Citation

Parker RN, Smith MD. Am. J. Sociol. 1979; 85(3): 614-624.

Copyright

(Copyright © 1979, University of Chicago Press)

DOI

10.1086/227052

PMID

unavailable

Abstract

The assumption that homicide is a unidimensional phenomenon has rarely been questioned in empirical research. Using newly available data, this analysis classifies homicide into two types, primary and nonprimary, based on the victim/offender relationship. Two models that have appeared in the literature are replicated, utilizing this classificatory scheme. State primary-homicide rates are found to be related to poverty and to the percentage of the population aged 20-34, while nonprimary homicide rates are significantly related only to the percentage of the state living in urban areas. Replication of the original models demonstrates that the failure to classify homicides in this manner results in the incorrect assessment of the relative size and importance of the various predictors of homicide included in these models.