. The data file was limited to only those victim records where the most serious offense involved the use or threat of force against the victim. Because the FBI provides no inherent offense hierarchy in the NIBRS file structure, NCJJ implemented the following offense hierarchy when selecting the most serious person offense against the victim: murder/nonnegligent manslaughter, kidnapping/abduction, forcible rape, forcible sodomy, sexual assault with an object, forcible fondling, robbery, aggravated assault, simple assault, and intimidation (definitions of each offense can be found in the glossary).

Finally, states that reported fewer than 1,000 domestic violence victims in 2016 were eliminated from the final data file (Alabama and Minnesota were the only states to meet this threshold). The resulting data file included more than 1.4 million victims of violence reported by more than 6,100 law enforcement agencies in 36 states and the District of Columbia in 2016. Note that NIBRS data do not track individuals across incidents; a victim who was involved in more than one incident will have multiple records.

Victims of Domestic Violence

In creating the victims of domestic violence extract, we first identified all victim records that included any of the following victim-offender relationships: Spouse, Parent/Step-parent, Sibling/Step-sibling, Child/Step-child, Grandparent, Grandchild, In-law, Other family member, Boyfriend/Girlfriend, Child of boyfriend/girlfriend, Homosexual relationship, and Ex-spouse. These relationship codes form the basis of our identification of a domestic violence victim.

Once all domestic violence victims were identified, the data file was further limited to only those victim records where the most serious offense involved the use or threat of force against the victim. As with the victims of violence extract, we implemented the following offense hierarchy when selecting the most serious person offense against the victim: murder/nonnegligent manslaughter, kidnapping/abduction, forcible rape, forcible sodomy, sexual assault with an object, forcible fondling, robbery, aggravated assault, simple assault, and intimidation (definitions of each offense can be found in the glossary).

Finally, states that reported fewer than 1,000 domestic violence victims in 2016 were eliminated from the final data file. This included Alabama, District of Columbia, and Indiana. The resulting data file included 620,043 victims of domestic violence reported by more than 5,600 law enforcement agencies in 34 states in 2016. Note that NIBRS data do not track individuals across incidents; a victim who was involved in more than one incident will have multiple records.

Offender information

The offender characteristics presented in this application vary according to the data extract. For both extracts, the records were limited to those with an offender at least 7 years of age or "missing".

Victims of Violence

In addition to describing victims of violence, this application provides demographic information (age, sex, race) on the first listed offender. The nature of the NIBRS data allow for multiple offenders to be involved in a single incident. However, this application includes only the demographic information of the first offender. Consider a situation where a woman is assaulted by her neighbor and her son, and the offenders are listed in this order in the Victim-level data file (i.e., neighbor listed first, followed by the son). In this example, the neighbor is the first offender, so their demographic information is retained for analysis purposes under the section labeled "Characteristics of First Offender" on the Victims of Violence tab. In 2015, 89% of all victims of violence reported via NIBRS involved a lone offender.

Victims of Domestic Violence

The approach to creating the victims of domestic violence extract is the same as above with the exception of looking for specific victim-offender relationships. For this extract, we took demographic information (age, sex, race) on the first listed offender with a domestic relationship to the victim. While multiple offenders can be involved in a single incident, this application includes only the demographic information of the first offender with a domestic/intimate relationship to the victim. Consider a situation where a woman is assaulted by her husband and her son, and the offenders are listed in this order in the Victim-level data file (i.e., husband listed first, followed by the son). In this example, the husband is the first offender with a domestic relationship to the victim, so his demographic information is retained for analysis purposes under the section labeled "Characteristics of First Offender With a Domestic Relationship" on the Victims of Domestic Violence tab. In 2015, 94% of all victims of domestic violence reported via NIBRS involved a lone domestic offender.