By Laura Mirsky

Part two in this series can be read at:
Part three in this series can be read at:

This is the first article in a series about family group conferencing (FGC), a restorative approach to problem-solving that involves the children, young persons and adults in families in making their own decisions. Originally developed in New Zealand, the family group conferencing process has taken root worldwide and is now known by several different names, including family group decision making and family unity meetings, among others. Family group conferencing began in the field of child welfare and youth justice, but is now used in mental health, education, domestic violence and other applications.

Families are more likely than professionals to find solutions which actively involve other family members, thus keeping the child within the care of the family, rather than transferring care of the child to the state.

Family group conferencing has acquired varied characteristics in the different locations where it is practiced, but certain common elements are evident, as well. In general, the philosophy underlying family group conferencing holds that families, when provided with the necessary pertinent information, are better able to devise plans to protect their own welfare than are professionals, because families know themselves — their problems, strengths and resources — better than professionals do. Young people need the sense of community, identity and stability that only the family, in its various forms, can provide, and families are more likely than professionals to find solutions which actively involve other family members, thus keeping the child within the care of the family, rather than transferring care of the child to the state. Also, when families are empowered to fix their own problems, the very process of empowerment facilitates healing.

The key features of the New Zealand FGC model are preparation, information giving, private family time, agreeing on the plan and monitoring and review. In an FGC, the family is the primary decision-maker. A wide definition of family applies, including extended family and close, concerned friends and neighbors. An independent coordinator facilitates the conference and refrains from offering preconceived ideas of the outcome. During private family time, the family, after hearing information about the case, is left alone to arrive at their own plan for the future of the child, youth or adult. The plan is evaluated by professionals with respect to safety and legal issues, and resources may be procured to help implement the plan. Professionals and family members monitor the plan's progress and often follow-up meetings are held.

A critical threshold for a family group conferencing program's inclusion in this article is the use of private family time. By taking the decision-making process out of the hands of professionals and governments and putting it back in the hands of those people who are directly affected, private family time indicates a crucial paradigm shift; one that is fully in tune with the International Institute for Restorative Practices's (IIRP's) definition of restorative practices and its goal of building a global alliance for family empowerment.

Mike Doolan, former chief social worker for the Department of Child, Youth and Family in New Zealand, helped develop FGC in New Zealand and has assisted FGC initiatives in the United Kingdom, Ireland, Denmark, Sweden and Israel. Doolan's report, "The Family Group Conference, 10 Years On," can be read at:

"The capacity of families to take control continues to astound us."
— Mike Doolan

New Zealand was formerly a "cosseted welfare state, too costly to maintain," where professional decision making dominated, said Doolan. New Zealand was also institutionally biased against the Maori people. A disproportionately high number of Maori children were in the care of the state, over-represented in social workers' caseloads and almost always placed with families of European descent. "The FGC process emerged because we were desperate to find alternatives to panels and courts staffed with people who seemed wealthy and racist in comparison to the people who appeared before them," said Doolan. At about the same as time as the welfare state began to break down, the Maori underwent a renaissance and wanted to take matters into their own hands.

In the Maori tradition, the community is responsible for its children and people face the community in cases of wrongdoing. Also, the smallest Maori social unit includes a person's every living relative. According to Doolan, the FGC concept of extended family derives from that tradition. In New Zealand, the law states that every member of an extended family is entitled to attend an FGC.

The Children, Young Persons and Their Families Act passed in New Zealand in November 1989. Said Doolan: "The act emphasizes that court proceedings, both civil and criminal, are a last resort, and encourages community-based solutions whereby families — take prime responsibility for their own children and young persons." Under the act, no court can make a decision on the disposition of a case unless an FGC has been held. More than 10 years later, the act has seen "outstanding results," said Doolan, adding, "The capacity of families to take control continues to astound us; families can make safe decisions for young people and are experts about themselves."

Doolan stressed the importance of mandated support for FGC. In New Zealand, he said, "we have a huge advantage," because FGC is the law, "not subject to policy-makers or practitioners." In other places, he said, "social work is having difficulty coming to terms with the notion of family led decision making. In the face of a growing body of knowledge — about the effectiveness of the family group conference partnership approach to problem resolution in child welfare, it still faces official and professional opposition where there is no legal mandate for the process."

Child welfare professionals across the world are grappling with this issue as Doolan has expressed it. Lisa Merkel-Holguin is director of the National Center on Family Group Decision Making (FGDM) within Children's Services at the American Humane Association. The center "provides training and technical assistance with FGDM all over the country and functions as a catalyst for good practice," said Merkel-Holguin. The center networks people with each other through conferences, its website and its online FGDM discussion group, and disseminates information about FGDM through videos and publications, including the AHA journal, Protecting Children. Two FGDM pilots are planned for 2004, with respect to adoption and pre-childbirth issues, to "expand the marketability of FGDM within child welfare," said Merkel-Holguin.

The center also creates mechanisms to research the effectiveness of FGDM to "standardize practices that are typically marginalized in this country," said Merkel-Holguin. The February 2003 issue of Protecting Children includes an extensive collection of family group conferencing research evaluations, a majority of which suggest a high level of satisfaction with family group conferencing by conference participants, both social service workers and families. A descriptive study on the center's website, co-sponsored by the center and spearheaded by Paul Nixon in the UK and Gale Burford in the U.S, asked anyone involved in with family group conferencing practice, research or administration to complete a simple, web-based survey.

"What started as an experiment in five communities in 1995 is now a widely recognized practice embraced by over 150 communities across the nation."
— Lisa Merkel-Holguin

Family group conferencing "has grown exponentially throughout the world," said Merkel-Holguin. In the United States, she said, "What started as an experiment in five communities in 1995 is now a widely recognized practice embraced by over 150 communities across the nation." England and Wales, she said, had only four pilot projects in 1994, but by 2002 had 97 family group conferencing initiatives running or under consideration by local authorities or non-governmental organizations (NGOs). But, she said, there is a difference between FGC in New Zealand, where it is a right, in the UK, where it is described as "good practice," and in the USA, where, "all too often, it's described as a 'tool' or a 'technique.' " She explained, "When practitioners view family group conferencing as a tool to be used on families and not as a process in which to engage them — they overlook the key preparation and follow-up steps that are critical to building community partnerships and increasing family involvement."

FGDM has three essential elements, said Merkel-Holguin. Quality preparation before the conference is fundamental, so that families, social workers and others invited have a clear understanding of their roles and what they're being asked to do. Private family time is key, because it turns decision making over to the family and allows them a sense of control. It is also imperative that the FGDM facilitator does not prescribe the outcome of the conference. Preparation, which takes 20–25 hours on average, is "the toughest sell," said, Merkel-Holguin, not because people don't believe it's important, but due to "a down-turned economy with already stretched budgets." Private family time, she said, "speaks not just to family empowerment but to family leadership." About the third element, Merkel-Holguin said, "Too often, we're seeing that conferences are not being used for decision making. Families are being asked to rubber-stamp an official outcome."

If these three elements are omitted, one negates the democracy-building potential of family group conferencing, said Merkel-Holguin, adding, "The work of Braithwaite [John Braithwaite, Crime, Shame and Reintegration, 1989] is helping us to think about FGDM within the context of the democratic process. It will be interesting to see if child welfare embraces FGDM as a democratic experiment."

Robert Tapsfield is chief executive of Family Rights Group (FRG), a registered charity established in the UK in 1974 to provide advice and support for families whose children are involved with social services. FRG was instrumental in developing FGC in England and Wales. In 1990, FRG invited a group of New Zealand FGC practitioners to the UK to speak about their experiences. FRG then implemented FGC child welfare pilots with six local authorities in England and Wales. All six cooperated with a landmark research project, Family Group Conferences in Child Welfare, by Peter Marsh and Gill Crow, Oxford University Press, 1998.

Today, FRG encourages and supports authorities and volunteer organizations in developing FGC initiatives in the UK, provides practice guidance, print and video resources and a website (, runs trainings and conferences, and lobbies nationally for greater use of family-based decision making in situations of family conflict. The FRG runs the Family Group Conference Network, which provides members with access to FGC information, training and consultation, clarifies core principles, disseminates good practice, develops models for project evaluation and publishes a tri-annual newsletter. Network membership fees support FRG's activities.

The UK Children's Act was passed in 1989. Tapsfield called it, "a unifying piece of legislation that brought together other bits of legislation for children in need" and provided that decisions be made according to a core set of principles, mandating that the state work in partnership with families. But, he said, "although the principles are sound, [the act] doesn't prescribe a mechanism that provides for the principles to be implemented." Since there is no legal mandate regarding practice, he said, professionals dominate the mechanisms and it's up to them to decide what to do.

"Given the lack of a legal mandate and financial difficulties for local authorities," said Tapsfield, FGC in the UK has been "an extremely wonderful achievement. Over 50 percent of communities in the UK are either using FGCs or considering doing so." On the other hand, he said, the decision to use FGCs is almost entirely at the discretion of local authorities. Families still do not have the right to ask for an FGC and FGCs "could disappear" in the UK. FGC survives, said Tapsfield, because of commitment in a range of voluntary and statutory organizations.

Private family time is "symbolic of the heart of FGC … it says that families are the leaders."
— Robert Tapsfield

Tapsfield said that the response of many UK social service providers to FGC has been very positive. They welcome it because it makes sense and gives them a way to use their expertise to be restorative. Providers in another group, however, have been reluctant to give up their traditional roles and their power. "Some people — on all levels — will be opposed; it doesn't matter what you do," he said. A third group of people may not understand — or may have chosen not to understand — FGC. They claim to be implementing the practice already — talking to relatives, etc. — without ever really learning about it. Researcher Peter Marsh invented the acronym "DATA" ("doing all this already") to describe them. Because most social welfare professionals have heard of FGC, they think they know what it is, and that can be dangerous, said Tapsfield. But, he said, "They don't understand that the mechanism is different from the way they're making decisions." The key element of FGC, said Tapsfield, is private family time, "not just because it works, but because it's symbolic of the heart of FGC … it says that families are the leaders."

There is increased interest in UK courts in FGC and the government has announced the publication of a "Green Paper" — a policy intention — on how to respond to children at risk of abuse or offending, which covers both child welfare and youth justice areas, said Tapsfield. FRG is lobbying the government about the Green Paper to commit to FGC as practice, because it works and because it strengthens communities and families. Tapsfield hopes for a government mandate ensuring families the right to access FGCs in situations where there are serious decisions to be made, and obliging the state to convene an FGC when it would otherwise act to permanently separate a child from its birth parents.

Joan Glode is executive director of Mi'kmaw Family and Children's Services of Nova Scotia, Canada. The agency is using FGC with children in the two First Nation (Aboriginal) tribes of Nova Scotia: the Mi'kmaw and the Maliseet. Glode, a member of the Mi'kmaw tribe, teaches an Aboriginal perspectives course at the Maritime School of Social Work at Dalhousie University in Halifax. Its curriculum concerns "building the past into the future." Said Glode: "We're doing the same thing with our agency."

When the agency was founded in 1985, there weren't enough First Nation social workers to staff it. "There were only two in the province of Nova Scotia," said Glode. The agency got a full mandate for its program in May 1990, along with a large group of First Nation social workers. At that point, said Glode, "we knew we should be looking at customary care [using traditional tribal methods]."

The agency's FGC initiative began, said Glode, partly as a result of her participation on a panel of the Law Commission of Canada, which examined government response to Aboriginal abuse in institutions. The panel found that the sole purpose of Canadian schools and orphanages regarding Aboriginal children had been to remove them from their culture, language and institutions. The schools had violated Aboriginal laws, by, for example, preventing mothers from passing down tribal traditions and knowledge to their daughters. For Glode, the panel's findings were an epiphany.

Glode attended the IIRP's international conference, in Toronto, Canada, August 2000. She received training in restorative practices and heard a plenary speech: "Aboriginal People and Justice Issues," by an Aboriginal judge, the Honorable Murray Sinclair, Associate Chief Judge, Manitoba. The talk traced the centuries-long, systematic destruction of Aboriginal resources, power and traditions by the Canadian government and stressed the importance of finding culturally appropriate approaches to the problems these policies had engendered.

FGC "fits with First Nation worldviews of respect, sharing resources, mutuality and interdependence, a family coming together and seeking its own solutions — the primary values in Aboriginal society."
— Joan Glode

On her return to Nova Scotia, Glode used the valuable information she had learned at the conference. A tribal chief was angry with the way Mi'kmaw Family and Children's Services had handled a child welfare case, said Glode, so the agency held its first FGC, with the family, agency workers and the chief attending. The meeting took five hours. By the end, the group had come up with a work plan for the child. Some weeks later, a second meeting was held and everything on the plan had been accomplished. After a third meeting, "that was the end of it," said Glode, "It's been going fine ever since … The family needed to be heard."

The FGC project was then in the pre-pilot stage. "We knew we needed to formalize what we were doing," said Glode. With help from D'Uma Young, a young lawyer with a degree in tribal justice, the agency began to adapt FGC and circles to child welfare and First Nation standards. Glode said that her agency uses the New Zealand FGC model. Private family time is used "if the people want it," she said, "When we finish up the formal piece, we come up with a plan. Sometimes we do it together; it depends on the level of anxiety or pain."

FGC is "so validating for us," said Glode, adding, "It fits with First Nation worldviews of respect, sharing resources, mutuality and interdependence, a family coming together and seeking its own solutions — the primary values in Aboriginal society." In times of stress, said Glode, the first language for many Mi'kmaw people is Mi'kmaw. "Translating into English is painful," she said, "FGC allows them to speak in whatever way they wish. If you give people permission to talk, they will. Things extraneous to the matter will come up — things the family needs to talk about but never has. It's important to make sure that people don't get shut out."

Glode's agency received one of five CND $25,000 grants from the Centers of Excellence in Child Welfare of Health Canada to implement an 18-month project to formalize FGC. The project, said Glode, will follow 30 child welfare scenarios over time, half using FGC, half using the "regular" approach. Glode's agency is also talking to the government of Nova Scotia about including FGC in child welfare standards. In under a year, she said, the agency will present their case to the province, which so far seems open to supporting FGC. Glode hopes that the province will amend legislation so that FGC and circles can be substituted for appearances before a family court justice. Currently, the law allows for mediation as a substitute, but, said Glode, "that doesn't work."

Rob van Pagee is CEO of Eigen-Kracht Centrale, the Netherlands, and is active in implementing FGC in that country. FGC is operating in five out of 12 provinces in the Netherlands. Van Pagee published a book, in Dutch, Eigen-Kracht FGC in the Netherlands: From Model to Implementation, the first book on FGC in the Netherlands. Eigen-Kracht recently received 22,500-euro award for Most Innovative Program in the Netherlands from the Union of Directors of Child Protection. Eigen-Kracht Centrale's website is:

We need to send a different message to society … "It’s not us who can help you, it’s you that can help you."
— Rob van Pagee

The results of traditional child welfare work are "not great," said van Pagee. Studies in the Netherlands showed that one third of child welfare cases were helped by the system, one third remained the same and one third were made worse. "If you did nothing, you'd have the same result," he said. In conventional child welfare work "you report child abuse anonymously and the case is put away in a secret file. This is the opposite of what you should do," he said, because, in perpetuating secrecy, you also perpetuate the problem.

The success of FGC in the Netherlands has been a revelation for child welfare workers, who were cynical about families and thought that it was going to be hard to get families to come to FGCs, said van Pagee. The reality, he said, is that the child welfare system has been a bigger problem for FGC than the families. "Many social workers feel they need to control safety issues," said van Pagee. When he taught FGC to second year students in social work school he faced a great deal of opposition. The problem lies in the way that society sees social workers — the image that social work sells to society. Social workers are regarded as saviors, so people become social workers to save people. We need to send a different message to society, said van Pagee: "It's not us who can help you, it's you that can help you." The beauty of the FGC model, said van Pagee, is that "I finally can do what I'm good at — facilitating, not making decisions for somebody else's life; that's not good for them or me."

