Comparing the correlates of county-level UCR and CDC homicide measures

• In Event: Population-Level Research on Homicide: US, International, and Cross-National Studies

Abstract

At the county level, the two primary U.S. homicide measures have historically been based on the Federal Bureau of Investigation's Uniform Crime Reports (now the National Incident Based Reporting System, NIBRS) and the Centers for Disease Control's National Vital Statistics System (NVSS). Each measurement approach has unique strengths and weaknesses, documented in previous literature. Among the most notable concerns are the incomplete coverage of the UCR county data (including the Supplemental Homicide Reports data), the distinction between place of residence and place of occurrence in the NVSS, and the differential handling of self-defense cases across both data sources. Researchers have also raised concerns about whether results for the impact of county-level independent variables on homicide outcomes using the two measures will differ. In this paper, we focus particularly on this latter set of concerns. In so doing, we examine the empirical link between several common county-level census variables and county-level homicides using both the UCR and the NVSS at several different time points. In order to provide useful information to homicide researchers, we analyze both data sets in terms of how best practices for researchers downloading data to begin a project. In providing systematic comparisons of the effects of independent variables on two sources of homicide data, we also address several key issues in the extant homicide literature including the: (1) very large number of zero-homicide counties; (2) highly skewed distribution of homicides among counties that have at least one homicide; and (3) widely disparate population sizes across counties. We conclude with implications for the use of counties and county-equivalents as a unit of analysis for homicides along with the measurement of aggregated homicide and its predictors.

Authors

• Gary LaFree, University of Maryland

• Yesenia Angelica Yanez, University of Maryland

• Robert Brame, University of Maryland