Transforming Community Supervision Through Evidence-Based Practice

by James Bonta

Abstract

In recent years, there has been added attention to training probation and parole officers in applying evidence-based supervision practices. The premier paradigm of offender rehabilitation is the risk-need-responsivity (RNR) model. However, research shows that adherence to the RNR principles during face-to-face supervision is less than adequate. The Strategic Training Initiative in Community Supervision (STICS) training model was developed to enhance officer adherence to RNR. Other training models influenced by STICS have also been developed, including Effective Practices in Community Supervision (EPICS), Staff Training Aimed at Reducing Re-arrest (STARR), and Skills for Effective Engagement, Development, and Supervision (SEEDS). The empirical evidence on these models is reviewed and critiqued. In general, research suggests that RNR-based training models of community supervision are associated with positive changes in officer behaviour and reductions in client recidivism. Lessons learned from the implementation of these models and their outcomes are drawn to guide further improvements in training community correctional staff to achieve effective supervision and reduce recidivism.

Acknowledgments

I would like to thank Amel Loza-Fanous for first suggesting that I write a review of STICS and similar training programs for officers supervising offenders in the community and supporting me throughout the project. Two other researchers at Public Safety Canada, Chloe Pedneault and Seung C. Lee, provided feedback on earlier versions of the manuscript and assisted with some of my effect sizes calculations. However, if there are any errors it is completely my fault. My sincere appreciation is extended to both of them for their help.

Author's Note

The views expressed are those of the authors and do not necessarily reflect those of Public Safety Canada. Correspondence concerning this report should be addressed to:

Research Division
Public Safety Canada
340 Laurier Avenue West
Ottawa, Ontario K1A 0P8
Email: PS.CPBResearch-RechercheSPC.SP@ps-sp.gc.ca

Introduction

The criminal justice system has many levers at its disposal to mitigate the occurrence and re-occurrence of criminal behaviour. One widely used lever to manage the recidivism of offenders is the administration of probation and parole. These forms of community supervision have many intentions that include ensuring that justice is done in accordance with the principle of proportionality to the crime, to deter recidivism, and to provide a context for delivering rehabilitation services to offenders. In Canada, there was an average of almost 90,000 adults under community supervision at any given point during 2018/19 (Malakieh, 2020). In the United States, the numbers are far larger, nearly 4.4 million in 2018 (Kaeble & Alper, 2020).

This report explores the use of community supervision, its effectiveness, and what can be done to improve it. In particular, the argument will be made that the probation and parole officers who are responsible for the supervision of offenders in the community could be more effective if given the proper training and tools. Part of the challenge is that their supervision practices have not closely followed the evidence with respect to what works in offender rehabilitation. However, the outlook is hopeful as many jurisdictions adopt models to train staff in evidence-based supervision. Central to these training models is the Strategic Training Initiative in Community Supervision (STICS). The report begins with setting the backdrop to the development of probation officer training models before summarizing the STICS research and similar training models. The report ends with a discussion of what the new training models mean for policy and practice in community corrections.

Part 1. History of Community Supervision and the Rehabilitation Context

Parole was the first legislated form of community supervision beginning with the introduction of a “marks system” in 1840 in Australia. Parole functioned as an incentive to maintain good behaviour in prison and earn early release. Canada introduced legislation creating conditional release in 1899 followed by New York state as the first US state to establish a parole system in 1907 (Patenaude-Harris, 2013). Probation, as an alternative to incarceration, was first practiced in 1841 in America by John Augustus. It was not until 1869 that the first probation statute was enacted in Massachusetts for youthful offenders with an official probation system established in 1878 (Eno Louden, 2019; Labrecque, 2018). Canada's probation system started later with youth in 1908 and adults in 1921 (Weinrath, 2013). An important aspect of both probation and parole is that they are assumed to rehabilitate offenders (Panzarella, 2002).

Empirical support for the rehabilitative value of community supervision was sporadic throughout the 1950s and 1960s. Then in the 1970s it received somewhat of a boost from an unlikely source, Robert Martinson himself, who concluded in his earlier review of the literature that rehabilitative efforts had no appreciable effect on reducing reoffending (Martinson, 1974). Although Martinson was associated with the “Nothing Works” movement, he and his colleagues argued against the abolishment of parole – advocated at the time by von Hirsch and his colleagues (von Hirsch & Hanrahan, 1978; von Hirsch & Albanese, 1979). Reviewing studies comparing inmates released on parole to inmates who are simply released at the end of their sentence, Martinson and Wilks (1977) found that parole was associated with a decrease in recidivism. Following their review, the research on the effectiveness of community supervision has been equivocal with some studies reporting positive effects (e.g., Lerner, 1977) and others negligible reductions in recidivism (e.g., Gottfredson et al., 1982).

In 2008, Bonta and his colleagues conducted a meta-analysis of the effectiveness of community supervision. Fifteen studies published after Martinson and Wilks' (1977) review were examined. The studies consisted of comparisons of either (a) inmates released without supervision or (b) longer probation supervision periods with shorter periods of supervision. The results showed a very small but significant association between community supervision and general recidivism (ɸ = .02, k = 26, 95% CI = .014, .030). Community supervision demonstrated no association with violent recidivism (ɸ = .00, k = 8, 95% CI = -.008, .016). The question then raised by the researchers was why did supervision not show a positive effect in light of the growing literature on the effectiveness of rehabilitation programs?

The RNR Model of Rehabilitation and the Effectiveness of Community Supervision

Lipton, Martinson, and Wilks' review of 231 offender “rehabilitation” programs was published in 1975 but before then Martinson (1974) pre-empted the findings with the conclusion that there is “very little reason to hope that we have found a sure way of reducing recidivism through rehabilitation” (p. 49). It was clear from the literature, however, that some treatments worked, some had no effect, and some were actually detrimental. In essence, conclusions varied depending upon the reviewer(s) viewing the bottle as half full or half empty. Even Martinson (1979), after a larger review of the literature, led himself to write “(I) reject my original conclusion” (p. 10).

The challenge was to understand under what conditions some treatments are effective. The first step toward this understanding came from a narrative review of the offender treatment literature. Andrews, Bonta, and Hoge (1990) proposed that effective programs followed four principles. First, there was the risk principle (match the level of service to the offender's level of risk). Second, the need principle stated that the goal of treatment should be to address the criminogenic needs or the dynamic risk factors of the individual. Criminogenic needs are psychological or behavioural features of the individual (e.g., attitudes, cognition, impulsivity, and self-control) that when addressed reduce the likelihood of reoffending. Third, addressing criminogenic needs is most effective when the intervention is cognitive-behavioural. That is, promote prosocial cognitions, reinforce the prosocial behaviours that follow and discourage procriminal thougths and behaviours. Teaching these new thoughts and behaviours should also be responsive to the individual's personality, motivation, and other personal characteristics (responsivity principle). Finally, there was the professional override principle that took into consideration the clinician's judgement in applying the best type of intervention for a specific client. This paper outlined what is today known as the Risk-Need-Responsivity (RNR) model of offender rehabilitation.

A few months after the publication of the 1990 paper outlining the RNR model, Andrews and his colleagues (Andrews et al., 1990) empirically tested the first three RNR principles. A meta-analysis of 80 treatment studies found that adhering to the principles was associated with lower recidivism rates with the lowest rates found when all three principles were followed. An expanded meta-analysis of 225 offender rehabilitation programs by Bonta and Andrews (2017) reaffirmed this pattern of results and demonstrated the importance of delivering interventions that follow the principles, especially in community settings where the effect size was double the effect size found for the custodial setting. The RNR model has evolved over time as new evidence has accrued, and Bonta and Andrews (2017) now count 15 principles.

In summary, treatment can work with an average 12 percentage point difference between treated and untreated offenders. Furthermore, the differences widen when the principles are closely followed and when the treatment is delivered in a community setting. The question then arises: why wasn't community supervision more effective in Bonta et al.'s 2008 review? The answer lies in examining what probation and parole officers actually do with their clients (from here on the acronym “PO” will be used to describe probation and parole officers).

Behind Closed Doors: Supervision and (the lack of) RNR Practices

Studies examining the daily activities of POs have typically focused on the legal and administrative requirements of the job. For example, POs spend a great deal of their time completing presentence reports (Bonta et al., 2005), attending court (DeMichele et al., 2011), and monitoring substance abuse testing (DeMichele & Payne, 2018). In general, more time is spent on administrative tasks than face-to-face meetings with clients (Bourgon et al., 2018). Interest in the nature of face-to-face contact during supervision, however, has emerged as an important locus of research.

Understanding how PO behaviour during supervision can influence client behaviour and recidivism owes much to the ground-breaking CaVIC (Canadian Volunteers in Corrections) studies by Andrews and his colleagues in the 1970s (Andrews, 1980; Andrews & Kiessling, 1979; Andrews et al., 1979). Twenty-six reports were published on various topics (e.g., training manuals, cost-effectiveness evaluation) but of relevance to the present report is the evaluation of the interactions of 60 volunteers and 14 professional POs with their clients. One hundred and ninety probationers were randomly assigned to either volunteers or professional POs for supervision, and supervision sessions were audio recorded.

Being supervised by either a volunteer or a professional PO made little difference in client outcomes; what mattered were the skills displayed by the supervisor. Supervisors who were interpersonally skilled were rated by the probationers as more helpful, and these officers were more likely to disapprove of procriminal statements and reward prosocial statements. Procriminal attitudes decreased under the supervision of the most skilled and prosocial supervisors, and recidivism during the probation period was also lower. An important finding from the CaVIC research was the identification of five key officer skills: 1) effective use of authority, 2) prosocial modeling and reinforcement, 3) problem-solving, 4) use of community resources, and 5) quality of the interpersonal relationship. The significance of these skills has been replicated in a meta-analysis (Dowden & Andrews, 2004) and has come to be known as Core Correctional Practice (CCP; Andrews & Kiessling, 1980).

