Hate Crime Investigations and Offender Profiles: A National Survey of U.S. Law Enforcement Agencies, United States, 2018 (ICPSR 38262)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

The purpose of this study was to gather nationally representative data on hate crime incidents by employing an LEA survey methodology developed by the University of New Hampshire (UNH) research center:

This project sought to address four primary research questions:

1. What are the rates of hate crimes coming to the attention of the U.S. criminal justice system, and what percentage of hate crimes reported to law enforcement end in arrest? How do rates using this study's methodology compare to data on hate crimes currently being compiled and reported by the FBI? How do rates vary by agency type and size?
2. Which agency-level policies in place across the country affect how decisions are made by law enforcement officers to identify and document bias crimes, and how do those policies vary by region and agency size?
3. Using a representative sample of hate crime incidents known to police collected from across the U.S., what can be learned about the nature of hate crimes being investigated by police? What offender typologies can be identified, and how do they vary by offender and incident characteristics and case outcomes? Who are the victims in these cases? What features mark youth-involved hate crime investigations?
4. What types of investigative practices are used to document bias in hate crime investigations, and which are associated with higher arrest and prosecution rates?

Study Design View help for Study Design

The NHCIS was designed to yield a nationally representative sample of hate crime cases using a five-frame stratified sample of LEAs (randomly selected by size and agency type, since hate crime investigations do not occur with equal probability among the more than 15,000 U.S. LEAs). The agency sample was drawn using the Bureau of Justice Statistics (BJS) Directory of Law Enforcement Agencies, a database available through the National Directory of Criminal Justice Data. This database includes an annually updated census of local, county, and state LEAs in the U.S. and provides geographic and other identifying information for each agency.

Agency-level survey procedures

Data were gathered using a multi-stage methodology:

1. Researchers used first class mail to send surveys, personalized cover letters, and business reply envelopes to the heads of local, county, and state LEAs in the sample.
2. Approximately three weeks after the initial mailing, researchers sent reminder postcards to all agency heads, asking them to complete and return the survey if they had not done so, and thanking them if they had.
3. Approximately six weeks after the initial mailing, researchers sent additional copies of the survey, personalized cover letters, and business reply envelopes to the heads of agencies who had not responded to date.
4. Due to COVID restrictions on bulk mailings that occurred during this time, instead of a third mailing, researchers followed up by telephone to collect agency-level survey data from non-responding agencies.

Case-level survey procedures

Case-level survey data were collected through telephone interviews using a Computer-Assisted Telephone Interviewing (CATI) system for 622 cases, and by abstracting data from case records provided by LEAs in 608 cases. Four trained interviewers conducted telephone interviews or abstracted data from case records. A research assistant programmed the CATI survey instrument, which included: 1) question and response series; 2) skip patterns; 3) interviewer probes and instructions; 4) range checks; and 5) special edit procedures.

Sample View help for Sample

Agencies: Researchers divided LEAs from the Directory of Law Enforcement Agencies into five sampling frames based on agency size and type. The first frame consisted of 1,162 municipal LEAs and sheriff's offices with 100 or more officers. All agencies in this frame were included with certainty. From these agencies, 7 percent (n=76) were ineligible to participate because they lacked jurisdiction to investigate hate crimes or the agency was a duplicate of another agency within that same or a different sampling frame. Seventy-three percent of eligible agencies (n=792) returned completed surveys or provided the information by phone, and 45 percent of responding agencies from this frame (n=359) reported one or more hate crime investigations in 2018.

The second frame consisted of all 50 U.S. state police agencies, which were included with certainty in the study. Of these agencies, 100 percent completed and returned mail surveys, but 26 percent (n=13) were ineligible to participate because they lacked jurisdiction to investigate hate crimes. From responding agencies, 49 percent (n=18) reported one or more hate crime investigations in 2018.

The third frame consisted of 1,187 municipal LEAs and sheriff's offices that had between 50-99 officers. Of these agencies, 50 percent (n=593) were randomly selected to participate in the study. Of the 593 third frame agencies that received mail surveys, 2 percent (n=14) were ineligible to participate because they lacked jurisdiction to investigate hate crimes or the agency was a duplicate of another agency within that same or a different frame. Of the 579 eligible agencies, 73 percent (n=425) completed and returned mail surveys, and 27 percent of responding agencies (n=113) reported one or more 2018 hate crime investigations.

