NCVS Redesign Information
About the Redesign
A pioneering effort when the National Crime Victimization Survey (NCVS) was begun in 1972, the survey was redesigned and the new methodology was systematically field tested and introduced starting in 1989. The first annual results from the redesigned survey were published for 1993.
Criticism of the earlier survey's capacity to gather information about certain crimes, including sexual assaults and domestic violence, prompted numerous improvements. Improved survey methodology enhances the ability of people being interviewed to recall events. Public attitudes toward victims have changed, permitting more direct questioning about sexual assaults.
How the Redesign Was Accomplished
An advisory panel of criminal justice policymakers, social scientists, victim advocates, and statisticians oversaw the work of a consortium of criminologists and social and survey scientists who conducted research on improved procedures. New questions were added to accommodate heightened interest in certain types of victimizations. Improvements in technology and survey methods were incorporated in the redesign. The survey now includes improved questions and cues that aid victims in recalling victimizations. Survey interviewers now ask more explicit questions about sexual victimizations. Advocates have also encouraged victims to talk more openly about their experiences. Together, these changes substantially improve reporting for many types of personal and household crime.
Results of the Redesign
Victims are now reporting more types of crime incidents to the survey's interviewers. Previously undetected victimizations are being captured. For example, the survey changes have substantially increased the number of rapes and aggravated and simple assaults reported to interviewers. For the first time, other victimizations, such as non-rape sexual assault and unwanted or coerced sexual contact that involves a threat or attempt to harm, are also being measured.
Break in Series
The NCVS uses an improved screening questionnaire, which determines whether the respondent has been the victim of any crime within the scope of the survey. The new "screener" uses extensive, detailed cues to help respondents recall and report incidents. As measured by controlled tests, the new screener elicits markedly higher reporting of some crimes, especially simple assault and sexual crimes. Consequently, the redesign has created a break in series, because some of the differences between data collected with the NCS instrument in 1973-1992 and the NCVS questionnaire in the 1992-1994 files and subsequent years are attributable to differences in survey procedures and not to differences in victimization experiences.
Because of the break in series caused by the NCVS Redesign, it is necessary to adjust NCS data (NCS-73-92) to examine victimization trends across the entire 25 year period of the survey. It is possible to make such adjustments, because the NCVS was introduced into half the survey sample in January, 1992. For a period of 18 months before the NCVS was fully implemented in July, 1993, the NCS and NCVS were each conducted in half the sample. This simultaneous administration of both versions of the survey allows study of method effects that result from differences in the questionnaires. BJS is currently studying procedures that will permit longitudinal comparisons that incorporate NCS data for 1973-1992 and NCVS data for 1992 through the present.
For longitudinal comparisons of data unadjusted for questionnaire differences, care must be exercised in selecting 1992 files. For comparisons to years before 1992, the NCS 1992 file should be used. For comparisons to NCVS data for 1993 and subsequent years, the 1992 NCVS files should be used.