Crjp Impact Restorative Justice Conferencing 63 Studies May 2002
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
www.rjp.umn.edu Center for Restorative Justice & Peacemaking An International Resource Center in Support of Restorative Justice Dialogue, Research and Training The Impact of Restorative Justice Conferencing: A Review of 63 Empirical Studies in 5 Countries Mark S. Umbreit, Ph.D. Robert B. Coates, Ph.D. Betty Vos, Ph.D. May 1, 2002 Conceptually, the restorative justice paradigm begins with the notion that crime is an act against people and a violation of relationships as well as a breaking of the law (Zehr, 1990). Restorative justice has become a framework for thinking about ways of humanizing justice, of bringing victims and offenders together in ways that provides opportunity for victims to receive explanation and reparation and for offenders to be accountable to the victim and the community, and of involving community members meaningfully in helping repair the wrong done to their neighborhoods. This shift in thinking away from the traditional punitive models of justice is also referred to as community justice (Griffiths and Hamilton, 1996; Stuart, 1995; Barajas, 1995), and restorative community justice (Young, 1995; Bazemore and Schiff, 1996). It remains to be seen whether these umbrella concepts are identical or whether significant shades of emphasis will emerge to further differentiate them. It is possible for example that some may be more victim driven and some more offender driven. It is also possible that certain components become more central in one framework compared with another. For example, Llewellyn and Howse (1998) identify voluntariness, truthtelling and a face-to-face encounter between victim and offender as the main elements of the restorative justice process. The face-to-face encounter between the victim and offender, a long standing centerpiece of victim offender mediation (although even there, in practice, there are numerous exceptions), does not seem to be as much of a priority in family group conferencing, and may be non-existent in some peacemaking and sentencing circle work. It is likely that conceptual underpinnings of restorative justice are being modified some by the experience of practice. Conceptual frameworks, hopefully, shape practice and well-documented practice will in turn bring about a refinement in conceptual frameworks. Here, we will use the term restorative justice conferencing which encompasses a number of practice approaches to justice. Perhaps best known are victim-offender mediation, family group conferencing, and peacemaking circles. In recent years, a blurring has taken place particularly among practitioners who are attempting to apply restorative concepts and processes to real life settings. The distinctions between and within the three modalities is less clear and from some points of view less important. Victim offender mediation which in its beginnings usually brought a single victim and a single offender together with a community mediator now works with as many victims, offenders and support persons (including parents) as seems relevant for the case. Family group conferencing brings together family members of victims and offenders as well as the victim and offender, and other community support members ©Regents of the University of Minnesota Center for Restorative Justice & Peacemaking School of Social Work, College of Education & Human Development may also be present. Circles explicitly intend to gather not only victim, offender, and family members but also a number of community members including some who are unknown to the victim and offender families. Increasingly, practitioners describe what they are doing as conferencing. The characteristics of the case and the nature of the underlying conflict as well as to a certain extent the desires of the victim and offender determine whether the conferencing process used resemble more the victim offender mediation, family group conferencing, or circle model. And any one case may lead to using each approach at different stages. Considerable empirical work has been done over the past twenty years or so to document the impact of programs attempting to implement restorative justice concepts. Here, we take a look at how this ongoing experiment with restorative justice conferencing is doing. We will consider client satisfaction, fairness, restitution, diversion, recidivism and cost. We will also reference a meta-analysis approach to some these questions recently carried out by a Canadian group (Lattimer, Dowden, and Muise, 2001) which offers considerable promise. A total of 63 empirical studies of restorative justice conferencing, from 5 countries, were reviewed. This included 46 studies of victim offender mediation, 13 family group conferencing studies, and 4 assessments of peacemaking circles. CLIENT SATISFACTION. Victim offender mediation proponents often speak of their efforts as ways of humanizing the justice system. Traditionally, victims were left out of the justice process. Neither victims nor offenders had opportunities to tell their stories and to be heard. The state somehow stood in for the victim, and the offender seldom noticed that his or her actions impacted real, live people. In addition, victims, too, were left with stereotypes to fill their thoughts about offenders. VOM, reformers believed, offered opportunities for both parties to come together in a controlled setting to share the pain of being victimized and to answer questions of why and how. This personalizing the consequences of crime, it was thought, would enhance satisfaction levels with the entire justice process. The vast majority of studies reviewed reported in some way on satisfaction of victims and offenders with victim offender mediation and its outcomes. Across program sites, types of offenders, types of victims, and cultures high levels of participant satisfaction were found. Before exploring the nature of this satisfaction further, it should be noted that forty to sixty percent of persons offered the opportunity to participate in VOM refused. Typically, these refusals came from victims who 1) believed the crime to be too trivial to merit the time required, 2) feared meeting the offender, or 3) wanted the offender to have a harsher punishment (Coates and Gehm, 1985; Umbreit, 1995). Ghem, in a study of 555 eligible cases, found 47% of the victims willing to participate (Gehm, 1990). Victims were more likely to participate if the offender was white, if the offense was a misdemeanor, and if the victim was representing an institution. Offenders were sometimes advised by lawyers not to participate (Schneider, 1986). And some simply didn't want "to be bothered" (Coates and Gehm, 1985). The voluntary nature of participating in VOM is a self-selection factor overlaying these findings. The high levels of satisfaction may have something to do with the opportunity to choose. Perhaps those who are able to choose among justice options are more satisfied with their experiences. Several studies noted that victim willingness to participate was driven by a desire to receive restitution, to hold the offender accountable, to learn more about the why of the crime and to share their pain with the offender, to avoid court processing, to help the offender change behavior, or to see that offender adequately punished. Offenders choosing to participate often wanted to "payback the victim" and "to get the whole experience behind them" (Coates and Gehm, 1985; Perry, Lajeunesse, and Woods, 1987; Umbreit, 1989; Roberts, 1995; Umbreit, 1995; Niemeyer and Shichor, 1996; Strode, 1997; Umbreit, Coates and Vos, 2001). Expression of satisfaction with VOM is consistently high for both victims and offenders across sites, cultures, and seriousness of offenses. Typically, eight or nine out of ten participants report being satisfied with the process and with the resulting agreement (Davis, 1980; Coates and Gehm, 1985; Perry, Lajeunesse, and Woods, 1987; Marshall, 1990; Umbreit, 1991; Umbreit and Coates, 1992; Warner, 1992; Roberts, 1995; Carr, 1998; Roberts, 1998; Evje and Cushman,2000). For example, a recent multi-site study of victim offender mediation in six counties in Oregon found an aggregate offender satisfaction rate of 76% and an aggregate victim satisfaction rate of 89% (Umbreit, Coates and Vos, 2001). Even in an England based study (Umbreit and Roberts, 1996) which yielded some of the lowest satisfaction scores among the studies reviewed, eighty-four percent of those victims engaged in face-to-face mediation were satisfied with the mediation outcome. For those individuals involved with indirect mediation, depending on shuttle mediation between parties without face-to-face meetings, seventy-four percent were satisfied with their experience. These findings were consistent with an earlier study based in Kettering where a small sub-sample of participants were interviewed indicating sixty-two percent of individual victims and seventy-one percent of corporate victims were satisfied (Dignan, 1990). About half of the offenders responding reported being satisfied. Participants involved in face-to-face mediation were more satisfied than those who worked with a go between. Victims often reported being satisfied with the opportunity to share their stories and their pain resulting from the crime event. A victim stated she had wanted to "let the kid know he hurt me personally, not just the money...I felt raped" (Umbreit, 1989). Some pointed to their role in the process with satisfaction. One victim said: "we were both allowed to speak...he (mediator) didn't put words into anybody's mouth" (Umbreit, 1988). Another female victim indicated, "I felt a little better that I've stake in punishment" (Coates and Gehm, 1985). Another indicated that "it was important to find out what happened, to hear his story, and why he did it and how" (Umbreit and Coates, 1992). Numerous victims were consumed with the need for closure. A victim of violent crime indicated that prior to mediation, "I was consumed with hate and rage and was worried what I would do when he got out" (Flaten, 1996). Of course not all victims were so enamored with the process. A male victim complained: "It's like being hit by a car and having to get out and help the other driver when all you were doing was minding your own business" (Coates and Gehm, 1985). A Canadian stated: "Mediation process was not satisfactory, especially the outcome. I was not repaid for damages or given compensation one year later. Offender has not be adequately dealt with. I don't feel I was properly compensated" (Umbreit, 1995). Offenders generally report surprise about having positive