Key points are not available for this paper at this time.
This article examines agreement between homicide estimates from the Supplementary Homicide Report (SHR) and the National Vital Statistics System (NVSS) at the county level for 1980 to 1988. NVSS and SHR estimates exactly agree in 22% of the counties (68% if agreement is defined as no more than a difference of four homicides), but in some cases, they differ substantially. Although the NVSS generally exceeds the SHR, the pattern is not uniform: 28% of the counties report more SHR homicides than NVSS homicides. Differences between estimates from the two systems are related to population size. Large-population counties often have substantial differences in homicide counts, but the deflating effect of dividing by population yields small rate differences. In contrast, dividing by population magnifies differences in small-population counties and produces discrepancies in the rates that are not present in the counts. The NVSS and SHR differ somewhat in their definition of cases, and other disagreements result from ambiguities in or failures to follow data collection procedures within each system.
Wiersema et al. (Wed,) studied this question.