FGC, said van Pagee, gives respect back where it belongs — to the families — who are willing and able to take responsibility for their problems. "It gives people energy to have a vehicle to deal with their problems," he said, adding, "It's really rare for people who are exposed to FGC not to like it." And, said van Pagee, people come in big numbers. The average number of people attending an FGC in the Netherlands is 16.7. They come, bring their children, make plans, bring their own resources, support their plans, and most of the plans they devise are accepted by referral workers as safe.

Eigen-Kracht has 60 trained FGC coordinators, drawn from the community — "citizens who like to do it" — not social workers, said van Pagee. Their task is to make sure the conference is organized the way the family wants it in terms of time and place, attendees, traditions (beginning the conference with a prayer, etc.), even type of food. The coordinator may also facilitate a person in the family — a natural leader, often an elder — to help with FGC arrangements. The coordinators should not have an interest in the conference's outcome, said van Pagee.

"The model is in place, we know it's working, now we can play with it," said van Pagee. The approach of restorative practices and FGC are so similar, he said, that he doesn't understand why people need to differentiate them. "It's just the organizations' need to define themselves and has nothing to do with the needs of individuals," he said. Conferencing turns the system upside down and puts the needs of individuals first. "If you understand that, then models aren't important anymore," he said, adding, "You can combine models and change them according to the needs of the family."

Van Pagee cited two cases where FGC and Real Justice scripted conference models were combined. In one, a 17-year-old boy had committed a crime and his family was set to have a Real Justice conference, but the victim wanted no part of it. So they had an FGC, in which 17 extended family members came up with a plan to prevent the boy from embarking on a life of crime. In another case, a father had murdered the mother of two young children and was in prison. The extended family was scheduled for an FGC to address the needs of the children, who had, in effect, lost both their parents. The two sides of the family, however, were not emotionally ready work together on the problem. Instead, a Real Justice conference was held, focusing on the feelings and thoughts around the murder, using scripted Real Justice questions. Proceeding around the circle, everybody had a chance to have his or her say. The family took a break for lunch, then came back and did an FGC about the needs of the children, which was very successful. The hard feelings the two families had for each other had been softened by the Real Justice conference, so that the FGC could work, said van Pagee.

Conferencing turns the system upside down and puts the needs of individuals first. "If you understand that, then models aren’t important anymore … You can combine models and change them according to the needs of the family."
— Rob van Pagee

Van Pagee hopes that FGC will become law in the Netherlands, where discussion about the possibility is already "on a pretty high level." He would like to see a law like New Zealand's, where citizens have the right to a conference before anybody else comes into the family. "We get the most complicated cases referred to us — the ones the social workers can't handle," he said, but "conferencing needs to be accessible to all. Why should the social workers decide which cases get referred? The same is true with Real Justice conferences. Why should police or prosecutors decide? Why shouldn't citizens have the right to decide?" These principles should apply in all situations in society where decisions need to be made, said van Pagee: elderly adults going into rest homes, in health care, domestic violence, cases of mental and physical handicaps. The key element is that the family has the first chance to deal with the problem. "That's the paradigm shift," said van Pagee. But you need vehicles to re-establish lost communities. That, he said, is why FGC and restorative conferences are important.

Karin Gunderson is a teaching associate at the Northwest Institute for Children and Families at the University of Washington School of Social Work, Seattle, Washington, USA. She manages grants, raises funds and oversees FGC studies in her state, where over 700 FGCs have been held. Gunderson collaborated on the largest long-term follow-up study of FGC published to date: "Long Term and Immediate Outcomes of Family Group Conferencing in Washington State (June 2001)." Seventy FGCs addressed the well-being of 138 children that had been in the child welfare system for over 90 days. Children who had an FGC experienced high rates of reunification or kinship placement and low rates of re-referral to Child Protective Services. These findings generally remained stable as long as two years post-conference. The paper may be read at:

Gunderson is involved in another New Zealand model-based FGC study with young people ages 11–18 in high-cost residential group care. Family members are attending conferences for all the youths, "even kids who'd burned all their bridges" with inappropriate behavior, said Gunderson. An average of six to eight family members attends the FGCs, two to three from the father's side. This is the average number of family members attending FGCs in every study done in the US, with the same distribution of members from the mother's and father's side, said Gunderson. The study will examine data 6 and 12 months after discharge from care. Interviews will be done with social workers, family members and youths to find out what worked and why. Preliminary findings suggest that, after FGCs, "the trajectory of the kids is changed," said Gunderson. Although they haven't necessarily been placed with family members, they have gone to "less restrictive placements, as opposed to juvenile justice facilities or the streets."

FGC provides resources, support, connections and improved relationships between families and professionals. Local family members may have antagonism for a child who has caused problems, but expanding the circle brings in other perspectives and knowledge about the child, for the benefit of both professionals and family members. As a result of FGCs, fewer kids are referred back into care. But, said Gunderson, FGC is not just a matter of family placement: "It's about reinvigorating family connections. Knowing who you are and where you belong are important ingredients in adult resilience."

There are only a few situations where FGCs aren't appropriate, said Gunderson, adding, "Sometimes there really is no family." FGCs are also not used in sexual abuse cases in the early stages of prosecution, because prosecutors fear that if everyone gets together in an FGC, the offender will manipulate the victim into withdrawing the charges. "Prosecutors don't want us to do to FGCs in any sex abuse cases," said Gunderson, "but we don't agree. The bottom line is, the child is the customer and we're going to do what's in the best interest of the child." If safety is a worry, the conference might be held in a firehouse or police station.

Many FGCs come about when social workers are faced with impossible dilemmas, said Gunderson, adding, "The older the child is, the more you see an impossible mess." And, she said, "every social worker has 30 of these cases." FGCs, said Gunderson, provide social workers with new thinking and new resources. For example, one family member can take the children on weekends; another can take them weekday afternoons. And, she said, relatives can ask for things that social workers can't.

FGC must "come in from the margins and be put into the hardwiring of child welfare."
— Karin Gunderson

"When a neutral person does preparation and families get together and start talking, so many unnecessary barriers become visible," said Gunderson. For example, FGC uncovered a gender bias in the system. Traditional practice understanding holds that fathers don't participate in child welfare. In FGC, however, not only do fathers come, but their fathers come, and their mothers, too. "Did we not invite the men," asked Gunderson, "or did fathers assume the process was going to make them feel worthless and so send the woman to take care of the problem?" Also, there's a widespread belief in child welfare that children in care "have no families, or that their families are dysfunctional, in jail or disgusted with their kids and won't get involved." The system reinforces this thinking and provides no help in finding relatives. But, said Gunderson, "If you think the aunt and grandmother are the child's only relatives, you've got another thing coming."

A critical problem in child welfare policy, said Gunderson, is that children placed with family members tend not to get the same level of support or resources as those placed in care outside the family. Funding streams are earmarked for foster parents, rather than kinship care. A foster family will get US $800 for three children; a relative, US $300. When a child goes to live with his grandmother, she is often single, older and poor, and she ends up having to give him up. The child then ends up back in care and stays there.

However, said Gunderson, there is now a large demographic of people reaching grandparent age in the US. Formerly, children who lived with their grandparents were mainly the poor and/or people of color and their needs were discounted, said Gunderson. Now, she said, many middle class white people are becoming caregivers to their grandchildren and demanding financial help to raise the children. But, said Gunderson, due to deeply entrenched values, the state won't give money to family caregivers, simply because the children are their relatives, even if the alternative costs much more money.

FGC must "come in from the margins and be put into the hardwiring of child welfare," and social workers must be required to name, search for and engage extended family, said Gunderson. She is working on a project to train social workers and lawyers in FGC together, so both can help clients participate in the FGC process. She also hopes to train judges in the value of FGC and to help remove legislative barriers to the involvement of extended families in child welfare. If you have supports in many places in the system, you have a much greater chance for success, she said.

Tight budgets are a concern now, however. The primary mandate in child welfare is to intervene to stop abuse and neglect, and Gunderson is afraid that will mean fewer FGC facilitators. Washington state's child welfare agency has been praised around the country for FGC, and every study recommends FGC, "yet we're losing funds for facilitators," she said. Still, child neglect cases are increasing across the country, especially chronic neglect due to substance abuse, and, as the economy declines, neglect cases increase. FGC is very effective in neglect cases, said Gunderson, especially substance abuse cases, because an extended family member can take a child while a parent gets help for substance abuse problems, but this rarely happens in foster care or adoption situations. The current US presidential administration has launched a big adoption drive, said Gunderson. But, she said, if you look for placement among relatives, it's much more organic.

Julia Hennessy is social work service manager for the Essex County Council, UK, which delivers FGCs in three areas of service: care and protection of children, young people who offend and adult mental health care planning. The Essex County Council Family Group Conference Project website is here.

The 1996 Essex Family Group Conference Project, a one year child welfare conferencing pilot program for children who were in state care, was so successful that "it is now mainstream policy," said Hennessy, adding, "The FGC process is embedded in Children's Services Practice and Policy across the county of Essex. As part of case planning in partnership with children, young people and their families, social workers consider and make a decision on using the process for family support, child protection and looked-after children, as well as prior to care proceedings applications." CPS has developed partnerships with the health and police departments. "The outcome of the partnerships has been exceptional," said Hennessy.

The New Zealand FGC model is employed. Good preparation is essential. The conference is held at a time and place convenient to as many family members and other close interested parties as possible. An independent coordinator facilitates. Private family time is used. Professionals agree to the family's plan, unless it places the child or youth at risk of significant harm. Various agencies and professionals negotiate resources. The plan is monitored and reviewed over time and review FGCs are built into the process, with dates arranged at the original FGC.

The real question professionals must ask themselves is: "Do you believe children should be with their families? And should people have a say in their own lives?"
— Julia Hennessy


FGC works because "it's about families being given good information to make decisions, children having the opportunity to have input," said Hennessy. In the past, she said, children were being put up for adoption because there was no opportunity to make other plans for them. Now, she said, families and communities are taking personal responsibility for their children. Involving the extended family in the conferences maintains a child's network, she said, and preserves its essential racial, community and family identity. Adopted children commonly become disconnected from these identities. Kinship care, a typical outcome of an FGC plan, preserves these identities.

Some health care professionals really like FGCs, said Hennessy, while some do not. "Individual power values are the hardest thing to overcome," she said, adding, "People are afraid of losing their power." But she said, the real question professionals must ask themselves is: "Do you believe children should be with their families? And should people have a say in their own lives?"

Essex County Council Child Protection Services (CPS) has seven senior practitioners and 20 independent coordinators and managers who deliver an average of 120 child welfare conferences per year in Essex County, said Hennessy. If a child is subject to care proceedings (i.e., a ward of the state), a lawyer asks for an FGC, but the case must go to court first. In the UK, said Hennessy, information on a child subject to care proceedings can't be disclosed without the court's permission.

An FGC can be accomplished in Essex County in two ways, said Hennessy. It can be voluntary when the parents retain full responsibility for the child. When an FGC is offered as an option the family must be in agreement and explanation must be given about the process. If the risk to the child is high, i.e., in cases of serious injury, sexual abuse, fractures or shaken baby, the court makes the initial care order and the local authority is the lead decision maker. These cases usually involve younger children where there's a strong possibility for adoption, said Hennessy, and the court can mandate that an FGC be held, even if the parents do not agree.

Hennessy cited the case of a family with two young children, both of whose parents had severe alcohol problems. The court determined that the children could not live with the parents and the referral came to CPS. The parents refused to attend the FGC, but the court mandated one anyway. The parents didn't attend the conference, but the rest of the family came together. "We knew there was an extended family out there," said Hennessy. The father had been married before and had older children whose information was important. In the end, both children were placed with a maternal aunt. This never would have happened without a court mandated FGC, said Hennessy.

Charles Clark is deputy chief constable of Essex County Police. He has been a police officer for 35 years and one of the lead chief police officers on youth justice issues in the UK for over 10 years. He has worked with the British Home Office on a range of policy development in relation to young people and was closely involved with the development of the Essex Family Group Conference Service.

The UK Crime and Disorder Act 1998 brought a statutory rigor to youth justice, expressing a clear and explicit aim toward prevention for the first time, said Clark. The introduction of Youth Offending Teams galvanized key players in the main statutory agencies — social services, probation, police, education and health — to work together on the prevention agenda. The Youth Justice Board of England and Wales was created to set national standards and targets and monitor performance.

The Essex police became interested in using FGC in youth justice because the model had proved successful in the child welfare arena. It was important that youth justice FGCs not be about punishment, but about changing people's behavior, said Clark. FGCs should not merely hold people to account, he said, but "set them up for a decent life in the future." However, Clark added, "When we asked for funding for the model, we not only wanted something that would feel good, but something that would give us hard-nosed results. The government held us accountable to reduce crime, too." The message, he said, is that FGCs "are not a soft option, but the toughest and most effective means of dealing with young people." It's much more difficult for youth to face up to what they've done than to simply be handed punishment, he said.

The Essex police, in conjunction with the Youth Justice Board, funded The Essex Family Group Conference, Young People Who Offend Project. The University of East Anglia provided a two-year independent evaluation of the project. Data was gathered from a small sample of 30 youth justice FGCs; the youth involved had offended at least three or four times and were at highest risk — on the verge of custodial sentences. The project's results were "spectacular," showing a "massive" reduction in reoffending rates, said Clark. Conventional youth justice approaches produce a thirty to sixty percent reoffending rate, while the project saw a reoffending rate of seven percent after two years.

FGCs "are not a soft option, but the toughest and most effective means of dealing with young people."
— Charles Clark

"I was skeptical when I first heard about the FGC process, but I was reassured and taken aback by the strength of the process," said Clark. He cited the case of a troubled young man who had burglarized residences of several elderly people. In foster care, the boy felt alienated and unloved and had fallen in with a bad crowd. Six months after the FGC, he was "very much on track," living with his aunt and uncle, attending school full time and speaking at FGC promotional events.

Private family time is an important part of the youth justice FGC model. The big issue, said Clark, is "What is the family going to do to help young people come through?" After the family comes up with a plan, it is regularly monitored to ensure that it is being followed. Families find solutions that are lasting and effective in an environment where facilitators are trained to bring people together, said Clark. Facilitators are drawn mainly from professional agencies and can include social workers, probation officers, police officers and others. The main qualification for a facilitator is his or her value system, said Clark: a genuine belief that the family is the best decision-maker.

In the youth justice FGC model, victims and their supporters attend conferences, along with youth offenders and their supporters. Of victims who participated in the project, 90 percent expressed satisfaction with the process and said they thought that FGCs should be offered to everyone in a similar situation. Clark said he was fascinated to see the interaction between victims and offenders. He expected a good deal of anger on the part of the victims, but found, instead, that most victims were very sympathetic toward the young offenders. Victims didn't condone the offenders' actions, he said, but were glad to be allowed to chastise them and tell how the offense had affected them.

It is important, said Clark, that the FGC process be marketed effectively. One way is by stressing that FGCs are much more cost-effective than conventional measures. Also, stringent evaluation is essential. The outstanding evaluation obtained from the Essex youth justice FGC project has been an effective marketing tool with government agencies. Clark said that there is now interest in youth justice FGCs within the British Home Office and the House of Lords, and among senior cabinet officials, coinciding with the government's interest in restorative justice.

Another article in this series about family group conferencing will be appearing soon on the Restorative Practices eForum.

Karin Gunderson, featured above, will be a plenary speaker at the IIRP conference, Building a Global Alliance for Restorative Practices and Family Empowerment, in Veldhoven, the Netherlands, in August 2003. For more information, go to:

By Laura Mirsky

This article is a follow-up to a presentation from “Dreaming of a New Reality,” the IIRP's Third International Conference on Conferencing, Circles and other Restorative Practices, August 2002. In that presentation Vidia Negrea, director of CSF Hungary, discussed her experience learning restorative practices in the USA and her plans to open a school in her country. The text of Negrea's presentation is available here.



