Victims' Experiences with, Expectations and Perceptions of Restorative Justice: A critical Review of the Literature

3. Victims Participating in Restorative Justice Programs

Victims of crime pass through various phases following their victimization. Immediately following the crime, victims often experience anger and fear. Later, most victims pass through a phase of searching for an explanation as to why the crime occurred. Consequently, Reeves (1989) argued that the length of time between the offence and the offer of a meeting is of critical importance for victims. Given the significance of time as a factor, the research presented below is divided into three groups based on the stage in the criminal justice process at which the research took place. Data are presented for programs operating in the pre-court phase (diversion), court-based programs and post-sentencing programs. The chapter concludes with a summary of the findings.

3.1 Diversion

In this section, a number of very different programs are discussed. The common factor among them is that each uses restorative justice practices in order to divert offenders out of the traditional criminal justice system. Typically, these programs take place at the police level and are directed at young offenders who have committed minor offences. However, a few programs, targeting adult offenders and moderately serious offences, have used diversion at a later stage in the criminal justice process. In addition, some of the studies discussed below examine both diversion and court-based programs.

3.1.1 Victim-Offender Mediation Programs

United States

In his book, Victim Meets Offender: The Impact of Restorative Justice and Mediation, Umbreit (1994) examined four different victim-offender mediation programs in the United States. The four programs in the study do not follow a single procedure. Each one is different and, within any one program, different procedures may exist. For example, at one site cases were referred to the program either as diversion (prior to adjudication), after the offender had entered a guilty plea (post-adjudication), or even after the disposition hearing. In all four programs, most of the cases were referred to at the pre-adjudication level (59% to 98%). In two of the three programs, the mediator would meet separately with the offender and the victim and then schedule a mediation session. At one site, the mediator had no prior contact with the parties and the court staff developed the cases for mediation.

All four programs are limited to juvenile offenders, especially young, first-time offenders. Most of the crimes are property offences (73% to 89%), although all four programs did include some minor violent offences. In all four programs, restitution was typically financial, although personal service and community service were also possible.

Umbreit’s research used three groups of respondents:

  1. those referred to mediation and for whom mediation was carried out (experimental group),
  2. those referred to mediation and for whom mediation was not carried out (1st comparison group), and
  3. cases not referred to mediation (2nd
    comparison group).

Umbreit also claimed to have pre-test as well as post-test data; however, pre-test data are not available for all three groups. It is available only for the mediation group (experimental group). Moreover, the pre-test did not take place before the mediation. Therefore, by and large, the design is a so-called “after only” experiment. This design makes it impossible to attribute any observed differences between groups to the treatment they received.

The data on victims are based on interviews. In all, interviews were held with 280 victims who participated in mediation (experimental group), 103 victims whose cases were referred to mediation but who did not participate in mediation (1st comparison group) and 157 victims whose cases were not referred to mediation (2nd comparison group).


Umbreit reported that among the victims participating in mediation, the primary expectation was both to recover their losses and to help the offender. This was followed in frequency by the opportunity to tell the offender the effect of the crime and, finally, getting answers to questions they had about the crime.

Before meeting the offender, one in four victims indicated that he or she was nervous about the pending mediation session with the offender, and nine out of ten victims believed that the mediation session would probably be helpful (it is important to bear in mind that these numbers are based only on those victims who had already agreed to participate in mediation).

For those 103 victims who were referred to a victim-offender mediation program but chose not to participate, their reasons for non-participation in mediation were broken down into three major themes:

  1. the inconvenience of mediation relative to the actual loss,
  2. a number of victims had already worked out a settlement with the offender, and
  3. the victim was too angry to meet the offender and/or disbelieved the offender’s sincerity.

Umbreit reported that across all four sites 91% of victims felt that their participation in mediation was voluntary. In other words, a small but important group of victims felt they had been coerced to participate in the process. One victim thought there was no other way to receive restitution. Another victim felt a burden of responsibility for the offender’s future. Umbreit claimed that feelings of revictimization are often experienced in connection with feeling that they were being coerced into mediation.

Feelings of revictimization can also result from a perception that the mediator is biased toward the offender. The mediator must remain neutral. While most victims (95%) were satisfied with the mediator, a small but significant group of victims was dissatisfied.