experiences. As one youth said, "He understood the mistake I made, and I really did appreciate him for it" (Umbreit, 1991). Some reported changes: "After meeting the victim I now realize that I hurt them a lot...to understand how the victim feels makes me different" (Umbreit and Coates, 1992). One Canadian offender stated his pleasure quite succinctly: "Without mediation I would have been convicted" (Umbreit, 1995). The following comment reflects the feelings of some offenders that victims occasionally abused the process: "We didn't take half the stuff she said we did; she either didn't have the stuff or someone else broke in too" (Coates and Gehm, 1995). Secondary analysis of satisfaction data from a US study and a Canadian study yielded remarkably similar results (Bradshaw and Umbreit, 1998; Umbreit and Bradshaw, 1999). Using step-wise multiple regression procedures to determine those variables most associated with victim satisfaction, the authors discovered that three variables emerged to explaining over 40% of the variance. In each study, the key variables associated with victim satisfaction were 1) victim felt good about the mediator, 2) victim perceived the resulting restitution agreement as fair, and 3) victim, for whatever reason, had a strong initial desire to meet the offender. The latter variable supports the notion that self-selection and choice are involved in longer run satisfaction. These findings also underscore the important role of the mediator, and, of course, the actual outcome or agreement resulting from mediation. These high levels of satisfaction with victim offender mediation also translated into relatively high levels of satisfaction with the criminal justice system. Where comparison groups were studied, those victims and offenders going through mediation indicated being more satisfied with the criminal justice system than those going through traditional court prosecution (Davis, 1980; Umbreit and Coates, 1992; Umbreit, 1995). Family Group Conferencing. Family Group Conferencing also yields fairly high satisfaction responses from participants. Less than six out of ten victims in New Zealand (Maxwell and Morris, 1993) were satisfied with their family group conferencing experience compared with more than nine out of ten in the United States (Fercello and Umbreit, 1998). These varying levels of satisfaction may reflect differences in culture within as well as across program. Authors of the New Zealand study, in particular, note differences among cultures within their sample (Maxwell and Morris, 1993). While only fifty-three percent of the victims were satisfied with the outcome of their cases, eighty-four percent of the offenders were satisfied. It should be noted that only forty-one percent of the victims attended the conference. Victims often attributed their lack of attendance to not having enough lead time to make the necessary arrangements. Although 95% of the cases in the New Zealand program reported some form of agreed upon outcome, victim comments regarding process and outcome were quite varied: "It is a soft option. He needs jail. It was very serious. I could have been killed." "It's lenient. He's only paying $20 a week. I had to pay out cash and lose interest. It's me that suffers -- it's not enough." "I got the ill feelings out of my system." "I felt able to understand the girl and her problems -- she was receptive." "The crime stinks, but the punishment stinks more." "The first family had up to twelve (support persons present for offender) and there was just me -- it was unbalanced." (Maxwell and Morris, 1993). In an Australian study, Daly (2001) noted that ninety percent of the offenders and 73% of the victims were either satisfied or very satisfied with how their cases were handled. In 74% of these cases, the victim was present for the conference. In an Indianapolis based study (McGarrel, et. al. 2000), over ninety percent of FGC victims reported satisfaction with how their cases were handled compared to 68% of victims in a control sample. There were few differences for youth and parents. Nine out of ten victims across three US based studies indicated satisfaction with the FGC process ( McCold and Wachtel, 1998; Fercello and Umbreit; and McGarrel et. al. 2001). The latter two studies also report offender satisfaction level at 90% and higher. The general rubric of satisfaction can tap any number of interest or needs of the participants. An Australian victim saw FGC as a means for underscoring the importance of people's rights. "The concept of other people's rights and their own responsibilities is very, very limited and this helps perhaps reduce that concept, in a very tangible, physical way." In contrast another victim in the same study stated, "I reckon it was just a put on." (Moore and Forsythe, 1995). In another Australian based study, victims participating in conferences were ten times more likely to receive some form of "repair" than their counterparts who went through the traditional court process. A family member of a victim alludes to this factor. "He (victim) would never have known the (offender) had to pay for what he did if the case had gone to court." (Strang and Sherman, 1977). Another victim from that study noted, "You realise they aren't the monsters you made them out to be...." Victims in a Minnesota study of FGC cited as most helpful in their experience the opportunity to "talk to the offender and explain effect of crime on them and to "hear the offender's explanation." The least helpful aspect of FGC was the "negative attitude of some parents." (Fercello and Umbreit, 1998). A victim from the Bethlehem, PA. study reflected a brand of skeptical optimism which pervades many of the comments across these studies: "I enjoyed taking part in this program! I do not feel that one meeting will change the offender's behavior. It was easy for the offender to predict what we wanted to hear. I'm not sure this program will be successful for all offenders. It's a great start though!" (McCord and Wachtel, 1998). A victim in a Wagga Wagga, Australia study focused on the less tangible outcomes: "At least the kids were made to front up to, you know, have to look a their parents and say, "Look, you know, I'm sorry I belittled you...." (Moore and Forsythe, 1995). In the Bethlehem, PA study, victim assessment of outcome while overwhelmingly favorable ranged from "this is an important community service" to "the court costs made the restitution paid inadequate in repairing the store's expenses." (McCord and Wachtel, 1998). A policeman interviewed for the Wagga Wagga study states clearly his view of how important satisfaction is to the practice of conferencing: "If the victim's satisfied; whether they get compensation; or a thank you...a sorry-letter, or just a straight out apology -- well that's the main thing." Circles. Fewer studies regarding participant response to restorative justice or peacemaking circles are available to us. Circles are most often imbedded in a broader community response to conflict. The earliest documented use of circles as a way of responding to offenders and victims come from various First Nation Communities in Canada. In those communities, there has been a reaching back to older traditions and spiritual heritages as a means of forming an holistic response to offenders and victims, their families, and the community at large. In other cultures, including many predominately Anglo communities in the United States, there has been a fascination with circles and their native heritage. There is an attempt in some of these communities to adapt aspects of the circle tradition while honoring its roots. Preliminary research efforts suggest that talking circles, healing circles, and sentencing circles have positively impacted the lives of those who have participated in them. An early evaluation of the Hollow Water First Nation Community Holistic Circle Healing approach to sex victimizers and others, their victims, families and the community pointed to positive outcomes as well as lingering concerns (Lajenunesse, T. and Associates Ltd. 1996). The Hollow Water Community is a group of four First Nation communities located one hundred and fifty miles northeast of Winnipeg, Manitoba. The intervention is explicitly holistic and spiritual. Traditional ways, such as the circle, are brought to bear on personal, intra family and community conflict. The intent is to work with the victimizer within the context of community rather than ship the individual to provincial or federal institutions. Some participants reported benefiting immensely from the circle process. Having a voice and stake in justice outcomes, mutual respect, and renewed community/cultural pride were cited as benefits of participation. On the other hand, lack of privacy, difficulty of working with family and close friends, embarrassment, unprofessionalism and religious conflict were cited by others as negative aspects of the circle process. More recently, the Native Counseling Services of Alberta conducted a cost-benefit analysis of the Hollow Water experience and concluded that "There are still criticisms from community members and outside sources about the healing process. However, in the minds of MASH [acronym for the four communities served by CHCH] members, CHCH stands clearly as a presence in the community that is good and desired. There is strong public acknowledgement of the strenuous, extraordinary work it has accomplished over 15 years." Victim satisfaction is cited as "very high" in the Healing/Sentencing Circles Program in Whitehorse, Yukon Territory (Matthews and Larkin, 1999). The original purpose of this effort was to develop and unify the community by involving community members in the process of working with individuals in trouble with the law in their own community. The healing circle process involves an application process, development of a "wellness plan," and monitoring the offender's progress. The program duration is six to eight weeks plus sentence which often is a probation term from six months to two years. Victims/families and offenders/families as well as community members participating in restorative justice circles organized by the South Saint Paul Restorative Justice Council in South Saint Paul, Minnesota, were very favorably disposed toward their circle experience (Coates, Umbreit, Vos (2001). This was equally the case with the circle participants in the South Saint Paul elementary and junior high schools. Two thirds of the Council's cases came from the South St. Paul Police Department and involved misdemeanors and low level assaults. Each of the thirty victim and offender participants indicated that they would recommend the circle