Assessing recidivism risk with criminal thinking and prior arrests: do risk factors accumulate or interact?

ABSTRACT

The purpose of this study was to determine whether two well-known risk factors for recidivism, criminal thinking and prior criminal arrests, were capable of achieving a combined effect, and if so, whether this effect was additive, interactive, or both. This hypothesis was tested in a group of 226 males incarcerated in a medium security federal prison who had participated in a criminal thinking intervention prior to their release from prison. The two main variables, criminal thinking (Exp(b) = 1.016 [1.007,1.024]) and prior arrests (Exp(b) = 1.062 [1.024, 1.102]), and their interaction (Exp(b) = 1.264 [1.002, 1.595]) were statistically significant according to the results of a Cox regression analysis. Moreover, the combined score displayed a small incremental increase in the area under the curve (AUC) of a receiver operating characteristic analysis (.660 [.583, .737]), whereas adding the interaction to the two-variable combination improved the AUC further (.668 [.592, .743]). These results suggest that risk factors for recidivism not only add predictive variance to a risk assessment they may also be capable of forming specific interactive effects with other risk factors, although the effects observed in the current study were small and in need of further evaluation.

KEYWORDS:

• Risk assessment
• additive effect
• interactive effect
• criminal thinking
• recidivism

Disclosure statement

Glenn D. Walters is the author of the Psychological Inventory of Criminal Thinking Styles (PICTS) and has received small remuneration from sales of the PICTS manual in the past.