Using small-area estimation (SAE) methods on data from the National Crime Victimization Survey (NCVS), this report estimates rates of violent and property crime for the 50 States and select metropolitan statistical areas.
The report compares SAE victimization rates for 1999 to 2013 with FBI crime rates from the Uniform Crime Reporting Program. Overall, there is close agreement between the two sets of 15-year estimates of violent crime and even closer agreement for property crime. The report cautions, however, that the NCVS (SAE) estimates are affected by sampling variability. The relatively large root mean square errors (RMSEs) associated with these estimates suggest that care is required when interpreting the estimates. Nonetheless, the NCVS (SAE) estimates provide insight into how crime, as measured by the NCVS, is distributed geographically. For States, the report presents results both for the most current period, 2011-2013, as well as for the longer 15-year period (1999-2013). Often the two results are similar for the same area; when the estimate for 2011-2013 is different, there is the possibility that the effect is transient or that it shows a trend that may persist for some time. The NCVS (SAE) county and Core Based Statistical Area (CBSA) estimates offer geographically detailed information that direct estimates from the NCVS may not provide for several years into the future, even though they have somewhat large RMSEs. For the four largest States, the NCVS (SAE) estimates improve the NCVS estimate by only modest amounts. Similarly, once the supplemental sample for the seven next largest States has accumulated data to publish direct estimates for 2014-2016, those estimates will be more accurate than any currently offered by the NCVS (SAE) for 2011-2013. Extensive tables and figures, 6 references, and appended trend figures for the 50 States and DC, for select counties, and for CBSAs