Bonta J., Bourgon G., Rugge T. et al.
Criminal Justice and Behavior: 2011, 38, p. 1127–1148.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Bonta at Jim.Bonta@ps.gc.ca. You could also try this alternative source.
For the first time this Canadian randomised study has shown that training probation officers in the risk-need-responsivity model of offender supervision can not only improve their skills and sharpen their practice, but also reduce the recidivism of the offenders they supervise, among whom substance use was a major issue.
Summary The risk-need-responsivity model has been very influential in guiding treatment interventions with criminal offenders. The model's three core principles are:
• Risk: providing intensive services to clients at higher risk of reoffending and minimal services to lower risk clients;
• Need: target criminogenic needs or the dynamic risk factors which underlie or drive criminal behaviour such as pro-criminal attitudes and substance use;
• Responsivity: match the style and mode of intervention to the abilities, motivation, and learning style of the offender; cognitive-behavioural interventions are generally the most effective. A synthesis of research findings found that as adherence to these principles increased, there was a stepwise increase too in the reduction in recidivism.
Paucity of evidence that supervising offenders in the community reduces recidivism may be related to the fact that probation officers do not implement these principles and rarely use cognitive-behavioural techniques. The featured study which took place in Canada was the first randomised trial to evaluate whether training for probation officers based on risk-need-responsivity principles changes how they deal with offenders and reduces their recidivism.
A three-day training programme featuring interactive teaching and role plays was developed which incorporated all three core principles, with a particular emphasis on targeting problematic attitudes and thoughts using cognitive-behavioural techniques. Audiotapes of offender supervision sessions were used to assess adherence to the principles targeted by the training and other features of the interaction between probation officers and offenders which might influence recidivism. Training was supplemented by a refresher course a year later and ongoing clinical support, including exercises based on audiotape extracts and feedback and coaching from the trainers.
Of the 80 probation officers who volunteered for the study, 51 were randomly assigned to be trained and 29 to an untrained control A group of people, households, organisations, communities or other units who do not participate in the intervention(s) being evaluated. Instead, they receive no intervention or none relevant to the outcomes being assessed, carry on as usual, or receive an alternative intervention (for the latter the term comparison group may be preferable). Outcome measures taken from the controls form the benchmark against which changes in the intervention group(s) are compared to determine whether the intervention had an impact and whether this is statistically significant. Comparability between control and intervention groups is essential. Normally this is best achieved by randomly allocating research participants to the different groups. Alternatives include sequentially selecting participants for one then the other group(s), or deliberately selecting similar set of participants for each group. group against whom findings from trained officers could be benchmarked. Following training, all 80 were asked to submit audiotaped sessions with accompanying information on new clients, but only 52 did so, contributing the 295 audiotapes analysed by the study. Of the offenders they supervised, 183 (56%) agreed to join the study but 40 later became unavailable, leaving 143 whose reconvictions were tracked for on average two and a fifth years from the first audiotape recording. Also obtained was the reconviction rate of a random sample of each officer's pre-training clients – 185 in all whose convictions records were obtained for on average three and a half years after their first sessions with their probation officers.
This data – available from 46 of the 52 officers – enabled the researchers to tell whether offenders whose officers had been trained went on to be reconvicted less often than the same officers' pre-training caseload, and (crucially) whether this reduction exceeded that among clients of untrained officers. Any excess reduction could be attributed to the effects of the training. If this happened, analysis of the tapes might shed light on what made the difference. Trained officers' uptake of post-training support was also monitored to check if this was related to their performance.
As well as identifying whether a topic was addressed at all in a session, each five-minute segment of each session was analysed and coded for the different topics addressed at least twice To eliminate casual references. within a segment. The tapes revealed that training had substantially affected how the officers conducted supervision. As intended, training had focused sessions on underlying causes of crime, while untrained officers spent more time on the probation conditions the offender had to comply with, and needs not identified as related to the offender's criminality.
In particular, pro-criminal attitudes were rarely raised by untrained officers (mentioned in just 2% of sessions and occupying 8% of session time) but were significantly more prominent in sessions by trained officers (35% of sessions occupying 13% of session time). On average, The main exception was criminogenic needs specifically related to employment and education, which untrained officers were more likely to raise in a session and spent more time on. trained officers spent more time in each session discussing the criminogenic needs (58% of segments v. 43%) and pro-criminal attitudes (13% of segments v. 3%) identified as possibly underlying the offending of that offender. Substance use as a criminogenic need was discussed in over 60% of sessions and in about a quarter of segments, but in this case no more often and for no longer after training.