process to others who were in similar circumstances. Offenders indicated that what they liked most about circles was "connecting with people in the circle," "changed attitude/behavior; "opportunity to payback victim and community," and "avoid court." Victims liked being able to "tell their story," "listening to others," and "connecting with people in the circle." Community representatives liked feeling that they "were giving something back to the community" and that "they were helping people." Criminal justice decision-makers support for circles ranged from enthusiastic to lukewarm. The process was regarded as far too time consuming by some and as only appropriate for minor cases and first time offenders. Others indicated that circle participation was an important way for involving local community members in the justice process and for letting victims and offenders know that the community cares about what happens to them. Fairness. Victim Offender Mediation. Related to satisfaction is the question of fairness. Many studies of victim offender mediation asked participants about the fairness of the mediation process and of the resulting agreement (Davis, 1980; Collins, 1984; Coates and Gehm, 1985; Strode, 1997 Umbreit, 1988, 1989, 1991, 1995; Coates and Umbreit, 1992; Umbreit and Roberts, 1996; Evje and Cushman, 2000; Umbreit, Coates, and Vos 2001). Not surprising, given the high levels of satisfaction, the vast majority of VOM participants (typically over 80%) across setting, cultures, and types of offenses reported believing that the process was fair to both sides and that the resulting agreement was fair. Again, these experiences led to feelings that the overall criminal justice system was fair. Where comparison groups were employed, those individuals exposed to mediation came away more likely feeling that they had been treated fairly than those going through the traditional court proceedings. In a study of burglary victims in Minneapolis, Umbreit found that eighty percent who went through VOM indicated that they experienced the criminal justice system as fair compared with only thirty-seven percent of burglary victims who did not participate in VOM (Umbreit, 1989). These positive satisfaction and fairness experiences have generated support for VOM as a criminal justice option. When asked, typically nine out of ten participants would recommend a VOM program to others (Coates and Gehm, 1985; Umbreit, 1991; Evje and Cushman, 2000; Umbreit, Coates, and Vos, 2001). Family Group Conferencing. Fairness is also an issue of concern for participants in family group conferencing and is often a focus of research. In a study of conferences in Adelaide and the towns of Port Augusta and Whyalla (Daly, 2001), eighty to ninety-five percent of victims and offenders reported that they were treated fairly and had a say in the agreement. In a brief paper on offender attitudes regarding deterrence, the RISE research team investigating Family Group Conferencing in selected Australian sites reports that seventy-four percent of the offenders felt the outcome of conferencing to be fair compared with 54% of comparison offenders prosecuted in the traditional courts (Sherman and Strang, 1997). Interestingly, the conference offenders were also more likely to feel that they would be caught if they reoffended. In three US based studies ( Fercello and Umbreit, 1998; and McCold and Wachtel, 1998; McGarrel et. al. 2001), about 95% of victims indicated the process/outcome was fair. Eighty-nine percent of the juvenile offenders in a Minnesota based study also indicated that the resulting conference agreement was fair. Over ninety percent of victims and offenders in the Bethlehem, PA study (McCold and Wachtel, 1998) and the Minnesota study (Fercello and Umbreit, 1998) would recommend the family group conferencing program to others. Nearly all victims in the Indianapolis based study (McGarrel et. al. 2001) indicated that they would recommend the program to a friend involved in a similar situation. This was compared to a quarter of the victims making such a recommendation in the control sample. Likewise 85% of the juvenile offender participants would recommend the program to friends compared with thirty-eight percent of those in the control group. RESTITUTION Victim Offender Mediation. Early on, restitution was regarded by many VOM program advocates as an important by-product of bringing offender and victim together in a face-to-face meeting. Restitution was considered somewhat secondary to the actual meeting where each party had the opportunity to talk about what happened. The form of restitution or what is called reparation in some jurisdictions is quite varied including direct compensation to victim, community service, work for victim, and sometimes unusual paybacks devised between victim and offender. Today, some jurisdictions see VOM as a promising major vehicle for achieving restitution for the victim. The meeting is necessary to establish appropriate restitution amounts and garner the commitment of the offender to honor a contract. Victims frequently report that while restitution was the primary motivator for them to participate in VOM what they appreciated most about the program was the opportunity to talk with the offender (Coates and Gehm, 1985; Umbreit and Coates, 1992). In many settings, restitution is inextricably linked with victim offender mediation. About half the studies under review looked at restitution as an outcome of mediation (Collins, 1984; Coates and Gehm, 1985, Perry, Lajeunesse and Woods, 1987; Umbreit, 1988; Galaway 1989; Umbreit, 1991; Umbreit and Coates, 1992; Warner, 1992; Roy, 1993; Evje and Cushman, 2000; Umbreit, Coates and Vos, 2001). Of those cases that reached a meeting, typically ninety percent or more generated agreements. Restitution of one form or another (monetary, community service, or direct service to the victim) was part of the vast majority of these agreements. Looking across the studies reviewed here, it appears that approximately 80-90% of the contracts are reported as completed. In some instances, the length of contract exceeded the length of study. One study was able to compare restitution completion between those youth participating in VOM with a matched group who did not (Umbreit and Coates, 1992.) In that instance, 81% of participating youth completed their contracts contrasted with 57% of those not in the VOM program, a statistically significant finding. In another study comparing an Indiana county whose restitution was integrated into victim-offender mediation with a Michigan county with court imposed restitution no difference in completion rates were found (Roy, 1993). Each was just shy of eighty percent completion. A study of victim offender mediation in six California counties showed a staggering increase in average obligated restitution paid. The increases ranged from +95% in Sonoma to +1000% in Los Angeles County (Evje and Cushman, 2000). Family Conferencing. Restitution or reparation is often part of the focus within family group conferences. Reparation agreements in the Indianapolis experiment (McGarrel, et. al., 2000) included the following elements: 1) apology, 62%; 2) monetary, 42%; 3) personal service, 36%; 4) community service, 24%; 5) other, 57%. Likewise a New Zealand study (Maxwell and Morris, 1993) found that apologies occurred in 70% of the cases; 58%, work in the community; 29% reparation (monetary payback). When victims were present for the conference, the work was more likely to be done for the victim than when they were not present, although this still happened in only two fifths of the cases. The authors conclude that victim presence had little impact on victim's receiving reparation. Reparation occurred 42% of the time when victims were present compared to 29% overall cases involving victims. Studies done on family group conferencing in Australia depict a more positive view of the FGC programs impact on restitution/reparation. Strang and Sherman (1997) conclude that victims processed through family group counseling were "ten times" more likely to receive repair than those processed through the traditional court. It should be noted that Strang and Sherman include apology, money, services, and other material compensation. Victims whose offender went through FGC were more likely to receive an apology (74%) than if the offender had been sent to court (11%). Moore and Forsythe ((1995) indicate that 87% of the conference agreements were largely completed. And Wundersitz and Hetzel (1966) reported an 86% compliance rate. In a Minnesota based study, Fercello and Umbreit (1998) indicate that 79% of victims and 92% of offenders indicated that an agreement had been successfully negotiated. DIVERSION. Victim Offender Mediation. Many VOM programs are nominally established to divert offenders into less costly, less time consuming, and frequently thought less severe options. Just as diversion was a goal lauded by many, others expressed concern about the unintended consequence of widening the net, that is, ushering in youth and adults to experience a sanction more severe than they would have if VOM did not exist. While much talk continues on this topic, there is a dearth of study devoted to it. Only a handful of the studies reviewed here address this question. One of the broadest studies considering the diversion question was that conducted over a three year period in Kettering, Northamptonshire, England (Dignan, 1990). Offenders participating in the VOM program were matched with similar non-participating offenders from a neighboring jurisdiction. The author concludes that at least sixty percent of the offenders participating in the Kettering program were true diversions from court prosecution. Jurisdictional comparisons also led him to conclude that there was a thirteen percent widening the net effect -- much less than local observers would have predicted. In a Glasgow, Scotland based agency where numbers were sufficiently large to allow random assignment of individuals between the VOM program and a comparison group going through the traditional process, it was discovered that forty-three percent of the latter group were not prosecuted (Warner, 1992). However, most of these pled guilty and were fined. This would suggest that VOM in this instance was a more severe sanction and indeed widened the net of government control. In a very large three county study of mediation in North Carolina, results on diversion were mixed (Clark, Valente, Jr., and Mace, 1992). In two counties, mediation had no impact on diverting offenders from court. In the third county the results, however, were quite dramatic. The authors concluded: "The Henderson program's