It wasn't until 2008 that the next direct observational study of PO supervision appeared (Bonta et al., 2008). Sixty-two volunteer POs from Manitoba audio recorded a sample of their supervision sessions and permitted access to their case files. The goal was to gauge the extent to which POs adhered to the RNR principles in their everyday supervision of probationers. The file reviews found relatively poor adherence to the risk principle with lower risk clients being seen only slightly less often than higher risk clients. With respect to the need and responsivity principles, only the criminogenic needs of family/marital and substance abuse were consistently discussed. Additionally, whereas the use of prosocial reinforcement was found in more than 50% of the audio recordings, other cognitive-behavioural techniques were much less frequent (e.g., behavioural practice in approximately 24% of recordings, prosocial modeling in 15%).

On the island of Jersey in the English Channel, Raynor and his colleagues have been examining the skills of POs using 95 video recordings (Raynor et al., 2014; Raynor & Vanstone, 2018). The Jersey Supervision Skills Project began in 2007 with two goals. The first goal was to develop a reliable and useful method of assessing skills through video recordings, and second, to evaluate the association between the RNR skills of the POs and client recidivism. A 63 item checklist measuring nine skill clusters was developed (Raynor et al., 2009). An assessment of the POs' video recordings found that cognitive restructuring and problem-solving received the lowest scores (Raynor et al., 2014). Raynor and his colleagues also reported a two-year reconviction rate of 53% for 36 probationers supervised by POs with lower skills level, whereas the reconviction rate for the 39 clients supervised by higher skilled POs was 31%.

Finally, Trotter and Evans (2012) introduced a third method for assessing PO supervision behaviour: direct, personal observation. Forty-six youth justice counsellors were observed during the first three months of supervision (119 supervision sessions). Observations were taken within. The observational coding manual consisted of 15 sections (e.g., problem-solving, prosocial modeling and reinforcement). Although the counsellors were relatively strong on relationship skills and prosocial modeling, they were weak on role clarification and cognitive-behavioural interventions. The researchers did not link skill level to recidivism, noting that the actual, real-life behaviours of the supervisors did not match well with the extant literature on CCP.

Part 2. Training POs in RNR-based Supervision Practices

By the mid-2000s there was wide agreement on the skills POs require to effect change in the behaviour of their clients (see Dowden & Andrews, 2004; Raynor & Vanstone, 2018). The previously reviewed studies exposed the fact that POs to a large extent were not displaying effective intervention skills. Detailed research on PO behaviours and the need for training became widespread enough to form a network of like-minded researchers in 2007 called the Collaboration of Researchers for the Effective Development of Offender Supervision (CREDOS; McNeil et al., 2010). Presently there are at least seven PO training programs derived from the RNR model of offender rehabilitation. We begin with the earliest of these models, Chris Trotter's pioneering work and his Prosocial Practice Model.

Prosocial Practice Model (PPM)

Trotter was influenced by the CaVIC studies described earlier, the research on problem-solving by Reid and Epstein (1972), and Jones and Alcabes' (1993) work on client socialisation. He developed what was called the Prosocial Practice Model (Trotter, 2006, p. 56). In an evaluation of the model, 30 POs were given five days of instruction on prosocial modeling, problem-solving, and empathy with ongoing training sessions (Trotter, 1996). Twelve officers continued to use the skills after training, confirmed by file reviews, while 18 reverted back to probation-as-usual. For the 93 clients of the POs who applied the skills from training, the four-year recidivism rate was 53.8%. For the 273 clients of the probation-as-usual POs the rate was 64.0% (p = .04).

Following the 1996 evaluation, Trotter (2006) expanded his prosocial practice model to include role clarification and relationship building. Trotter partially replicated his 1996 study by using a modified model developed for working with families. (CFW: Collaborative Family Work; Trotter, 2013a, 2017). In the first study, 50 child protection workers attended a two-day training workshop in the model and their clients were compared to workers who did not attend training (Trotter, 2002). Early case closures were unrelated to skill level, although client (parental) satisfaction and ratings of family progress were significantly higher for the workers who reported greater use of prosocial skills.

In a more recent evaluation, Totter and his colleagues (2020) trained counsellors of juvenile offenders and their families in CFW. Recidivism outcomes were compared for offenders from four groups: a) completed the program (n =31), b) non-completers (n = 14), c) refused the program (n = 46), and d) never offered the program (n = 40). Recidivism outcomes were in the expected direction, i.e., completers had the lowest recidivism rate at six and 24 months, but the differences were not statistically significant. However, the time to recidivism was significantly longer for program completers (p < .01). To further support PPM, a number of “black box” investigations have shown positive outcomes associated with the skills central to the model. These studies include the model's relevance to women offenders (Trotter et al., 2012), families of delinquents (Trotter et al., 2020) and clients of mental health services (Trotter, 2006).

The Strategic Training Initiative in Community Supervision (STICS)

Spurred by findings that PO behaviour during supervision was at odds with the RNR principles, steps were taken in 2006 to develop a PO training model to improve practice. Researchers at Public Safety Canada (PSC) were guided by the available treatment research at the time (particularly CCP), Trotter's model, and the treatment programming delivered at the Rideau Correctional Treatment Centre outside of Ottawa. A deliberate decision was made that training would avoid a modular approach focusing on the different criminogenic needs. For example, a module on how to address substance abuse, another to deal with family/marital conflict, one for anger control, etc. The opinion was that to deliver multiple modules in a matter of a few days would be insufficient to raise skills to a level that would facilitate changes in the clients of POs. Rather, it was decided to focus on addressing problematic attitudes since they underlie the criminogenic needs of offenders (e.g., “there are easier ways to make money than working” or “I had a tough week and deserve to get drunk”).

There are two components to the STICS model: a training component and an ongoing professional development component. The training was to run over three days, and it was designed to target medium and high risk offenders (risk principle). Following the need principle, POs were to be trained to identify the procriminal attitudes and other problematic thoughts of their clients. Finally, cognitive-behavioural intervention techniques were taught to assist clients to see that their thinking leads to behaviour and procriminal thoughts need to be replaced with prosocial thoughts if they are to act more prosocially (responsivity principle). A summary of the 3-day training event is shown in Table 1.

Note. Adapted from Bourgon et al., 2010

The implementation of new practices requires ongoing skill development to maintain and improve upon skills learned in a training workshop (Walters et al., 2005). Thus, the second component of the STICS model is the requirement for POs to engage in continuous professional development activities. Three types of clinical support were provided to POs. First, a one or two day refresher was offered on an annual basis. Second, a PO was selected from each office, given additional training, and assigned the role of a coach who guided structured monthly meetings with staff (see Table 2). Finally, POs were expected to submit two audio recordings of supervision sessions once a year for expert feedback.

After developing the STICS teaching modules and in-class exercises, the next step was to formally evaluate it. Through the Federal/Provincial/Territorial Heads of Community Corrections Table, a call was put out asking the provinces and territories if they would participate in an experimental evaluation of a new training program meant to teach and enhance RNR-based community supervision. In 2006, three provinces indicated their willingness to participate (BC: British Columbia, SK: Saskatchewan, and PEI: Prince Edward Island).

Front-line POs from the three provinces were explained the goals of the project (improve RNR-skills during supervision), the requirements of the study (audio record a sample of their session soon after intake, at three months and six months, and participate in clinical support activities), and that they would be randomly assigned to STICS training or to a control group. Eighty POs agreed to participate, and 51 were randomly assigned to receive STICS training and 29 to a control group that received a half-day lecture on the RNR model and the research protocols. The control group was also told that if the STICS model produced benefits that they would also be trained at a future date.

Training began in 2006, and after attrition, 33 trained POs and 19 control POs submitted 299 audio recordings on 143 probationers (Bonta et al., 2011). Analyses of the audio recordings clearly showed that the STICS trained POs were more likely than the control POs to discuss the criminogenic needs of their clients and to use cognitive-behavioral intervention techniques. A two-year reconviction follow-up also found a lower recidivism rate for the clients of the experimental POs (25.3%) compared to the clients of the control POs (40.5%). The application of cognitive intervention techniques was significantly more frequent among the STICS POs. In addition, for the POs who participated more fully in the clinical support activities, the reconviction rate was lower for their clients compared to POs with lower participation (22% vs. 27.1%; Bonta et al., 2010). This latter finding suggests that not only do intervention skills in general improve with structured practice, but cognitive intervention techniques may be particularly important. More will be said about this later.

The promising results from the STICS evaluation (Bonta et al., 2010, 2011) garnered attention from jurisdictions that wanted to apply and evaluate the STICS model within their specific agency. The findings also brought interest internationally from researchers and practitioners who looked to incorporate aspects of STICS into new training programs. Probation and/or parole services in Alberta, BC, Ontario, the Correctional Service of Canada, Sweden, and Denmark requested the assistance of Public Safety Canada researchers to conduct either a pilot of STICS or actually implement STICS across a community supervision system. As will be described shortly, many replication and implementation attempts brought challenges but also yielded new learnings. Research on other STICS-like models also added to the understanding of effective community supervision.

Replications of STICS

Edmonton, Alberta

Alberta Community Corrections was interested in adopting STICS as their model of community supervision. However, before doing so, it was decided to conduct an experimental evaluation of STICS in three probation offices in Edmonton. As in the original study, POs volunteered to participate in the same training and research protocols (i.e., audio recordings, and participation in clinical support activities). Thirty-six volunteers were randomly assigned to STICS training or probation-as-usual in 2012. After attrition and the fact that a few POs did not submit any data, Bonta et al., (2019) were left with 15 POs in the experimental group and 12 POs in the control group. An improved evaluation feature in the study was that clients were randomly assigned to POs. In the 2011 study, POs selected the clients for audio recording and, thus, there was the possibility of “cherry picking” clients for the experiment.