As well as the content of each session, raters judged the skill with which the officers addressed these and other issues within four categories. Under all four headings, trained officers were judged more skilful, in three cases significantly so. The three were how well they structured a session (for example, checking for crises and prioritising needs), fostering a good client-officer relationship (for example, by actively listening to the client), and the application of cognitive-behavioural techniques, especially when targeting pro-criminal attitudes. Also non-significantly better were their 'behavioural' techniques including use of reinforcement and disapproval to shape the offender's behaviour. Together these four summed to significantly better correctional skills among the trained officers.
Training had focused especially on applying cognitive-behavioural techniques to pro-criminal attitudes and ways of thinking, so these skills were further analysed. As expected, many more trained officers – three quarters versus just one in ten – had discussed pro-criminal attitudes with their clients, and 70% versus just 5% had employed cognitive techniques.
The first set of recidivism analyses focused on the subset of offenders for whom there were two years of follow-up data on reconvictions. Over these two years, 25% of offenders supervised by trained officers had been reconvicted. Before training, 47% of clients of the same officers had been reconvicted chart. This post-training reduction in reoffending was statistically significant. In contrast, among untrained officers, in both time periods the reconviction rate was just over 40%. Though differences between trained and untrained officers were not statistically significant, these figures reveal a post-training reduction in offending apparent among trained but not untrained officers, a finding suggesting that training had made the hoped-for difference.
A second set of analyses adjusted Also adjusted for were the offenders' age and risk of re-offending. for different follow-up periods, so could include the reconviction records of all the offenders. Again, the proportion of offenders supervised by trained officers who were reconvicted was significantly lower after they had been trained, but there were no significant difference between them and offenders supervised by control group officers.
Another way to segment the offenders was to contrast those supervised by officers after they had been trained, with offenders supervised by the same or different officers not (or not yet) trained. Over two years the difference in the proportions reconvicted was a statistically significant 25% versus 44%. This finding was confirmed by an analysis which adjusted Also adjusted for were the offenders' age and risk of re-offending. for different follow-up periods, so could include the reconviction records of all the offenders.
These analyses suggest that reoffending was reduced when offenders were supervised by trained officers. Digging deeper, the researchers probed what it was about how the officers behaved which might have made the difference. Using cognitive techniques to alter pro-criminal attitudes was the sole factor. In the two years after training, 10% of the 42 offenders (all but one supervised by trained officers) subject to these techniques had been reconvicted compared to 37% of the 70 offenders with whom they had not been used. Also, the more these techniques had been used, the less likely was an offender to be reconvicted. No such links were found with the other types of skills assessed by the study – forging relationship, structuring sessions, or behavioural techniques.
For the first time at the level of the offender being supervised one-to-one by a probation officer, the study has confirmed that adherence to risk-need-responsivity principles is associated with decreased recidivism, and that correctional staff can effectively be trained to apply these principles during offender supervision. The probation officers in the study improved their targeting of criminogenic needs, especially pro-criminal attitudes, and learnt and used specific cognitive-behavioural intervention skills to help clients change their attitudes, beliefs and ways of thinking, which in turn decreased the likelihood of a return to crime.
Compared to their untrained colleagues, trained officers spent more time and/or addressed these issues in more sessions, while untrained officers spent more time on non-criminogenic needs and probation conditions. It meant that after training, the balance of the sessions was more in tune with risk-need-responsivity principles. Moreover, training improved skills needed to effectively address these issues, including building a relationship and the use of cognitive-behavioural techniques. These techniques were evident in the work of only one of the untrained officers but 7 in 10 of those trained.
Random allocation of officers to be trained or not means it can confidently be concluded that these improvements were due to the training. However, crediting the training with reducing recidivism is complicated by the fact that monitored offenders were selected by the officers, who chose to include them in the study by submitting audiotapes and data. But if this biased the findings, there is no reason to believe any such bias would have differed between trained and untrained officers, yet lower recidivism was only apparent among offenders supervised by trained officers. Consistent findings across several different types of analyses all indicate that training officers reduced recidivism among their supervisees, findings in line with the association between risk-need-responsivity principles and recidivism in studies which observed real-world practice.
