Predictive validity of the Personality Assessment Inventory (PAI) for identifying criminal reoffending following completion of an in-jail addiction treatment program.

Abstract

We examined the predictive validity of the personality assessment inventory (PAI; Morey, 2007) for criminal reoffending over a 44-60 month follow-up period among incarcerated offenders (N = 124) who completed an addiction treatment program. With this study, we sought to test whether PAI scores conceptually relevant to Andrews and Bonta's (2010) central eight and responsiveness factors were predictive of general and specific reoffending in a sample with criminal involvement and substance-related impairment. PAI scores tapping antisocial personality style and the anger/hostile component of antisocial cognitions had the most robust relationships with the timing and occurrence of reoffending. Hostile dominance and discrepancy scores of observed versus predicted drug and alcohol problems were also predictive of reoffending. None of the other variables examined were consistently associated with reoffending. We discuss these results in relation to the use of the PAI to inform clinical decision making for individuals involved in the criminal justice system.