Vidia Negrea (second from left), director of CSF Hungary, poses with staff members Laszlo Gupcsi (far left) and Erika Bognar (far right), and two students.

A new Community Service Foundation (CSF) school has just been launched in Hungary — its goal: to implement restorative practices with delinquent and at-risk youth in that country. The school, which opened January 6, 2003 in Budapest, is funded by grants from CSF and the International Institute for Restorative Practices (IIRP) in the United States and the Hungarian Ministry of Children, Youth and Sports. The Hungarian Ministries of Justice and of Social and Family Affairs are also supporting the school as “a model for institutions to work with children with problems or at risk,” said Vidia Negrea, the school’s director.The CSF school represents an important turning point for Hungary. As a matter of course, delinquent youth in that country have been removed from their home environment and housed in reformatories or other special institutions, put on probation or jailed. Negrea explained: “There is no law in Hungary supporting day treatment.” She hopes that the school will give troubled youth a better chance to reintegrate into society by providing an alternative approach in their own community. A new law will take effect in June acknowledging the need for services provided by NGOs (non-government organizations), in effect encouraging entities like CSF Hungary.

Working as a psychologist at the largest, oldest boys’ reformatory in Europe, in Aszod, Hungary, Negrea saw that the methods in use there weren’t working. Research data on 500 boys showed that six months after leaving the reformatory, 75 percent had reoffended. Data analysis revealed that the boys who had most closely followed reformatory rules had reoffended sooner and more often than others.

Boys at the reformatory told Negrea that they were scared to go home, afraid of being labeled for life as “bad” and terrified of meeting the victims of their crimes and their families. To address this fear, she asked 100 students in her care to write an imaginary letter to their victims. Their heartfelt, remorseful letters had Negrea and her colleagues in tears. Some boys wrote real letters to their victims, apologizing for the harm they had done to them and their families and offering any help they might provide. This was the first time Negrea felt she came close to fulfilling people’s needs, both victims and offenders, instead of just administering punishment.


My friends and colleagues at home deserve a place where they are happy to come to work, even with the most difficult students.
- Vidia Negrea

Negrea received her first training in restorative practices when Beth Rodman and Paul McCold of the IIRP came to Hungary. She was thrilled to finally find a framework for dealing with troubled youth similar to her experience with the letter-writing project at the reformatory. She later spent a year doing hands-on training in restorative practices at CSF schools in Pennsylvania and was determined to bring the fruits of her training back to Eastern Europe.

She also appreciated the restorative work environment created for the staff, where mutual support was a workplace priority. Said Negrea: “My friends and colleagues at home deserve a place where they are happy to come to work, even with the most difficult students. Here at CSF, I realized it was possible.”

At first, Negrea wasn’t confident that she could lead such an effort. Near the end of her stay in the United States, CSF/IIRP president Ted Wachtel suggested that Negrea try being a supervisor. When she served as a substitute supervisor at two CSF schools, she gained confidence working in new settings. “It was a great opportunity to see what I had learned and how my learning would apply elsewhere,” she noted.

The key to the work, Negrea discovered, “was not about being knowledgeable about everything happening in each particular school and having all the answers, but about using the same way of working with people and being responsible and respectful with people. The structure and philosophy were the same.” Negrea realized that her work in Hungary would be based on implementing the same structure and philosophy.

When she went home to Hungary in 2001, Negrea quit her two jobs: as a psychologist at the boys’ reformatory, and at the National Institute for Family and Children, and began to try to introduce restorative practices. Some of Negrea’s colleagues told her that Hungary wasn’t ready for those methods. “It’s not America,” they said. But Negrea understood that restorative practices were “not based on American culture, but on relationships between people.”

Negrea drew on her existing relationships with people in social services — students, probation officers, teachers, administrators, police — and “made a lot of noise” about why restorative practices are important. She was then asked to do trainings for government institutions that wanted to make changes. “I used to be sent in as an expert to say why things weren’t going well,” said Negrea, adding, “They expected me to report on how to change or close the institutions.”

Instead, Negrea did two-day Introduction to Restorative Practices trainings with all the employees at several institutions, “not just teachers and counselors, but the gatekeeper, the cook, the director.” Negrea told them: “If you really want to change, you have to think about how you can work together, otherwise, they’re going to close you down.”

Negrea was struck by how quickly people understood the practices she introduced. “Very simple people can understand in hours that it’s their responsibility to change the institution,” she said. The institutions stayed open and are now using restorative circles. In one big group home, a semi-secure unit “where children are sent because nobody else can deal with them,” the staff is using circles for themselves and with the children. At first it was strange for the staff to discuss their problems with each other. They were accustomed to blaming each other and relying on a hierarchy. But now they are expressing their issues openly and the children are developing their own group “norms” — standards for behavior.

Negrea also did restorative practices presentations for prosecutors, judges, lawyers, probation officers, schoolteachers and administrators — people who will refer children to her CSF school. She used IIRP training videos and overhead projections, which had been translated into Hungarian, as well as interactive exercises.

To elucidate the social discipline window, which illustrates the concepts of TO, FOR, NOT, and WITH, she tried an exercise at one of her presentations that was originally developed in a CSF Professional Learning Group. Dividing the participants into four groups, she gave each group the same simple task to perform (drawing a flower). She then did NOT do anything and completely neglected one group (little support or control); did the entire task FOR another (high support, little control); dictated TO another group exactly what to do (high control, little support); and gave the WITH group the help they needed, but let them perform the task themselves.

By the end of the exercise, the participants had obtained a thorough, visceral understanding of the restorative paradigm. People in the neglected group felt hurt because they were being ignored. Those in the FOR group sat there and did nothing, upset that they had no input. The TO group members were angry and resentful at being told what to do. The people in the WITH group were happy and productive, having been treated restoratively: that is, with both high support and high control.

In all, the restorative practices presentations went extremely well. “I was amazed,” said Negrea, adding: “Nobody wanted to break for lunch; they wanted to keep going.” Instruction in Hungary is traditionally very intellectual and theoretical, said Negrea. The restorative practices trainings, in contrast, were dynamic and interactive, based on real practice and understanding through action and reflection.

The trainings were a revelation. One teacher said that, deep inside, she had always known about the concepts Negrea had shared, but had never had the words for them, never known how simple and obvious they could be. Said Negrea: “I wanted the authorities to realize that they needed the program.”

Negrea’s strategy has been working. At her last presentation for police, some of them realized that they had been talking on the phone about the same students for a long time without ever meeting each other. They talked about their need to start working with each other, and not just in a formal way. “Like a small but good virus, it’s spreading,” said Negrea. The idea is to disseminate restorative practices everywhere. “Children at the new school will take what they learn home with them and it will affect their families,” said Negrea.

Negrea hired Erika Bognar, a young Hungarian woman, as her first CSF staff member. Bognar had six weeks of training at the CSF school in Bethlehem. Despite initial language difficulties, she felt that she learned a great deal about restorative practices to bring back to her country, and was able to work with Negrea on plans for the opening of CSF Hungary. A third staff member, Laszlo Gupcsi, was hired more recently as a counselor and math, science and art teacher, and there are plans to hire more staff as the school expands.

It's all working. Kids are confronting each other and taking responsibility for themselves and each other.
- Vidia Negrea

The school now has four students, with more on the way. The four are ages 13 – 16, all children at risk who were expelled from school: two who had used drugs, one with a criminal record and one with truancy problems. All are doing very well at CSF Hungary. These are kids who never wanted to come to school, but now they come on time every day, even during heavy snowstorms. They help each other get up in the morning by calling each other on the phone. “It’s a tough group,” said Negrea, but she feels that the kids and the staff are making great progress together.

“We’re doing exactly the same things we did at the CSF school in Bethlehem,” said Negrea: “The same schedule, the same group structure, the same questions, the same ‘feelings poster.’ ” And it’s all working. Kids are confronting each other and taking responsibility for themselves and each other. “I thought it would be more difficult,” said Negrea. But, she said, “the tough kids feel respected, so they have no reason to be disrespectful to the staff or to each other.”

Feedback from parents has been extremely positive. Negrea delighted parents by inviting them to attend intake interviews, along with the students and caseworkers. One parent exclaimed: “You mean we are going to be in the same room with our children and their social workers?” Said Negrea: “From the beginning, they have seen restorative practices and have been part of the program. Everybody had the chance to speak up.” It’s also a new thing for parents to be called with positive news from their children’s school, a “good surprise” for them, said Negrea. One mother said that she never thought that her boy would go to school — that he never went before without being pushed.

Feedback from caseworkers has been positive, as well. Most caseworkers didn’t believe students would come to the program. Now, said Negrea, “every day we are being viewed more and more seriously.” Students are being referred to the school from group homes all over Budapest. Soon, they will have from 12 to 15 kids. Negrea wants to keep the maximum at 15 for now.

Negrea is hopeful that restorative practices will be good for Hungary. “Like every other country that was under the socialists for so many years,” she said, “somebody always decided FOR you what you had to do, or told you what TO do.” For that reason, she thinks, the notion of doing things WITH people will be especially beneficial for Hungary. (Socialism is not a bad thing in and of itself, Negrea believes, but it became an unsound system because people abused power.) “Restorative practices will be a good tool to help people develop fair relationships between people,” she said. “Hungary needs that,” she said, “and not just Hungary!”

A "feedback group," shown here, is a structured opportunity for community members to tell one another how their behavior affects them.

“Some people are waiting for a miracle,” said Negrea, “but I need to remember that it’s not all me, it’s up to them, too. The referral sources — everyone.” In the candid spirit of restorative practices, “We’ll be open for everybody,” she said, adding, “The Ministry of Justice is already planning to visit.” Ultimately, sustained by the knowledge that she has the full support of CSF and the IIRP behind her, Negrea declared: “I’m sure it’s going to work.”

The IIRP’s Benjamin Wachtel is traveling to Hungary from Pennsylvania this month to help set up the school’s computer network, train the staff and kids to use it, and instruct them in digital video and photography. He’s looking forward to the experience, even though he doesn’t speak Hungarian and the kids don’t speak much English. Negrea and her staff are teaching the kids English, a little at a time, and Wachtel will no doubt provide assistance with that, as well.

The kids at CSF Hungary are proud to be the first ones in a new program, and to finally really belong somewhere. Said Negrea: “We have kids with a lot of new dreams.” We at the IIRP have great hopes for Negrea and CSF Hungary, and we’ll keep you informed of their progress.

By Laura Mirsky

Caption Text Test Stuff For FunNancy Riestenberg is a prevention specialist with the Minnesota Department of Children, Families and Learning (CFL), which oversees funding and programming for public schools in that state. She works with violence prevention education, restorative schools, safe and drug-free schools and coordinated school health programs. She was interviewed by journalist Laura Mirsky at IIRP's Third International Conference on Conferencing, Circles and other Restorative Practices in August 2002.

Q: How are restorative practices being implemented in public schools in Minnesota?

A: The whole range of restorative practices is happening in the schools: Family Group Conferencing, Restorative Group Conferencing, Circles to Repair Harm, Circles of Understanding, Victim Offender Dialogue. All of those practices are being tried to varying degrees in schools. The activity is at a relatively high level.

-Nancy Riestenberg

It's hard to keep track of it because people go to trainings and then they may do things that are difficult to see or to track. For instance, a teacher may go to a training and that may influence the way they conduct their classroom, or an administrator may go to a training and then handle discipline differently than what he or she did before. In some instances, school districts have taken a more holistic approach or a sort of institutional approach—they've actually changed some policy and decided to proactively train people. Activity is happening at all levels: at a kind of grass roots level, at a building level and in some instances at a district level.

Q: How have restorative practices changed the schools where they're being used?

A: It's hard to quantify that. I would say that they've changed the way people do their job and their attitudes toward their jobs. It's not uncommon for me to hear administrators say, "I like my job more," "I feel more confident that I'm really getting to the root of the problem," "I feel as though by doing this I make better connections between students and teachers," "I feel as though I have more connections with the students that I work with." The atmosphere in the building might feel better, more comfortable, more respectful. Others feel that their teaching has improved, that they're making connections with children, seeing children being empowered. It's always fun when a kid can ask for a talking piece and hold a circle in the corner of the playground with his friends and feel as though they have taken care of their problems themselves. When kids learn a problem-solving process and they practice it, it becomes their own; they figure out how they can do it themselves.

Sixth-grade students in a circle at Kaposia Elementary School, South St. Paul, Minn.

Q: What appeals to you about restorative practices?

A: I have always worked in the fields of prevention and education. I have dealt with social issues. One of the issues that intrigued me was bullying. There has always been a question in my mind about the school's response to this and how they might do it differently. If we suspend bullies or make them sit out from recess, how does that help them or make them change their behavior? How can you help people figure out another way of behaving so they don't hurt other people? I think a good restorative process attempts to get at that and gives people some ideas. It is also a place where people can begin to reflect on what they're doing and why they do it.

The other thing that appealed to me about restorative practices involves empathy. I spent a lot of time working in the area of child sexual abuse prevention education. One of the things that struck me is that one of the reasons why offenders do what they do – this most horrific act – is because they either don't know how to or have decided not to empathize with someone. If they were to stop for a moment and think about what this would be like, they just could not go down that road. I know that the skills of empathy can be taught. The parts of empathy can be taken apart, explored and looked at, but you can't make someone empathize with someone else. With restorative practices, you create a safe place, people are prepared, they're supported, they're there with their friends or their family and they know that they can express their feelings. Everybody is part of figuring out how to solve a problem. I think it creates an environment where empathy can happen. So you can lead them to the water and hope that they will drink in the process. In terms of prevention, that is such a key element. If we can get people to be able to empathize with each other, they're so much less likely to hurt one another. It's a place where you can both teach it and hold out the possibility of people being able to actually practice it.

Q: You did a three-year evaluation of the restorative practices in your school districts. Could you tell us about it?

A: We had money from the legislature to evaluate the implementation and results of using restorative practices in four different school districts. We had an urban kindergarten through eighth grade building, a suburban school district with three buildings, and a consolidated rural district that had three buildings and a rural high school. This was an evaluation. It was not research, so we didn't have control groups and all those sorts of things. This was really about gathering information, about telling the story of how they went about doing this and what kind of preliminary outcomes they saw. There were a number of things that emerged from this evaluation. I think one of the most important, and for some people reassuring, things that we found out is that to make change happen in a school you need to have at least two to three years. Just because somebody gets money in July they're not going to be able to implement immediately in September and then start testing it in January. Nothing happens that fast in a school. That's a very important thing for people to remember if they are going to sustain any kind of energy in trying to create change in a building or among people. It takes time. Even with people who have the heart, soul, energy, resources and desire, it still takes time.

Passing the "talking piece" at Kaposia

Another thing that we found is that it's easy to gather information about offenders, about what kind of an impact a restorative process might have had on them and whether or not they reoffended. It's a little bit harder to gather information about the impact on the person who was hurt or on the community. It doesn't show up as numbers, so you don't have it on the discipline track. "We had five victims last month and this month we only have two" –nobody keeps numbers like that. How would you define them anyway? That's kind of hard. That has to be defined by the person. Therefore, you have to use other ways of figuring out how to evaluate. You have to ask people questions and you have to go with their perceptions and feelings. That gives you a richer piece of information, but it's more qualitative than quantitative. The other thing I think we learned from this it that you really need to put the two together. You have to have numbers and you have to have the stories. You need to have both of those things together to get a clear understanding about what happened and to figure out what you can learn from that.

In one building in particular there were very strong quantitative results. It appears that with the institution of Circles to Repair Harm, along with circles used in the classroom for building community, they went from about seven incidents of violence a day to around one a day in the course of three-and-a-half years. That was a significant drop. There was also a significant drop over three years in terms of overall behavior referrals to the office. They went from somewhere in the thousands to 450, like 1600 to 450 over the course of three years. That drop was an all school effort. It was not the effort of just one person. It was the administrative team making decisions to do things differently, as well as quite a number of staff people deciding that they were going to include this community-building activity in their classroom.