The experimental POs, compared to the control POs, were more likely to address procriminal attitudes during supervision and use cognitive intervention techniques. Both are important features of STICS training. Although the findings regarding the targeting of criminogenic needs were in the right direction, the magnitude of the changes resulting from training paled in comparison to the 2011 findings. For example, the proportion of a session devoted to procriminal attitudes was .05 in the 2019 Edmonton evaluation, but it was almost three times higher in the 2011 study (0.13). The mean score on the use of cognitive techniques was significant but small (1.19 out of a maximum score of 4); it was 0.00 for the control group. Unfortunately, direct comparisons to the 2011 study were not possible due to differences in the construction of the cognitive techniques measure.

Consistent with the less than impressive results related to changes in PO behaviour, no difference was found in the 2-year recidivism rate of the clients from the two groups (52% for 79 STICS clients and 49% for 41 control clients). However, an important replicated finding was the importance of cognitive intervention techniques. The use of such techniques was associated with a longer time to recidivism and, after controlling for age and risk, the recidivism rate for clients exposed to cognitive interventions was significantly lower (43%) than for probationers not exposed to cognitive interventions (54%). The failure to completely replicate the 2011 results was explained by the authors as a failure of the experimental POs to engage in the clinical support activities and to demonstrate high levels of cognitive intervention skills.

Before the results were all in, Alberta community corrections still wanted to pursue the adoption of STICS across the service. However, PSC had devoted all their available resources to the system-wide implementation of STICS in BC (to be described shortly) and was not in a position to assist the province with a STICS implementation. As a consequence, the adult and youth branches decided to contract with the University of Cincinnati to train their POs in the EPICS model. EPICS shares many similarities to STICS and will be explained later in this report. Training began in June 2016 and included training a subgroup of staff who could train others (Warner et al., 2019). By the end of March 2021, approximately 95% of 400 staff had been trained.

Sweden Pilot

In March of 2012, a Memorandum of Understanding (MOU) was signed by PSC with the Swedish Prison and Probation Service to provide STICS training and assist in the design of an evaluation of STICS pilot study. Thirty-six POs from across Sweden volunteered for the evaluation, and 16 were randomly selected to attend four days of training, and the other 20 attended a half-day information session on RNR and the research requirements of the study. These two events were held in September 2012. The trained POs were asked to audio record supervision sessions, according to the previously established protocol used in the original STICS experiment and the Edmonton pilot, and to participate in all the clinical support activities. There was two Swedish POs who acted as coaches for the experimental POs. These coaches attended a previously held training in Canada (March 2012) and were given additional training from PSC.

In Sweden, STICS is referred to as KRIMSTICS (“KRIM” refers to “Kriminalvården” or the Swedish Prison and Probation Service). The evaluation had two goals. The first was to assess the trained POs' adherence to the STICS model and work in agreement with the principles of need (target procriminal attitudes and other criminogenic needs) and responsivity (apply cognitive-behavioural interventions). In accordance with the risk principle, only moderate to high risk clients were selected for the study. The second goal was to assess the potential impact on client recidivism resulting from training in KRIMSTICS.

The evaluation was conducted independently from PSC, and the findings showed that KRIMSTICS-trained POs, in comparison with control officers, followed more closely the RNR principles (Swedish Prison and Probation Service, 2018). Although PSC provided the audio coding manual and some advice, PSC did not have direct oversight as to how the recordings were coded. Thus, comparisons to the Canadian STICS studies should be viewed cautiously. Based on the 319 audio recordings, the experimental POs spent a greater proportion of the sessions discussing criminogenic needs and the amount of time spent on procriminal attitudes was more than five times greater than the time spent by the control POs. With respect to intervention techniques, the KRIMSTICS-trained POs were more likely to display cognitive-behavioural interventions than the control POs.

Although changes in officer behaviour were demonstrated, client reconviction rates did not differ between the experimental (48.8%) and control (49.4%) groups. The failure to find a reduction in recidivism was puzzling for the authors of the report. A number of explanations were forwarded (e.g., STICS may not transfer well to the Swedish context, the quality of the cognitive techniques remained modest even after improvements following KRIMSTICS training). Due to the various explanations, restraint was advised in drawing a definitive conclusion on the STICS model. Nevertheless, and prior to having the recidivism data, Sweden decided to implement KRIMSTICS across the probation service in 2014.

CSC Women's Parole Supervision

In 2011, the STICS research caught the attention of the Deputy Commissioner for Women of CSC. A small pilot study was planned with parole officers from the Women Supervision Units (WSU). An experimental design was originally planned, but after forwarding a request for participants, only 18 parole officers volunteered for the study. Therefore, a pre-post evaluation design was conducted (Bonta & Bourgon, 2015). The officers came from WSUs offices in BC, Alberta, Manitoba, Ontario, and Nova Scotia.

The participants were asked to gather baseline data for six new clients and audio record their supervision sessions at three points in time (start of supervision, three months, and six months later). During baseline, the officers also submitted information on their client's risk and need factors to allow the construction of a risk scale (at the time CSC did not have a uniformly applied risk-need scale for women under community supervision). The client selection criteria were as follows: 1) volunteer to participate, 2) had a minimum of 12 months of supervision, and 3) the client was a new case for the parole officer. The recruitment of clients during baseline was extremely slow and by December 2012, only 17 clients were recruited. Therefore, the criterion related to the length of supervision was reduced to eight months in December 2012 in order to boost recruitment. Despite these efforts, the recruitment of six clients per officer was never realized (37 clients provided 68 audio recordings). The baseline period ran from June 19, 2012, to November 4, 2013.

Over the course of the baseline period, six parole officers dropped out of the study. Consequently, only 12 parole officers with six of their supervisors were given three days of STICS training in November 2013. During the post-training phase, as with the baseline period, the engagement of the parole officers and recruitment of clients was challenging. A year after training and with a refresher course in April 2014, only eight parole officers submitted any recordings for analysis (27 recordings for 17 clients). Monthly teleconferences and meetings were also poorly attended. As a result, the project was terminated in November 2014.

With very limited data (68 baseline recordings and 27 post-training recordings), a few cautious observations could be made. First, a comparison of the present baseline findings to the control group in the original STICS experiment (Bonta et al., 2011) found the parole officers to have shorter recording sessions (19.4 minutes vs. 24.4 minutes for the provincial POs), and the untrained federal parole officers behaved no differently than their untrained provincial counterparts both in terms of the content of their discussions and the skills used. Addressing procriminal attitudes and the use of cognitive intervention techniques were almost non-existent (see Table 3 and 4). With training, the parole officers showed significant improvements in spending less time on noncriminogenic needs and more time on criminogenic needs and attitudes. They also showed improvements in the relationship building cognitive intervention skills. However, the skill levels reached by the parole officers following training did not reach the level found in Bonta et al. (2011). There was no change in behavioural techniques, but this may be due to prosocial modeling and reinforcement already being performed reasonably well by the parole officers.

Notes. * p < .05, ** p < .01

System-wide Implementations of STICS

There have been five attempts to implement STICS across community corrections systems: the provinces of BC, Ontario, PEI, and in the countries of Denmark and Sweden. Two of the jurisdictions (Sweden and BC) also conducted formal evaluations of the implementation of STICS. But first, we summarize the efforts in PEI, Ontario, and Denmark.

Prince Edward Island

PEI community corrections participated in the 2011 experiment with 11 volunteer POs. This is a very small jurisdiction with a total of 15 POs in the adult system at the time. Since 11 POs had already been trained in STICS (the control POs were also trained later), then it was relatively easy from a resource perspective to training the remaining four POs. It was also decided to include the training of five POs in the youth system, making it the first time that STICS was applied to a youth clientele.

With the goal of ensuring the sustainability of STICS without support from PSC, a STICS Mentor Model was established in 2012. Under the guidance of one of the STICS developers, two POs were selected to function as mentors to their colleagues. The mentors facilitated monthly meetings, provided feedback on six audio recordings per year, and delivered the annual refresher workshop. They were trained over a one-year period by PSC and monitored in the second year as the mentors delivered the various STICS activities. By 2014, PEI required no further training support from PSC. However, no outcome evaluation of the efficacy of the implementation has been conducted.

Ontario

Ontario expressed their intention to train over 800 POs in the STICS model. The training of POs was to be led by a STICS trainer at PSC while on secondment from Ontario. This trainer was part of the STICS training team for the pilot project in Alberta and the implementation of STICS in BC. At the beginning of implementation in Ontario (2013), PSC researchers provided limited assistance with the trainings in the Eastern Region of Ontario and complete control of the training was handed-over to the Ontario trainer in January 2018.

An important modification to the STICS model was that training would consist of two separate training events. The first training was two days and covered Modules 1 to 5 (recall Table 1). These modules taught the background skills needed for applying the more complex cognitive-behavioural intervention skills and these foundational skills take place early in supervision. POs would come to the first training, return to their offices, and practice the skills before coming back to a three day training approximately four months later to learn Modules 6 to 10. It was judged that the difficult cognitive interventions were more appropriate later in supervision when rapport has been established and the client more responsive to the idea that thoughts lead to behaviour and, to change criminal behaviour, procriminal thoughts need to be replaced by prosocial thoughts.

Near the end of 2018, approximately 300 POs were trained, but with the election of a new provincial government, further training was suspended and only ongoing clinical support for the extant POs was provided. In 2020, the Ontario Ministry of the Solicitor General announced a halt to further STICS training, and a redesign of PO training in Ontario community supervision is planned.

Denmark

In 2013, three staff from the Danish Probation Service attended a STICS training in BC and subsequently began applying the STICS model across the Danish Probation Service. However, PSC researchers were unavailable to provide training and direct assistance as they were committed to training and evaluating the system-wide implementation of STICS in BC. It was agreed between the STICS developers and Denmark that the service could use the STICS material and adapt it for their use but not call it STICS. The result was a training model labelled MOSAIK. The roll-out of MOSAIK began in 2015. Early evaluation studies have reported that, although POs express satisfaction with the model, they did not use the skills often (especially the cognitive interventions) and spent too much time with low risk offenders. No recidivism outcomes are available at this time.