effect on trials was impressive; it may have reduced trials by as much as two-thirds." Mediation impact on incarceration was explored in an Indiana-Ohio study by comparing consequences for seventythree youth and adults going through VOM programs with those for a matched sample of individuals who were processed in the traditional manner (Coates and Gehm, 1985). VOM offenders spent less time incarcerated than did their counterparts. And when incarcerated, they did county jail time rather than state time. The length and place of incarceration also had substantial implications for costs. Family Group Conferencing. The impact of family group conferencing on diverting offenders from the formal system or to a less severe sanction remains unclear. McCold and Wachtel (1998) indicate that FGC had left police and courts largely unaffected. As with many of the programs that are studied, one might question how realistic it is to expect that interventions small in size will have any significant impact on large formidable justice institutions. Moore and Forsythe (1995) conclude that the introduction of family group conferencing as part of community policing in the so called Wagga Wagga model is "associated with a substantial decrease in the total number of police interventions involving young people, and with a substantial increase in the number of those cases dealt with by way of 'caution' rather than in court." New Zealand's Children, Young Persons and Families Act of 1989 established new procedures for state intervention into families and the lives of children and young people. The Act provided new roles for victims and a voice for the young people and their families. These altered the ways of police and court processing and provided for a "new decision making forum, the Family Group Conference"(Maxwell and Morris, 1993). The impact was dramatic. Before the Act there were up to 13,000 court cases each year. In 1990 there were 2,587. While this constitutes massive diversion, Maxwell and Morris point out that only three out of five youth who appeared in court previously received any formal penalty. For those offenders now going through FGC 95% receive a penalty or make an apology. The authors conclude, "Thus the total number who now receive some form of penalty is almost certainly greater than in the past – in other words, the net appears to have widened." Circles. The Hollow Water First Nation Community Holistic Circle Healing Process was designed, in part, as a way of keeping victimizers in the community (Native Counseling Services of Alberta, 2001). Over a ten year period, ninety-four individuals, including sixty-eight adult males, seven adult females and nineteen youth were diverted within the four communities making up Hollow Water. Forty-one of these persons had assault charges and thirtyseven had sexual assault charges. An additional seven adult males came to the program from other reserves. According to the authors of this study, one hundred and one individuals were diverted from the provincial or federal justice system. We will consider the cost savings generated by this diversion effort in a later section of this document. RECIDIVISM Victim Offender Mediation. While recidivism may be best regarded as an indicator of society's overall response to juvenile and adult offenders, it is a traditional measure used to evaluate the long term impact of justice programs. Accordingly, a number of studies designed to assess VOM have incorporated measures of recidivism. Some simply report re-arrest or reconviction rates for offenders going through the VOM program understudy (Carr, 1998; Roberts, 1998). Since no comparison group or before/after outcomes are reported, these recidivism reports have local value, but offer very little meaning for readers unfamiliar with typical rates for that particular region. One of the first studies to report recidivism on VOM was part of a much larger research project regarding restitution programs (Schneider, 1986). Youth randomly assigned to a Washington, DC VOM program were less likely to have subsequent offenses resulting in referral to a juvenile or adult court than youth in a comparison probation group. These youth were tracked for over thirty months. The results were 53% and 63%; the difference was statistically significant. A third group, those referred to mediation, but who refused to participate, also did better than the probation group. This group's recidivism prevalence was 55%. The study based in Kittering, England ( Dignan, 1990) compared recidivism data on the VOM offenders who went through face-to-face mediation with those who were exposed only to "shuttle mediation." The former group did somewhat better than the latter: 15.4% and 21.6%. As with satisfaction measures reported earlier, face-to-face mediation seems to generate better results both in the short run and in the longer run than the less personal indirect mediation. In a study of youth participating in VOM programs in four states, youth in mediation had lower recidivism rates after a year than did a matched comparison group of youth who did not go through mediation (Umbreit and Coates, 1992). Overall, across sites, eighteen percent of the program youth reoffended compared to 27% for the comparison youth. Program youth also tended to reappear in court for less serious charges than did their comparison counterparts. The Elkhart and Kalamazoo county study (Roy, 1993) found little difference in recidivism between youth going through the VOM program and the court imposed restitution program. VOM youth recidivated at a slightly higher rate, 29% to 27%. The author noted that the VOM cohort included more felons than did the court imposed restitution cohort. A study of 125 youth in a Tennessee VOM program (Nugent and Paddock, 1995) reported that these youth were less likely to reoffend than a randomly selected comparison group: 19.8% to 33.1%. The VOM youth who did reoffend did so with less serious chargers than did their comparison counterparts. A sizeable cohort of nearly eight hundred youth going through mediation in Cobb County Georgia between 19931996 was followed along with a comparison group from an earlier time period (Stone, Helms, and Edgeworth (1998). No significant difference in return rates was found: 34.2% mediated to 36.7% non-mediated. Threequarters of the mediated youth returned to court did so because of violation of the conditions of mediation agreements. Wynne and Brown (1998) report on a longstanding study of the Leeds Victim Offender Unit which began in 1985. Of the ninety offenders who met in face-to-face mediation from 1985- 1987, 87% had had previous convictions before mediation. Sixty-eight percent had no convictions during a two year follow-up post mediation. In another English study focused on seven varying restorative justice schemes across England, Miers et. al. (2001) contend that "the only scheme that routinely involved victims (West Yorkshire) was for the most part both lower cost and more effective than the other schemes." And this same program had a "significant impact on reoffending, both in terms of the offence frequency and offence seriousness." Stone (2000) compared youth going through Resolutions Northwest's Victim Offender Mediation Program in Multnomah County Oregon with a comparison group. Eighty percent of the youth processed through VOM did not recidivate during a one year follow-up period while 58% of the comparison group did not reoffend during a year of follow-up. In a Lane County Oregon study, Nelson (2000) took a different tack. One hundred and fifty youth referred to VOM from July of 1996 to November 1998 in that county were also followed for a year after referral. Comparing their referral frequencies the year prior to the referral to VOM with the year after, all referred youth had 65% fewer referrals to the system in the subsequent year. Juvenile referred to VOM but refusing to participate had 32% fewer referrals; youth who met with their victims had 81% fewer referrals that the preceding year; and juveniles who fully completed their agreements had 76% fewer referrals compared with 54% fewer referrals for those youth who did not complete any part of the agreement. Recidivism data was gathered on VOM programs in two additional Oregon counties in the study conducted by Umbreit, Coates and Vos (2001). These data reflect one year before intervention comparisons of number of offense with one year after. For the group of youth in the Deschutes County program there was a 77% overall reduction in reoffending. Similarly, for the group of juveniles going through the victim offender program in Jackson County there was an overall 68% reduction in recidivism. In a six county study in California conducted by Evje and Cushman (2000), one of the victim offender mediation programs experienced a 46% higher rate of recidivism than its comparison group. In the other five counties, the VOM groups ranged from 21% to 105% less recidivism than their comparison groups. Nugent, Umbreit, Wiinamaki and Paddock (1999) conducted a rigorous reanalysis of recidivism data reported in four studies involving 488 VOM youth and 527 non-VOM youth. Using ordinal logistical regression procedures the authors determined that VOM youth recidivated at a statistically significant lower rate than non-VOM youth and when they did reoffend they did so for less serious offenses than the non-VOM youth. Family Group Conferencing. Two hundred and eighty-one juvenile cases going through the family group conferencing model employed by the Woodbury Police Department, in Woodbury, Minnesota between 1995 and 1999 were compared to a group of non-conferencing youth in 1993 (Hines, 2000). Thirty-three percent of the conferencing youth reoffended compared to 72% of the non-conferencing youth. Sixteen percent of the first time offenders who went through conferences re-offended contrasted with 52% of repeat offenders going through conferences. Maxwell and Morris (1993) report that 48% of those referred for an FGC in their New Zealand study reoffended within six months. There was some variation by region. In the Wagga Wagga experiment with introducing FGC into community policing there was little change in reapprehension patterns nine months before and after FGC. Youth going to court were more likely to be reapprehended, 35.6% compared with 18.7% (Moore and Forsythe, 1995). In a recent work by Maxwell and Morris (2001), the authors take a most important step toward attempting to discern what factors influence a youngster's propensity to reoffend or not. In 1996 they were able to contact some 108 young people (accounting for 67% of the original sample) and 98 parents who had participated