That was another key thing we learned from the evaluation. If you just have a restorative intervention, that will get you someplace. If you just have classroom management skills that are cognitively based and are about problem solving rather than using power and control over kids, you will get someplace. If you get the two of them together, you will go so much farther in a quicker period of time because the whole school then becomes congruent. There are lots of classroom management approaches that are cognitively based. They're about problem solving. They're about helping individual children learn how to make amends, use conflict resolution, etc. That's been around for a long time and there's been good research on it indicating that it's very helpful, useful and that it makes the classroom a better place. But when you have the inevitable fight, which is going to happen no matter what, then what do you do? Are you going to be able to continue to carry the philosophy that you built up so nicely in the classroom into the principal's office or are you going to have to go back to suspensions and expulsions? So restorative interventions help to complete a kind of circle of support, if you will, for children within the school. The message is: we recognize that people are going to make mistakes, but that doesn't mean that you have to leave the community. We have this other way to hold you accountable and help you fix the problem that you made. That was a very interesting observation—when you put the two together, you just got further faster.

Q: Do you have any particular stories that you would like to share?

A: There's one story that I like to tell about four third-grade boys: three African-American boys and one European-American boy. This was in a suburban district, so the African-American boys were very much in the minority, about 10% of that school were kids of color. The white boy called the other three boys a racial slur. In this district that was considered a bottom-line behavior, which meant it was racial harassment and he could have been suspended for probably two or three days.

The boys were familiar with the circle process and the administrators were willing to consider a restorative response to the offense. The boys all agreed that they wanted to sit in a circle and talk about what had happened. The restorative justice planner in that school facilitated the circle. The significant thing that came out of the circle was that the three African-American boys had an opportunity to tell the boy what that slur meant. For one boy, it was the word that a white man used when he shot his uncle in the head. The second boy said, "it's the word those men in the white sheets use in the movies when they go to burn down my people's houses." The third boy said when he hears that word, "it just hurts my heart so much I just have to leave; I have to get away." I think the offending boy knew it was a powerful word, but I kind of want to believe that in third grade he didn't know just how powerful it was. He does now. He certainly knows now.

I thought the boy received a gift from those other boys. They had the courage to share that with him. What they wanted from this kid, to make amends for what he had done, was for them to be friends. So for the rest of that year, they played together on the playground. Three years later, the woman who ran that group said that they still played together. They were still friends. That's one of my favorite stories.

In another school incident, a fight broke out among about four or five boys. This fight happened, of course, in a context. It happened a couple of weeks after a boy in that school died in a tragic car accident. One boy made some disparaging remarks about the deceased–that what happened was probably his fault. Maybe he wasn't wearing his seat belt, or he was driving too fast, or he was impaired in some way. He was blaming the person who had died, in a way. Some of the friends of the boy who had died heard this and were incensed. It was very recent. Grief was still very high in everybody's mind and they jumped him. They all got into a fight.

The boys would normally have been suspended for at least three days for a fight. But, in this particular instance, they all agreed to go to the restorative justice planner in the school. They wanted to have a circle to talk about what happened. As they talked about how their behavior had affected themselves and other people, they all came to the conclusion that the person they had harmed the most was the boy who had died. To make amends for this, they all got into the restorative justice planner's car and they went to the graveyard and apologized to the gravestone. It's that kind of creativity that is so compelling for me, where you can have a connection between the true harm and the consequence. That is profound. Not only is it profound, I'm sure that it was also therapeutic for these boys. I bet it was probably healing for them. I think that it helped them to appreciate more what they had lost. That's the kind of connection we need as human beings. That's what being a human being is about. It's not about the recipe of the student conduct book. Those are some of my favorite stories.

Q: What advice would you give someone who wants to bring restorative practices into their school?

A: I guess the first thing that I would advise people is that if they are in a position to do it and they believe in it, they should just do it. There's a lot of autonomy amongst the adults in a school, and wherever you can find it, even if it is just with the low-level stuff, it is a good way to operate. In doing that, you can do that grassroots kind of building where people hear about things and they become intrigued. They come because of their own interest to try to find out about something. If you are a person who does have a position of power, then maybe what you want to do is try to go at the top end. I think that you just need to decide what your sphere of influence is and start there.

Another piece of advice is part of the restorative philosophy: This is a process that should be voluntary. Just let go of the idea that everybody in a school is supposed to do this and that every incident needs to be handled this way and everybody needs to be a believer and everybody needs to participate. That's a road to exhaustion. Look for people who are friends. Look for people who are compatible. Look for people who would be advocates with you. Go where the strengths are. Do that in a respectful way. That's what the philosophy calls people to do.

I think the other thing that is important in trying to do something in a school is that if you get to a position where you can do training, involve kids in the training. It makes a big difference. It gives people a different perspective on kids to see them in a different setting. They offer an enormous amount of wisdom and perspective.

Q: What do you see, hope, dream is the future of restorative practices in schools?

A: I hope that this is not a fad like a lot of other things. I'm concerned that people don't co-opt the term just to appease people, that they don't just call certain things restorative when they're not. I think a good example of that is community service. Community service is a wonderful thing. It can be an extraordinary way for kids to learn. It's a great way to teach people. It can be part of a restorative agreement where you use community service as a means to make amends. But when you tell somebody to do community service, just because it's community service doesn't mean it's restorative. You're missing those steps of coming together, talking about who was affected and then deciding together what would be some ways to repair this harm. If one of the things that they figure out is, "Oh, it would be cool if you did community service. You took time away from education, so why don't you tutor in the classroom? You're good at math." That's very different than saying, you took time away in the hallway, so now I want you to go to community service.

A celebratory moment

I hope I come to a point where I will be able to ask school people if they have policies attending to the needs of victims in their school and they will be able to answer yes. They will be able to articulate what those are. There will be things offered to kids when they have been harmed, harassed, bullied, or part of a fight–the opportunity to talk to someone, to get education, to be able to ask for a restorative process. I would hope that we could come to a time when the school is not just focused on the person who did the harm, but is equally focused on the person who was harmed. Pairing punishment with restorative processes is perhaps problematic. People do it because it satisfies both sides. Certainly, even if you suspend the child, doing the restorative process pays attention to the victim, but I hope that we would get to a place where we would not have to do the two of them together, that people would be satisfied with restorative consequence.

I hope that people make the connection between restorative interventions and the way the staff is trained to talk to kids, to manage their classroom and to try to help the kids with their behavior. I would like to see them make the connection between restorative interventions and the health curriculum they teach about problem solving and decision making. I would hope for school people to have enough time somewhere in their lives where they can stop and breathe and see the larger picture.

A video of this interview can be found here.

This page includes a summary and link to the full report on the findings of a three-year study of the Thames Valley Police restorative justice program, the first large-scale restorative justice program in the UK. The study was conducted by Carolyn Hoyle, Richard Young and Roderick Hill, researchers from the Centre for Criminological Research at Oxford University.

Papers and reports on the Reintegrative Shaming Experiments (RISE) for restorative community policing. RISE is a project of the Law Program, Research School of Social Sciences, Australian National University, Canberra.

By Laura Mirsky

Len Wildman and Tom Dwyer work for the Rochester Police Department in Rochester, New York, U.S.A. Len is the manager of the Family and Victims Services section. Tom is the coordinator of that section’s Juvenile Accountability Conferencing (JAC) program. They were interviewed by reporter Laura Mirsky at IIRP’s Third International Conference on Conferencing, Circles and other Restorative Practices in August 2002.

Len Wildman is manager of the Family and Victims Services section, Rochester Police Department, Rochester, N.Y.
Tom Dwyer is coordinator of Juvenile Accountability Conferencing, a program of the Rochester Police Department’s Family and Victims Services section.

Q: How did you get involved in integrating conferencing into the police department?

A (Len): Before Tom came to the police department to coordinate the project, we had heard about conferencing within our department, the juvenile section. We were interested in focusing on juveniles and developing a diversion program. We knew we needed more and better diversion programs. We just didn’t know in which direction we should go. We had eliminated a few programs because our research was telling us that they weren’t effective. One of our commanding officers had heard about something called restorative practices. She really didn’t know much about it and she asked me to pursue it. I called a number of people.

One of them directed me to Bethlehem, Pennsylvania. I met the folks at Real Justice [an IIRP program] on the phone. After a number of conversations and some faxed material, I sat down with my supervisor and told her what I thought the program was about. It was a unique idea and I thought it would work for police officers in our department. She was convinced that it was a good thing to try. She found some money for us. We hired Real Justice to come and train 20 of our police officers and about 10 of our civilians in my section to be conference facilitators. As an experiment, we chose a junior high school. All of our junior and senior high schools have school police officers. We picked a few officers and began the process. We did about 40 conferences in a year. We thought it was very successful. We wrote a grant, won the award and were able to hire Tom to coordinate the program. This was the official beginning of our program which has now operated for three years.

Q: Do you feel that the program has worked well?

A (Len): We are asked that by our grants people and by our chief. We measure our recidivism rate by noting whether or not the juvenile enters back into the juvenile system. We don’t track them after they become adults and in New York state you become an adult when you are 16. We have had a 93% success rate thus far. Again, by our definition, this means they don’t come back into the system. I think that mirrors some of the research that has been done, so we are glad that we are on target and that helps us define this as successful. I was a little frightened in the beginning because most of the police departments with such programs were small and served communities of less than 50,000 people. And I was a little concerned about how this process would work in a large police department with a more diverse community. I would go around at conferences asking, “How big is your community?” and, “Is there somebody around who is doing this in a community of 250,000 people that is multi-ethnic and has varying backgrounds?” I didn’t meet anybody. There were a few attempts in private agencies, but I didn’t meet anybody within the police community. That is in this country (USA). I think in other countries, it is different. So it was a concern.

I don’t have that concern now. It is a bit bureaucratic. It’s a little bit more difficult because of the largeness of our organization. However, I don’t think that’s the fault of restorative practices, I just think it is the problem of trying to implement a program in a large organization.

Q: What kinds of offenses are you dealing with in this program?

A (Len): Based on the advice of a few people and my own notions about starting new programs, I tried not to make too many hard and fast rules. We said we would target juveniles, 15 and younger. We decided we would look for minor crimes, where children have minimal involvement with the police or none, but that we would not keep ourselves from felonies or other situations outside that definition if we thought they were appropriate. We also decided that we wanted to hold conferences pre-court rather then post-court, again with a few exceptions.

Also, the fire department been having difficulty with arson and juveniles. Aside from a very small percentage of children who set fires who have mental health issues, most kids set them because they are playing with matches. It’s accidental, but it’s very scary, obviously. In New York state, all fires, regardless of who sets them or what the intent is, are felonies. They don’t feel comfortable charging 9-year-olds with felonies. So they were at a loss for several years on what to do and how to manage this. I suggested to Tom to go talk to the arson investigator about using conferencing as an alternative tool. He has, in the last three months, gotten a lot of conferences from that.

Q: Tom, how would you describe your experience with this program?

A (Tom): We’ve had 140 conferences to date and I get goose bumps from the 140th one just like I did from the first one. We see that the process is really working and, in turn, there are some positive outcomes that come from conferences. It’s exciting to be a part of this and it just keeps getting better for me.

Q: Are you personally involved with each conference?

A (Tom): Yes, I go to just about all of the conferences as the coordinator. If I have facilitators that come in, I like to go through the conference afterwards with them, to debrief them and give them some feedback on their facilitating. I also get feedback from the participants. At the end of each conference, we pass out surveys to all the participants and get their feedback. We value their input as well and file each survey.

Q: Can you tell me the kind of feedback you have received?

A (Tom): It’s been exceptionally positive in just about all cases. Victims, offenders and their supporters really like this process. They feel there should be more of it and they wonder why it didn’t happen years ago. They really embrace the process. Everyone needs to sign the conference agreement at the end, so if they’re not satisfied, they need to work that through during the conference. There is a dialog that occurs. Most of our questions come from a script. However, a lot of our conferences, after all of the questions are asked in the script, it’s open to dialogue - people sharing more of their story and more about their experience with each other.

Q: Can you tell me about a particular conference that really had an affect on you or really stood out for you in some way?

A (Tom): There’s been many. There is such a dynamic about each one that is very special. It’s really difficult to find just one. (To Len) Which one should we pick?

A (Len): Tom, I don’t attend very many anymore, but I used to, and I used to facilitate them. One that Tom facilitated, and I’m sure he remembers, was after Columbine and during the school shootings that were occurring around the country. We had a 14-year-old boy who had a replica of a Beretta handgun. It was not a red-handled gun, so it looked very real. He was waving it around in front of the school and pointing it at children and passersby.

The officer that arrested him was very concerned that this kid just didn’t get it. He didn’t get the distress that he had caused all these people. The officer was more concerned about the kid understanding this than he was about the punishment. The officer called Tom and asked, “What can we do about this?” Tom suggested a conference. In the course of the conference, the child heard from teachers, passersby and other students about how his actions affected them, seeing that gun, imagining something terrible was happening in their community like they had seen on TV. Particularly, I remember one man who was on the second floor of a museum across the street who could not hear anything but could see what was happening. He was too frightened to come to the conference - still. But he wrote a letter that he asked to be read. The child began to break down and cry, feeling the impact of what he had done. To me, that was very powerful - and still is.

Q: Was his family there?

A (Tom): His mother was there and the boy’s employer came to the conference. The victim, the police officer and school officials came to express how it affected them.

Q: Has there ever been another problem with this child?

A (Tom): No. He’s done well. He found a lot of support at the conference. He realized that people weren’t out to get him, but just wanted him to know the effect of his actions on other people. That was really important for those that came, that was the most important thing.

Q (Len to Tom): Do you remember the one about the wastepaper basket?

A (Tom): Yes. This incident occurred at a middle school when a boy, on a dare, threw a wastebasket over a railing and hit a girl down below. This could have been a tragic incident.

I first heard about it through the victim’s godfather, who knew about restorative practices and knew about what the police department was doing. The victim’s family was very religious. They were seeking answers as to why this happened to their daughter. These two students had never met each other.

The conference brought all of those people together to share how they were affected by the incident and what they wanted as an outcome. One of the outcomes of the conference was that this boy, as a community service project, would go out to the schools with a police officer to tell other young people about peer pressure, about accepting a dare and the consequences of his actions. Interestingly, the young lady wanted to join him and talk about her feelings and how she was affected. That was completely voluntary and wasn’t planned going into the conference. That was from the stories and people talking about their feelings. Everyone embraced this. In fact, the offender, out of his own personal fund, brought her flowers before the conference even occurred. Those were presented to her before the conference even began. I had no idea that was going to take place. It just showed the profound effect it had on the offender, as well as the victim and their families.

Q: Tom, did you do other police work before you started this sort of facilitating, were you a police officer?

A (Tom): No. Actually, I worked in the finance area, working with police grants. I also did some volunteer work with the victim assistance unit within the Rochester Police Department, so I knew a lot of the personnel in the department. I had a lot of human services experience. The job and the philosophy behind restorative practices, it was something that was a part of who I was. I think it’s part of my belief system: forgiveness, reconciliation, healing, positive communication and people taking responsibility for their actions. I feel that is the most important thing about working in this position: that your belief system is consistent with restorative practices.

Q: Do you both feel that the work you are doing is having an impact on the community at large?

