Treating Offenders with Substance Abuse Problems: Implicit Beliefs about Addiction and Failed Drug Tests

ABSTRACT

Rehabilitating substance-abusing offenders is difficult. And while correctional interventions that target problematic attitudes and cognitions have proven to be successful at reducing recidivism, the roster of attitudes and cognitions that should be targeted is often vague. This study examines the relationship between “implicit beliefs” about addiction – those beliefs that offenders might not even be aware that they hold, but that can nonetheless undermine successful treatment – and failed drug tests among a sample of offenders in a community-based treatment program. Results indicate a significant bivariate association between implicit beliefs and failed drug tests. The relationship holds up in a multivariate context, yet the patterns of magnitude and statistical significance differ depending upon how implicit beliefs are measured. The findings point to the potential importance of targeting for change offenders’ implicit beliefs and indicate that caution should be used when considering treatment models that emphasize the notion that offenders are “helpless” when it comes to their substance use addiction.

KEYWORDS:

• Implicit beliefs
• drug testing
• community supervision

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. This difference of proportion effect size is generally referred to as the “absolute risk reduction” (ARR) effect size estimate (Schechtman, 2002), and is common in the corrections and criminological literature (Cullen, Pratt, & Turanovic, 2016).

2. Recent work indicates that such models are preferable to nonlinear techniques like logit and probit (see Woolrige, 2002), since linear estimation can provide more interpretable and reliable probabilities without the additional step of transformations (Gomila, 2019).

3. We conducted collinearity diagnostics on each of the estimated models. None of the variance inflation factors (VIFs) exceeded a value of 2.0, and none of the condition index values exceeded 30, the traditional cutoff points for the presence of harmful collinearity (Fox, 1991).