Criminally oriented communications : factors identifying authorship and prediction of risk

Abstract

This study investigated criminally oriented communications on two levels; firstly, the possibility of identifying an author and secondly, assessing the risk of author approach. To examine the first aim 84 criminally oriented communications were assessed in terms of: Victim demographics, communication features, linguistic factors and thematic content. The results revealed that offender characteristics of gender, ethnicity, victim-offender relationship, and victim-offender geographic location could be established from characteristics within the communication. A predictive model was developed which indicated a high precision rate but low predictive rate. To examine the ability to link crime actions to offender characteristics the methodology of offender profiling was explored. Multivariate analysis found no association between a cluster of actions and a cluster of characteristics. A bivariate analysis demonstrated associations between individual actions and offender characteristics. The homology assumption of offender profiling was validated by the bivariate analysis. Indications of approach risk were examined in a sample of 114 cases, which revealed no authors exhibited approach behaviours. This result supported previous research of Calhoun (1998) those that 'Howl', do so to merely express while those that 'Hunt' move directly to action. Examination of a sub-set of communications used in the commission of, or after, a crime revealed that differences exist between authors that write to merely express their disquiet and those that write with an intention of action. The findings demonstrated that criminally oriented communications are committed by a distinct offender group, who pose a low risk to following through with action after their communication. It is possible to predict offender characteristics from the communication with a reasonably high degree of precision.

Date of Award	2013
Original language	English