Skip to main content

An evaluation of the implementation and effectiveness of an initiative in restorative cautioning

An evaluation of the first large-scale restorative justice programme in the UK ‘Restorative justice’ seeks to encourage offenders to take active responsibility for repairing the harm caused by crime.

Written by:
Carolyn Hoyle, Richard Young and Roderick Hill
Date published:

In 1998, Thames Valley Police took the bold decision to transform its cautioning practices through its restorative cautioning initiative. Proceed with caution documents the findings of a three-year study conducted in the Thames Valley, and provides a unique insight into the development and achievements of the first large-scale restorative justice programme in this country.

The report considers possibilities for improving restorative cautioning using action research that takes into account procedural fairness. It discusses both participants’ views of the restorative session and the longer-term aims of the project. The report draws attention to the difficulties the police experienced in implementing the restorative justice model and to the improvements in practice during the project.

By the end of the project, implementation was much better, although still not always good. Nonetheless, offenders, victims, and their respective supporters were generally well satisfied with the fairness of the process and the outcomes achieved. In addition, restorative cautioning appeared to be more effective than ‘old-style’ cautioning in reducing the risk of re-offending.


In 1998, Thames Valley Police launched a restorative cautioning initiative, whereby police officers administering cautions were meant to invite all those affected by the offence, including victims, to a meeting. The police officer used a script to facilitate a structured discussion about the harm caused by the offence and how this could be repaired. Richard Young and Carolyn Hoyle of Oxford University helped the police to implement this new way of cautioning, and carried out a formal evaluation of the process and the outcomes achieved. The researchers found that:

  • Thames Valley Police was largely successful in transforming its cautioning practice. The restorative justice script was used in over two-thirds of cautions.
  • Over the first three years of the initiative, 1,915 restorative conferences took place at which victims were present. In a further 12,065 restorative cautions, the views of any absent victims were relayed by the cautioning officer. This is the largest-scale restorative justice programme in the United Kingdom to date.
  • Implementation of the restorative cautioning model in individual cautions was often deficient. Police facilitators sometimes sidelined the other participants and occasionally asked illegitimate questions.
  • By the end of the research project implementation was much better, although still not always good.
  • Offenders, victims and their respective supporters were generally satisfied with the fairness of the process and the outcomes achieved.
  • Cautioning sessions that adhered most closely to restorative justice principles tended to produce the most positive outcomes.
  • Restorative cautioning appears to be significantly more effective than traditional cautioning in reducing the risk of re-offending.


A police caution is the formal disposal of a criminal case without the involvement of prosecutors or the courts. In recent years, about a third of all criminal cases resulting in a criminal record have been disposed of by a police caution rather than a court conviction.

According to Home Office guidelines, in a traditional police caution an officer is supposed to warn an offender that the disposal will be kept on record and can influence future prosecution and court decisions. Young offenders are required to have an appropriate adult with them in a cautioning session, but victims and others affected by the offence are not invited.

Previous research had shown that cautioning sessions were sometimes used to humiliate and stigmatise offenders. Thames Valley Police officers interviewed for the present study confirmed that in traditional cautions the usual aim was to give offenders a "bollocking" and to make them cry. Officers said that there had been no training in how to administer a caution, no supervision of practice, and no expectation of consistency.

In 1998, Thames Valley Police decided to transform its cautioning practices, for both juveniles and adults, through a restorative cautioning initiative. Police officers administering cautions were trained to use a script in order to facilitate a structured discussion about the harm caused by the offence and how this could be repaired. Under the new model, they were meant to invite all those affected by the offence, including any victims, to the cautioning session. As part of the process of helping the police to implement the model successfully, the researchers studied 23 cases in depth for an interim evaluation in 1999, and a further 56 cases for a final evaluation a year later.

An impressive transformation

Thames Valley Police was largely successful in transforming its practices. Its records showed that the restorative justice script was used in over two-thirds of cautions. In each of the initiative's first three years, over 600 restorative conferences took place at which victims were present, as well as several thousand restorative cautions in which the cautioning officer relayed the views of any absent victims. This is the largest-scale restorative justice programme in the United Kingdom to date.

The police achieved an impressive degree of success in engineering a shift away from low visibility, idiosyncratic, sometimes overtly stigmatising police behaviour, towards more consistent practice under conditions of greater visibility and accountability, according to definite aims and standards.

Deficient implementation

Implementation of the restorative cautioning model in individual cautions was often deficient, particularly in the cases in the interim evaluation. Police facilitators tended to dominate the discussion, prioritising their own agendas rather than those of the other participants.

In the worst examples, officers reinvestigated the offence, sought admissions to prior offending, and asked questions that appeared to be attempts to gather useful criminal intelligence. More frequently, facilitators behaved as if the offender had to account to them personally, with the other participants reduced to little more than passive observers. Officers also overstepped their remit by extracting apologies from offenders in a fairly coercive way and by pursuing their own ideas about reparation rather than enabling a discussion with those affected by the offence.

Improving facilitators' practice

The report of the interim evaluation documented these deficiencies in implementation. It was widely disseminated within Thames Valley Police and led to top-up training and a revised script which exhorted facilitators not to pursue a policing agenda within the cautioning session.

The final evaluation found that this remedial activity resulted in a much greater use of the restorative justice script, although significant deficiencies in facilitators' practice remained evident. Whereas in over half of the cases in the interim study the facilitator asked the offender irrelevant or improper questions, this occurred in only a quarter of cases in the final evaluation. The views of participants in the final evaluation are considered below.