A (Tom): I know it is. I think the people make it possible and the police officers that give us the referrals. You can’t do this work alone. It’s other people identifying with the same issues as I just explained, the same personal issues. I think we have all been victimized at one point or another in our lives and this is how we would want to deal with it in our own personal situations. It’s very helpful that it is out there because most people really don’t want to go to court. I see it as a win-win situation for the police department and the community to have a process that includes all the people that are affected and gives everyone equal ability to express their emotions. It’s the kind of program I see building from the bottom up. No one really embraces the program until you really experience it and see it for yourself. Len and I both do presentations and talk about juvenile accountability, but it’s not until you participate, in some role, in the conference that you see the dynamics and sparks really happening. You see the receptivity of the community and how they rally around it. It’s a growing process. The best way to become educated about it is to take part in the process. We also bring in a lot of community members that just want to observe our conferences and they’re more than welcome to come and see any of our conferences. It’s a very open format.

One of the things that has additionally come out of conferencing and marketing is that the Rochester Police Department has contracted out with many agencies in the Rochester community. These agencies work with offenders when victims have asked that as a consequence for the crime or violation that the offenders give back to the community. So a lot of the victims and their families are asking that there be a community service project for that youth. That’s another way for the community to know more about restorative practices. In fact, we do not enter into a contract with agencies until we spend some time educating them about restorative practices. We want them to know what a conference experience was like for that young person and some of the issues that may have come out of that conference. Additionally, we share the conference agreement with the community service agency, so they know the responsibilities that the young person has in carrying out the conditions that were set up for him or her.

Some of our young offenders are working with painters, painting houses inside and out. They are working with carpenters and learning carpentry skills along with doing their community service project. The offenders really embrace some of the skills they learn along with the projects. They come to know of the resources that agency can offer them. We have an open door mission, so the kids serve food to the homeless and less fortunate. All these kids are supervised in their community service projects. It’s another form of outreach for us where the community gets involved and becomes part of the process.

A (Len): In the beginning, Tom would have to scurry around and find an agency that might do community service. We would have to explain what we were about and what kind of a community service project we wanted the youth to do. We weren’t interested in children picking up papers on the side of the road. We wanted them to get something out of it, an added value, something beyond the actual work. In the second year of our grant, we talked about finding agencies that would work with us. If the agency was in the neighborhood where the child lived, we would go to them and ask, “Can you take this child? What could you do?” For example, in the third year of the grant, Tom came to me and told me about two 7-year-old kids who had broken some windows. Their community service was to work in a voluntary vegetable garden in the community. They apparently enjoyed this because after their service was up, I think it was 10 hours, they asked the volunteer coordinator of that agency if they could stay for the rest of the summer. So the agency called Tom and asked, “Is it OK if these kids stay? They want to learn about gardening and we want to teach them about adult relationships.” Tom said, “Yeah, if they want to!” I think the idea is, “Why not take the next step?” It seems natural. It allows them to connect with the child and bring the child back into the community.

Q: Do you see restorative practices expanding to other departments within the police department? Would you like to see that?

A (Len): Yes, I would like to see that. I have some notions that I’ve been kicking around about how restorative practices can be part of ethical practices and how to integrate that into the way we operate as a police department or, for that matter, any organization. I also see that it could become part of management, the way we manage and supervise each other. I’ve been experimenting a little within my organization, with my own section. Also, there is a group of people who have formed something called the Finger Lakes Restorative Justice Consortium, of which Tom and I are members. They are more interested in promoting restorative justice in its many forms, not just the forms we use. We are a part of that effort. Also, the county of Monroe, which we are in, has about one million people and there are many different governmental agencies—school districts and police departments—who have inquired about what we are doing and how it works. The juvenile detention center, which is county-operated, is asking us how they could use that practice in their detention center. We are very willing to talk about what we do, what we believe in and in keeping with the practice of restorative justice, let them decide what’s best for them.

Q: Has this begun to spread to the schools in your area?

A (Len): In our area, the schools utilize our program. They like it because Tom can be there in a few hours. He can usually set up a conference in less than a week, as opposed to slower court or school processes. They love that Tom gets involved in addressing issues they’re having with the child, when they don’t think the school process will work fast enough or well enough from their perspective. They have all seen what the conferences can do. Something that I thought 3 years ago I think is beginning to happen now. It has snowballed. Now there are times when Tom may conference 5 to 10 a week, as opposed to 5 to 10 a month, to the point that we’ve hired a part-time person to work with Tom to help him with the process. We still want to keep to the belief that face-to-face contact with each of the participants and sitting down with them and talking to them about their roles and expectations of the process is better than phone conversations. We are on the brink of becoming overwhelmed as people the process as an effective tool. I’m very glad for The Finger Lakes Consortium because they can help in directing people. There are so many ways of doing this. It doesn’t have to be our way or a particular model. Our model works for us, but there are other models that may work for other people. Also, I know two districts outside the Rochester City School District have already explored training some people and doing their own conferencing.

A (Tom): Additionally, I want to make mention of one middle school that has endorsed us completely. It is the collaborative effort of many folks. In lieu of short term suspensions, they agree to juvenile accountability conferencing in place of a suspension, which they feel is punitive and doesn’t serve the needs of their students or the families that are involved. We have a school resource officer, two of them actually, in one middle school. The administrators in that school have come to endorse and believe in restorative practices. In that school, the teachers know us. There are many instances of bullying, different low-level assaults and theft at that school. We have a little location within the school. We’re allowed to use some of the equipment in the school to make copies of our agreements, to use the phones and to work with some of the staff in educating them about restorative practices. We go where we’re wanted and where we can be a resource to the community.

Q: What advice would you give to a police department that wants to get into restorative practices?

A (Len): You have to find a few people who really believe in it. We eventually targeted one school and one section of the police department—we have seven—to see if we could create some interest. So it was a slow start. We require, with a few exceptions, that the arresting office be a part of the conference. So we have to arrange for that person to be there at a time that is convenient for everybody. As they saw the process, they thought it was really good. They would go out and tell a couple more police officers. One person who was a supporter held the rank of commander, a high-level management position. She believed in it. And because she believed in the idea, she allowed me to do it and gave me the opportunities to do it. When she was promoted and no longer in my particular bureau, she still supported it. I always felt I had an advocate. Also, it was important to find someone like Tom Dwyer, who is very sincere, very patient and flexible. I hired Tom primarily because when he asked me what the rules were, I told him, “I don’t know.” When he asked me what the parameters of the program would be, I said, “I don’t quite know yet.” He liked that. Most of the other people I interviewed had social work backgrounds. They were very competent and probably, in some ways, more qualified than a person with an academic background in finance and accounting. But they wanted to know what the rules were and what the parameters were. I didn’t want to set those yet because I didn’t know how it would grow. He’s comfortable with that. That is the advice: don’t get too comfortable with the rules. Let this play out. It’s kind of organic. It won’t hurt anybody. I had a professor in college when I was in the master’s program in counseling who said, “No one ever died from counseling, so don’t worry about it.” No one is going to die from this process. We learn from mistakes we make or we learn from the process. People teach us, I think, all the time.

Q: Anything else you would like to say?

A (Tom): I’m just anxious to get back to Rochester to do more conferencing. Someone else in the conference earlier mentioned that she was anxious to get back to do circles. That’s where I really get my satisfaction. That’s what excites me, running the conferences and meeting all these families that have the same concerns that I would with my family.

A (Len): I have been employed with the department for many years as I mentioned. I have done many different projects and I am thinking seriously about retiring in 6 months. So, for me this has been a very satisfactory way to leave my agency. I feel that this is a wonderful piece, to see it grow and live. It certainly will live well beyond my employment in the department. I think when I do retire, between fishing trips and gardening, I might write about my experiences. It feels very satisfying to have been a part of this contribution to the community and see how the community has reacted to it. I think the best way to leave someplace you work is when you feel good about it and I feel good about it.

If you would like to learn more about the Juvenile Accountability Conferencing program or have any questions you may contact Len Wildman or Tom Dwyer at (585) 428-7236.

For more information about the IIRP’s Real Justice program and available trainings go to:


This reserach brief of a study by Paul McCold, Ph.D., reports on the effect of restorative practices on reoffending, attitudes, and program completion among students at the CSF Buxmont schools in southeastern Pennsylvania. Dr. McCold presented this paper at the American Society of Criminology annual meeting, November 13-16, 2002, Chicago, Illinois.

This executive summary of a study by Paul McCold, Ph.D., reports on the effect of restorative practices on reoffending, attitudes, and program completion among students at the CSF Buxmont schools in southeastern Pennsylvania. Dr. McCold presented this paper at the American Society of Criminology annual meeting, November 13-16, 2002, Chicago, Illinois.

By Paul McCold

Paper presented at the American Society of Criminology
annual meeting, November 13-16, 2002, Chicago, Illinois.

The IIRP gratefully acknowledges the assistance of the juvenile probation departments of Northampton, Bucks, Montgomery, and Lehigh counties, Pennsylvania, for providing access to court data used in this analysis.

The Community Service Foundation (CSF) and Buxmont Academy operate six school/day treatment programs (abbreviated as “CSF Buxmont schools”) in southeastern Pennsylvania. They are non-secure community treatment settings for adjudicated delinquent and at-risk youth.[1] Additionally, CSF operates three auxiliary programs—the residential, intensive supervision, and home and community programs. The residential program provides group home services to a portion of students attending CSF Buxmont schools. The intensive supervision program provides additional support to youth released from drug treatment facilities. Finally, the home and community program provides a variety of support services for youth experiencing family, emotional, and drug and alcohol related crises. Some students attending CSF Buxmont schools participate in multiple programs simultaneously.

All of these programs utilize what is broadly termed “restorative practices.” Restorative practices provide both high levels of control and high levels of support to encourage appropriate behavior.[2] The philosophy underlying these practices holds that human beings are happier, more productive, and more likely to make positive changes in their behavior when those in positions of authority do things WITH them, rather than TO them or FOR them. This hypothesis maintains that the punitive and authoritarian TO mode and the permissive and paternalistic FOR mode are not as effective as the restorative, participatory, and engaging WITH mode.

CSF Buxmont has developed a culture in which “restorative” characterizes not only staff interaction with youth, but staff-to-staff and student-to-student relationships as well.[3] This researcher has coined the term “restorative milieu” because CSF Buxmont culture is comprised of many restorative techniques and processes and not just isolated restorative interventions.

Within this restorative milieu, youth are held accountable for their actions while being given the social and emotional support necessary to make changes. Restorative practices empower the young person and the group to develop their own behavioral standards and actively confront misbehavior. The young people act as a micro-community of support for each other, consciously building interdependency and a sense of responsibility to the community. This process is facilitated by CSF Buxmont staff.

CSF Buxmont schools have an excellent reputation for empowering youth to make positive changes. However, since their founding in 1977, they have never been formally evaluated. This analysis tests the effectiveness of restorative practices in encouraging positive changes among adjudicated delinquent and at-risk youth. However, correlation does not mean causation. To conclude that any observed changes in attitudes or behavior are the result of participation in CSF Buxmont’s restorative milieu, four conditions must be met:

1) The attitudes and behavior of youth must significantly improve.

2) These improvements must be positively related to the length of participation.

3) These improvements must remain after controlling for known risk factors, such as age, gender, race, and prior offending.

4) These improvements must remain after controlling for the effects, if any, between these risk factors and early program discharge (case selection bias).



This analysis presents the outcome experiences of the 919 youth discharged from CSF Buxmont schools between June 1, 1999 and August 30, 2001. The 334 youth discharged during the 1999-2000 school year and the 445 youth discharged during the 2000-2001 school year include all youth discharged during these school years. The 140 youth discharged during school year 1998-1999 include only those discharged during the final two weeks of that school year.

The total number of youth discharged from CSF Buxmont schools between the 1999-2000 school year and the 2000-2001 school year increased by 33% (from 334 to 445). Reflecting the expansion of CSF Buxmont programs generally, this increase in the number of discharges was greatest for school referrals, with a one year increase of 69% (from 104 to 176), compared to 18% among probation referred (from 192 to 226), and 13% for children and youth referrals (from 38 to 43).

Youth are referred to CSF Buxmont schools from three sources: juvenile probation (56% or 516), children and youth services (10% or 88), and individual school districts (34% or 315). Probation referrals are overrepresented among discharges prior to September 1999 (70%), and children and youth referrals are somewhat underrepresented (5%) during this time period.

Demographics. The demographic distribution of cases is presented in Table 1.

Table 1
CSF Buxmont Schools Sample

1) Gender

Overall, girls represented 28% (260) of all discharged youth, most of whom were referred by probation. A majority of children and youth referrals were girls (52%), compared to one-fourth of probation and school referrals (both 26%).

2) Age

Overall, 55% of youth were 16 years old or older upon entering CSF Buxmont schools. Probation referred youth were older than those referred by schools or children and youth. Most of probation referred youth (62%) were 16 years old or older when entering CSF Buxmont schools, compared to less than half of school and children and youth referrals (47% and 41% respectively). The average age on entry into CSF Buxmont schools was 15.8 years old for probation referred youth, 15.5 for school referred youth, and 15.3 for children and youth (F = 12.2; p < .001).

3) Race

Most youth referred to CSF Buxmont schools were white (77%), followed by black (11%) and Hispanic (7%). The race of youth differed significantly by referral source. The great majority of youth referred by probation and schools were white (78% and 79% respectfully), while a smaller majority of those referred by children and youth were white (64%). Hispanic youth totaled 17% of children and youth referrals, compared to 7% for probation and 6% for school referred youth. Black youth constituted 14% of those referred by children and youth, 11% of probation, and 10% of school referrals (ChiSq = 16.3, df = 6, p < .05).

4) County of referral

Youth are primarily referred by three counties: Bucks (57%), Montgomery (25%), and Northampton (15%). Three percent of referrals were from Lehigh, Monroe, and Chester counties (26, 3, and 1 referrals respectively). Most of those referred by probation were from Bucks (77%), with fewer from Montgomery (12%), and Northampton (7%). The largest percent of school referred youth were from Montgomery (43%), followed by Bucks (31%), and Northampton (23%). Among children and youth referrals, the largest percent were from Montgomery (39%) and Northampton (34%), with fewer from Bucks (26%).

Overall, race was unrelated to gender, with whites constituting 79% of boys and 75% of girls. Likewise, age was unrelated to gender, with the mean age upon entering CSF Buxmont schools being 15.7 years old for boys and 15.6 for girls. Hispanic youth were significantly younger (15.2), than were white youth (15.7), black youth (15.6), or Asian youth (15.5) (F = 2.69, df = 4, p < .05). Also, Hispanic youth tended to be younger with 62% under 16 at program entry, compared to 43% of whites, 48% of blacks, and 43% of others. Children and youth referrals tended to be younger, were twice as likely to be girls, and nearly three times more likely to be Hispanic than youth referred by probation or school. Because of the overrepresentation of Hispanic and younger youth referred by children and youth, this age by race relationship became insignificant once referral source was taken into account.

A higher proportion of youth from Northampton County were minorities (38%), than Montgomery (28%), or Bucks (13%). One-quarter of Northampton County referrals to CSF Buxmont schools were Hispanic, compared to 3% for Bucks County and 6% for Montgomery County. Compared to other counties, a higher proportion of youth referred from Bucks were white, from Northampton were Hispanic, and Montgomery were black.

Delinquency. Criminal history information was gathered during intake interviews on 666 youths, including 364 referred by probation, 228 by schools, and 74 by children and youth. Among these, all of the probation referrals, 39% of the school, and 32% of the children and youth referrals had been arrested prior to entering CSF Buxmont schools.

Among those arrested, drug or alcohol related offenses were the most common (29%), followed by assault-related offenses (26%), theft (22%), violations against the public order (12%), and other offenses (12%). Of those arrested for assault-related offenses, one-third involved aggravated assault or robbery (44 of 133). Less than 1% (5) were placed for a sexual offense.

Among those with previous arrests, probation referred youth were more likely to have committed drug and alcohol related offenses or theft related offenses (31% and 24%) than school (25% and 17%) or children and youth referred (13% and 19%). School referred youth with prior arrests were more likely to have committed an assault (32%) than probation (25%) or children and youth referrals (26%). Offense type did not differ between race or gender.

