To produce estimates on criminal victimization, the Bureau of Justice Statistics’ National Crime Victimization Survey (NCVS) collects information from a sample of U.S. households that represents the nation. The sample design consists of two stages.
The first stage of sampling involves the definition, stratification, and selection of primary sampling units (PSUs), which are defined as a large metropolitan area, county, or group of counties. First-stage sampling occurs once every ten years to account for shifts in the population identified through the most recent decennial census. When BJS discusses sample redesign, it is referring to the first-stage sample.
The first-stage sample of the NCVS was most recently redesigned in 2016 with two goals:
- to reflect changes in the U.S. population based on the 2010 decennial census
- to make it possible to produce direct state- and local-level victimization estimates for the largest 22 states and specific metropolitan areas within those states.
For more information about the most recent NCVS sample redesign to account for changes in the population, see National Crime Victimization Survey, 2016: Technical Documentation and Criminal Victimization, 2016: Revised. For more information about expanding and redistributing the NCVS to sample produce state and local estimates of victimization, see NCVS Subnational Estimates.
The next NCVS sample redesign to reflect changes in the U.S. population is scheduled for 2026, based on the 2020 decennial census.
The second stage of sampling occurs every year for housing units and every 3 years for group quarters. Within the PSUs selected, the sampling process identifies addresses to be included in the sample and interviews are conducted with persons and households at those addresses. Once selected, households remain in the sample for 3.5 years, and eligible persons in these households are interviewed every 6 months, for a total of seven interviews. If the residents of a given household move during this period, the new residents at that address will be interviewed. For more information, see National Crime Victimization Survey, 2016: Technical Documentation.
Frequently asked questions
First-stage sampling occurs once every ten years to account for shifts in the population identified through the most recent decennial census. When BJS discusses sample redesign, it is referring to the first-stage sample. Once data from the decennial are collected, statisticians at the Census Bureau begin work for this sampling stage. Historically, new sample was introduced in the field in 2016, 2006, 1996, and 1986. The next sample redesign is scheduled for 2026.
A redesign of the NCVS sample is necessary to account for shifts in the U.S. population identified through the most recent decennial census. The update ensures that the sample reflects the current population distributions.
The two-stage sample design reduces survey costs and allows for consistent data collection from trained interviewers within each primary sampling unit area.
The sample design defines the method of sampling households in the NCVS. The NCVS sample is a two-stage stratified sample of housing units and group quarters within 542 sample areas. Every ten years, first-stage sampling is conducted to form and select primary sampling units based on population data from the most recent decennial census. This process of first-stage sampling that occurs every decade is referred to as a sample redesign.