Citation

Hall HV. J. Threat Assess. 2001; 1(3): 1-39.

Copyright

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

DOI

10.1300/J177v01n03_01

PMID

unavailable

Abstract

Empirically-based violence risk analysis has evolved considerably in recent years. The measures presented in this article, numbering over a dozen distinct systems involving over 100,000 subjects, will hopefully be revised in concordance with new findings and new methodologies. Certain findings regarding risk analysis are likely to remain from this study. Among these is that empirically-based decision analysis is superior to clinical judgment, reversing a trend in forensic mental-health practice over the last century. The practice guidelines for violence prediction, which are largely independent of theoretical views and specialized concerns, should remain durable as they pertain to the positive features of any risk analysis. Regardless of which measures are utilized, violence prediction is a process which always involves questions regarding the comprehensiveness of the database, whether deliberate deception and voluntary distortion were taken into account, if basal history was present, and the presence of recent factors associated with violence, such as triggers and opportunity factors. In addition to keeping abreast of the empirical literature, which should modify and refine one's evaluation of risk, the above approach will serve to keep the violence predictor humble. As a standard practice when testifying in court on a risk analysis, the readership is encouraged to share with the trier of fact one's own percent of true positives and negatives in previous cases involving violence prediction.