. "Appendix C: Procedures and Operations of the National Crime Victimization Survey." Surveying Victims: Options for Conducting the National Crime Victimization Survey. Washington, DC: The National Academies Press, 2008.
The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Surveying Victims: Options for Conducting the National Crime Victimization Survey
representing” (SR) because each is assigned to its own stratum and each is automatically selected for the sample. Decennial census data on “geographic and demographic characteristics” are used to group smaller PSUs into strata.2 These smaller PSUs are “non-self-representing” (NSR) because only one PSU per stratum (selected with probability proportional to population size) is drawn for inclusion in the sample.
Second stage: Within each sample PSU, “clusters of approximately four housing units or housing unit equivalents” are drawn from each of four nonoverlapping frames: (1) the unit frame, a listing of housing units for the PSU from the decennial census Master Address File (MAF); (2) the group quarters frame, a listing of nonhousehold units such as college dormitories and rooming houses derived from the decennial census roster of such places; (3) the area frame, consisting of census blocks; and (4) the permit frame, a listing of addresses compiled from building permit data. The use of the permit data allows for housing constructed after the most recent decennial census to be included in the NCVS.
For NCVS data collected in 2005, the sample design consisted of 93 SR PSUs and 110 NSR strata. The original design based on 1990 census files included 152 NSR strata, but 42 were eliminated due to a sample cut in 1996 (see Box 1-2).
The various units, strata, and frames used in the sample design of the NCVS are keyed to the decennial census; however, the transition to files based on a new decennial census is not as rapid as might be imagined. Sample based on addresses from the 2000 census Master Address File only began to be phased in starting in 2005; as of 2007, the NCVS remains a hybrid of 2000- and 1990-census-based sample. Likewise, both 1980- and 1990-based sample were used through 1997, with 1990-only sample only beginning in 1998.
With respect to the group quarters coverage of the NCVS sample, it is important to note that BJS and the Census Bureau exclude some major types of nonhousehold, group quarters types from eligibility in the survey. In particular, “institutionalized persons, such as correctional facility inmates,” are not included in the NCVS, nor are seaborne personnel on merchant vessels or armed forces personnel living in barracks.
(e.g., townships) or the similarly sized census county divisions that the Census Bureau defines for other states—may also be grouped into strata. Hence, PSUs need not be made up of whole counties. This is particularly so in New England and Hawaii, where stratification PSUs “could consist of one or more” of these subcounty areas.
“These characteristics include geographic region, population density, rate of growth, population, principal industry, and type of agriculture” (U.S. Census Bureau, 2003:A1-9).