Despite an increasing emphasis on reentry services for individuals leaving prison, recidivism rates remain high, and policymakers are searching for ways to help parolees make more successful transitions from prison. One strategy is to incorporate interventions into the parole supervision process. This paper presents findings from the Changing Attitudes and Motivation in Parolees (CHAMPS) study, which examined the implementation of a pilot of one parole-based intervention, known as the Next Generation of Parole Supervision (NG).

MDRC conducted the study, which was funded by the National Institute of Justice (NIJ). Drs. Caleb Lloyd and Ralph Serin developed the NG model with funding from NIJ, and the National Institute of Corrections developed the NG curriculum for parole officers. The Bureau of Justice Assistance funded the implementation of NG in the three study sites: Dallas, Texas; Denver, Colorado; and Des Moines, Iowa.

NG is intended to improve parolee outcomes by enhancing parole officers’ knowledge and the strategies they use during their regular supervision meetings with parolees. Building on existing literature about best practices in parole supervision, the NG curriculum focuses on desistance — a process through which individuals who have been involved in crime change their self-perceived identity and cease participating in crime — and helps parole officers to use parolee-centered conversations to identify and reinforce a parolee’s strengths and to identify potential stabilizing and destabilizing influences in the individual’s life.

The study focuses on understanding how NG was implemented as it was piloted in the three sites and assesses NG-trained parole officers’ fidelity to the NG model. In order to better understand NG’s implementation and the business-as-usual practices it was intended to replace, the study also included a second group of parole officers who were not trained in NG. The groups were not randomly assigned. This paper uses a range of qualitative and quantitative data to describe how NG was implemented, including assessments of the knowledge and skills of the parole officers in the two groups over a nine-month period.

The results show that, while there was some variation across sites, parole officers in the CHAMPS sites generally already knew many of the concepts that were part of the NG training, as many of the topics were included in trainings that they had attended previously. The quantitative data suggest that changes to officers’ supervision practices were limited; video observations provided evidence that only in Dallas were NG-trained parole officers exhibiting practices that were substantially different from those observed among untrained officers. In the other two sites, the practices of the trained and untrained officers were largely similar.

These results suggest that a parole agency’s supervision culture and parole officer training can affect both the implementation of an intervention such as NG and the likelihood that it will lead to real change in parole officer behavior. For example, if parole officers are already receiving a good deal of training in practices similar to the new intervention being introduced and if the supervision culture is already aligned with the new intervention’s approach, it may be unlikely that additional training in a curriculum will result in significant change. This scenario appears to have been the case in Des Moines, where officers already received training in motivational interviewing and the agency’s culture was rehabilitation focused. There, officers who did not receive NG training exhibited skill in NG-related strategies at medium to high levels, similar to officers trained in NG.

In contrast, parole officers in Dallas started out with the lowest levels of skill and least amount of experience. The Dallas parole agency’s shift toward a rehabilitation culture and training in motivational interviewing was more recent. In that site, the supervision practices of NG-trained parole officers did appear to change over time to become more consistent with the NG curriculum.

The results in Dallas also suggest that coaching may be important to successfully implementing an intervention that involves changing parole officers’ skills and practices. While NG-trained parole officers exhibited small changes in their supervision practices early in the study period, there were more noticeable changes once coaching was introduced. It is not surprising that the implementation of some strategies, such as changing the way a parole officer responds to a comment from a parolee, requires ongoing practice and reinforcement. The other two sites engaged coaches for the entire study period, but little change was observed in parole officer behavior, suggesting that the presence of a coach itself is not a guarantee for change in parole officer behavior when officers are already familiar with the skills the coach is supporting. Implementing coaching was challenging at all sites. Logistical issues, such as officers being out of the office and in the field or in training, made it difficult for coaches and parole officers to meet frequently.

Overall, this study shows that parole officers are amenable to training and coaching to help them improve their supervision practices, but that consistent implementation can be challenging.