in family group conferencing in 1990-91. All the young people were in their twenties. Twenty-ninety percent of these young people had never been reconvicted. Twenty-eight percent had been persistently reconvicted. Several multivariate analyses were conducted to sort out predictors of reconviction and pathways to reoffending. The authors conclude "that family group conferences can contribute to lessening the chance of reoffending even when other important factors such as adverse early experiences, other events which may be more related to chance, and subsequent life events are taken into account. Critical factors for young people are to have a conference that is memorable, not being made to feel a bad person, feeling involved in the conference decision-making, agreeing with the outcome, completing the tasks agreed to, feeling sorry for what they had done, meeting the victim and apologizing to him/her, and feeling that they had repaired the damage." The authors point out that "these factors reflect key restorative values, processes and outcomes." Preliminary recidivism patterns in the Canberra Reintegrative Shaming Experiments (Sherman, Strang, and Woods, 2000) yield mixed results. Cases were randomly assigned to court or conferencing across four experiments identified by the types of offenses handled. Recidivism rates are based on one year before and after comparisons. For the youth violence offenders, the court group rate of offending fell by 11% in the year before and after comparisons. The FGC youth reoffending rates fell 49%. This difference constitutes a 38% reduction in the conference group relative to the change in the court group. Rate of reoffending for the drunk driving offenders showed a slight increase for both groups, that is, those who were referred to court and those who went through FGC. There was no significant differences between groups of offenders processed for juvenile property-shoplifiting or for those processed for juvenile property-with personal victims. McCold and Wachtel (1998) also report on a one year follow-up of youth randomly assigned to treatment (FGC) or control group in their Bethlehem, PA study. Their treatment group was divided in two: a group that went through FGC and a group that did not because either the victim or the offender chose not to participate. The declining group had a larger number of "violent" crimes. It should be noted, however, that it remains unclear how these offenses were so labeled, given that most offenders charged with violent crimes were not eligible for the FGC diversion option in the first place. The recidivism results are as follows. For property offenders, 32% of the conference youth reoffended compared with 35% of those who declined to participate, and 21% of those who were in the control group. For violent offenders, 20% of the conference youth reoffended compared to 48% of those who decline to participate, and 35% of the control group. The authors conclude that conferencing positively "affects recidivism by resolving conflict between disputing parties." The Restorative Justice Conferencing Experiment in Indianapolis also relied on a random assignment experimental design (McGarrel, 2000). Two hundred and thirty two youth went to the restorative justice family group conferencing program and two hundred and twenty-six went to other diversion programs. Recidivism was measured by contact with the court during a six or twelve month period since the initial incident, and contact with the court after completion of assigned diversion program. During the six months after the initial incident, 79.6% of the conference youth had no further court contact contrasted with 58.8% of the control youth. That difference is significant at the .01 level using Chi-square. For the twelve month period, 69.2% of the conference youth had no further court contact compared with 58.8% of the controls. That difference is significant at the .05 level using Chisquare. For the six months post diversion program completion, 87.7% of the conferencing youth had no contact with the court compared with 77.3% of controls. Again, using a Chi-square the researchers found the difference to be significant at the .05 level. Circles. While recidivism is not a primary focus of any the circle studies surveyed here, it was mentioned inn two of the reports. Matthews and Larkin (1999) note that an internal self-study was completed for the Healing/Sentencing Circles Program at Whitehorse, Yukon Territory by an outside consultant. Over a two year period the program served sixty-five clients. Follow-up tracking "indicated an 80% decrease in recidivism." Also, the Hollow Water study conducted by the Native Counseling Service of Alberta reported that only two clients (approximately 2%) over the ten years had re-offended. They suggest that typical "recidivism rates for sex offenses is approximately 13% and for any form of recidivism the figure rises to approximately 36%." It remains unclear if these latter comparative figures refer to provincial data , federal data or both. All in all, recidivism findings across a fair number of sites and settings, suggest that restorative justice conferencing approaches are at least as viable at recidivism reduction as traditional approaches. And in a good number of instances, youth going through conferencing programs are actually faring better. COSTS Victim Offender Mediation. The relative costs of correctional programs is difficult to assess. Several studies reviewed here addressed the issue of costs. Cost per unit case is obviously influenced by the number of cases handled and the amount of time devoted to each case. The results of a detailed cost analysis in a Scottish study were mixed (Warner, 1992). In some instances, mediation was less costly than other options and in others more. The author notes that given the "marginal scope" of these programs it remains difficult to evaluate their cost if implemented on a scale large enough to impact overall program administration. Evaluation of a large scale VOM program in California led the authors to conclude that the cost per case was reduced dramatically as the program went from being a fledgling to being a viable option (Niemeyer and Schichor, 1996). Cost per case was $250. An alternative way of considering the cost impact of VOM is to consider broader system impact. Reduction of incarceration time served can yield considerable savings to a state or county (Coates and Gehm, 1985). Reduction of trials, such as in Henderson County, North Carolina where trials were reduced by two-thirds, would have tremendous impact at the county level (Clarke, Valente Jr., and Mace (1992). And researchers evaluating a VOM program in Cobb County, Georgia point out that while they did not do a cost analysis, per se, time is money (Stone, Helms, and Edgeworth, 1998). The time required to process mediated cases was only a third of that needed for nonmediated cases. The potential cost savings of VOM programs when they are truly employed as alternatives rather than as showcase add ons is significant. Yet a cautionary note must continue to be heard. Like any other program option, these programs can be swamped with cases to the point that quality is compromised. And in the quest for savings there is the temptation to expand the eligibility criteria to include individuals who would not otherwise penetrate the system or to take on serious cases that the particular program staff are ill equipped to manage. Staff and administrators must be prepared to ask, "Cost savings at what cost?" Circles. A cost-benefit analysis was the cornerstone of the Native Counseling Services of Alberta study of the Hollow Water's Community Holistic Circle Healing Process (2001). Efforts were made to track the cost that would have occurred if the ninety-four victimizers participating in the program had not been diverted but rather would have proceeded on to the provincial or federal justice systems. Estimates of pre-incarceration, incarceration, and parole costs were derived. These were compared to the costs of the CHCH. It is estimated that the total costs to provincial and federal governments without CHCH in place would have ranged from $6,212,732 to $15,902,885. The authors conclude that given the "very low recidivism rate…it is appropriate to state that the value of services to both the government and community has been significantly understated." META-ANALYSIS Increasingly the field of social science is witnessing the emergence of meta-analyses. These are methods of research synthesis across a set of empirical studies. Meta-analysis will typically involve reviewing the relevant literature, including published journal articles, books and perhaps less well known research monographs. Data are extracted from these studies and are aggregated for further statistical analysis. Three such studies are reported on here. Nugent, Umbreit, Wiinamaki and Paddock (2001) conducted a rigorous reanalysis of recidivism data reported in four previous studies involving a total sample of 1,298 juvenile offenders, 619 who participated in VOM and 679 who did not. Using ordinal logistical regression procedures the authors determined that VOM youth recidivated at a statistically significant 32% lower rate than non-VOM youth and when they did reoffend they did so for less serious offenses than the non-VOM youth. In a forthcoming work, Nugent, Williams and Umbreit have expanded their effort to include fourteen studies to compare the prevalence rate of subsequent delinquent behavior of VOM participants with that of adolescents who did not participate in VOM. This analysis relied on a combined sample of 9,037 juveniles. The results "suggested that VOM participants tended to commit fewer reoffenses … [and] tended to commit less serious reoffenses." (Nugent, Williams and Umbreit, forthcoming)" In another large meta study, Latimer, Dowden and Muise (2001) reviewed eight conferencing and twenty-seven victim-offender mediation programs. In order to qualify for inclusion in this analysis the study had to have evaluated a restorative justice programs, ie, "restorative justice is a voluntary, community-based response to criminal behaviour that attempts to bring together the victim, the offender and the community in an effort to address the harm caused by the criminal behavior;" used a control group or comparison group that did not participate in the restorative justice program; reported on at least one of the following four outcomes—victim satisfaction, offender satisfaction, restitution compliance, and/or recidivism; and provided sufficient statistical information to calculate an effect size. Some of the major results of this analysis are: *Victim