Arrest history distributed in much the same way as current offense in that probation referred youth had more serious criminal histories. Among those who had been arrested, 67% of probation, 46% of school, and 39% of those referred by children and youth had been arrested more than once. Among those with multiple arrests, the number of prior arrests did not differ between referral source, age on admission, gender, or race.

Among the youth with prior court adjudications, probation referred were more likely to have been first adjudicated before age 14 (25%) than school or children and youth referrals (both 12%). Age at first adjudication was unrelated to race or gender.


In April 1999, CSF Buxmont implemented the evaluation protocol of the Program Development and Evaluation System (ProDES), an outcome-based information system that tracks delinquents for the duration of his or her involvement with the juvenile justice system. This system was developed by Temple University’s Crime and Justice Research Center. Its primary focus is program development, providing a continuous flow of intermediate (changes during the program) and ultimate (recidivism and community adjustment) outcome information. This performance monitoring system was designed to track delinquent youth for Philadelphia Juvenile Court. Since 1993, ProDES has gathered data on more than 20,000 juvenile delinquents placed in more than 78 programs.[4]

ProDES collects information on each youth’s current offense, prior record, socio-demographic information, and family structure. Also, ProDES collects information through risk and needs assessments, and self-esteem and values assessments at three points: program entry, program discharge, and six months following discharge. This system’s scales have been empirically derived and have high reliability. Follow-up data is gathered by checking official court records, through the self report of juveniles and interviews with the juvenile’s caregivers. ProDES has demonstrated the utility of official records, parent reports, and self report data in monitoring program and system performance.

All new admissions were to be interviewed within one month and assessed not sooner than a week nor later than a month following admission. New admissions are students entering CSF Buxmont schools for the first time and students returning after an absence of three months or more (excluding summer break). Students admitted prior to April 1999 were treated as “grandfathers” and were not interviewed unless they were returning after an absence of three months or more. Students returning with less than three months of interruption (excluding summer break) are treated as a continuous stay with a period “out of program.”

Overall, matched intake-exit interviews were collected among half (49.8%) of the 919 discharged youth. Compliance with the protocols for the 665 youth admitted after August 1999 was 94% for assessments, 84% for intake interviews, and 80% for exit interviews. Eighty-eight percent of all intake interviews were conducted within the 30 day window.

Independent Variables

CSF Buxmont School/Day Treatment Exposure. Youths attend CSF Buxmont schools for differing lengths of time depending on a variety of factors. Most are returned to public school after a few months when their referral source and the CSF Buxmont staff deem that the youth’s behavior has sufficiently improved. Others are discharged when they turn 18, graduate, or complete their GED. Some youth are discharged for a variety of behavioral problems, most within a short period of time, but some after extensive participation.

The primary factor determining the length of exposure was discharge type. The two discharge types are defined as:

1) Normal Discharge—Youth successfully completes the program by returning to public school, graduating or receiving GED, or upon turning 18 years old.

2) Behavioral Discharge—Youth is discharged prior to completing program for failure to attend, running away, showing repeated noncompliance with program directives, violating conditions of their probation, or for an arrest resulting in a new placement.

Since the start of a new school year or semester presents minimally disruptive opportunities to return youth to public schools, many are discharged at these times. Because CSF Buxmont admits youth continuously over the course of the year, some near the end of the school semesters, length of stay for youth receiving normal discharges varies widely. Lastly, some youths attend CSF Buxmont schools for multiple school years, but most do not.

Overall, the mean length of exposure to CSF Buxmont schools was 104 school days (20.9 weeks), with a minimum of none (for 2 youths) and a maximum of 531 days (106 weeks). Half of all youth spent 87 days (17.4 weeks) or more. Among normal discharges, the average length of exposure was 118 school days (23.7 weeks), with half attending for 103 days (20.6 weeks) or longer. Among behavioral discharges, the average length of stay was 77.2 school days (15.4 weeks), with half being discharged within 58 school days (11.6 weeks). While probation referred youth were less likely to receive a behavioral discharge than those referred by schools or children and youth, the difference in length of exposure becomes insignificant, once type of discharge was taken into account. Among behavioral discharges, the length of program exposure did not significantly differ by the type of behavior problem.

Auxiliary Program Exposure. As stated above, CSF also operates a residential care program providing group home services, an intensive supervision program for youth returning to the community from residential drug rehabilitation programs, and a home and community program providing other various support services to youth in crisis. These programs operate as auxiliaries to CSF Buxmont schools.

Youth utilize various combinations of these programs as deemed necessary by their referral source, themselves, and CSF Buxmont staff. Youth referred by schools generally do not participate in these auxiliary programs. Some probation referred youth participate in a combination of all four programs.

Among the 88 children and youth referrals, 72% spent some time in residential care and the remainder did not participate in any auxiliary program. Among those that participated in residential care, the mean length of residential stay was 179 days, with half spending less than 166 days (min = 8, max = 546). The mean length of stay in residential care for those referred by children and youth was shorter for those receiving behavioral discharges (123 days) than it was for normal discharges (222 days) (F = 12.4, df = 1, p < .001).

Among the 516 youths referred by probation, 63% (325) participated in one or more auxiliary program. Among those that participated in only one of these programs, 11% (54) participated in residential care, 10% (52) in the home and community program, and 6% (29) in the intensive supervision program. Among those that participated in two or more auxiliary programs, 24% (124) participated in both residential care and intensive supervision, 4% (22) participated in residential care and home and community, and 3% (14) in both intensive supervision and home and community. Among probation referred youth, 6% (30) participated in all three auxiliary programs.

Among the 230 probation referred youth spending some time in residential care, the mean stay was 147 days with half spending less than 140 days (min = 2, max = 453). Among the 170 participating in the intensive supervision program, the mean length participation was 172 days with half spending less than 166 days (min = 1 max = 455). Among the 87 in the home and community program, the mean length of participation was 114 days with half participating less than 88 days (min = 3 max = 407). In any of these auxiliary programs, the length of participation for probation referred youth was not significantly different between normal and behavioral discharges.

Some of the probation referred youth participated in the intensive supervision program or home and community program prior to entering CSF Buxmont schools and some participated in one of these programs following discharge. Among the 516 probation referred youth, 5% (25) participated in intensive supervision prior to attending CSF Buxmont schools, and 26% (136) continued in the intensive supervision program after they were discharged. Also, 11% (54) participated in the home and community program prior to attending and 10% (53) participated in home and community following discharge from CSF Buxmont schools. The mean length of exposure to an auxiliary program prior to attendance was 47 days for intensive supervision (median = 8, min = 1, max = 252), and 69 days for home and community (median = 43, min = 1, max = 387). The mean length of participation in these programs following discharge is 71 days for intensive supervision (median = 62, min = 1, max = 214), and 75 days for the home and community program (median = 64, min = 8, max = 324).

There is an important caveat limiting the results presented in this report. Not all youth in CSF’s intensive supervision program or home and community program attended a CSF Buxmont school. Data was not collected for these youth. Since these youths may differ from those included in this study in a number of ways, the results of this evaluation can not be generalized to apply to these CSF auxiliary programs in their whole.

Complicating the analysis beyond the varying mixes of auxiliary program exposure, for youth attending a CSF Buxmont school, is that the length of this exposure is related to the length of attendance. This creates a high level of time-dependent multi-colinearity among the independent variables, making unbiased estimates of the independent effects of these programs problematic. This analysis focuses on the effect of the restorative milieu at CSF Buxmont schools on participating youth, taking into account the mix of auxiliary program exposure, but without attempting to isolate those effects. Therefore, results from this study need to be interpreted within the context of an evaluation of CSF Buxmont schools combined with CSF auxiliary program services.

Performance Measures

The performance of CSF Buxmont schools was evaluated using three distinct measures: program completion, youth attitudes, and offending following discharge.

Program Completion. Restorative practices demand that youth take an active role in maintaining the behavior and safety of the community, so it is expected that CSF Buxmont would retain youth at rates higher than other comparable programs. Therefore, the capacity of any program to retain their clients to normal discharge (i.e. successful program completion as defined above) is an important measure of effectiveness. CSF Buxmont schools admit youth with widely disparate personal and criminal histories. Yet despite this variety of risk factors, CSF Buxmont staff take pride in the fact that they work hard to retain the majority of participating youth to normal discharge/successful program completion.[5] Policy only requires behavioral discharge for behavior that clearly endangers the safety of others in the program.

Youth Attitudes. A successful program will demonstrate measurable positive changes in youth attitudes as a result of program exposure. Changes in youth attitudes were measured by comparing each youth’s response to a series of interview questions designed to measure social values and self-esteem administered shortly after entering the program to those concurrent with their discharge.

Offending Following Discharge. Finally, changes in each youth’s behavior were measured by reviewing juvenile and adult court records and noting any petitions for arrests occurring during the six months following discharge from CSF Buxmont schools. Arrests occurring while attending and resulting in discharge are not included in this measurement. The great majority of youth that offend after discharge do so within six months. While court petitions may not be the best measure of behavioral change, or even recidivism, it is a generally accepted and consistent measure of program success.


Program Completion

Figure 1
Reason for discharge by referral source for full school years only.
Figure 2
Percent of normal discharges by referral source and offense groups.

The capacity of any program to retain their clients to successful program completion/normal discharge is an important measure of effectiveness. Youth who complete a program are more likely to be compliant and therefore are expected to have more positive outcomes than youth receiving behavioral discharges. This potential for selection bias means a program’s completion rate is an important intervening variable between the length of program exposure and the effects of that exposure. Knowledge of the variables that explain noncompletion is critical in order to test and statistically control for this likely source of outcome bias.

Reasons for discharge from CSF Buxmont Schools for the 919 youth were as follows:

1) Youth who completed the program and returned to public school, graduated high school, or completed their GED (66%).

2) Youth discharged because of a new arrest or violation of probation, which resulted in a new placement order for that youth (7%).

3) Youth discharged because of extended absences or who ran away (12%).

4) Youth who otherwise repeatedly failed to respond to the program (16%).

Among the 919 youth in the sample, those referred by probation were significantly more likely to receive a normal discharge (72% of 516) than were those referred by schools (58% of 315) or by children and youth (60% of 88) (ChiSq = 16.4, df = 2, p < .001). A distribution of discharge type more representative of a typical school year is arrived by restricting analysis to data from full school years only (see Figure 1). Thus, successful completion rates for CSF Buxmont schools were 67% of probation, 53% of school, and 57% of children and youth referrals.

Among the 313 youth receiving behavioral discharges, 32% of probation, 12% of school, and none of the children and youth referrals were discharged because of a new arrest. School referred youth were more likely to be discharged for failure to attend (40%) than were those referred by probation (32%) or children and youth (20%). Children and youth referrals were more likely to be discharged for failure to respond (80% of 35) than probation (36% of 147) or school referred youth (48% of 131) (ChiSq = 36.9, df = 4, p < .001).

A multivariate discriminate analysis was conducted on the probability of completing the program successfully in order to determine characteristics that distinguish youth discharged for misbehavior. Results revealed that probation referrals and those charged with a drug offense were more likely than others to complete the program and receive a normal discharge. Also, black girls were less likely than others to receive a normal discharge (Wilks’ Lambda .98, .96, and .97 respectively, all p < .001, with 67% correctly classified).


Figure 3
Percent of normal discharges by race and gender.
Figure 4
Cumulative weeks of participation by reason for behavioral discharge.
Figure 5
Distribution of months participation by referral source (behavioral discharges).
Figure 6
Distribution of months participation by referral source (normal discharges).

As seen in Figure 2, a greater proportion of youth charged with a drug offense successfully completed to normal discharge. Also, probation referrals completed at a rate of 7-8% higher than others. Youth charged with a drug offense have an additional 10% higher normal discharge rate among probation referrals and 6% among those referred by other sources.

Figure 3 shows the race by gender interaction effect on discharge type. All race/gender groups except one have normal discharge rates above 60%. The exception was black girls, of whom only 41% were discharged normally.

How a youth is discharged determines, at least in part, the length of their participation. The overall median length of participation was 87 days, but 103 days for normal discharges and 58 days for behavioral discharges (Kolmogorov - Smirnov Z = 4.6, df = 1, p < .001).

A majority of behavioral discharges occurred within the first three months. As shown in Figure 4, youth discharged for failure to attend were more likely to fail during their first week (10% of 63) than nonresponsive youth (2% of 114) or arrested youth (2% of 144) (ChiSq = 11.0 df = 2, p < .01). Half of those discharged for failure to attend had participated for less than 8 weeks, compared to 12 weeks for other behavioral discharges. Differences between the reason for behavioral discharges were not statistically significant beyond the first week.

Within each type of discharge, the differences in the length of CSF Buxmont school exposure between referral types were insignificant (see Figure 5 and Figure 6). Once the higher completion rates of probation referred youth were accounted for, the length of exposure was determined more by the type of discharge than by the referral source. Comparing the two figures reveals that the shape of the distribution of exposure time was different for normal discharges than it was for behavioral discharges (approximating first-order and second-order delay functions, respectively).

A multivariate survival analysis was conducted on the number of days of exposure for the 311 behavioral discharges. Only one variable was predictive of time until behavioral discharge—youth who were also in CSF residential care remained in CSF Buxmont schools significantly longer than other youth (B = .200, Chi Sq = 10.1, df = 1, p = .0015). As shown in Figure 7, among those receiving a behavioral discharge, youth in residential care remained in the program longer than others. Among those youth not participating in residential care, one-quarter received behavioral discharges by 4 months, half by 9 months, and three-fourths by 17 months. Comparable proportions for residential care participants were 7 months, 16 months, and 29 months respectively. None of the demographic or delinquency variables were significantly related to the time until behavioral discharge. Differences in survival rates were not significant between referral sources, demographic groups, or prior delinquency.

Since residential care is not provided to school referred youth, there is a potential interaction effect between referral source and participation in this auxiliary program. To test whether the increase in survival rates for youth in residential care was simply a proxy for not being a school referred youth, the effect of residential care was tested among probation referrals only. Among this group, youth participating in residential care still had significantly greater survival rates (B = 2.16, ChiSq = 4.9, p = .027).

Youth Attitudes

Among the 723 intake interviews, 88% were conducted within 30 days and 91% within 6 weeks of the start of attendance at a CSF Buxmont school. The remaining intake interviews were excluded from the sample as not accurately reflecting the youth prior to admission, leaving 656 cases with a valid intake interview. Among these cases, 414 (63%) had a matching exit interview. Of these valid matched interviews, the minimum time between measures was 34 days and the maximum was 639 days, or 21 months (one youth with 774 days was excluded from the sample as an outlier). There were 365 (88%) with one year or less between measures and 49 (11%) with more than one year between measures. These interviews measured changes in social values and self-esteem.

Social Values

ProDES utilizes the 38-item Jessness Values Scale to measure changes in a youth’s attitudes toward social values. The scale ranges from 1 for the least to 10 for the most positive values score. The scale was found to be reliable on both the intake and exit interviews (36 items: Standardized alpha = .838 and .866, respectively). As shown in Figure 8, social values on program entry were normally distributed, with a mean of 5.34 and a median and mode of 5 (skewness = -.08).


Figure 7
Cumulative weeks of participation by CSF residential care (behavioral discharges).
Figure 8
Frequency distribution of social values score on entry.
Figure 9
Change in social values by type of discharge.
Figure 10
Frequency distribution of self-esteem score on entry.
Figure 11
Change in self-esteem by type of discharge.

A youth’s value score on entry to a CSF Buxmont school was a good predictor of their score on exit (r = .692, df = 412, p < .001). A few participating youth became less positive (9%), most remained fairly stable in their values (57%), while more than a third developed more positive social values (34%). A majority of all youth showed some improvement between measures. As shown in Figure 9, the net improvement was greater for normal discharges than it was for behavioral discharges. Also, behavioral discharges were nearly twice as likely to become more antisocial than normal discharges (ChiSq = 5.974, df = 2, p = .050).