The fourth frame consisted of 12,837 municipal LEAs and sheriff's offices with less than 50 officers. Of these agencies, 12 percent (n=1,540) were randomly selected to participate in the study. Of the fourth frame agencies that received mail surveys, 3 percent (n=38) were ineligible to participate because they lacked jurisdiction to investigate hate crimes, the agency no longer existed, or the agency was a duplicate of another agency within that same or another frame. Of the 1,502 eligible agencies, 69 percent (n=1,035) completed and returned mail surveys, and 5 percent (n=55) of responding agencies reported one or more 2018 hate crime investigations.

The final, fifth frame consisted of 2,110 university and college police agencies of which 17 percent (n=357) were randomly selected for inclusion. Overall, 11 percent (n=41) were ineligible to participate because they lacked jurisdiction to investigate hate crimes. Of the 316 eligible agencies, 63 percent (n=199) completed and returned mail surveys, and 15 percent reported one or more 2018 hate crime investigations.

Cases: Of the 2,488 agencies that responded to the mail survey, 575 (23 percent) reported a hate crime investigation in 2018. These 575 agencies reported a total of 3,530 hate crime investigations, with 359 first frame agencies reporting 2,888 cases, 18 second frame agencies reporting 181 cases, 113 third frame agencies reporting 288 cases, 55 fourth frame agencies reporting 96 cases, and 30 fifth frame agencies reporting 77 cases.

From the 3,530 cases reported by responding LEAs, 5 percent (n=173) were ineligible for the study (either because they were not investigated in 2018 or were determined to have no indicators of possible hate or bias motivation). Among eligible cases, a little over half were randomly sampled to be included in the study. From the 1,935 cases included in the sample, researchers sought detailed case information, either from telephone interviews with investigators or from case reports provided by LEAs. Information was successfully gathered on 1,230 cases (64 percent of cases selected into the sample), with telephone interviews conducted in 51.6 percent (n=622) of these cases and data abstracted from case reports in 49.4 percent (n=608).

Time Method View help for Time Method

Universe View help for Universe

United States law enforcement agencies and hate crime case investigations in 2018.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Description of Variables View help for Description of Variables

Agency Data variables provide the size of the population the agency serves, number of full-time sworn officers, total number of hate crime cases reported in the mail survey, written policy guidelines and procedures for investigating hate crimes, whether the agency participates in NIBRS, and whether the agency conducted outreach to community or local advocacy groups on hate crimes.

The Case Data file contains information about hate crime cases and crime scene details, including type of bias indicated, types of evidence collected, number and types of witness interviews conducted, use of internet or online communication by suspects in affiliation with the incident, arrests, charges, and outcomes of criminal cases, including juvenile court outcomes. Suspect and victim demographic information includes sex, gender identity, age, race/ethnicity, religious affiliation, sexual orientation, disability status, and immigration status. Suspect information also includes criminal background, and victim information includes relationship to the suspect.

Response Rates View help for Response Rates

The overall response rate for the mail survey was 71 percent of eligible agencies (n=2,488). Forty-three percent of agencies (n=1,070) who provided the number of 2018 hate crime investigations also completed the policy section of the survey.

Original Release Date View help for Original Release Date

2024-12-10

Version History View help for Version History

Weight View help for Weight

The data are not weighted. However, each dataset includes weight variables that must be used in any analysis. Agency-level and case-level weights and additional sampling variables were constructed to enable unbiased estimation and to reflect complex sample design information pertinent to variance estimation. First, each case was given a sampling weight to account for the probability of selection to both the mail survey and telephone interview samples. The sampling weights were adjusted for agency non-response, case-level non-response, and duplication of cases among agencies. Second, a primary sampling unit (PSU) variable and a secondary sampling unit (SSU) variable were created to account for clustering of cases (i.e., the SSUs) within agencies (the PSUs). Third, a stratification variable was created based on the different sampling strategies for each frame. Finally, finite population correction factors account for the sample being selected without replacement, which can be used in variance estimation.