Satisfaction. In all but one of the thirteen restorative programs studied, victims were more satisfied than those in traditional approaches. The authors indicate that "VOM models tended to yield higher levels of victim satisfaction rates than conferencing models when compared to the non-restorative approaches." They suggest that this result may be explained by the conferences typically having more participants and thus it may be more difficult to find as much satisfaction with an agreement. *Offender Satisfaction. Initial analysis shows "no discernible impact" on offender satisfaction. However when an outlier program is removed, "moderate to weak positive impact on offender satisfaction " is noted. *Restitution. "Offenders who participated in restorative justice programs tended to have substantially higher compliance rates than offenders exposed to other arrangements." *Recidivism. "Restorative justice programs, on average, yielded reductions in recidivism compared to nonrestorative approaches to criminal behaviour." The authors discuss and consider the issue of self-selection bias, that is, victims and offenders choose to participate in these programs. They note that McCold and Wachtel (1998) attributed apparent differences in recidivism to the effect of self-selection bias. Latimer, Dowden and Muise conclude: "Notwithstanding the issue of self-selection bias, the results of this meta-analysis, at present, represent the best indicator of the effectiveness of restorative justice practices (i.e. those individuals who choose to participate in restorative justice programs find the process satisfying, tend to display lower recidivism rates and are more likely to adhere to restitution agreements." CONTINUING ISSUES Restorative justice conferencing, in its various forms, has been studied empirically over twenty-five years in numerous countries and with a wide range of populations. It has probably been examined as extensively, if not more so, as any justice or correctional reform. Studies have ranged from small exploratory undertakings, to quasiexperimental designs, to meta studies. Research on restorative justice conferencing began with a focus on how victim and offender participants experienced being part of these efforts to involve them in the process of justice-making. This was a reasonable beginning point, given the emphasis within restorative justice frameworks to include victims and their wishes as well as giving offenders opportunities for making things right and getting on with their lives. If studies were to repeatedly show that the bulk of offenders and victims were dissatisfied or felt additional harm was heaped upon them, then it wouldn't matter if the programs were effective or not at increasing rates of restitution completion or decreasing rates of recidivism. Taken as a whole, the studies reviewed here reflect remarkably consistent levels of victim and offender satisfaction with conferencing strategies. Furthermore, it appears that conferencing does increase the likelihood that restitution contracts will be paid. It is suggested that this is likely because the criminal violation and its consequences have been made more personal so that any resulting agreement becomes more personal. And even crime reduction, measured by recidivism, seems to be happening for a significant number of offenders who are processed through restorative justice conferencing approaches. Still, issues remain. Here we will consider a few policy and research issues. Policy Issues 1. Overcoming the myth that nothing works. For over a quarter century correctional philosophy and wisdom has been engrained with the dictum "nothing works." Rightly or wrongly a veneer of skepticism and cynicism often shrouds policymaker and practitioner alike. Each is often able to point to individual cases which bear out the generalization while the accounts of success are relegated to "stories." This skepticism leads to low expectations: "nothing mattes more than anything else." Or it may lead to justifying rote experimentation: "what's new this year? It won't make any difference, but we need to keep trying." Skepticism is often accompanied by considerable heart and compassion. And this combination frequently leads to frustration and disillusionment which in turn leads to a self-fulfilling prophecy that "nothing works." But some things do work for some individuals and probably always have. Single approaches to justice are unlikely to work equally well with everyone. In cases where victims meet with offenders, it is likely that those who were motivated to seek such a meeting in the first place may be more likely to be satisfied with the experience and the outcomes. This will not be the result for every victim who sought such a meeting. Some will be disappointed; some will feel like it was a waste of time. Even within the realm of conferencing, some individuals and cases may be better suited for victim offender mediation, some for family group conferencing, and some for circles. Matching limited resources with particular cases and individuals is a perplexing problem for justice and corrections as a whole; such will also be the case for conferencing options. The research community not only has a continuing responsibility for conducting studies which are as rigorous as possible given the practice circumstances. It also has a responsibility for articulating the meaning of findings in ways that inform educators, policymakers and practitioners in clear and reasonable ways. As researchers employ more sophisticated statistical tools, including meta-analysis approaches, the onus remains on the research team to explain the results so their policy and practice implications are evident. Slogans and aphorisms such as "nothing works" are too easy, but it is also too much to expect that all individuals concerned about directions in justice and corrections have a thorough understandings of betas, canonical variables, and effect sizes. 2. Restorative justice conferencing: what is it? Conceptual thinking about restorative justice continues to evolve as do attempts to develop processes grounding those concepts and principles into practice. As noted at the beginning of this chapter, an encounter of all the parties with a stake in a particular crime incident is regarded by many as being at the core of restorative justice (Llewellyn and Howse, 1998), along with voluntariness of participation and truth telling. At a recent seminar for practitioners and policymakers held at the University of Minnesota, Howard Zehr, a widely recognized leader in the movement toward restorative justice, indicated an uneasiness with the centrality of direct encounters in restorative justice definitions. He has moved toward a broader conceptual understanding reflected in the following: "Restorative justice is a process to involve to the extent possible, those who have a stake in a specific offense to collectively identify and address harms and obligations in order to heal and put things as right as possible." (Zehr, 2002) Likewise the term "conferencing" has emerged as a generic descriptor to include victim offender mediation, family group conferencing and circles as well as some reparation boards. Without attempting to sort out these conceptual and definitional issues here, it does seem important to highlight some pluses and minuses of the lack of a clear, widely agreed upon definition. A major plus is that there exists considerable room for innovation. There is no lockstep program which is expected to work with all individuals. Conceptual breadth allows for policymakers and practitioners to fit restorative justice conferencing to the particular needs of their communities. A minus, however, is conceptual blurriness which hints at few parameters. A common criticism in corrections over the years is its tendency to rewrap old programs under new names and banners without changing much, if anything. For practitioners, policymakers and researchers alike, the continuous evolution of concepts and definitions can make implementation and evaluation difficult at best. And it can lead to frustration, particularly for the practitioner who has worked hard and "moved mountains" to implement a program only to find out that the "field" no longer considers the effort to be "state of the art" or "best practice." 3. Realistic expectations regarding system wide impact. There seems little doubt that restorative justice conferencing can impact in transforming ways the lives of at least some victims and offenders. What are reasonable expectations regarding system wide impact? Many administrators point to the time demands of conferencing approaches and state flat out that there will never be enough resources to handle all the cases in conferencing approaches. A continuing debate within systems that are taking restorative justice seriously is what kinds of cases and participants are best suited for conferencing. Some will argue that the least serious cases are best suited while others will contend just the opposite is true. Perhaps it is unrealistic to expect any massive shift in the numbers of cases being processed through conferencing strategies, but if restorative justice is indeed a process then it should be expected that system wide approaches to processing and interacting with offenders and victims should be noticeably different from the way they were. A useful dialogue, it seems to us, is an ongoing discussion among policymakers, practitioners and researchers on what we would expect to change and how to measure such change at the system level. Given our limited experience of raising these questions in a few communities, we expect that responses will range from changed philosophies and attitudes, to changed methods of processing and documenting cases, to changed ways of working with offenders and victims. That thinking and discussion is likely to lead to a continuum of restorative responses or services which is anchored in the context of security and available resources. 4. Widening the net. As with any good idea or well intentioned program or process, there continues to be concern that the program or process is not simply being used to enlarge the system's social control capacity. Many conferencing programs are supposed to divert individuals, often youth, from the formal processing of the traditional system. The question remains whether the majority of these participants suffer more sanction under the diversionary programs than they would have if such programs had not existed. A corollary question also must be answered: to what extent, if any, is there substantial positive service and outcome to both the offender and victim which balances off any net widening effect. Obviously, to some observers nothing would balance off net widening. 