Sweden

As described earlier, an experimental evaluation of a pilot was conducted in 2012. That evaluation found improvements in officer RNR-based behaviours but no difference in the recidivism of STICS clients and control clients. The final recidivism results from the pilot had not yet become available when the Swedish probation service decided to fully implement KRIMSTICS in 2013 with training beginning in the following year. During the next four years, over 700 POs were trained in the model with support by a coach in each office.

The first evaluation report of the KRIMSTICS roll-out focused on data collected in 2016 (Starfelt Sutton et al., 2021). A second report will present the recidivism findings. Two hundred and forty-eight POs were trained that year, and 188 agreed to participate in the evaluation. Three hundred and five case files were used to evaluate adherence to the risk principle, and audio recordings of 88 clients were examined to assess adherence to the need and responsivity principles. The findings were mixed. First, adherence to the risk principle was inconsistent; there was no difference in face-to-face contacts between medium and high risk probationers. There was a small improvement in focusing on criminogenic needs after training, but with time this discussion content focus returned to baseline levels. Following training, the use of cognitive-behavioural intervention techniques did improve but was described as poor in quality.

The authors of the evaluation concluded that the Swedish probation service needs to reallocate resources to support POs in their adoption of RNR-based skills. Starfelt Sutton and her colleagues also suggested to better engage community service providers to work with high risk-high need clients, thereby maximizing adherence to the risk principle.

British Columbia

The decision to implement STICS across the community corrections division was based upon the promising results of the 2011 experiment and projections of significant cost savings if reductions in recidivism were realized. BC's Ministry of Public Safety and Solicitor General entered into a partnership with PSC to train front-line staff, coaches, and STICS coordinators in the STICS model and undertake an evaluation of the implementation. A MOU was signed, and training began in September 2011. PSC researchers delivered the trainings and conducted the evaluation. The training continued until February 2015 with approximately 350 POs being trained.

There were two general questions posed by the evaluation (Bonta et al., 2021). First, was there a change in PO behaviour following training and, second, if the change did occur, was it related to client recidivism? To answer the first question, POs were asked to audio record four supervision sessions prior to attending training. Two of the recordings were to be conducted early in the period of supervision (within three months) and two later in supervision (three to six months). Initially, the POs were allowed to choose the client for recording, but starting mid-2013, the client was randomly selected. After attending STICS training, POs were assigned randomly selected new clients and were requested to record supervision sessions at the beginning of supervision, and three and six months later. By comparing pre-training (or baseline) recordings with post-training recordings, changes in PO behaviour could be assessed.

Over 900 baseline recordings and 1,750 post-training recordings were submitted. From these two sets of recordings, 283 baseline and 953 post-training recordings were randomly selected for coding. The main analysis of behavioural changes was based on 201 POs who submitted both baseline and post-training recordings, ensuring that the comparisons were based on the same POs.

The analyses of the recordings found significant improvements in the content of discussions emphasized by STICS. POs spent more of their sessions discussing their clients' criminogenic needs and less on their noncriminogenic needs and probation conditions. In line with the importance that STICS training places on addressing procriminal attitudes and cognitions, there was a large increase in the average proportion of a session devoted to this criminogenic need after training (from .02 prior to training to .18 after training). With respect to intervention skills, scores for each of the five intervention techniques examined, as well as their composite score, significantly improved following training. Thus, the answer to the first question regarding improvements in RNR-based behaviours following training was an unqualified yes.

To assess the impact of STICS on client outcomes, the 798 audio recorded post-training clients were compared to a random sample of 396 non-recorded clients supervised by the POs prior to STICS training. The probationers supervised by the STICS-trained POs had a significantly lower 2-year general reconviction rate (43.0%) than the probationers who were supervised prior to STICS training (61.4%). The 2-year violent reconviction rate was also lower for clients exposed to STICS (14.9% vs. 21.2%). Moreover, examining the influence of the various intervention skills on recidivism, only cognitive intervention skills added to the prediction of recidivism. One lacunae in this study, compared to other STICS studies, was that the researchers could not evaluate the influence of ongoing clinical support activities as record-keeping of these activities was incomplete.

The authors of the study concluded that, based on the findings, the implementation of STICS in the community supervision services of BC was relatively successful. Changes in PO behaviour after training were in the expected direction (e.g., improved focus on procriminal attitudes, use of cognitive intervention techniques). In addition, decreases in general recidivism were observed, and for the first time in a STICS evaluation, violent recidivism was also assessed and showed a decrease.

STICS Influence Spreads: STARR, EPICS, and SEEDS

In 2008, the researchers at PSC hosted a one-day meeting with two professors from the University of Cincinnati's Corrections Institute to discuss the STICS model. Approximately a year later, a representative from London Probation Trust in the UK visited the STICS developers in Ottawa for the same purpose. The STICS training manual was shared with them as well as the research protocols for the 2011 STICS evaluation, which was ongoing at the time. None of them directly observed a training session or any of the clinical support activities. All three of them returned to their respective university/agency, and all three of them developed training models that were influenced by STICS.

Effective Practices in Community Supervision (EPICS)

EPICS was developed at the University of Cincinnati and subsequently modified by staff at the Corrections Institute for use by POs supervising adults and youth and for the use of families of delinquents (Latessa et al., 2013). Similar to STICS, three days of training is provided to POs to improve their targeting of criminogenic needs of medium to high-risk clients using cognitive-behavioural interventions. Monthly coaching is delivered by university staff, unlike the STICS model where selected POs deliver the coaching within their office. There is no annual refresher or feedback on audio recordings in EPICS. The training and coaching are provided on a fee-for-service by the Corrections Institute.

An early evaluation of EPICS was conducted with a small sample of 10 POs (Smith et al., 2012). Six POs were assigned to receive training by the manager, and four continued with supervision as-usual (i.e., assignment was non-random). Analysis of 93 audio recordings from 52 adult and youth probationers (31 experimental and 21 control clients) demonstrated that training was associated with a greater likelihood of targeting criminogenic needs, spending more time on addressing these needs, and challenging procriminal thoughts. The POs also attended 10 monthly coaching sessions, and their skills improved with participation. No recidivism data was collected.

The next study employed a larger sample of POs (N = 44), 24 of whom were trained in EPICS (Labrecque, Schweitzer, et al., 2013). The trained officers also attended 24 monthly coaching sessions. Seven hundred and fifty-five audio recordings were submitted by 38 POs for analysis (the number of experimental and control recordings was not reported). The EPICS trained POs were more likely than the comparison group to show changes in many skills (cognitive restructuring, prosocial modeling, problem solving, effective disapproval, and structured learning). Again, recidivism outcomes were not gathered.

Thus far, there is one recidivism outcome evaluation of EPICS (Latessa et al., 2013), and it is a follow-up to Labrecque, Schweitzer, et al.'s (2013) study just described. Forty-one adult and youth POs were randomly assigned to EPICS training (with 24 monthly coaching sessions) or routine supervision. Client selection involved medium to high risk probationers and excluded individuals with serious mental illnesses and those with a history of sex crimes. Sessions were audio recorded at the beginning of supervision, three months, and six months later. The audio recordings were scored on a 3-point scale measuring adherence to EPICS (i.e., 0 = had the opportunity to use the skill but did not, 0.5 = used the skill but missed some major steps, and 1 = proficient use of skill). Three recidivism outcomes were reported (reincarceration, rearrest for a new crime, and new technical violation). There were no differences between the experimental and control probationers on any of the outcomes. Moderator analysis by risk level or by EPICS fidelity scores also found no statistically significant differences between groups. However, when risk and fidelity were considered together, high risk offenders supervised by “high fidelity” officers (EPICS adherence score of 0.5 or more) showed lower recidivism rates on all three measures.

One final comment on the EPICS research is warranted. It is commonly assumed in STICS and similar training models that criminal behaviour decreases as a result of the offender adopting more prosocial attitudes. Labrecque, Smith, et al. (2013) examined this assumption by exploring a subset of POs (N = 37) who supervised 238 probationers. The probationers were administered the Criminal Sentiments Scale (CCS), a measure of procriminal attitudes. The researchers found that, for the clients of the 17 EPICS trained POs, their scores on most of the CSS subscales were lower than for the clients of the untrained POs. However, the results were not statistically significant.

Staff Training Aimed at Reducing Re-arrest (STARR)

One of the University of Cincinnati's researchers who had met with the STICS developers in Ottawa subsequently left to join the US Federal Probation Service. It was there that he, along with others, developed STARR. STARR is available only to US federal POs and, by late June 2020, over 3,000 of approximately 4,000 federal POs and another 906 coaches have been trained.

STARR's training curriculum follows many of the modules and exercises found in STICS, but the emphasis is more on learning skills. Targeting procriminal attitudes and other criminogenic needs receives less attention. Three and one-half days of training is provided, followed by “booster” trainings. There have been two published evaluations of STARR, on the same sample, but none of the large-scale implementation. In the first study (Robinson et al., 2011, 2012), most of the 88 volunteer POs were randomly assigned to STARR training or to routine supervision (two districts would not collaborate with the random assignment). After attrition, there remained 41 POs in the experimental group supervising 295 medium to high risk clients. There were 26 POs in the control group supervising 218 medium to high risk clients.

The evaluation methodology followed that of the 2011 STICS experiment (Bonta et al., 2011) that included audio recording supervision sessions and a one-year recidivism follow-up. Over 700 audio recordings were submitted for analysis. US federal POs supervise both pretrial and post-conviction clients. Failure to appear in court was the measure of recidivism used for the pretrial clients and a new arrest was used for the post-conviction clients. Interestingly, despite two types of clients supervised by the POs with different recidivism measures, they were combined into one failure rate. Both the results from the audio recordings and the recidivism outcomes were favourable to STARR training. The trained POs were more likely to demonstrate cognitive-behavioural skills than the control POs (an analysis of the content of discussions has not been reported in the STARR publications). The one-year recidivism rate for clients of the trained officers was 26% and it was 34% for clients of the control officers, with the difference accounted largely by the medium risk clients (16% for the experimental clients and 32% for the control clients). A follow-up report on the two-year recidivism rate appeared to show that the experimental clients maintained their gains (Lowenkamp et al., 2014). The experimental clients had a two-year recidivism rate of 28%, and the control clients had a rate of 41%, but the difference was not statistically significant overall nor for the moderate risk offenders in particular. However, they did find that, for high risk offenders under the supervision of STARR and POs trained in Motivational Interviewing (MI), the recidivism rate was statistically significantly lower at the one-year mark but not at two years.

