Journal of Quantitative Criminology

, Volume 15, Issue 4, pp 407–422

Murder, Nonnegligent Manslaughter, and Spatial Autocorrelation in Mid-South Counties

  • F. Carson Mencken
  • Cynthia Barnett

DOI: 10.1023/A:1007592124641

Cite this article as:
Mencken, F.C. & Barnett, C. Journal of Quantitative Criminology (1999) 15: 407. doi:10.1023/A:1007592124641


In this paper we explore to what extent county murder and nonnegligentmanslaughter rates in the midsouth are spatially autocorrelated, usinga variety of spatial autocorrelation tests. The data are 3-year averages ofUCR murder and nonnegligent manslaughter rates from the 383 midsouthcounties. Moran's I statistics show a statistically significant amountof spatial autocorrelation in the murder and nonnegligent manslaughter ratesamong the 383 midsouth counties. G statistics, however, fail to detect aglobal pattern in this region. We also compute Gi statistics and localMoran's I statistics with these data and detect some patterns oflocalized spatial autocorrelation. We estimated and compared an MLE spatiallag model and an OLS model with constructs informed by socialdisorganization theory. The regression analysis failed to detect anysignificant spatial effects for the midsouth counties.

murder nonnegligent manslaughter spatial autocorrelation midsouth counties 

Copyright information

© Plenum Publishing Corporation 1999

Authors and Affiliations

  • F. Carson Mencken
    • 1
  • Cynthia Barnett
    • 2
  1. 1.Department of Sociology and AnthropologyWest Virginia UniversityMorgantown
  2. 2.Research, and Development, CJISFederal Bureau of InvestigationUSA

Personalised recommendations