5. Equal justice and opportunity. A justice and corrections response is typically faced with how to apply scarce resources to a broad population. Why and how are some offenders offered "innovative programs" and more opportunity for resources to come to bear on their family and community networks than others? As long as restorative justice conferencing approaches are not sufficiently abundant to offer to each offender or victim, then that decision-making process determining who has the opportunity and who does not needs to be scrutinized carefully by policymakers, practitioners and researchers to be certain that bias does not become a determining factor. Research Issues 1. Factors that foster satisfaction. Understanding participant satisfaction with the conferencing process and outcome is central to determining to what extent conferencing is meeting some of the fundamental goals of restorative justice. Researchers, along with practitioners, need to continue ferreting out what factors contribute to participant satisfaction. On the whole, victims as well as offenders are satisfied with their experiences in these programs. We know that what is important to one participant may not be as salient to another so there are many factors which potentially contribute to satisfaction levels. It seems to us that continued exploration of these factors may contribute to further program development and refinement. If it is important, within a restorative justice framework, to remain sensitive to victim, offender and community needs and responses, then satisfaction and the factors influencing it is a fruitful research path to follow. 2. Recidivism. The study of recidivism is rife with danger. There are the familiar problems of what recidivism actually measures (offender behavior, police and court practices, overall societal response to offenders and their families and communities) and what measures should be used (frequency of lawbreaking, seriousness of offense, level of sanction). In addition to these, two more are apparent. First, if a particular program does not show a positive reduction in further offenses, however measured, there is the danger that the program will be scrapped while it is meeting other markers of success. Second, if a particular program does show positive reduction in further offenses, there may be a tendency to back off from other restorative goals such as victim and offender involvement in working out a solution, or voluntariness, or victim centeredness. In other words, if recidivism is regarded as the most important desired outcome, it may become the only desired outcome and a "restorative" program may over time be stripped of those qualities that make it restorative and that contribute to reduction in further offending. We have long believed that simple "they did or they didn't recidivate" thinking overlooks examining the strengths and weaknesses of programs and of offenders and their social networks. A more important question from a policy and program perspective is "what factors contribute to an offender's likelihood of reoffending or not reoffending?" Attempting to sort out program impact, offender characteristics, family impact, community impact and interaction with justice decision-makers will often require longitudinal, time consuming study, large samples and use of sophisticated multivariate statistical techniques (Coates, Miller and Ohlin, 1978). We believe that more work like that of Maxwell and Morris (2001) can offer much in addressing these pivotal questions. 3. Describing the black box. There is a continuing need to describe the proverbial black box, that is, the actual program under study. Programs may be called victim offender mediation, family group conferencing, or circles, but we cannot assume that these labels mean that programs under each rubric are the same. One of the strengths of meta analysis is that it enables researchers to increase sample size by aggregating across a large number of programs. And these meta analyses have much to contribute to the field. However, they cannot replace the ongoing evaluative studies upon which they depend. And these latter studies must continue to provide rich qualitative description of what is actually happening in the program as well as tell the stories of victims and offenders. If we lose the story of the program, of the victim and of the offender, then we have lost the heart of restorative justice. 4. Diversion, widening the net, costs. As noted above under policy issues, questions regarding diversion, widening the net and costs need to be pursued. Some may argue that conferencing is so inherently valuable that one need not be concerned about wider systemic impact. We believe that policymakers and administrators have a legitimate responsibility to raise these critical questions and to pursue answers. Data is sparse and mixed at this point. 5. Setting up programs. An area of study seldom pursued is that of the process of establishing restorative justice conferencing programs. Yet, it is just this kind of information that is desired by groups and jurisdictions thinking about setting up a victim offender mediation program, or a family group conferencing program, or a community circle council. In a recent study of six Oregon counties (Coates, Umbreit and Vos 2001), it was this program development lens that showed how each program worked with similar yet very different local conditions and expectations to establish a victim offender mediation program. The resulting programs, in some instances, were strikingly different, reflecting those contrasting local conditions. Yet each county had a viable victim offender mediation option available to the court/corrections system. CONCLUSION Just as interest in restorative justice conferencing is growing within the justice arena so is the body of empirical knowledge collected to evaluate, shape and refine conferencing. Involving victims and offenders and community members in talking about the impact of the crime and developing a plan to repair the harm is yielding, for the most part, positive responses from participants. The vast majority of participants find the experience satisfactory, fair and helpful. In a number of jurisdiction rates of restitution completion have climbed. And offenders going through conferencing approaches often have lower levels of re-offending than they did before or than compared with a similar group of offenders who did not go through conferencing. Studies reviewed here range in rigor from exploratory to experimental random assignment designs. More questions need to be pursued and broadened, but given the empirical evidence generated over the past twenty-five years or so and across many countries, it seems reasonable to say that restorative justice conferencing strategies do contribute to increased victim involvement and reparation, to offenders taking responsibility for their behaviors, and to community members participating in shaping a just response to law violation. REFFERENCES Barajas, (1995). Topics in Community Corrections: Community Justice. Longmont, CO: U.S. Department of Justice, National Institute of Corrections Information Center. Bazemore, G. and M. Schiff (1996). “Community Justice/Restorative Justice: Prospects For a New Social Ecology for Community Corrections.” International Journal of Comparative and Applied Criminal Justice Vol. 20(2) pp. 311335. Bradshaw, W. and Umbreit, M. (1998). "Crime victims meet juvenile offenders: Contributing factors to victim satisfaction with mediated dialogue," Juvenile and Family Court Journal Vol 49 (Summer) pp. 17-25. Carr, C (1998). VORS Program Evaluation Report. Clarke, S., E. Valente, Jr., and R. Mace (1992). Mediation of Interpersonal Disputes: An Evaluation of North Carolina's Programs. Chapel Hill: Institute of Government, University of North Carolina. Coates, R. and J. Gehm (1985). Victim Meets Offender: An Evaluation of Vicitm-Offender Reconciliation Programs. Valparaiso, IN: PACT Institute of Justice; Coates, R. and J. Gehm (1989) "An Empirical Assessment" in M. Wright and B. Galaway (eds.) Mediation and Criminal Justice. London: Sage, pp. 251-263. Coates, R., A. Miller, L. Ohlin (1978). Diversity in a Youth Correctional System: Handling Delinquents in Massachusetts. Cambridge, MA: Ballinger. Coates, R., M. Umbreit and B. Vos (2000). Obstacles and Opportunities for Developing Victim Offender Mediation for Juveniles: The Experience of Six Oregon Counties. Center for Restorative Justice and Peacemaking, University of Minnesota. Coates, R., M. Umbreit and B. Vos (2000). Restorative Justice Circles in South Saint Paul, Minnesota. St. Paul, MN: Center for Restorative Justice & Peacemaking, University of Minnesota; Umbreit, M., R. Coates, and B. Vos (2002). "Restorative Justice Circles: The Impact of Community Involvement," Perspectives Vol. 26 (Winter) pp. 36-40. Daly, K. (2001). "Conferencing in Australia and New Zealand: Variations, Research Findings, and Prospects." In A. Morris and G. Maxwell, Restorative Justice for Juveniles. Oxford: Hart Publishing. Davis, R. et. al. (1980) . Mediation and Arbitration as Alternative to Prosecution in Felony Arrest Cases, An Evaluation of the Brooklyn Dispute Resolution Center. New York, NY: VERA Institute of Justice. Dignan, (1990). Repairing the Damage: An Evaluation of an Experimental Adult Reparation Scheme in Kettering, Northamptonshire. Sheffield: Centre for Criminological Legal Research, Faculty of Law, University of Sheffield. Evje, A. and R. Cushman, (2000). A Summary of the Evaluations of Six California Victim Offender Rehabilitation Programs. San Francisco, CA: Judicial Council of California, Administrative Office of the Courts. Fercello, C. and M. Umbreit, (1998). Client Evaluation of Family Group Conferencing in 12 Sites in 1st Judicial District of Minnesota. St. Paul, MN: Center for Restorative Justice & Mediation. Fercello, C. and Umbreit, M. (1999). "Client Satisfaction with Victim-Offender Conferences in Dakota County, Minnesota. St. Paul, MN" Center for Restorative Justice & Mediation, University of Minnesota. Flaten, C. (1996). "Victim Offender Mediation: Application with Serious Offences Committed by Juveniles." In: B. Galaway and J. Hudson (eds.), Restorative Justice: International Perspectives. Monsey, NY: Criminal Justice Press, pp. 387-401. Galaway, B. (1989). "Informal Justice: Mediation between Offenders and Victims." In: P. Albrecht and O. Backes (eds.), Crime Prevention and Intervention: Legal and Ethical Problems. NY: Walter de Gruyter, pp. 103-116. Galaway, B. (1995). "Victim-Offender Mediation by New Zealand Probation Officers: The