Skills for Effective Engagement, Development, and Supervision (SEEDS)

Informed by STICS, Trotter's PPM, and the work of others, SEEDS was developed in England and Wales. The SEEDS training manual devotes a significant amount of time, approximately a day, to MI (National Offender Management Service, 2012). Intervention skills taught are problem-solving, prosocial modelling, and the “challenging” function in MI as expressed by reflection. There are some clear similarities to STICS, such as session structure and the teaching of the Behaviour Sequence.

Two pilots were launched in 2011-12 (Rex & Hosking, 2013). The first pilot involved three days of training along with coaching. Training was delivered across eight sites, and POs reported that the training was valuable and important (Ns were not reported). The second pilot took place at an additional eight sites and examined the organizational culture needed to support SEED. Managers were interviewed and participated in workshops to discuss how resources can be allocated and practical steps taken to support staff. These activities resulted in the Reflective Supervision Model; that is, POs need to adopt a more “reflective style of supervision” (p. 337). The two perspectives were then combined as SEEDS with one part being the training and the second part being managerial involvement and support.

A “consumer satisfaction” study was conducted by the two pilots (Sorsby et al., 2013). Seventy-two POs completed over 300 satisfaction questionnaires administered during the course of a year (there were four follow-up refresher trainings after the initial 3-day workshop). Approval ratings of the different training elements were all in excess of 70%, and 87% of the POs thought it was important to their supervision practice. However, as time progressed, the percentage of POs who felt that cognitive-behavioural techniques were useful fell quite dramatically from 45% to 4%. No explanation was given by the authors for the drop.

The intention was to implement SEEDS into 33 of 35 sites in England and Wales, but a change in government put a pause on this plan, and the new government introduced a privatization initiative in 2014. However, the initiative was a failure (Raynor, 2018), and a new version of SEEDS was introduced (SEEDS2). No reconviction study of the impact of SEEDS has been conducted. However, Sorsby and her colleagues (2017) examined the impact of SEEDS on a short-term outcome: compliance during supervision. Three sites of SEEDS-trained POs were compared to two sites where supervision was conducted in the usual manner. Compliance outcomes were gathered over a one-year period. Adjustments were made on risk scores and other factors to control for differences between the two client groups (e.g., clients from the SEEDS group had higher risk violence scores and were more likely to have committed a domestic violence offense). The evaluation found that the odds of a successful outcome were 1.6 times higher for the SEEDS clients. It was 1.8 times higher when a new offense was the reason for the noncompliance.

Holistic RNR-based Probation Supervision Programs

There are two PO training programs that encompass aspects of RNR beyond the specific training in intervention skills to be used with all offenders during supervision. The Proactive Community Supervision (PCS) model was developed by the state of Maryland with the assistance of researchers from nearby universities (Taxman et al., 2006). The goal of the model is to encourage and support RNR-based supervision. There are different components to the model (e.g., risk/need assessment, developing case plans that match services to offender needs), but one important component is to facilitate learning prosocial behaviours among the clients.

Different probation/parole districts had their staff trained in cognitive interventions (identify “triggers”), MI, and other RNR-based skills. Client outcomes from PCS areas were compared to clients from the non-PCS districts (Taxman, 2008). High risk clients were randomly selected from the PCS areas (N = 274) and matched to 274 clients from the probation-as-usual areas. The behaviour of POs was measured from file information (e.g., number of telephone contacts, family contacts). The outcome measures were rearrest, technical violations, and drug use. For the PCS clients, reductions were found on rearrest (30% vs. 42% for the control) and technical violations (34.7% vs. 40.1%), but there was no difference in positive drug tests. There has been no further research on this promising program.

A second similar program in some respects is the Citizenship program developed in the United Kingdom (Bruce & Hollin, 2009). Citizenship combines PO supervision practices with the effective use of community resources to target medium and high risk clients. POs deliver specific treatment modules (e.g., drug misuse) to address the needs of their clients. Community agencies are contracted to augment the services provided by POs. In evaluations of Citizenship, information on what the POs actually did with their clients is not reported, only how many modules and referrals to community agencies were completed. An evaluation of a pilot study of 170 clients (85 Citizenship vs. 85 control clients) found a positive but nonsignificant trend (reoffending rate of 30% for Citizenship and 48% for the controls; d = 0.26, p = .09). However, a subsequent and larger evaluation (3,819 offenders in the Citizenship group and 2,110 in the comparison group) by Pearson et al. (2011) found a significantly lower 2-year recidivism rate for Citizenship (41%) compared to the comparison (50%; d = 0.20). Finally, a more recent evaluation demonstrated no statistically significant differences in recidivism between Citizenship and a comparison but did show a significant reduction in recidivism for high risk offenders (Pearson et al., 2016).

Part 3. Consolidation

The research on RNR-based training models has flourished since Trotter's (1996) seminal work. There have been a number of narrative reviews on the subject (Bourgon et al., 2020; Raynor & Vanstone, 2018; Trotter, 2013b) and one meta-analysis (Chadwick et al., 2015). All of the reviews paint a favourable picture with respect to effectiveness in reducing recidivism, with Chadwick and his colleagues (2015) finding a large effect size of 0.82 (Cohen's d calculated from odds ratio of 1.48 reported in the article). However, a difficulty in interpreting the effect size from the meta-analysis is the inclusion of three studies that are problematic. The study by Raynor at al. (2014) is actually a black box study and did not include PO training. The remaining two studies (Pearson et al., 2011; Taxman, 2008) combine PO training with other aspects of the model (e.g., risk-need assessment, use of community resources; modular presentations to address criminogenic needs) and it is impossible to disentangle the effect of the training from other components. In fact, in the Citizenship program (Pearson et al., 2011, 2016), POs essentially deliver structured one-on-one lessons to clients based on RNR principles and they are not trained directly on one-to-one supervision practices.

Next, we consolidate the findings from the STICS studies and their implications. Then, we turn to adding observations from other training models and inquire as to how the information from both areas of research may inform future developments in improving community supervision.

There are four Canadian evaluations of STICS. These studies provide a unique opportunity to learn from the results because the studies use almost exactly identical research protocols. Many times replications are conducted by independent researchers who, although they claim to use the same measures, we cannot know how closely their measurements mirror the original methods. For example, the Swedish studies used the same audio coding manual as the STICS 2011 study, but we do not know if the coding results would have been the same if conducted by the Canadian researchers. Granted there are some advantages to independent replications, but there are also disadvantages.

Tables 5 and 6 summarize the audio recording findings from the four Canadian studies. Although the same audio coding manual was used across studies, there was a difference in how the discussion areas and skills were constructed. The original 2011 evaluation of STICS did not choose the items to construct a scale based on the criteria of requiring an Intraclass Correlation Coefficient (ICC) of at least .60. The post-2011 studies used this minimum ICC criterion for the discussion topics and intervention skills. However, this also means that the number of items included in skill will vary slightly by sample. For example, in the BC 2021 study, only two of the five items that comprised the relationship skill met minimum the ICC criterion and, in that report, “relationship” was relabelled “collaborative”. Therefore, caution is advised in interpreting the similarities and differences across the evaluations.

When it comes to discussion topics (Table 5), an improvement in targeting procriminal attitudes following STICS training was consistently found in all four studies. This is important because, if there is one thing that STICS training emphasizes, it is the need to address procriminal attitudes. However, there is considerable variability in how much time is spent on procriminal attitudes across studies. The Edmonton site did show a statistically significant increase in discussing procriminal attitudes, but it was the smallest increase of the four studies. The BC 2021 study showed the expected pattern of results for criminogenic and non-criminogenic needs, such that, after training, POs spent more time on criminogenic needs and less time on non-criminogenic needs. There was no other apparent pattern in the results for criminogenic and non-criminogenic needs. Decreases in the proportion of a session addressing the conditions of probation were found in all studies except the CSC evaluation, but this may be explained by the very small sample size in this study.

Note. * p < .05; ** p < .01; *** p < .001.

A major goal in STICS is to train POs to use cognitive interventions to change procriminal attitudes. All studies found improvement following training in the application of cognitive interventions (Table 6). Also noteworthy is that, without training, the use of cognitive interventions is almost non-existent in community supervision. Behavioural interventions (e.g., role playing, prosocial modeling) demonstrated no differences due to training except in the BC 2021 implementation study. The other uniform findings are the increases in relationship skills and overall effective correctional skills. However, the improvement on the latter was to be expected as it is a composite measure based on the other skills.

Note. * p < .05; ** p < .01; *** p < .001.

Table 7 displays the recidivism results for three STICS evaluations. These studies used the same definition of recidivism (reconviction) and coded the outcome in exactly the same way. The Edmonton evaluation showed no significant difference between the experimental and control clients. This nil finding may partly be explained by the relatively low frequency use of cognitive interventions compared to the other two studies. Evidence suggests that cognitive techniques may be critical for facilitating change.

When we consider the main training model features of not only STICS but the other models reviewed, a number of similarities stand out (Table 8). Most models require at least three days of initial training, with PPM and EPICS for Families having the longest training (five days). All of these programs have ongoing skill development activities reflecting the recognition that a single training event is insufficient to adequately learn RNR-based behaviours. Two programs (EPICS and PPM) also target a youthful delinquent population (STICS has been modified for training youth POs in PEI, but it was never applied elsewhere). Finally, EPICS and PPM can be used with families of troubled youth. In the case of EPICS, POs are required to go through the standard EPICS training before receiving two additional days of training, including family visits with a coach. Presently, EPICS with Families is used with delinquents and their families, but there are plans to use it with adults (Labrecque et al., 2014). Similarly, Trotter and his colleagues (2012, 2017) has also adapted PPM to working with families in both a juvenile justice and child protection context.