Youth receiving behavioral discharges had lower social values scores on intake (mean = 4.97), compared to those receiving normal discharges (mean = 5.44) (F = 5.3418, df = 411, p < 0.05). There was an increase of 10.6% in positive social values among youth receiving normal discharges (from 5.44 to 6.01) (paired-t = 7.72, df = 327, p < .001). The 5.7% increase in mean social values for the 84 behavioral discharges, although it was in a positive direction, was not statistically significant (from 5.00 at entry to 5.29 at exit, paired t = 1.97, df = 83, p = 0.053).

Three questions were thought to be particularly useful in testing the effectiveness of restorative practices. These questions were related to acceptance of responsibility. At intake, 90% of all youth in CSF Buxmont schools agreed with the statement, “When I get into trouble, its my own fault”, compared to 100% at exit (paired-t = 6.65, df = 409, p < .001). Also, all youth that participated in CSF Buxmont schools became less likely to agree with the statement, “I am blamed at home for things that are not my fault” (from 56% to 48%; paired-t = 2.87, df = 411, p = .004). Further, normal discharges became less likely to agree with the statement, “I do not get a fair chance” (from 26% to 19%; paired-t = 2.31, df = 328, p = .022), while the change was insignificant for behavioral discharges (from 36.9% to 38.1%).

There were also significant improvements in youth attitudes toward authority as a result of attending CSF Buxmont schools. Youth were more likely to agree with the statement, “Police try to help” (from 47% to 100%; paired-t = 21.22, df = 406 , p < .001). Further, normal discharges were less likely to agree that “People in authority are bossy” (from 59% to 53%; paired t = 2.17, df = 324, p = .031), compared to a nonsignificant increase among behavioral discharges (from 63% to 66%; paired-t = -.60, df = 82, p = .552).


ProDES utilizes a 10 item self-esteem scale to measure changes in a youth’s self-image. The scale ranges from 1 for the lowest to 10 for the highest. The self-esteem scale was found to be reliable on both the intake and exit interviews (Standardized alpha = .873 and .888, respectively). As shown in Figure 10, self-esteem scores on program entry were highly skewed, with a mean of 7.88, a median of 9 and a mode of 11 (skewness = -1.22). Only 30% percent of youth entering CSF Buxmont schools recorded a self-esteem score below 8 and only 13% a score below 5.

A youth’s self-esteem score on entry was a good predictor of their score on exit (r = .6456, df = 412, p < .001). A few youth had lower self-esteem at exit (10%), most remained relatively stable (66%), and nearly a quarter of youth had substantial increases in their self-esteem (24%). As shown in Figure 11, normal discharges were nearly twice as likely to have a large increase in self-esteem (27%) than behavioral discharges (14%).


Table 2
Self-esteem Mean Scores at Intake and Exit by Type of Discharge


Among normal discharges, self-esteem scores improved by 8.8% (from 7.89 to 8.58; paired-t = 6.26, df = 326, p < .001). There was essentially no change in self-esteem scores for youth receiving behavioral discharges (from 7.87 to 7.88, p = .851). Normal discharges also had statistically significant increases on eight of the ten self-esteem questions, while behavioral discharges had no significant change on any of the self-esteem items (see Table 2).

Trends in Attitudes by Program Exposure

For youth receiving normal discharges, there was a statistically significant relationship between the change in youth attitudes and the length of program exposure (values r = +.157, df = 326, p = .004; esteem r = +.122, df = 326, p = .027). There was no relationship between program exposure and change in social values or self-esteem for youth receiving a behavioral discharge (r = -.012 and -.032, respectively).


Figure 12
Scattergram of change in social values by weeks between measures
(normal discharges).
Figure 13
Scattergram of change in self-esteem by weeks between measures
(normal discharges).

As shown in Figure 12, the longer youth who were normally discharged were exposed to CSF Buxmont schools, the more positive their values became. Difference scores for social values increased an average of .0183 for every week of participation. Beginning with the mean social values score on entry of 5.44, this corresponds to 0.34% improvement per week, 1.35% per month, 4.37% after three months, and 8.75% after six months.

There was a corresponding increase in self-esteem scores for youth receiving normal discharges as shown in Figure 13. Self-esteem scores increased an average of .0217 for every week of participation. Beginning from the intake average of 7.89, this change corresponds to 0.28% improvement per week, 1.10% per month, 3.58% after three months, and 7.15% after six months. These improvements in social values and self-esteem were unrelated to referral source, gender, age, race, risk, or to participation in a CSF auxiliary program.

To explore a possible nonlinear relationship between the changes in youth attitudes and the length of exposure to CSF Buxmont schools, mean scale scores for each month between measures were computed for youth discharged normally. Half of these youth had 23 weeks or less between measures (min = 4 weeks, max = 91 weeks). Mean scores for any specific week are very unstable, since most weeks have fewer than 10 cases. A simple smoothing technique was employed to increase the number of cases for any given week by averaging the weighted scores across multiple weeks. The means for all weeks are based upon more than 20 cases when moving averages are computed over a 16-week period.

The trends produced by this 16-week smoothing reveal the relationship between the length of participation in CSF Buxmont schools and youth scores at intake and exit. As shown in Figure 14, both social values and self-esteem measures reveal a similar pattern over time. Youth with greater time between measures were more likely to have lower intake values and slightly higher exit scores. Because of the great variation in intake scores and the lack of trends in exit scores, neither are statistically different from zero. Youth with 33-52 weeks between measures had unusually high intake scores which were not reflected in the exit scores, with mean social value and self-esteem scores more than one-half of a point higher than the overall mean (0.510 and 0.671, respectively).

The trends of the difference scores (exit minus entry means) represent an estimate of the effect CSF Buxmont schools have on youth attitudes. As shown in Figure 15, there was nearly a 0.5 point mean improvement in both scales for youth with 9-36 weeks between measures. The effect appears to decline from 20-42 weeks, which is solely the effect of the unusually high intake scores of the youth with 35-52 weeks between measures. The effect on attitudes was most dramatic for youth spending a second year in program (52+ weeks between measures), but no obvious nonlinear trend emerges.

Exit interviews were not conducted with any youth participating for less than 30 days and matched interviews were only available on 22% of normal discharges with 30-60 days in attendance. For these reasons, the data are unable to estimate any tipping point in the shape of the relationship between length of exposure and attitude changes, as it must have occurred within the first 9 weeks following program entry.


Figure 14
Entry and exit attitudes by weeks between measures, 16 week weighted moving averages, and linear trends (normal discharges).
Figure 15
Attitude changes by weeks between measures, 16 week weighted moving averages, and linear trends (normal discharges).

Still, the overall trends were toward more positive social values and higher self-esteem the longer youth had participated CSF Buxmont schools. Linear least squares estimates of these trends with zero intercepts have slopes slightly higher than estimated from the zero-order correlations above (.0217 for social values and .0240 for self-esteem). These correspond to a 10.5% increase in social values and a 7.9% increase in self-esteem after 6 months of participation in CSF Buxmont schools.

Offending Following Discharge

Court records were available on 96% of the 919 youth discharged from CSF Buxmont schools. Overall, 10.4% had a new court petition and an additional 5.6% had a petition for a violation of probation filed in juvenile court during the six months following discharge. Behavioral discharges were nearly twice as likely (15.1% of 305) as normal discharges (7.9% of 581) to have a new court petition following discharge (ChiSq = 11.0, df = 1, p = .0009).

School referred youth were slightly more likely to have a new court petition (11.6% of 303) than probation (10.1% of 496) or those referred by children and youth (8% of 87). Also, boys were more likely than girls to have a new court petition (11.3% to 8.0%), but this difference was not statistically significant. Probation referred youth were, of course, more likely to be petitioned for a violation of probation (8%) than school (3%) or children and youth referrals (1%) (ChiSq = 12.8, df = 2, p = .002).

Black youth were nearly twice as likely to receive a new court petition (18% of 91) than other youth (10% of 795) (ChiSq = 5.6, df = 1, p = .017). Youth entering CSF Buxmont schools who were 16 years old or younger were more likely (12% of 621) than older youth (7% of 244) (ChiSq = 4.3, df = 1, p = .038), and those with a prior arrest were more likely (14% of 364) than others (8% of 286) to have a new court petition (ChiSq = 4.1, df = 1, p = .042).

Overall, the proportion of youth with a new court petition during the six months following discharge from CSF Buxmont schools was related to the weeks of participation. The rate of new petitions was 16.2% of the 277 youth who participated less than three months, compared to 7.7% of the 609 participating three months or longer (Lambda = .9878, df = 884, p = .001). For youth receiving a normal discharge, the reduction in offending following discharge held for both new petitions (Lambda = .9915, df = 579, p = .0265) and for all petitions (new petitions and probation violations combined) (Lambda = .9921, df = 579, p = .0317). Length of stay was not significantly related to offending following discharge for youth receiving a behavioral discharge.

As shown in Figure 16, those that attended for three months or longer and received normal discharges, representing more than half of all youth, had the lowest rate of new court petitions (6.1%). Behavioral discharges with less than three months of participation had the highest rate of offending following discharge (17.6%), followed by normal discharges with less than three months (14.5%), and behavioral discharges participating three months or longer (12.5%) (ChiSq = 20.6, df = 3, p < .001).


Figure 16
Percent with new court petitions by type of discharge and length of participation.
Figure 17
Percent with new court petitions by weeks of participation, 16 week weighted moving average, and linear trend (normal discharges).
Figure 18
Percent with new court petitions by weeks of participation, 16 week weighted moving average, and nonlinear trend (normal discharges).
Figure 19
Percent with new court petitions by length of participation and risk category (plus all court petitions for youth participating in intensive supervision following discharge).

A linear estimate of the relationship between length of exposure to CSF Buxmont schools and new court petition rates for youth receiving normal discharges is shown in Figure 17. Beginning at a 15% rate, the rate of new court petitions was reduced by 1% for every three weeks of exposure, for up to 40 weeks. The rate of new petitions was reduced to 11.4% for youth participating for three months, 7.4% at six months, 3.5% at nine months, and finally to 1.8% at 10 months.

As shown in Figure 18, a nonlinear estimate of the relationship between length of exposure and new court petitions produced a more accurate estimate (r = -.1279, df = 581, p = .002) than a linear estimate (r = -.092, p = .027). The largest reduction was for youth attending for more than 3 months (12-20 weeks). This nonlinear trend reveals a clear tipping point occurring after 3 months of participation with the effect leveling off beyond 5 months.

A multivariate analysis revealed a significant interaction related to offending following discharge between age on admission and existence of a prior arrest history. A risk variable was created to capture this interaction effect to include three categories: 1) low risk includes youth that were over 16 years old upon entry and all youth without prior arrests, 2) medium risk includes youth that were 16 years old with a prior arrest, and 3) high risk includes youth younger than 16 years old with a prior arrest. This risk variable was a strong predictor of new court petitions, with 8% of 641 low risk youth, 13% of 109 medium risk youth, and 19% of 136 high risk youth petitioned for a new offense within six months of discharge (ChiSq = 15.4, df = 2, p = .00045).

Among the risk categories, low risk black youth (19.4% of 62) were nearly three times as likely to be petitioned for a new offense as other low risk youth (6.9% of 579) (ChiSq = 11.6, df = 1, p = .00065). Race was unrelated to the probability of a new court petition among the medium risk (p = .08) and high risk youth (p = .54).

High risk youth were slightly more likely to receive behavioral discharges (41%) than medium risk (35%) or low risk (33%), although these differences were not significant (p = .18). Among normal discharges, risk category was strongly predictive of new petitions following discharge (low = 5%, medium = 15%, high = 18%, ChiSq = 21.1, df = 2, p = .00003), weakly predictive for probation violation petitions (5%, 7%, and 13% respectively, ChiSq = 6.2, df = 2, p = .0444), and strongly for all petitions (20%, 23%, and 30% respectively, ChiSq = 26.5, df = 2, p < .00000). Among youth receiving behavioral discharges, risk category was not predictive of new petitions (p = .19), probation violations (p = .92), or all petitions (p = .35).

A generalized least-squares multivariate analysis revealed that the effect of exposure to CSF Buxmont schools on the reduction of new court petitions (beta = -13.0%, t = 3.2, p = .0014 ) was still statistically significant after controlling for risk (beta = 17.2%, t = 4.3, p < .0000) and race (beta = 9.9%, t = 2.449,p = .0146) (change in R2 = .0460, F = 13.945, p = .000). Additionally, youth who participated in intensive supervision following discharge were less likely to have been petitioned for a new offense than other youth (beta = -8.6%, t = -2.1, p = .0344; change in R2 = .0073, F = 4.494, p = .034).

To isolate the independent effect of intensive supervision, a separate analysis was conducted among the 515 probation referred youth in the sample, 26% of whom participated in intensive supervision following discharge from CSF Buxmont schools. New petitions were filed for 1.5% (2 of 136) of those participating in intensive supervision following discharge, as compared to 13.3% (48 of 360) of youth that did not participate (ChiSq = 15.3, df = 1, p < .001). However, those that participated in intensive supervision following discharge were more likely to be petitioned for a probation violation (14.7% to 5.6% respectively, ChiSq = 11.1, df = 1, p < .001). While the combined new court petition and probation violation rate was somewhat lower among those participating in intensive supervision following discharge (16.2% vs. 18.9%), this difference was not significant (ChiSq = 0.5).

The effect of the CSF Buxmont schools on reducing new court petitions was independent of the other factors related to offending (Wilk’s Lambda RISK = .98, IPAFTER = .95, LRBLACK = .94, and E(RECID) = .96, all p < .001, with 61% correctly classified).[6] As shown in Figure 19, youth who participated for more than three months had lower new petition rates across all risk categories. There were also fewer new court petitions after three months for behavioral discharges and fewer petitions for a probation violation among those participating in intensive supervision following discharge.


The CSF Buxmont restorative milieu produced positive results in all three performance measures: program completion, youth attitudes, and offending following discharge.

The majority of youth referred to CSF Buxmont schools successfully completed the program with rates of 66% for probation referrals, 53% for school referrals, and 57% for referrals from children and youth. These rates are very high when compared to programs serving similar populations. The average completion rate for programs serving youth placed by Philadelphia Juvenile Court is estimated at 35% for community-based programs and 55% for non community-based programs (excluding boot camps). About 33% of nonadjudicated youth in Philadelphia’s delinquency prevention programs complete their program and 40% simply stopped attending [7] Thus, the completion rates for CSF Buxmont schools are very respectable for a community-based program.

These high completion rates are especially significant because CSF Buxmont schools have a very open referral policy and accept youth with a wide range of offense seriousness and related risk factors.[8] Further, rather than screening out high need youth, CSF Buxmont’s restorative milieu tends to retain them. Youth with the lowest social values and self-esteem scores upon entry were retained longer than others. It was these youth that showed the greatest improvements. This is especially impressive in light of the fact that the proportion of youth receiving a behavioral discharge did not differ by age, gender, race, prior offending, or other risk factors.

Youth were more likely to have more positive social values and higher self-esteem at discharge than at program entry. These improvements were greater for youth completing the program than for youth receiving behavioral discharges and were positively related to the length of participation in CSF Buxmont schools. Youth completing the program had more positive social values and higher self-esteem scores upon discharge regardless of their scores at intake.

Offending during the six months following discharge was reduced by 58% for those youth who completed the program successfully with more than three months participation. Youth in all risk categories were less likely to offend following discharge, including those discharged unsuccessfully, if they participated in the program for more than three months. Finally, the greatest reduction in offending occurred among youth with the highest risk factors for offending.

These results present sufficient empirical evidence to conclude that the changes in attitudes and behavior of youth were the result of participating in CSF’s restorative milieu because:

1) There were significant improvements in both attitude and behavioral measures.

2) These improvements were positively related to the length of participation.

3) The relationships between improvement and participation remained statistically significant after accounting for relevant risk factors.

