Predictive validity of self-reported self-control for different forms of recidivism

Purpose - This study aims to investigate the predictive validity of self-control (SC) for several forms of criminal recidivism (general, property, violence, sexual). Design/methodology/approach - In total, 1,838 male prisoners were interviewed while serving a prison sentence. Personality traits known to be related to SC served as operationalization of SC. Cluster analyses identified three clusters of SC-related traits: Emotion regulation, Self-assertion, and Effortful control. Survival-analyses predicted recidivism, which was assessed using official data. The follow up period amounted to 72 months. Findings - The SC-related trait clusters significantly predicted general and violent reoffending, after controlling for established risk factors for recidivism (age, age at first offense, social status, previous youth detention, out-of-home placements, and length of imprisonment). However, trait clusters did not predict reoffending with a property offense. Offenders with violent or sex offenses in their criminal history showed different profiles on the trait clusters. Originality/value - The paper shows that SC is an important risk factor for violent recidivism. SC-related trait clusters should not be combined to form a single score, because essential information for risk profiles would be lost.