Evaluations of EPICS and STARR, not surprisingly given the influence of STICS on their development, have used many of the same research protocols as the original STICS experiment (Table 9). Except for the implementation projects in BC and Sweden, POs are all volunteers and are assigned to training or probation-as-usual. In a number of cases, the POs are randomly assigned to group conditions (Bonta et al., 2011, 2019; Lowenkamp et al., 2014; Robinson et al., 2011, 2012; SPPS: Swedish Prison and Probation Service, 2018). Although it may be difficult to conduct a randomized experimental evaluation, as attested by the evaluations of the other models, it is possible. That being said, even the experiments had limitations. In addition to the fact that most of the evaluations included PO volunteers (a self-selection bias), there was also the possible influence of a client selection bias. In some studies, the POs were allowed to select the clients to be audio recorded and, of course, the client her/himself also must agree to be recorded, introducing another source of bias.

Evaluations of EPICS and STARR also used audio recordings to assess changes in PO behaviour. The developers of both models did have access to the STICS coding manual although it is unclear to what extent the manual was used in the evaluations. Except for the early study of EPICS where the number of criminogenic needs and the time spent on them were coded (Smith et al., 2012; Table 3), all future evaluations of EPICS, and STARR reported only changes in PO skills. Furthermore, the skills were not subjected to ICC analysis or any other test for reliability, which may threaten the validity of the skill constructs. Although the research consistently shows that training in EPICS and STARR leads to improvements in skills, without knowing if the skills are targeting the appropriate criminogenic needs it is difficult to assess the relationship of training to recidivism.

Three models have been evaluated experimentally: STICS, EPICS, and STARR. However, the experimental evaluation of EPICS (Labrecque et al., 2017) did not examine recidivism outcome, only PO intervention skills. This is unfortunate in that the findings would have strengthened (or weakened) a conclusion of a causal relationship between training and recidivism. The remaining models (PPM and SEEDS) used non-experimental evaluation strategies, and PPM conducted an indirect assessment of officer behaviour (file review) while SEEDS had no measure of officer skills.

Note. nr = not reported; V = volunteer; M = mandated; audio = audio recorded

The effects of training on PO skills and recidivism are summarized in Table 10. Significant main effects on recidivism have been found for STICS and PPM. However, when moderators were considered for EPICS (fidelity and risk level) and STARR (MI and risk level) significant effects were found.

Note: * Calculated from raw numbers.

Part 4. Conclusions, Lessons, and Policy Implications

1. The Design of Initial Training

All of the training models reviewed attest to the influence of the RNR model. STICS, STARR, and EPICS unequivocally state that the training is aimed at medium to high risk offenders in accordance to the risk principle. Trotter's PPM model and SEEDS were not explicitly developed for medium to high risk offenders, and evaluations do not assess the impact of these two models according to risk. Thus, it is unclear how well PPM and SEEDS adhere to the risk principle. With respect to the need principle, all models purport to follow the need principle, but from the evaluations, only STICS consistently measures such adherence. The STICS model elevates procriminal attitudes above the other criminogenic needs; nevertheless, the PO's success at targeting other criminogenic needs are examined in the Canadian and Swedish studies.

The strength of all the models is their consideration of the responsivity principle, namely the teaching of cognitive behavioural skills. There is some variation in terms of emphasis (e.g., PPM trains prosocial modeling, problem solving, and relationship skills, SEEDS adds MI training). A closer inspection of the research suggests that training in certain aspects of cognitive-behavioural interventions may be more important than others. Cognitive techniques, in particular, appear important in reducing recidivism. Other more behavioural interventions, such as the use of reinforcement and relationship skills may be less important. With respect to relationship skills, many POs already function at a high level either because they are hired for these qualities or have received MI training.

An important point needs to be made about the coordination of the need and responsivity principles. Training POs to be proficient in delivering cognitive intervention skills requires that POs use these skills to target criminogenic needs. Without monitoring what POs are attending to runs the risk of intervention techniques being applied to increase, for example, self-esteem (a noncriminogenic need). This point has been raised before (Bourgon et al., 2020) but needs to be emphasized.

2. Target population

It may be self-evident, but most models have not been very clear except in a general sense as to whom the model works best for and how the model is modified for the specific population. Trotter has been the most transparent in the application of PPM to youth and child protection settings, writing books on this practice. EPICS has been applied to youth and families but, without comparing the training manuals to the adult manual, we do not know how the emphasis on skills and criminogenic needs has changed and we certainly have no outcome evaluations. As these training models continue to expand, the effectiveness of training with other offender populations (e.g., racialized offenders, sex offenders) become fertile grounds for future research. Some of this work has begun, with PPM and EPICS exploring applications with offenders who have mental health issues (Thompson & Schweitzer, undated; Trotter, 2006).

3. Ongoing skill development

Every single model recognizes the importance of ongoing training to improve PO skills. The activities include coaching, booster or refresher sessions, and expert feedback on audio recordings or in vivo observations. As with the above comment on the effective ingredients of cognitive-behavioural skills, we can ask a similar question: Is one type of clinical support activity more important than another? At this point, we have evidence from STICS and EPICS that coaching improves skill development and outcomes, and we need research to tell us whether other supports add value to coaching alone. This is important for our understanding of the effectiveness and the efficient allocation of resources. Most models also attempt to ensure that the clinical supports are provided for long periods (e.g., 2 years in the case of EPICS).

4. Build capacity

As the researchers and program developers withdraw their involvement after evaluations of pilots, organizations that choose to implement a model across their community supervision service are left with the task of ongoing training for new employees and the maintenance of skill development activities. This involves the allocation of resources to train a few staff to sufficiently high levels of expertise so that they can train others and support ongoing clinical activities. The transition from pilot to full implementation should take advantage of the expertise of the model's developers and researchers to plan how best to conduct the implementation of the model.

Policy and Organizational Support

5. Policies consistent with the RNR model of offender rehabilitation

The successful adoption of any of the training models reviewed requires policies, structure, and the resources that support RNR-based practice (assuming that there is agency buy-in for RNR). Producing the appropriate policies means: a) understanding the RNR model, b) conducting a review of existing policies and procedures to identify discrepancies in policy with respect to RNR, and c) replacing the problematic policies with new policies that are synchronized to RNR. This may require consultation services from experts in RNR to advise the organization, given some of the complexities of the RNR model.

6. Structure and the resources to support successful implementation of the training model

The RNR-based training models require not only the appropriate policies but also establishing agency contingencies that support the new supervision approach of POs. This will require a shift in the values of the organization. Rewarding surveillance and control (e.g., zero tolerance for breaches) to rewarding the PO's helping efforts will need to change. POs need to be encouraged by management to have more face-to-face contact with the higher risk client and be defended when a low risk case has an adverse outcome. There is research that too much of a PO's time is spent on paper work and not on actually helping clients (e.g., Bourgon et al., 2018). The implementation research can be informative in identifying important structures and practices that lead to success (take the example of BC's two levels of POs with one level devoted to medium and high risk clients). Of course, creating new structures in regards to staffing or management practices will require additional resources, and cost-benefit studies can be conducted to help secure these resources.

Management must display its commitment to RNR. This can be partially accomplished through the usual communication channels available to organizations (e.g., newsletters, town halls). However, this commitment can be expressed more strongly by other means. At a minimum, it is recommended the managers attend and fully participate in the initial training. It demonstrates to staff by their participation that the training is important, and they can better appreciate the difficulty in learning new skills. Management can even go further and include as part of annual staff performance reviews an assessment of their use of skills with clients.

7. Monitoring and Evaluation

When BC was implementing STICS across the community corrections division, as part of the research, there was a system in place to monitor staff buy-in to the model. This was fortunate as it alerted the researchers and senior management that there were problems with the implementation. Thanks to the commitment of management to STICS (see #6 above), there was a pause on the implementation until the obstacles were resolved before resuming the implementation process. In the absence of a monitoring and evaluation component to the implementation, the successful outcomes found in the evaluation, in all likelihood, would not have materialized.

Future Research

When it comes to future research, the list of possible areas to study is almost endless. We will mention only the more obvious. We know little about the differential effectiveness of RNR-based training on offender subpopulations (women, racialized offenders, young offenders, sex offenders, etc.). Replications of the training models are few and far between. Long-term recidivism outcomes, including violent re-offending, are needed to evaluate EPICS for Families, CWF, and SEEDS. PO behaviours directed to criminogenic needs require further study with the non-STICS training models, while SEEDS and PPM should examine the skills actually used by POs.

The research demonstrates that conducting randomized experiments is possible. It is understandable that some evaluations use weaker methodologies (e.g., comparison groups of convenience, quasi-experimental), but every effort needs to be made to experimentally evaluate these training models. It is the only way to establish cause and effect. Finally, implementation research would benefit the field immensely. This research can identify what works and why to inform organizational readiness to adopt a training model and steer the organization through the many hurdles that large scale implementation brings.

Closing Comments

This paper reviews the various PO training models, with the STICS model central to the paper. STICS was developed and evaluated by researchers at PSC, but it has also had a significant influence on many of the other models reviewed. Similarities and differences among training features of the models were summarized and, more importantly, the research on the effectiveness of the different training programs was examined. The STICS model appears as the most comprehensive and researched of all the models reviewed. It closely follows the RNR principles, includes procedures to facilitate skill development, and has demonstrated that it can be successfully implemented on a large scale.

There have been three randomized experimental evaluations of the STICS model with various degrees of success, but even failures to replicate can be instructive. Two large implementation projects have also been evaluated. One implementation in BC finds very positive results, while the implementation in Sweden has only been partially evaluated as we await recidivism outcome results. No other large scale implementation of the other models has been evaluated. SEEDS comes the closest, but we have no information on PO behaviours that mediate any impact on client recidivism.

