<p>Personality disorders are common among high-risk offenders. These disorders may have relevance for their risk of offending, and they are likely to present barriers to their engagement in rehabilitation programmes. Co-morbidity between personality disorders - and the high frequency of clinical disorders in general - in offender samples complicate research on personality disorder in offender rehabilitation. One approach to understanding this heterogeneity is to use cluster analysis (CA). CA is an empirical strategy which is used to identify subgroups (clusters) of individuals who have similar scores on the variables used in the analysis. It has been used to empirically identify different patterns of personality and clinical psychopathology among incarcerated offenders. Two profiles frequently emerge in cluster analytic research on offender psychopathology profiles: an antisocial/narcissistic profile and a high-psychopathology profile. However, previous research has not empirically examined whether the identification of these profiles has clinical relevance for offender rehabilitation; that is, whether the profiles are simply descriptive, or whether they can provide useful information for the management and rehabilitation of offenders. In the current research, I used data collected from high risk offenders entering prison-based rehabilitation programmes to investigate the clinical utility of psychopathology clusters. Using a self-report measure of personality and clinical psychopathology - the Millon Clinical Multiaxial Inventory III - I identified three clusters: a low-psychopathology cluster (26% of the sample), a high-psychopathology cluster (35% of the sample), and an antisocial/narcissistic cluster (39% of the sample). The high-psychopathology and antisocial/narcissistic clusters in particular resembled high risk clusters found in previous research. To determine whether the three clusters had clinical relevance, I investigated cluster differences in criminal risk, treatment responsivity, and self-report predictive validity. I found evidence for cluster differences in criminal risk: men in the high-psychopathology and antisocial/narcissistic clusters had higher rates of criminal recidivism after release compared to men in the low-psychopathology cluster. However, I found that regardless of psychopathology, men in all three clusters made progress in treatment, and there was little evidence that clusters that reported more psychopathology were less engaged, or made less progress. In the final study I examined cluster differences in self-presentation style and the predictive validity of self-report. Results indicated that offenders who reported high levels of psychopathology had a more general tendency for negative self-presentation, and their self-report on risk-related measures was highly predictive of criminal recidivism. Combined, the results of this research show that cluster analysis of self-reported psychopathology can generate a parsimonious model of heterogeneity in offender samples. Importantly, the resulting clusters can also provide information for some of the most important tasks in offender management: assessment and treatment. The results suggest the highest risk offenders tend to report higher levels of psychopathology, and that offenders who report extensive psychopathology also have highly predictive risk-related self-report. Perhaps one of the most reassuring findings of the current research is that even offenders who report high levels of psychopathology appear to benefit from rehabilitation.</p>
Gene Cluster Analysis Method Identifies Horizontally Transferred Genes with High Reliability and Indicates that They Provide the Main Mechanism of Operon Gain in 8 Species of γ-Proteobacteria