Among the victims who participated in mediation, most (90%) were satisfied with the outcome of the mediation session. It is interesting that victims can be dissatisfied with an outcome to which they had freely agreed. This raises the question of whether victims truly felt they had a choice to either accept or reject the offender’s offer.

Comparing pre-test and post-test data for the mediation group, Umbreit found that victims’ priorities did change somewhat. After the mediation session, negotiating restitution, receiving answers and information from the offender, and telling the offender about the effect of the crime were significantly more important to victims than at the time of the pre-test. No significant changes were found for the importance of receiving restitution and receiving an apology.

In terms of the emotional impact of mediation, Umbreit noted that compared to the pre-mediation data, victims who participated in mediation were significantly less often upset about the crime after the mediation session. Similarly, victims were less often afraid of being revictimized by the offender (23% versus 10%). These findings suggest that mediation might have a positive impact on victims’ emotions. However, due to the absence of a control group, it is not possible to be certain that the observed changes are due to mediation and are not caused by something else. It is possible that the observed reduction in fear is simply due to the passage of time.

Regarding satisfaction with how the justice system responded to their case, 79% of the victims in the mediation group indicated satisfaction, 57% in the referred-but-no-mediation group and 57% in the non-referral-to-mediation group indicated satisfaction. Similarly, victims in the mediation group were more likely to state that they had experienced "fairness" in the processing of their case than victims in the comparison groups (83% for the mediation group, 53% for comparison group and 62% for comparison group 2). Umbreit attributed the observed differences between groups to mediation. However, as stated above, his design does not allow him to attribute these or any other observed differences to mediation. It may be that victims in the comparison groups had more negative attitudes toward the justice system to begin with.

Great Britain

In 1988, Marshall and Merry (1990) conducted a largely descriptive study of mediation programs in the United Kingdom. They described this type of research as action research, which examines what happened, versus evaluation research, which looks at what works. Their study, entitled Crime and Accountability: Victim/Offender Mediation in Practice, includes both police-based and court-based schemes. It is based on several sources of information, including interviews with victims: 33 victims who participated in the police-based schemes and 60 victims who participated in the court-based schemes. While the design does not allow the researchers to make causal references, the study does provide interesting insights into victims’ experiences in the early experiments with mediation in England. Marshall and Merry (1990) presented the findings for the police-based and the court-based schemes separately; however, their conclusions refer to both types of programs. As there is a great deal of overlap in victims’ expectations and experiences in both types of programs, here, the findings for both types of programs are presented together.


Victims’ reasons to participate in mediation include:

  • a sense of social responsibility,
  • a desire to help,
  • curiosity to learn why the offender had committed the offence, and
  •  seeking reparation or an apology.

The authors stressed that the expectation of reparation or an apology accounted for only small numbers. However, they pointed out that compensation orders have been available in the English criminal justice system since the 1970s. Therefore, mediation programs focus on reconciliation. Marshall and Merry (1990) asserted that this is quite different from the situation in the United States, and that this might explain why restitution is less important in British programs than in American ones. However, they also noted that compensation orders are under used. The reason judges often fail to impose them is due to ambiguity regarding the amount of the damages. Marshall and Merry mentioned that while parties will probably have more information regarding the extent of the damages available to them than judges, if there is any disagreement about the damages, a wrangling between parties might not be the best way to solve it. They argued that victims would find it inappropriate to enter into personal negotiations in this way and would rather leave it to the court.

The authors emphasized the importance of accurate and realistic information for victims. They reported that some victims may have been (mis)led to believe they had a high chance of compensation. Obviously, when expectations are not met this can lead to disappointment.

In addition, they reported that victim participation was significantly higher in the police-based (79%) versus the court-based programs (51%). The researchers suggested that this might be because the police-based programs dealt solely with juvenile offenders. Victims may have a greater sense of social responsibility when the offender is a juvenile rather than an adult.


Victims were generally appreciative of the open and tactful way they had been approached by the project workers. Participation was voluntary and most victims said they did not feel they had been pressured to participate.

Marshall and Merry reported that most victims who had met their offender were satisfied with the experience and felt that they had done something useful for the offender.