In conclusion, community supervision agencies can take heart from the findings of the research. There is scientific literature that can guide organizations to be more effective in the protection of the community through offender rehabilitation. Although determination is required from agencies and their staff, they can learn and use with a degree of sophistication new skills that can make a difference in the lives of the people they supervise.

References

References indicated with * are STICS publications NOT cited in the text

Alexander, M., Lowenkamp, C. T., & Robinson, C. R. (2013). A tale of two innovations: Motivational interviewing and core correctional practices in United States probation. In P. Ugwudike & P. Raynor (Eds.), What works in Offender Compliance: Internatinoal perspectives and evidence-based practice (pp. 242-255). Palgrave Macmillan.

Andrews, D. A. (1980). Some experimental investigations of the principles of differential association through deliberate manipulations of the structure of service systems. American Sociological Review, 45(3), 448-462. https://doi.org/10.2307/2095177

Andrews, D. A., Bonta, J., & Hoge, R. D. (1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17(1), 19-52. https://doi.org/10.1177/0093854890017001004

Andrews, D. A., & Kiessling, J. J. (1979). An introduction to the reports in the CaVIC series and a summary of selected CaVIC contributions and findings. Ministry of Correctional Services, https://www.ncjrs.gov/pdffiles1/Digitization/66865NCJRS.pdf

Andrews, D. A., & Kiessling, J. J. (1980). Program structure and effective correctional practices: A summary of the CaVIC research. In R. R. Ross & P. Gendreau (Eds.), Effective correctional treatment (pp. 441-463). Butterworth.

Andrews, D. A., Kiessling, J. J., Rusell, R. J., & Grant, B. A. (1979). Volunteers and the one-to-one supervision of adult probationers: An experimental comparison with professionals and a field-description of process and outcome. Ministry of Correctional Services.

*Bonta, J. (2012). From evidence-informed to evidence-based: The Strategic Training Initiative in Community Supervision (STICS). In D. Evans (Ed.), What works: Defeating recidivism – keys to making it happen (pp. 345-359). American Correctional Association.

Bonta, J., & Andrews, D. A. (2017). The psychology of criminal conduct (6th ed.). Routledge.

Bonta, J., & Bourgon, G. (2015). Report on the Correctional Service of Canada's women parole supervision STICS pilot project [Unpublished manuscript]. Public Safety Canada.

Bonta, J., Bourgon, G., Jesseman, R., & Yessine, A. K. (2005). Presentence reports in Canada. Public Safety Canada, https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/prsntnc-rprts-cnd/prsntnc-rprts-cnd-eng.pdf

*Bonta, J., Bourgon, G., & Rugge, T. (2018). From evidence-informed to evidence-based: The Strategic Training Initiative in Community Supervision (STICS). In P. Ugwudike, P. Raynor, & J. Annison (Eds.), Evidence-based skills in community justice: International perspectives on effective practice (pp. 169-191). Policy Press.

*Bonta, J., Bourgon, G., Rugge, T., Gress, C., & Gutierrez, L. (2013). Taking the leap: From pilot project to wide-scale implementation of the Strategic Training Initiative in Community Supervision (STICS). Justice Research and Policy, 15(1), 17-35. http://doi.org/10.3818/JRP.15.1.2013.17

Bonta, J., Bourgon, G., Rugge, T., Pedneault, C. I., & Lee, S. C. (2021). A system-wide implementation and evaluation of the Strategic Training Initiative in Community Supervision (STICS). Journal of Criminal Justice, 74(May/June), 1-12. https://doi.org/ 10.1016/j.jcrimjus.2021.101816

Bonta, J., Bourgon, G., Rugge, T., Scott, T. L., Yessine, A. K., Gutierrez, L., & Li, J. (2010). The Strategic Training Initiative in Community Supervision: Risk-need-responsivity in the real world. Corrections Research User Report (2010-01). Public Safety Canada, https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/2010-01-rnr/2010-01-rnr-eng.pdf

Bonta, J., Bourgon, G., Rugge, T., Scott, T. L., Yessine, A. K., Gutierrez, L., & Li, J. (2011). An experimental demonstration of training probation officers in evidence-based community supervision. Criminal Justice and Behavior, 38(11), 1127-1148. https://doi.org/10.1177/0093854811420678

Bonta, J., Rugge, T., Bourgon, G., & Wanamaker, K. (2019). A conceptual replication of the Strategic Training Initiative in Community Supervision (STICS). Journal of Experimental Criminology, 15(3), 397–419. https://doi.org/10.1007/s11292-019-09371-4

Bonta, J., Rugge, T., Scott, T., Bourgon, G., & Yessine, A. (2008). Exploring the black box of community supervision. Journal of Offender Rehabilitation, 47(3), 248–270. https://doi.org/10.1080/10509670802134085

*Bourgon, G. (2013). The demands on probation officers in the evolution of evidence-based practice: The forgotten foot soldier of community corrections. Federal Probation, 77(2), 30-35.

*Bourgon, G., Bonta, J., Rugge, T., & Gutierrez, L. (2010). Technology transfer: The importance of on-going clinical supervision in translating “What Works” to everyday community supervision. In F. McNeil, P. Raynor, & C. Trotter (Eds.), Offender supervision: New directions in theory, research and practice (pp. 88-106). Willan Publishing.

Bourgon, G., Bonta, J., Rugge, T., Scott, T. L., & Yessine, A. K. (2010). The role of program design, implementation, and evaluation in evidence-based “real world” community supervision. Federal Probation, 74(1), 2–15.

Bourgon, G., Chadwick, N., & Rugge, T. (2020). Beyond core correctional practice: Facilitating prosocial change through the Strategic Training Initiative in Community Supervision. In J. S. Wormith, Craig, L. A., & Hogue, T. (Eds.), Handbook of what works in violence risk management: Theory, research and practice (pp. 505-526). John Wiley and Sons.

*Bourgon, G., Chadwick, N., Rugge, T., Bonta, J., & Gutierrez, L. (2015). The Living Laboratory report phase 1: Insights into community supervision practices at Edmonton Alberta central intake probation office. Public Safety Canada.

*Bourgon, G., & Gutierrez, L. (2012). The general responsivity principle in community supervision: The importance of probation officers using cognitive intervention techniques and its influence on recidivism. Journal of Crime and Justice, 35(2), 149–166. https://doi.org/10.1080/0735648X.2012.674816

*Bourgon, G., Gutierrez, & Ashton, J. (2012). The evolution of community supervision practice: The transformation from case manager to change agent. Irish Probation Journal, 8(October), 2011, 28-48. Also reprinted in Federal Probation, 76(2), 27-35.

Bourgon, G., Rugge, T., Chadwick, N., & Bonta, J. (2018). The Living Laboratory studies: Providing insights into community supervision practices. Federal Probation, 82(1), 3-12.

Bruce, R., & Hollin, C. (2009). Developing citizenship. EuroVista, 1(1), 24-31.

Chadwick, N., Dewolf, A., & Serin, R. (2015). Effectively training community supervision officers: A meta-analytic review of the impact on offender outcome. Criminal Justice and Behavior, 42(10), 977-989. http://dx.doi.org/ 10.1177/0093854815595661

DeMichele, M. T., & Payne, B. (2018). Taking officer time seriously: A study of the daily activities of probation officers. Probation Journal, 65(1), 39-60. http://doi.org/ 10.1177/0264550517748358

DeMichele, M. T., Payne, B. K., & Matz, A. K. (2011). Community supervision workload considerations for public safety. Office of Justice Programs.

Dowden, C., & Andrews, D. A. (2004). The importance of staff practices in delivering effective correctional treatment: A meta-analysis of core correctional practices. International Journal of Offender Therapy and Comparative Criminology, 48(2), 203-214. http://doi.org/10.1177/0306624X03257765

Eno Louden, J. (2019). Probation. In R. D. Morgan (Ed.), The Sage Encyclopedia of Criminal Psychology (pp. 1113-1116). Sage.

Gottfredson, M. R., Mitchell-Herzfeld, S. D., & Flanagan, T. J. (1982). Another look at the effectiveness of parole supervision. Journal of Research in Crime and Delinquency, 19(2), 277-298. https://doi.org/10.1177/002242788201900209

Jones, J., & Alcabes, A. (1993). Client socialization: The Achilles Heel of the helping professions. Praeger

Kaeble, D. & Alper, M. (2020). Probation and parole in the United States, 2017-2018. Bureau of Justice Statistics, https://www.bjs.gov/content/pub/pdf/ppus1718.pdf

Labrecque, R. M. (2018). Probation in the United States: A historical and modern perspective. In O. H. Griffin & V. Woodward (Eds.), Handbook of corrections in the United States (pp. 155-164). Routledge.

Labrecque, R. M., Schweitzer, M., & Smith, P. (2013). Probation and parole officer adherence to the core correctional practices: An evaluation of 755 offender-officer interactions. Advancing Practice, 3, 20-23. University of Cincinnati.

Labrecque, R. M., & Smith, P. (2017). Does training and coaching matter? An 18-month evaluation of a community supervision model. Victims & Offenders, 12(2), 233–252. https://doi.org/10.1080/15564886.2015.1013234

Labrecque, R. M., Smith, P., & Luther, J. D. (2015). A quasi-experimental evaluation of a model of community supervision. Federal Probation, 79(3), 14-19. https://www.uscourts.gov/sites/default/files/79_3_3_0.pdf

Labrecque, R. M., Smith, P., Schweitzer, M., & Thompson, C. (2013). Targeting antisocial attitudes in community supervision using the EPICS model: An examination of change scores on the Criminal Sentiment Scale. Federal Probation, 77(3), 15-20. https://www.uscourts.gov/sites/default/files/fed_probation_dec_2013.pdf

Labrecque, R. M., Luther, J. D., Smith, P., & Latessa, E. J (2014). Responding to the needs of probation and parole: The development of the effective practices in a community supervision model with families. Offender Programs Report, 18(1), 1-16.

