Recidivism And Pcl-R

Can Hare’s Psychopathy Checklist predict recidivism?

It appears as if the PCL-R is an adequate predictor of both general and violent recidivism. A large-scale meta-analysis relating the Hare measures of psychopathy to antisocial conduct which used a broad definition of recidivism found an r = .27 (Leistico, Salekin, DeCoster, and Rogers, 2008). Two meta-analytic reviews found similar average correlations between future general recidivism and PCL-R scores (r = .27, Salekin, Rogers & Sewell, 1996; r = .28, Gendreau, Little & Goggin, 1996). Additionally, similar results were found in three meta-analyses on violent recidivism, with two reporting an identical r of .27 and the third reporting an r of .32 (Campbell, French & Gendreau, 2009; Hemphill,

Many of the factor two items are, in themselves, well known predictors of recidivism (e.g. Early behavioural problems, juvenile delinquency and impulsivity). Both Hemphill et al. (1998) and Leistico et al. (2008) conducted meta-analyses to investigate the factorial loadings of recidivism predictability. Both reviews found Factor 2 (antisocial and lifestyle facets) a better predictor of general recidivism than Factor 1 (interpersonal and affective facets). This indicates that Factor 1 items may have little use in criminal risk assessment. Additionally one Canadian study found that, Factor 2 scores, but not Factor 1 scores, had a moderate predictive efficacy for nonviolent and general recidivism over 3, 5, 10, and 20 fixed year follow-ups, and predicted violent recidivism at shorter follow-ups (INSERT CITATION –Canadian ARTICLE). Interestingly this study also found that while Factor 2 scores were significantly negatively correlated with all age variables (except age at last conviction); Factor 1 scores were mostly positively correlated with the age variables. The authors concluded that the PCL-R’s performance in risk assessment varies with the length of the prediction period and the type of offending. It has better predictive efficacies for general and nonviolent than violent recidivism, and for shorter rather than longer follow-ups with mostly small effect sizes (INSERT CITATION –Canadian

The defendant, who was on trial for multiple capital murders, was assessed by a psychologist (it should be noted here that the psychologist in question never actually interviewed the defendant) using the PCL-R and attained a score of 36 (which is above the recommended cut off score of 30). The prosecution purportedly argued that the defendant would represent a "continuing threat" to society even while serving a life sentence in a prison with a 23-hour-per-day lockdown. Despite some early evidence that PCL-R score was correlated with violent disciplinary infractions (Forth, Hart & Hare, 1990), there isn’t any evidence which even remotely supports this decision. Base rates of violent institutional misconduct with criminals who Hare’s checklist identified as psychopaths differ only marginally from non-psychopathic prisoners in some studies, while others have found no significant difference (Cunning & Reidy, 1998; Edens, Skeem, Cruise, & Cauffman, 2001 as cited in Edens et al., 2001). Additionally, none of these studies were conducted in prisons similar to the 23-hour-per-day lockdown facility that the defendant would have been sent to (Edens et al., 2001). It is not hard to see how the “expert opinion” of the psychologist and the general stigma around a psychopathy diagnosis might have influenced the