Doubts About the Assumption that Chronic Offenders = Violent Offenders
Although numerous analyses have found high correlations between offending frequency and violent criminal behavior, an important empirical question is whether a group of chronic, but wholly nonviolent offenders exists, and if that group is substantial enough in size to be useful in comparisons with violent offenders. Reports by Piquero suggest not (e.g., Piquero, 2000; Piquero et al., 2007).
In other studies, however, there is ample evidence that chronic, nonviolent- only offenders do exist. In a Dutch cohort, Blokland, Nagin, and Nieuwbeerta (2005) found that chronic offenders were largely persistent, petty thieves. In risk prediction studies, the number of nonviolent acts does not seem to matter. Instead, a history of violent behavior is the best predictor of future dangerousness and the “probability of assaultiveness increases with each additional previous act of violence” (Brizer, 1989, p. xv). In a Canadian study, Carrington, Matarazzo, and DeSouza (2005) reported that the most serious charge for persistent offenders was a person crime in 61.9% of cases for male offenders and 64.1% for female offenders. This means that over 35% of persistent offenders in their sample had not been charged with a violent crime. In a classification study, Chaiken et al. (1994) confirm that violent recidivists can be distinguished from nonviolent ones. In a cohort of male releasees from a Louisiana juvenile correctional facility, those who had committed only property offenses were 40% less likely to be charged with a violent offense as an adult, controlling for number of juvenile convictions, among other things (Thomas, Thomas, Burgason, & Wichinsky, 2014).
In a meta-analysis of studies designed to predict violence, Derzon (2001) laments the low-magnitude association between indicators of antisocial behavior and later violence. He estimates that 24% of those with no history of prior criminal activity went on to commit a crime against a person and that those who engage in antisocial or substance-using behaviors are “unlikely” to engage in later violence. In this study and virtually all others, while correlations between violent and recidivistic nonviolent offending are statistically significant, they are far from perfect.