The data suggest two possible links between how the training affected the officers and criminality among probationers. First, use of cognitive techniques was associated with lower recidivism and seems to have been one mechanism via which training reduced crime. Second, but in the opposite direction, the more that probation officers discussed probation conditions, the higher the recidivism rate. It seems that preoccupation with the conditions of probation, or the enforcement role of the probation officer, interferes with the establishment of a therapeutic relationship, obstructing more directive intervention. Trained officers spent less time on this enforcement role.
Finally, the findings suggest that ongoing skill development is important. Officers who received more clinical feedback and attended more of the monthly meetings and the refresher courses were more likely to demonstrate the skills taught in training and to focus their discussions on matters of importance, such as criminogenic needs.
The fact that none of the recidivism differences between offenders supervised by trained and untrained officers reached statistical significance was probably due to small sample sizes, but this would need confirming in a larger study. It should also be borne in mind that the officers volunteered for the study and therefore too for the training. Whether such training would have a similar impact if routinely implemented remains to be seen.
commentary The featured study should be distinguished from those of specialist drug or alcohol interventions for offenders such as the ASRO group programme for problem substance users (itself based on cognitive-behavioural principles) or referral to specialist treatment. Instead it is about improving core probation supervision by the general run of officers. While the study was not specifically about substance use or users, the fact that substance use as a contributor to offending was discussed in over 60% of sessions and in about a quarter of session segments suggests it was a common and important issue among this caseload.
The authors are commendably cautious about attributing recidivism reductions to the consequences of the training, but the fact that offenders seen by untrained officers were reconvicted at the same rate before and after other officers had been trained, yet these trained officers saw a near halving in the proportion reconvicted, is strong evidence that training was the active ingredient. Training seems to have embodied effective methods including interactive exploration of the topic, role play practice, feedback on actual performance, and continued support 'pushed' to the officers rather than left for them to access on their own initiatives.
While fewer reconvictions can fairly confidently be claimed to be the consequences of training the officers in the study, the reservation that they self-selected participation is potentially an important limit on the generalisability of the findings. Training which looks effective among trainees already in tune with its messages can look less effective if it falls on stonier ground in the form of unreceptive trainees. However, the Canadian province of British Columbia was sufficiently convinced to implement the training tested in the featured study across its probation service.
The conclusion that cognitive-behavioural methods were the key mechanism via which training improved probation's impact on crime rests on shakier foundations than the impact of the training itself. Since with just one exception only trained officers used these methods, their use was effectively coincident with training overall. This makes it impossible to disentangle the impact of training overall with the impact of this specific element in the training. Similarly, the fact that the more often officers used these methods, the more recidivism was affected, could reflect the probability that it was these officers who had benefited most from training overall or were best able due to their predispositions and work situations to put it in to effect. The same factors may have improved their recidivism outcomes regardless.
Even when in a controlled study a cognitive programme been found effective, this has not necessarily been maintained in a larger scale roll-out, as found in British prisons. Interventions for offenders are, it has been argued, highly context-specific; what works in one culture at one time may well be ineffective in other settings and at other times. The intervention tested by the featured study may have sidestepped this limitation to a degree, because the training was not about implementing a set 'programme', but applying broad principles and ways of fostering change.
The featured study's characterisation of the relatively ineffective untrained officers as focusing more on probation conditions is worryingly reminiscent of the findings in a British study of services probably sampled in the mid-2000s. Perfunctory brief encounters focused on dose, prescribing and dispensing arrangements, attendance records, and regulatory and disciplinary issues were the main features of keyworking service offered by some British criminal justice teams to offenders on opiate substitute prescribing programmes.
Thanks for their comments on this entry in draft to Tim McSweeney of the Institute for Criminal Policy Research at Birkbeck College in London. Commentators bear no responsibility for the text including the interpretations and any remaining errors.
Last revised 11 April 2012. First uploaded
DOCUMENT 2013 Sometimes best to break the rules
REVIEW 2006 My way or yours?
REVIEW 2005 The motivational hallo