Fairness of the process

Over a third of the participants were not contacted directly by the facilitator prior to the restorative cautioning session. This meant that they had little chance to think about who they might like to come to the session, what they might want to get out of it, how they might feel when faced with the other participants, and what they might say or ask. Victims who chose not to participate rarely made a fully informed decision. Many offenders and their supporters had little understanding of the process they were entering and felt that they had no choice but to participate. Some victims and their supporters were also confused about the purpose of the meeting.

Despite these problems, most participants reported that they were satisfied with various aspects of the restorative process. In particular, they felt that the process was fair because they were given the opportunity to say what they wanted to say. It was particularly important to a number of offenders that they were provided with the same opportunity to speak as everyone else, and that they were listened to with a degree of respect. Two-thirds of the participants felt that the meeting was well facilitated; victims and their supporters were particularly satisfied.

Outcomes achieved

The cautioning sessions were generally successful in achieving the multiple short-term aims of the restorative cautioning initiative. By a large majority, participants believed that the meeting helped offenders to understand the effects of the offence and induced a sense of shame in them. By an even larger majority, offenders confirmed that this was so. In most cases, apologies were offered or arranged and these gestures were mostly seen as a manifestation of genuine remorse. Over half of the participants gained a sense of resolution and felt better because of the restorative session, and four-fifths saw holding the meeting as a good idea. As one victim put it:

Before, yeah, I'd love to have met them down a dark alley, sort of thing, but now you've had the meeting with them and you've spoke about it, you feel totally different. You actually find out what they're like as people. You thought they were animals to start off with, but you find out they're actually normal people same as anyone else. (Victim)

However, a significant minority of apologies were perceived to have been coerced, and nearly two-fifths of offenders said that they felt the meeting made them feel like a bad person - something the process was meant to avoid.

In a substantial minority of cases, a link was found between the restorative cautioning session and longer-term positive outcomes. Notably, 28 per cent of offenders reported at least some improvement in their relationships with friends or family, and a few victims had clearly gained a great deal from the experience of meeting their offender. Almost a third of offenders entered into a formal written reparation agreement at the restorative caution. Within a year, the majority had been completely fulfilled and only three remained completely unfulfilled.

Linking process to outcome

High-quality facilitation produced the most restorative results. Participants exposed to high-quality facilitation were most likely to feel that they had experienced a fair process. They were also most likely to believe that the meeting made offenders feel ashamed of their criminal behaviour and helped them to understand the effects of that behaviour on others. Looking just at offenders, high-quality facilitation appeared to be linked to improved relationships with the police, and to reductions in criminal behaviour.


The Thames Valley model was based in part on the theory that the best way to control crime is to induce in offenders a sense of shame for their actions, while maintaining respect for them as people. This was seen as being best achieved by exposing offenders to the emotionally charged views of those whom they most care about, such as parents, partners and friends. The researchers used self-report instruments to measure actual offending in the year before and the year following the restorative caution. They also examined detected offending for these periods using data from the Police National Computer. About a quarter of offenders were helped by the restorative session to stop or reduce their actual offending. Of the 56 offenders aged ten to 17 in the final evaluation sample, just 14 per cent were re-sanctioned within a year of the restorative session. Based on a comparison with a study of the effectiveness of traditional cautioning, this suggests that restorative cautioning halved the likelihood of re-sanctioning within a year. A larger-scale re-sanctioning study of restorative cautioning is required to confirm the validity of this finding.


Thames Valley Police largely succeeded in transforming its cautioning practices from traditional cautioning to restorative cautioning. In particular, it eradicated much of its earlier poor practice in a relatively short period of time between the interim study and the final evaluation. While there was considerable room for further improvement, the findings suggest that even restorative sessions that were less well facilitated were a substantial improvement on traditional cautions. However, the best results were achieved in cases in which facilitators adhered most closely to the restorative justice script. The results of the evaluation therefore provide a new incentive for Thames Valley Police to intensify the process of securing full implementation of its model.

Nearly all participants came away from a restorative session feeling that it was a good idea to meet with the other participants in the presence of a facilitator trained in restorative justice thinking. Participants generally saw restorative cautioning as fair, and believed it to be successful in most of its short-term aims. During the twelve months following the restorative caution, they remained broadly positive and a substantial minority reported longer-term benefits.

In a period of great change for the criminal justice system, restorative measures are being introduced at various levels to various degrees. Following implementation of the Crime and Disorder Act 1998, and the Youth Justice and Criminal Evidence Act 1999, a few other police services and all youth offender teams have introduced restorative justice measures into their responses to youth crime. Other agencies are running experimental schemes for adult offenders. The main messages of this research, concerning the importance of neutral facilitation, procedural fairness and adequate preparation before any intervention, apply to all of these different programmes and initiatives under the restorative justice banner.

About the project

The fieldwork for this study took place in Aylesbury, Banbury and Reading. The interim evaluation was based on the observation and tape-recording of 23 restorative cautions and conferences held in 1999 and on 135 interviews with the participants in these cases. The final evaluation in 2000 studied a further 56 cases in an identical way, with 483 more interviews. The key participants were spoken to just after the restorative session, a few months later and, finally, after a year had elapsed. The cases studied in depth were found to be reasonably representative of practices taking place in Thames Valley Police over this period.


542.pdf (44.95 KB)
1859353819.pdf (339.17 KB)