Latessa, E. J., Smith, P., Schweitzer, M., & Labrecque, R. M. (2013). Evaluation of the effective practices in community supervision model (EPICS) in Ohio [Unpublished manuscript]. University of Cincinnati. http://doi.org/10.13140/RG.2.1.2115.7362

Lerner, M. J. (1977). The effectiveness of a definite sentence parole program. Criminology, 15(2), 211-224. http://doi.org/10.1111/j.1745-9125.1977.tb00062.x

Lipton, D., Martinson, R., & Wilks, J. (1975). The effectiveness of correctional treatment: A survey of Treatment Evaluation studies. Holt Rinehart & Winston.

Lowenkamp, C. T., Holsinger, A., Robinson, C. R., & Alexander, M. (2014). Diminishing or durable effects of STARR? A research note on 24-month re-arrest rates. Journal of Crime and Justice, 37(2), 275-283. http://doi.org/10.1080/0735648X.2012.753849

Malakieh, J. (2020). Adult and youth correctional statistics in Canada, 2018/2019 (Report No. 85-002-X). Statistics Canada. https://www150.statcan.gc.ca/n1/pub/85-002-x/2020001/article/00016-eng.htm

Martinson, R. (1974). What works? Questions and answers about prison reform. The Public Interest, 35, 22-54.

Martinson, R. (1979). New findings, new views: A note of caution regarding sentencing reform. Hofstra Law Review, 7(2), 1-16.

Martinson, R., & Wilks, J. (1977). Save parole supervision. Federal Probation, 41(3), 23-27.

McNeil, F., Raynor, P. & Trotter, C. (Eds.) (2010). Offender supervision: New directions in theory, research and practice. Willan.

National Offender Management Service (2012). Skills for effective engagement, development and supervision (SEEDS) Part 2: The engaging practitioner trainers' manual version one. Ministry of Justice.

Panzarella, R. (2002). Theory and practice of probation on bail in the report of John Augustus. Federal Probation, 66(3), 38-42. https://www.uscourts.gov/sites/default/files/66_3_6_0.pdf

Patenaude-Harris, L. (2013). Parole and conditional release. In J. Winterdyk & M. Weinrath (Eds.), Adult corrections in Canada (pp. 295-323). De Sitter Publications.

Pearson, D. A. S., McDougall, C., Kanaan, M., Bowles, R. A., & Torgerson, D. J. (2011). Reducing criminal recidivism: evaluation of Citizenship, an evidence-based probation supervision process. Journal of Experimental Criminology, 7(1), 73-102. http://doi.org/10.1007/s11292-010-9115-3

Pearson, D. A. S., McDougall, C., Kanaan, M., Torgerson, D. J., & Bowles, R. A. (2016). Evaluation of the Citizenship evidence-based probation supervision program using a stepped wedge cluster randomized controlled trial. Crime & Delinquency, 62(7), 899-924. http://doi.org/10.1177/0011128714530824

Raynor, P. (2018). From “nothing works” to “post-truth”: The rise and fall of evidence in British probation. European Journal of Probation, 10(1), 59-75. http://doi.org/10.1177/20662203187664719

Raynor, P., Ugwudike, P., & Vanstone, M. (2009). The Jersey Supervision Interview Checklist:Version 7C. St. Helier: The Jersey Crime and Society Project.

Raynor, P., Ugwudike, P., & Vanstone, M. (2014). The impact of skills in probation work: A reconviction study. Criminology & Criminal Justice, 14(2), 235-247. http://doi.org/ 10.1177/1748895813494869

Raynor, P., & Vanstone, M. (2018). What matters is what you do: The rediscovery of skills in probation practice. European Journal of Probation, 10(3), 199-214. http://doi.org/ 10.1177/2066220318820199

Rex, S., & Hosking, N. (2013). A collaborative approach to developing probation practice: Skills for effective engagement, development and supervision (SEEDS). Probation Journal, 60(3), 332-338. http://doi.org/ 10.1177/0264550513499002

Reid, W. J., & Epstein, L. (1972). The task-centered system. Columbia University Press.

Robinson, C. J., Lowenkamp, C. T., Holsinger, A. M., VanBenschoten, S., Alexander, M., & Oleson, J. C. (2012). A random study of staff training aimed at reducing re-arrest (STARR): Using core correctional practice in probation interactions. Journal of Crime and Justice, 35(2), 167–188. https://doi.org/10.1080/0735648X.2012.674823.

Robinson, C. J., VanBenschoten, S., Alexander, M., & Lowenkamp, C. T. (2011). A random (almost) study of staff training aimed at reducing re-arrest (STARR): Reducing recidivism through intentional design. Federal Probation, 75(2), 167-188. https://doi.org/10.1080/0735648X.2012.674823

*Rugge, T., & Bonta, J. (2014). Training community corrections officers in cognitive-behavioral intervention strategies. In R. C. Tafrate & D. Mitchell (Eds.), Forensic CBT: A Handbook for Clinical Practice (pp. 122-136). John Wiley & Sons.

*Rugge, T., Bourgon, G., & Bonta, J. (2019). The Strategic Training Initiative in Community Supervision. In R. D. Morgan (Ed.), The SAGE encyclopedia of criminal psychology. Sage.

*Rugge, T., Bourgon, G., Chadwick, N., Wanamaker, K., & Grace, R. (2017). The Living Study phase 1: Insights into community supervision practices of four eastern Ontario offices. Public Safety Canada.

Smith, P., Schweitzer, M., Labreque, R. M., and Latessa, E. J. (2012). ‘Improving probation officers' supervision skills: an evaluation of the EPICS model. Journal of Crime and Justice, 35(2), 189-199. https://doi.org:10.1080/0735648X.2012.674826

Sorsby, A., Shapland, J., Farrall, S., McNeill, F., Priede, C., & Robinson, G. (2013) Probation staff views of the Skills for Effective Engagement Development (SEED) project. Sheffield: Centre for Criminological Research. http://shef.ac.uk/polopoly_fs/1.293093!/file/probation-staff-views-seed.pdf

Sorsby, A., Shapland, J., & Robinson, G. (2017). Using compliance with probation supervision as an interim outcome measure in evaluating a probation initiative. Criminology & Criminal Justice, 17(1), 40-61. https://doi.org/10.1177/174889516653992

Starfelt Sutton, L. C., Dynevall, M., Wennerholm, J., Åhlén, S., Rugge, T., Bourgon, G., & Robertsson, C. (2021). Evaluation of the implementation of a risk-need-responsivity service in community supervision in Sweden. Criminal Justice and Behavior, 48(5), 617-636. https://doi.org/10.1177/0093854820958744

Swedish Prison and Probation Service. (2018). Utvärdering av Krimstics försöksverksamhet 2011–2013 (Report No. 7289). Norrköping: SPPS.

Taxman, F. S. (2008). No illusions: offender and organizational change in Maryland's Proactive Community Supervision efforts. Criminology & Public Policy, 7(2), 275–302. https://doi.org/10.1111/j.1745-9133.2008.00508.x

Thompson, C., & Schweitzer, M. (undated). Use of the EPICS model to reduce recidivism in mentally disordered offenders [PowerPoint Slides]. Corrections Institute, University of Cincinnati.

Trotter, C. (1996). The impact of different supervision practices in community corrections: Cause for optimism. Australian & New Zealand Journal of Criminology, 29(1), 1–19. https://doi.org/10.1177/000486589602900103

Trotter, C. (2002). Worker skill and client outcome in child protection. Child abuse Review, 11(1), 38-50. https://doi.org/10.1002/car.719

Trotter, C. (2006). Working with involuntary clients: A guide to practice 2nd ed. Crows Nest, Australia: Allen & Unwin.

Trotter, C. (2013a). Collaborative family work. A practical guide to working with families in the human services. Sydney: Allen & Unwin.

Trotter, C. (2013b).Reducing recidivism through probation supervision: What we know and don't know from four decades of research. Federal Probation, 77(2), 43-48. https://www.uscourts.gov/sites/default/files/77_2_7_0.pdf

Trotter, C. (2017). Working with families in youth justice. Probation Journal, 64(2), 94–107. https://doi.org/10.1177/0264550517692057

Trotter, C., & Evans, P. (2012). An analysis of supervision skills in youth probation. Australian & New Zealand Journal of Criminology, 45(2), 255-273. http://doi.org/ 10.1177/0004865812443678

Trotter, C., Evans, P., & Baidawi, S. (2020). Collaborative family work in youth justice. Australian Social Work, 73(3), 267-279. http://doi.org/10.1080/0312407X.2019.1618886

Trotter, C., McIvor, G., & Sheehan, R. (2012): The effectiveness of support and rehabilitation services for women offenders. Australian Social Work, 65(1), 6-20. http://dx.doi.org/10.1080/0312407X.2011.641985

Von Hirsch, A., & Albanese, J. S. (1979). Problems with abolishing parole release – the New York case. Criminal Law Bulletin, 15(5), 416-435.

Von Hirsch, A., & Hanrahan, K. J. (1978). Abolish parole? Center for Policy Research.

Walters, S. T., Matson, S. A., Baer, J. S., & Ziedonis, D. M. (2005). Effectiveness of workshop training for psychosocial addiction treatments: A systematic review. Journal of Substance Abuse Treatment, 29(4), 283-293. http://doi.org/10.1016/j.jsat.2005.08.006

Warner, J. J., Rorem, D., Cooke, L., Pealer, J., & Schweitzer Smith, M. (2019). Cross national collaboration in implementing evidence-based practices. Corrections: Policy, Practice and Research. Advance online publication. https://doi.org/10.1080/23774657.2019.1663563

Weinrath, M. (2013). Probation. In J. Winterdyk & M. Weinrath (Eds.), Adult corrections in Canada (pp. 85-110). De Sitter Publications.

Date modified:

2023-06-26