Possibilities and the Reality." Mediation Quarterly Vol. 12, pp. 249-262. Gehm, J. (1990). "Mediated Victim-Offender Restitution Agreements: An Exploratory Analysis of Factors Related to Victim Participation." In B. Galaway and J. Judson (eds.), Criminal Justice, Restitution, and Reconciliation. pp. 177-182. Griffiths, C.T. and R. Hamilton (1996). “Sanctioning and Healing: Restorative Justice in Canadian Aboriginal Communities.” In B. Galaway and J. Hudson (eds.) Restorative Justice: International Perspectives. Monsey, NY: Criminal Justice Press, pp. 175-192. Greenwood, J. and M. Umbreit (1998). "National Survey of Victim Offender Mediation Programs in the US," VOMA Connections (Winter). Hines, D. (2000). The Woodbury Police Department Restorative Justice Program Recidivism Study. Woodbury, MN: Inter-faith Ministries. Hughes, S. and A. Schneider, (1990). Victim-Offender Mediation in the Juvenile Justice System. Washington, D.C.: Office of Juvenile Justice and Delinquency Prevention. Lajeunesse, T. & Associates (1996). Evaluation of Community Holistic Circle Healing. Hollow Water First Nation. Volume 1: Final Report. Submitted to the Solicitor General. Latimer, J., C. Dowden and D. Muise (2001). The Effectiveness of Restorative Practices: A Meta-Analysis. Department of Justice, Canada: Research and Statistics Division Methodological Series. Llewellyn, J. and R. Howse (1998). Restorative Justice: A Conceptual Framework. Ottawa: Law Commission of Canada. Marshall, T. (1990). "Results of Research from British Experiments in Restorative Justice." In B. Galaway and J. Hudson (eds.), Criminal Justice, Restitution, and Reconciliation. London: Sage, pp. 83-107. Matthews, S. and G. Larkin (1999). Guide to Community-based Alternatives for Low Risk Juvenile Offenders. Topeka, KS: Koch Crime Institute. Maxwell, G. and A. Morris, (1993). Family, Victims and Culture: Youth Justice in New Zealand. Wellington, New Zealand: Social Policy Agency and Institute of Criminology, Victoria University of Wellington. Maxwell, G. and A. Morris, (2001). "Family Group Conferencing and Reoffending." In Morris, A. and G. Maxwell (eds.), Restorative Justice for Juveniles: Conferencing, Mediation & Circles. Oxford: Hart Publishing. McCold, P. and B. Wachtel, (1998). Restorative Policing Experiment: The Bethlehem Pennsylvania Police Family Group Conferencing Project. Pipersville, PA: Community Service Foundation. McGarrel, E. et. al. (2000). Returning Justice to the Community: The Indianapolis Juvenile Restorative Justice Experiment. Indianapolis, IN: Hudson Institute. Miers, D. et. al (2001). An Exploratory Evaluation of Restorative Justice Schemes. Executive Summary. Crime Reduction Research Series Paper 9. Home Office, Crown Copyright. Moore, D. and L. Forsythe, (est. 1995). A New Approach To Juvenile Justice: An Evaluation of Family Conferencing in Wagga Wagga. Wagga Wagga, Australia: Centre for Rural Social Research. Native Counseling Services of Alberta (2001). A Cost-Benefit Analysis of Hollow-Water's Community Holistic Circle Healing Process. Solicitor General of Canada: Aboriginal Corrections Policy Unit. Nelson, S. (2000). Evaluation of the Restorative Justice Program. Eugene, OR: Lane County Department of Youth Services. Niemeyer, M. and D. Shichor (1996). "A Preliminary Study of a Large Victim/Offender Reconciliation Program," Federal Probation Vol. 60 ( September) pp. 30-34. Nugent, W. and J. Paddock (1995) "The Effect of Victim-Offender Mediation on Severity of Reoffense," Mediation Quarterly Vol. 12, Summer, pp.353-367. Nugent, W., M. Umbreit, L. Wiinamaki, and J. Paddock (1999). "Participation in Victim-Offender Mediation and Severity of Subsequent Delinquent Behavior: Successful Replications?" Journal of Research in Social Work Practice. Nugent, W., M. Williams and M. Umbreit (forthcoming). "Participation in Victim-Offender Mediation and the Prevalence and Severity of Subsequent Behavior." University of Tennesee. Perry, L., Lajeunesse, T. and Woods, A. (1987). Mediation Services: An Evaluation. Manitoba Attorney General: Research, Planning and Evaluation. Roberts, L. (1998). Victim Offender Mediation: An Evaluation of the Pima County Juvenile Court Center's Vicitm Offender Mediation Program (VOMP). Available from Frasier Area Community Justice Initiatives in Langley, British Columbia. Roberts, T. (1995). Evaluation of the Victim Offender Mediation Project, Langely, BC: Final Report Victoria, BC: Focus Consultants. Roy, S. (1993) "Two Types of Juvenile Restitution Programs in Two Midwestern Counties: A Comparative Study," Federal Probation Vol. 57, pp. 48-53. Schneider, A. (1986). "Restitution and Recidivism Rates of Juvenile Offenders: Results from Four Experimental Studies," Criminology Vol. 24, pp. 533-552. Sherman, L., H. Strang, and D. Woods (2000). Recidivism Patterns in the Canberra Reintegrative Shaming Experiments (RISE). Centre for Restorative Justice, Australian National University. Stone, K. (2000). An Evaluation of Recidivism Rates for Resolutions Northwest's Victim-Offender Mediation Program. Masters Thesis. Portland, OR: Portland State University. Stone, S., W. Helms, and P. Edgeworth (1998). Cobb County Juvenile Court Mediation Program Evaluation. Strang, H. et. al. (1997). Rise Working Papers 1-4: A Series of Reports on Research in Progress on the Reintegrative Shaming Experiments (RISE) for Restorative Community Policy, Institute of Advanced Studies, Australian National University. Strode, E. (1997). Victims of Property Crime Meeting Their Juvenile Offenders: Victim Participants' Evaluation of the Dakota County (MN) Community Corrections Victim Offender Meeting Program. Masters Thesis. Northampton, MA: Smith College School of Social Work. Stuart, B. (1996). “Circle Sentencing: Turning Swords into Ploughshares.” In B. Galaway and J. Hudson (eds.) Restorative Justice: International Perspectives. Monsey, NY: Criminal Justice Press, pp. 193-206. Umbreit, M. (1988). "Mediation of Victim Offender Conflict," Journal of Dispute Resolution, pp. 85-105. Umbreit, M., (1989). "Violent Offenders and Their Victims," In M. Wright and B. Galaway (eds.) Mediation and Criminal Justice. London: Sage, pp. 99-112. Umbreit, M., (1989). "Crime Victims Seeking Fairness, Not Revenge: Toward Restorative Justice," Federal Probation (September 1989) pp. 52-57. Umbreit, M. (1991). "Minnesota Mediation Center Produces Positive Results," Corrections Today (August) pp. 194-197. Umbreit, M. (1995). "Mediation of Criminal Conflict: An Assessment of Programs in Four Canadian Provinces. St. Paul: Center for Restorative Justice and Mediation; Umbreit, M. (1996) "Restorative Justice Through Mediation: The Impact of Programs in Four Canadian Provinces, " in B. Galaway and J. Hudson, (eds.) Restorative Justice: International Perspectives. Monsey, NY: Criminal Justice Press Umbreit, M. (1999) "Victim Offender Mediation in Canada: The Impact of an Emerging Social Work Intervention," International Social Work. Umbreit, M. and W. Bradshaw (1997). "Victim Experience of Meeting Adult vs. Juvenile Offenders: A CrossNational Comparison," Federal Probation Vol. 61, (December) pp. 33-39. Umbreit, M. and W. Bradshaw (1999). "Factors that Contribute to Victim Satisfaction with Mediated Offender Dialogue in Winnipeg: An Emerging Area of Social Work Practice," Journal of Law and Social Work. Umbreit, M., W. Bradshaw, and R. Coates (1999). Victims of Severe Violence Meet the Offender: Restorative Justice Through Dialogue. St. Paul: Center for Restorative Justice & Mediation, University of Minnesota. Umbreit, M. and K. Brown (1999). Victims of Severe Violence Meet the Offender in Ohio. St.Paul: Center for Restorative Justice & Mediation, University of Minnesota. Umbreit, M. and R. Coates, (1992). Victim Offender Mediation: An Analysis of Programs in Four States of the US. Minneapolis: Minnesota Citizens Council on Crime and Justice. Other publications related to this study include: Umbreit, M. & R. Coates, "The Impact of Mediating Victim Offender Conflict: An Analysis of Programs in Three States," Juvenile & Family Court Journal (1992), pp. 1-8; Umbreit, M. & R. Coates, "Cross-Site Analysis of Victim-Offender Mediation in Four States," Crime & Delinquency Vol. 39, (Oct. 1993), pp 565-585; Umbreit, M. "Juvenile Offenders Meet Their Victims: The Impact of Mediation in Albuquerque, New Mexico," Family and Conciliation Courts Review Vol. 31, (Jan. 1993), pp.90-100; Umbreit, M. Victim Meets Offender. Monsey, NY: Criminal Justice Press, 1994; Umbreit, M. "Crime Victims Confront Their Offenders: The Impact of a Minneapolis Mediation Program," Research On Social Work Practice Vol. 4 (Oct. 1994), pp. 436-447; Umbreit, M. "Restorative Justice through Mediation: The Impact of Offenders Facing Their Victims in Oakland," Journal of Law and Social Work Vol. 5 ( Spring 1995) pp. 1-13; Umbreit, M. "Restorative Justice Through Victim-Offender Mediation: A Multi-Site Assessment," Western Criminology Review Vol.1 (1998). Umbreit, M., R. Coates, and A. Roberts, "Impact of Victim-Offender Mediation in Canada, England and the United States," The Crime Victims Report (January/February 1998) pp.20-92. Umbreit, M., R. Coates, and B. Vos (2001). Juvenile Victim Offender Mediation in Six Oregon Counties. Salem, OR: Oregon Dispute Resolution Commission. Umbreit, M. and C. Fercello, (1998). "Family Group Conferencing Program Results in Client Satisfaction," Juvenile Justice Update, (December/January). Umbreit, M and A. Roberts (1996). Mediation of Criminal Conflict in England: An Assessment of Services In Coventry And Leeds. St. Paul, MN: Center for Restorative Justice and Mediation and in Roberts, A. and M. Umbreit (1996). "Victim Offender Mediation: The English Experience," MEDIATION UK Vol 12, Summer. Warner, S. (1992) Making Amends: Justice for Victims and Offenders. Aldershot: Avebury. Wiinimaki, K. (1997). Doctoral Dissertation. School of Social Work, University of Tennesee. Wynne, J. (1996). "Leeds Mediation and Reparation Service: Ten Years Experience of Victim-Offender Mediation." In: B. Galaway and J. Hudson (eds.), Restorative Justice: International Perspectives. Monsey, NY: Criminal Justice Press, pp. 443-461. Wynne, J. and I. Brown (1998). "Can Mediation Cut Reoffending?" Probation Journal Vol. 45 pp. 21-26. Young, M. (1995). Restorative Community Justice: A Call to Action. Washington, DC: National Organization for Victim Assistance. Zehr, H. (1990). Changing Lenses: A New Focus for Crime and Justice. Scottsdale, PN: Herald Press. Zehr, H. (2002). Listening to Victims: Implications for Restorative Justice Practice in the U.S. and Abroad. Seminar sponsored by the Center For Restorative Justice & Peacemaking, University of Minnesota.