Citation

Meloy JR, Davis B, Lovette J. J. Threat Assess. 2001; 1(1): 3-16.

Copyright

(Copyright © 2001, Informa - Taylor and Francis Group)

DOI

10.1300/J177v01n01_02

PMID

unavailable

Abstract

A nonrandom sample (N = 59) of "obsessional followers" charged with the crime of stalking and related offenses were studied to determine risk factors for violence. Sixty percent of the sample were physically violent toward person and/or property. Six dichoto-mous variables were selected a priori to determine their relationship, if any, to violence risk based upon previous research: prior sexual intimacy with the victim, major mental disorder, explicit threat toward the victim, personality disorder, chemical abuse/dependency, and prior criminal history. Those who were violent toward person and/or property were significantly more likely to have had a sexually intimate relationship with the victim, to have no major mental disorder, and to have made an explicit threat. A forward stepwise logistic regression found that only one variable-prior sexual intimacy-was necessary to predict violence, with an overall correct classification rate of 90%, sensitivity of 97%, specificity of 78%, positive predictive power of 89%, and negative predictive power of 93%.