But a “significant minority” (p.152)[2] was dissatisfied. Reasons for dissatisfaction include:

  • high expectations of compensation,
  • feeling they were being asked to do something for the offender rather than vice versa and resenting this, and
  • not being informed regarding the progress of events.

Interestingly, Marshall and Merry (1990) found that victims who were primarily interested in compensation tended to be less satisfied than victims who had acted out of a sense of social responsibility and for whom compensation was merely a token or symbol representing the reconciliation between parties.

Another important source of dissatisfaction was failure by the offender to follow through with agreements. The researchers stressed the importance of follow-up activities for victim satisfaction. Follow-up is important both in terms of reparation (did the offender meet his or her agreement and pay restitution?) and the psychological impact of the programs. A small group of victims identified issues that they considered unresolved after the meeting with the offender. Marshall and Merry suggested that follow-up discussion or counselling by the schemes would help victims cope with lingering uncertainties. However, they found that follow-up of cases among the projects was rare.

Besides reparation, restorative justice programs offer victims psychological benefits. Most victims said that at the time of the offence they felt angry, shocked, depressed or were left feeling foolish. By the time of the interview, virtually all of them felt less strongly or were not concerned at all any more with the offence. While such feelings are likely to diminish over time, many victims said that the scheme had been helpful in this respect.

When asked what it was about the scheme that had proven most valuable, 37% focused on the meeting. However, when asked if they would have been happy to have the meeting only, without the apology or reparation, two-thirds felt this would not have been sufficient. Similarly, the idea of an apology or reparation without meeting was rejected by 65% as inadequate. Apparently, the combination of the two aspects is what seems to make the experience valuable for most of the participants.

Marshall and Merry (1990) found that there is no single time that is good for everyone to meet the offender. They suggested that mediation could be a significant part of the support given to victims in the resolution of their problems. It should be offered at a time when feelings are neither extremely high nor diminished altogether, but should be part of a more general plan of support for emotionally affected victims.


Recently, the Research and Statistics Division of the Department of Justice published a meta-analysis on the effectiveness of restorative justice practices. The study, by Latimer, Dowden and Muise (2001), is based on an analysis of 35 program evaluations from around the world. The study included a wide variety of programs. The study is discussed here because the majority (77%) of the programs are victim-offender mediation programs. While most (57%) programs included several points of entry into the criminal justice system, the largest single group was at the pre-charge stage (20%).

According to the authors, only studies that used a control group or a comparison group that did not participate in a restorative justice program were included in the study. The methodology used for this study is known as a post-test only design with non-equivalent groups (Cook and Campbell, 1979). This design is relatively common; however, it does not permit reasonable causal inferences. According to Cook and Campbell:

Its most obvious flaw is the absence of pre-tests, which leads to the possibility that any post-test differences between groups can be attributed either to a treatment effect or to selection differences between the different groups. The plausibility of selection differences in research with non-equivalent groups usually renders the design uninterpretable. (1979: 98-99)

In programs like mediation, which are based on voluntary participation, the likelihood of selection differences is great. Therefore, it is not possible to draw causal inferences based on studies using this design. While the authors addressed the issue of self-selection bias and suggested that future research include pre-test measures to avoid this problem, they nonetheless drew conclusions about the effects of the interventions.


Latimer et al. compared research findings concerning victim satisfaction. They concluded that “participation in a restorative justice program resulted in higher victim satisfaction ratings when compared to a comparison group in all but one of the thirteen programs examined” (2001: 12). They found this difference to be significant. Unfortunately, the authors did not specify which 13 studies were used to draw this conclusion, making it difficult to verify. Also, given their selection criteria regarding the designs of the studies, it is plausible that the designs of many, if not all of the studies, did not meet the necessary rigorous requirements to draw causal inferences. In order to attribute any observed differences between groups to their participation in a program, one would have to rule out the possibility, for example, that the persons with more favourable attitudes were more likely to participate in restorative justice programs. It would appear that the authors drew conclusions about the effects of participation in restorative justice programs based on studies that did not meet the above selection criteria.

[2] This study is qualitative and the authors did not specify exact number or percentage of victims. See Marshall and Merry, 1990: 152.

Date modified: