BJS: Bureau of Justice Statistics

Home  |  About Us  |  Contact Us  |  Help  |  A-Z Topic List
Bureau of Justice Statistics (BJS)
Alcohol and Crime: Data from 2002 to 2008


Comparison of the Different Data Sources

The Alcohol and Crime data are from four different data sources, including the Bureau of Justice Statistics’s (BJS) National Crime Victimization Survey (NCVS) for 1997 through 2008; the Federal Bureau of Investigation’s (FBI) 2007 National Incident Based Reporting Program (NIBRS);  and BJS’s 2004 Survey of Inmates in State and Federal Correctional Facilities (SISFCF)  and the  2002 Survey of Inmates in Local Jails (SILJ).  These sources differ in terms of the nature of the crimes collected, the populations covered, and the methods used to obtain information.  In particular, these sources vary as to whether they include violent crimes—both known and unknown to law enforcement authorities— the coverage of crimes involving juvenile victims, the characteristics of the geographical areas covered by the data source, the ways in which multiple events of the same type are captured by the systems, and the manner in which alcohol involvement is ascertained.  As a result, users of this information should consider these differences when comparing similar findings from the different data sources as the findings may not be directly comparable.

NIBRS includes only those violent crimes reported to law enforcement agencies which in many ways reflect  the severity of the crime and other existing avenues for handling the matter (e.g., violence committed by juveniles at school could be handled by school authorities).  Moreover, the NIBRS system includes law enforcement agencies that serve 24% of the population of the United States and very few large jurisdictions are represented.  Also, when multiple crimes are committed during the year, NIBRS reports them as separate events and includes all reported violent crimes including those in which the victim was under age 12.

While NIBRS is a police administrative data collection, the National Crime Victimization Survey (NCVS) is a representative sample of the residential population of the United States. It includes all non-fatal violent crimes (e.g., homicides are excluded) described by respondents age 12 and older whether the crimes are reported or not reported to the police.  The survey interviews persons twice a year and at each interview asks the respondents to report the victimizations that have occurred in the last 6 months. The data from each of the two 6-month periods are aggregated to produce the yearly statistics. Consistent with estimates published in the BJS Criminal Victimization bulletins, the NCVS estimates presented here exclude “series” victimizations—crimes that are similar in nature that people experienced 6 or more times within the previous 6 months. These crimes represented about  4% of all violent crimes in 2008. If these offenses were reported to the police they would be counted in the NIBRS data, but not fully in the NCVS. 

The prison (SISFCF) and jail (SILJ) inmate surveys collect data from inmates on their alcohol use at the time they committed their crimes.  The jail inmates included in this analysis were limited to those who had been convicted of a crime, while the prison inmates included those sentenced to incarceration and held in an adult correctional facility.  The prison and jail inmates represent a selected subset of offenders from those covered in the victim survey or police administrative data.

Determining alcohol involvement also differs across these data collections.  Alcohol use in the NIBRS data is based primarily on the law enforcement officer’s assessment of the situation.  In contrast, in the NVCS alcohol involvement is based on the victim’s judgment.  The prison and jail data are self-reported by inmates.   

The NIBRS data may reflect a conservative estimate of the rate of alcohol involvement as the officers may limit the designation of alcohol involvement to cases in which alcohol use is more readily apparent.  While NIBRS and NCVS are not directly comparable, the rate of alcohol involvement in crimes from NIBRS for 2007 was 10.7% compared to the 22.1% from the 2007 NCVS (including all responses – see Table 26). This supports the assumption that NIBRS may undercount alcohol-involved crimes and focuses on crimes in which alcohol use was substantial.  In the correctional data, the recollection of the inmate is used to determine alcohol involvement.  The inmate is in a much better position to report whether he or she had been drinking at the time of the crime than the police or the victim, but they may have an incentive not to admit to alcohol use or to overstate the role of alcohol in the crime.

The full impact of the differences in the coverage of violent crimes and the determination of alcohol involvement among the four data collections is unknown.  The limitations of the data sources should be kept in mind when reviewing these statistics.  To the extent that findings are consistent across the collections, the observed relationship between alcohol and crime may not be a product of the differences between the data collections.  However, some observed differences across data sets may be due to the different methodologies used. For example, when the NIBRS data indicate that a higher proportion of alcohol-involved events take place in a residence than do the NCVS data, this could plausibly be due to over representation of more rural places in the NIBRS data, the inclusion of crimes against young children, the exclusion of repeat victimizations, and differences in the types of crimes reported to authorities.

Questions on Alcohol and Drug Use by the Offender

The NCVS incident reporting form is available at

Two questions are used to identify whether the victim perceived the offender was under the influence of alcohol.

64b. Was the offender drinking or on drugs, or don’t you know?

Yes (drinking or on drugs) – Ask 65
No (not drinking/not on drugs) – Skip to 66
Don’t know (if drinking or on drugs) – Skip to 66

65. Which was it? (Drinking or on drugs?)

On drugs
Both (drinking and on drugs)
Drinking or on drugs – could not tell which

Additional information on the NCVS methodology is available at

The Surveys of Inmates in Correctional Facilities and Jails
Following are links to information on the two data collections:

Survey of Inmates in Local Jails, 2002

Survey of Inmates in State Correctional Facilities, 2004

Survey of Inmates in Federal Correctional Facilities, 2004

Questions on Alcohol and Drug Use By the Offender

Both surveys utilize a similar question strategy:

Had you been drinking any alcohol at the time of the (Insert stored offense)?
(1) Yes
(2) No
Respondents may enter a don’t know or refused response

The location of the question differs according to the survey:

Jail Survey, 2002: S8Q4c, p.136

Survey of Inmates in State and Federal Correctional Facilities, 2004, S8Q3a., p. 137

Back to Top