BJS home page

Contents of Intimate Partner Violence in the U.S.

· Definitions

Additional Information about data

Data sources | National Crime Victimization Survey Methodology | Additional sources

Data Sources

The National Crime Victimization Survey (NCVS) gathers data about crimes using an ongoing, nationally representative sample of households in the United States. NCVS data include information about crime victims (age, gender, race ethnicity, marital status, income, and educational level), criminal offenders (gender, race, approximate age, and victim-offender relationships) and the nature of the crime (for example, time and place of occurrence, use of weapons and nature of injury). NCVS victimization data include incidents reported and not reported to police.

The tables on nonfatal intimate partner victimization in this report include data from the re-designed National Crime Victimization Survey (NCVS) for 1993-2004. Interviews are conducted with households and individuals regarding their experiences with crime. The victims recall the incidents and the data are not verified through other data sources. The survey relies on the victim's ability to accurately recall the characteristics of each incident.

The experiences and estimates of intimate partner violence reflect those of the individuals residing in households. It does not capture the experiences of homeless individuals or those living in institutional settings such shelters for homeless or battered persons.

The FBI's, Supplementary Homicide Reports (SHR) of the Uniform Crime Reporting Program collects data about homicide incidents submitted monthly by local law enforcement agencies with details on victim and offender characteristics. These reports include information on the month and year of an offense, on the reporting agency and its residential population, county and Metropolitan Statistical Areas (MSA) codes, geographic division, and population group, on the age, race, and sex of victims and offenders, victim/offender relationship, weapon use, and circumstance of the crime. For the years 1976-2005, contributing agencies provide supplemental data for 538,210 of the estimated 594,277 murders. Supplemental data are also reported for 597,359 of the estimated 659,862 offenders. For more on methods used with these data, see Homicide Trends About the data.

To the top

National Crime Victimization Survey Methodology

Data on rape, sexual assault, robbery, aggravated assault, and simple assault come from the National Crime Victimization Survey (NCVS). The NCVS gathers data on crimes against persons age 12 or older, reported and not reported to the police, from a nationally representative sample of U.S. households. The NCVS provides information about victims (age, gender, race, ethnicity, marital status, income, and educational level), offenders (gender, race, approximate age, and victim-offender relations) and the nature of the crime (time and place of occurrence, use of weapons, nature of injury, and economic consequences).

In 2005, 77,200 households and 134,000 individuals age 12 and older were interviewed. Between 1993 and 2005, response rates varied between 91% and 96% of eligible households and between 84% and 93% of eligible individuals.

Standard error computations

Comparisons of percentages and rates are tested to determine if observed differences are statistically significant. Only those comparisons that are statistically significant are reported. Differences described as higher, lower, or different pass a hypothesis test at the .05 level of statistical significance (95% confidence level). The tested difference is greater than twice the standard error of that difference. For comparisons that are statistically significant at the 0.10 level (90% confidence level), "somewhat," "slightly," or "marginally" is used to note the nature of the difference.

In some instances, the sample size used to generate an estimate is small. While the estimate is reliable, it is also likely associated with a relatively large confidence interval and should be viewed with caution.

Caution is required when comparing estimates not explicitly discussed. What may appear to be large or small differences may or may not test as statistically significant at the 95% or the 90% confidence level. Significance testing calculations are conducted at the Bureau of Justice Statistics using statistical programs developed specifically for the NCVS by the U.S. Bureau of the Census. These programs take into consideration many aspects of the complex NCVS sample design when calculating generalized variance estimates.

For additional information about NCVS methodology visit:

http://www.ojp.usdoj.gov/bjs/cvict.htm#ncvs
http://www.ojp.usdoj.gov/bjs/pub/pdf/cvus/cvus04mt.pdf

Series crimes

This report includes data on series victimizations. A series crime incident is defined by the NCVS as a crime in which a respondent experiences at least six similar victimization incidents during the given reference period (previous six months) but can not report the date and details of each clearly enough to report them separately. These incidents are recorded as one incident and included in all tables. The characteristics of the incident are based on those represented in the last incident in the series.

Series data are included for analyses where repeat victimization is an important aspect of the subject being analyzed. This report includes series victimizations in estimates of intimate partner violence, counting a series as one victimization.

Marital Status

Marital status is based on the respondent's situation at the time of the interview, not necessarily at the time of the victimization. The survey is not able to determine whether a victim's marital status changed between the victimization and the interview, not necessarily at the time of the victimization. The survey is not able to determine whether a victim's marital status changed between the victimization and the interview.

Caution is warranted when interpreting intimate partner violence and marital status in the NCVS because marital status may be related to a respondent's willingness or ability to disclose violence by an intimate partner. For example, a married woman may not view, may not wish to view, or may be unable to report the behavior of her partner as violent or criminal. That same woman, if separated or divorced, may view or may be able to report the same behavior as violent.

Race and Hispanic Origin Categories

In 1997 the Office of Management and Budget (OMB) introduced new guidelines for the collection and reporting of race and ethnicity data in government surveys. These methodological changes were implemented for all demographic surveys as of January 1, 2003. Individuals are now allowed to choose more than one racial category. In prior years they were asked to select a single primary race.

Racial categories presented in this report now consist of the following:

• White only,
• Black only,
• American Indian/Alaska Native only
• Asian/Pacific Islander only (if only one of these races is given),
• Two or more races (all persons of any race indicating two or more races).

About 0.9% of persons in the NCVS sample and about 2.6% of victims of violence identified two or more races. Individuals are now asked whether they are of Hispanic origin before being asked about their race, and are now asked directly if they are Hispanic or Latino.

Victimization rates as compared to victimization counts

These analyses incorporate rates per 1,000 persons age 12 or older. Because populations may vary over time, using rates takes into account the size of the population being examined. This allows for accurate comparisons of groups of differing sizes as well as changes in populations over time. Examining counts alone ignores these differences in populations and can lead to misleading conclusions.

Estimating change in the reporting of intimate partner victimizations to police

Since 1995, the NCVS has undergone sample reductions because of the escalating costs of data collection. At the same time, the rate of violence has continued to decline. The combination of fewer survey respondents and less crime has resulted in a diminished ability to detect statistically significant year-to-year changes in rates of relatively rare crimes such as intimate partner violence against males.

To the top

Additional sources

This list is in order of the most recent publication first. Additional titles are listed on other topical pages and a comprehensive list is contained on the BJS publications page. To see a full abstract of a publication with links to electronic versions of the publication, click on the title below.

Homicide Trends in the United States

Intimate Partner Violence, 1993-2001. NCJ 197838 (February 2003)

Reporting Crime to the Police. NCJ 195710 (March 2003)

Intimate Partner Violence and Age of Victim, 1993-1999. NCJ 187635 (Oct. 2001)

Intimate Partner Violence. NCJ 178247 (May 2000)

Violence by Intimates. NCJ 167237 (March 1998)

Violence against Women: Estimates from the Redesigned Survey. NCJ 154348 (August 1995)

Violence between Intimates. NCJ 149259 (November 1994)

Bureau of Justice Statistics
www.ojp.usdoj.gov/bjs/
Send comments to askbjs@usdoj.gov

OJP Freedom of Information Act page
Privacy Policy
Page last revised on December 19, 2007