4) The reduction in offending following discharge after three months participation remained significant after controlling for relevant risk factors and reason for discharge.

Very few programs for delinquent and troubled youth can demonstrate similar positive outcomes. These results provide very strong empirical support that restorative practices are effective and that youth in a restorative milieu will become more positive in their social values, develop an improved self-image, and will be less likely to offend in the future.


1. For more information on CSF Buxmont schools go to:

2. Wachtel, T. and McCold, P. (2000). Restorative justice in everyday life. In J. Braithwaite and H. Strang (eds.), Restorative Justice in Civil Society. New York: Cambridge University Press; and P. McCold and T. Wachtel (2002). Restorative justice theory validation. In E. Weitekamp and H-J. Kerner (eds.), Restorative Justice: Theoretical Foundations. Devon UK: Willan Publishing.

3. Harris, P., Jones, P., Naiburg, J., and Washnock, J. (2000). Objectives and Outcomes of Community Service Foundation Programs: Results of the Evaluability Assessment Study. Philadelphia, PA: Crime and Justice Research Institute. “The programs of the Community Service Foundation share several common objectives and a mission that includes building among its clients greater self control, a belief in the capacity of people to change, an appreciation of the value of community, and life-styles in which sobriety is the norm. ... The values of believing in people’s capacity to change, the importance of community, the need for open confrontation and the establishment of a few important rules of behavior dominate the agency’s culture.” p.14. Also see Karp, D. and Breslin, B. (2001). Restorative justice in school communities. Youth & Society, 33(2), 249-272.

4. Jones, P., Harris, P. , Grubstein, L. and Fader, J. (2000). Juvenile Justice Trends and Projections. Program Development and Evaluation System. Philadelphia, PA: Crime and Justice Research Institute. Also see: Crime and Justice Research Institute. (1997). ProDES: The Program Development and Evaluation System Reference Packet (rev. 3/19/98). Philadelphia, PA: Author

5. Harris, et al., op. cit.

6. An independent analysis by Peter Jones, Criminal Justice Research Institute, Philadelphia, Pennsylvania, June 27,2002 confirmed a significant reduction in offending for youth receiving normal discharges. Jones identified several variables which, when combined, created a useful risk classification. He found a significant relationship between this a priori risk classification and offending (ChiSq = 31.6, df = 2, p < .001) with an additional effect for program stays of more than 3 months (ChiSq = 9.0, df = 1, p < .003). In a multivariate logistic regression, the odds ratios for time in program show a 65% reduction in the probability of offending when you move from program time of less than 3 months to 3+ months. Separate one-way ANOVA analysis shows that the effect of time in program is significant in one specific type of interaction - high risk youth. These general results were also confirmed by a discriminate analysis. Repeating the analysis on several community-based programs in the PRoDES database failed to reveal a similar effect for any of these programs.

7. Jones, et al., op. cit., and Jones, P., Harris, P., Poulin, M., and Moss, M. (2001). Delinquency Prevention Program Trends and Projections. Prevention Outcome Monitoring Information System (PROMIS). Philadelphia, PA: Crime and Justice Research Institute.

8. Harris, et al., op. cit.

By Lorenn Walker

This article originally appeared in the Federal Probation Journal, Volume 66, No. 1, June, 2002.

Author's Note

Lorenn Walker, J.D., M.P.H., is a health educator who designs violence prevention and resiliency development programs. She teaches administration of justice for the University of Hawaii Honolulu, and is a former trial lawyer who represented the State of Hawai‘i in civil claims and later defended indigent juveniles for law violations. She coordinates programs for the Hawai’i Friends of Civic and Law Related Education. Address correspondence to Lorenn Walker, P.O. Box 489, Waialua, Hawai’i, 96791. E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it..


The Honolulu Police Department conducted an experimental diversion project for first time juvenile offenders in the City and County of Honolulu. Juveniles were diverted to restorative justice conferences instead of traditional diversion programs. Conferences are based on the assumption that crime damages relationships between people. Conferencing is a group reconciliation process that is best used when someone who has caused harm admits wrongdoing. Offenders and those most hurt by the wrongdoing, gather in a circle at conferences to discuss how they have been affected, and collectively decide how to repair the harm. This study analyzed the effects of conferencing on participant satisfaction, offender agreement compliance, and recidivism. Results show that victims were highly satisfied with the process and conferenced juveniles arrested for non-violent offenses did not escalate to arrests for violent crimes, while juveniles who participated in traditional programs had a significantly higher arrest rate for subsequent violent crimes.


The man was seated in a circle with seven other people. He bent down to the eye level of a 10-year-old boy sitting across from him and said, “I want you to know that my son will never touch you again.” The tense looking boy sighed, “o.k.” The boy’s mother and father, who were seated along side him, nodded their heads in agreement and relaxed their faces. The man’s words to the boy were reassurance that his 16-year-old son would not hurt him again. A month before, the 16-year-old held the 10-year-old upside down against his wishes. Because the older boy would not admit that he touched the boy, this conference turned out to be the least successful compared to 84 others held for first time juvenile offenders in Honolulu. Yet, even with its shortcomings, the conference provided the 10-year-old and his parents assurance that the offender would not touch him again.

Meeting the needs of victims, and providing them with the opportunity to be assured that they will not be hurt again by offenders, is not something traditional criminal justice systems provide. Conferencing attempts to meet victims’ needs, which in the end benefits not only the victims, but also offenders, and the community affected by crime. This study evaluates how effective conferencing was in meeting victims’ needs and its effects on recidivism in a recent Honolulu juvenile diversion program.

Historical Development of the Conferencing Process

Conferencing is a generic term for a group reconciliation process. Conference groups are facilitated by a neutral third party, and reach decisions by consensus. Western governments are using conferencing in criminal and child protective services cases (Hudson, Morris, Maxwell & Galaway, 1996). Conference models include family group conferencing (Maxwell & Morris, 1993), community conferencing (Cameron & Thorsborne, 1999), family group decision-making (Graber, Keys & White, 1996), and Real Justice conferences (O’Connell, Wachtel & Wachtel, 1999).

Conferences incorporate the conflict resolution practices of many indigenous people, including the Maoris of New Zealand, Hawaiians, North American Indians, and Africans (Shook, 1985; Schiff, 1998; Some, 1999; Zion, 1998; Walker, 2001). In 1989, New Zealand enacted legislation that required diversion of all juvenile offenders to family group conferences rather than traditional criminal justice processes (Maxwell & Morris, 1993). The New Zealand family group conference model is based on the Maori conflict resolution practice called whanau. Although the conference model that New Zealand developed was not meant to replicate the Maori process, “it seeks to incorporate many of the features apparent in whanau decision-making processes” (Maxwell & Morris, 1993).

Daly (2001) argues that conferences do not “reflect [and are not] based on indigenous justice practices”; however, New Zealand’s legislative history clearly shows that the Maori practice influenced that country’s mandated conferencing process. In 1986 New Zealand had rejected legislation and conferencing that was “monocultural” because it failed to include the “cultural identity of the tangata whenua (the people indigenous to or belonging in an area)” and did not “involve parents, family groups, whanau, hapu and iwi in developing solutions to the problem situations” (Hassall, 1996). After these two issues were addressed and included in the 1989 legislation, New Zealand enacted the law requiring that juveniles be diverted to family group conferences.

While conferencing did not develop in the West as a restorative justice process, but as a multi-disciplinary team approach in social work, it is a restorative process (Hassall, 1996; McCold, 1999). Restorative justice is an “alternative approach to criminal justice” which began in response to what its proponents viewed as the ineffectiveness of our current system (Pranis, 1996). Western justice is based primarily on retributive values where: “Crime is a violation of the state, defined by lawbreaking and guilt. Justice determines blame and administers pain in a contest between the offender and the state directed by systematic rules” (Zehr, 1990, p. 181). In contrast, restorative justice is based on values that hold “Crime is a violation of people and relationships. It creates obligations to make things right. Justice involves the victims, the offender, and the community in a search for solutions which promote repair, reconciliation, and reassurance” (Zehr, 1990, p. 181).

Emphasis on reconciliation should be an important part of our criminal justice system where in the United States most convictions come from guilty pleas. “More than 90 percent of all felony cases are disposed of by guilty pleas.” (Hall, 1996). The percent of juveniles who plead guilty or admit to offenses before court intervention is unknown, but it is assumed to be higher than the number of adults. During the six years that this author defended juveniles for law violations, all of them admitted responsibility for an offense.

Conferencing is a reconciliation process that addresses the needs of those most affected by crimes to a greater extent than traditional adversarial and autocratic processes can. The father of the 16-year-old offender addressed the need of the 10-year-old victim and his parents that the youngster be safe from the older boy in the future. While the offender’s father could not guarantee that his son would not harm the victim again, the father’s assurance was more than that offered by the criminal justice system. Even though the offender did not fully admit his wrongdoing, the conference benefited the victim. This could not be accomplished with our current justice processes, which do not provide the opportunity for victims and their loved ones to openly communicate with offenders. Without victim participation and the presence of the offender’s father at the conference, the victim’s needs could not be addressed.

The Real Justice Conferencing Model

This pilot project used the Real Justice conferencing model. In Real Justice conferences, participants sit in a circle without a table between them. Participants include victims, offenders, and representatives of the community most affected by crime. The community includes supporters of the victims and offenders (family and/or friends), and others who have been affected. For example, if a crime occurred on a school campus, school staff may be a part of the affected community and participate in a conference.

In some cases, victims do not want to meet with offenders. In these cases, victims may send a representative. Of the 102 juveniles who had conferences, 30 were for assaults and of those, 21 victims participated in conferences. In the remaining nine cases, representatives were from the victims’ families or the schools where many of the assaults occurred.

Conferences are facilitated by a neutral third party who does not participate in decision making. Real Justice facilitators attend a two-day training program. Thirty facilitators from Honolulu were trained, 12 of whom facilitated most of the conferences. Their backgrounds included housewife, sales clerk, college student, retired social worker, and contractor. They were paid $150.00 for each conference. Facilitators also convened and found a conference location convenient for the parties. The main scheduling priority was the victim’s convenience.

Real Justice conference facilitators follow a script developed by Terry O’Connell, a former Australian police officer. In 1990, O’Connell developed protocols from what he learned about New Zealand’s process. These protocols are followed in the Real Justice script (O’Connell, 1998). The script maintains a specific speaking order for conference participants. The facilitator begins a conference by reading a preamble, which creates an atmosphere of respect, and subtly establishes the ground rules. The offender speaks next, before the other participants, allowing him or her to take responsibility for the bad behavior immediately. Having the offender take responsibility at the beginning of the conference gives the victims some emotional relief by knowing denial is not an issue. The offender answers a round of questions that cause him or her to consider the consequences of the bad behavior.

Although some find the use of a script distasteful or stilted, conferences conducted with them have been thoroughly researched and evaluated. These studies have consistently found high rates of participant satisfaction, perceptions of fairness, and offender compliance with agreements (McCold & Wachtel, 1998; Moore & Forsythe, 1995; Umbriet & Fercello, 1998, 1999; O’Connell, Wachtel, Wachtel, 1999). Perhaps the most beneficial reason for using the script is it ensures that the participants will maintain control of the conference outcome, not the facilitator. Participant control during group process has been found to generate more cooperative relationships than autocratic group process (Lewin, 1997). The Real Justice script can prevent facilitators from becoming autocratic and defeating this consensus-based process.

Real Justice Conference Parts

There are four phases to a Real Justice conference. First, offenders admit what they did. They explain what they were thinking when they committed the offense, what they have thought about since then, and whom they think has been affected by their actions. Second, the other individuals in the group discuss how they have been affected by the offender’s behavior. Third, the group discusses and decides what can be done to repair the harm to make things right. Finally, a written agreement is decided upon by the group, which all participants sign. The conference ends with the participants eating together—a ceremonial breaking of bread—which allows for further reintegration, closure, and healing.

The Study

In 1999 the Honolulu Police Department (HPD) received a federal grant through the Hawai‘i State Office of Youth Services to divert first-time juvenile offenders to restorative conferences rather than traditional police diversion programs. The project was designed by the Hawai‘i Friends of Civic and Law Related Education (Hawai‘i Friends), a non-profit educational organization.


Between March and September 2000, 102 first-time juvenile offenders participated in conferences instead of traditional police diversion programs in the City and County of Honolulu. Eighty-five conferences were held for the 102 offenders (co-defendants participated together in single conferences).

Selection of Cases for Conferencing

The juveniles who participated in the 85 conferences were initially selected randomly, but after several weeks, shoplifting and runaway cases were selected out. Theft cases were avoided when a major retailer in Honolulu, frequently a juvenile shoplifting victim, refused to participate in conferences. Without the retailer’s participation, conferencing its theft cases could not be meaningful.

Runaway cases were also avoided after the program began because they often concern complicated family issues and not the resolution of a specific wrongdoing. Some juveniles run away because of abuse or neglect. The Real Justice conference model does not address these complex problems and is more effective for clear cases of wrongdoing.

A total of 160 juveniles were selected for conferences. Of those originally selected, 102 juveniles participated in a conference. Fifty-eight of the 160 juveniles originally selected for conferences did not participate. Twenty-five of those denied that they were responsible for the crime. Most of them were charged with violent offenses. The conferencing process is only used when the offender admits wrongdoing.

The next most common reason for not participating in a conference was that the juveniles’ families either never returned phone calls about the program or did not have telephones. In a few of the cases where there was no telephone, facilitators went to the homes and arranged a conference. The remaining 16 juveniles selected for conferences did not participate for a variety of reasons.

Nature of Conference Agreements

The majority of agreements, 73 percent or 61 out of 83 cases sought purely symbolic (i.e., apology) or a combination of symbolic and service reparations (i.e., apology with service including counseling for the offender), See Table 1. Fourteen cases, 17 percent, sought purely service reparations where juveniles agreed to repair damage or provide other community service. One juvenile agreed to service, symbolic, and monetary reparation. Only 7 percent of the offenders (8 juveniles out of the 100 whose agreements were reviewed) were required to pay monetary restitution. This study confirms what many believe: most victims want to know offenders are remorseful (Tutu, 1999).

Result: Participant Satisfaction

A total of 405 participants in 831 conferences were surveyed immediately following conferences for their satisfaction with the process. Although there may be concern that immediate surveys may have a “bubble effect,” where individuals are more inclined to report satisfaction than they would over time (McCold, 1998), one study shows that this does not apply to conferencing satisfaction survey results (Palk, Hayes & Prenzler, 1998).

Recent research of 35 fully-restorative, mostly-restorative and non-restorative conflict resolution programs throughout the world found a significantly higher satisfaction level with conferencing which is a fully-restorative process (McCold & Wachtel, 2000). The fully-restorative programs (conferencing and circles) had a 91.3 percent level of satisfaction while the mostly- restorative programs (victim offender mediation) had an 81.6 percent level of satisfaction. The non-restorative programs (boot camp and scared programs) had only a 55.6 percent level of participant satisfaction. The differences in the satisfaction levels reflect the differences in participation of stakeholders whose needs are addressed during the process. For example, in conferencing all the key stakeholders (i.e., victims, offenders and their communities of care) participate in the process that addresses their needs. In the mostly-restorative processes, victim offender mediation, usually only the victim and offender participate, their communities of care are not included in decision making nor are their needs addressed. Finally, in the non-restorative processes, the boot camp and scared straight programs, offenders reluctantly participate, and the victims and communities of care are excluded. McCold & Wachtel found that the more key stakeholders participating in the process, the higher the level of satisfaction with the process. Conferencing includes the main stakeholders, who are all encouraged to meaningfully participate, which the other programs lack.

Participant satisfaction with the conferencing process here was extremely high. Only seven of the 405 participants who completed surveys at the 83 conferences indicated that they did not believe the process served justice. Three of these participants were from the single conference that was incorrectly conducted after the 16-year-old denied full responsibility. The other four participants who did not believe the conference served justice were two offenders and two offenders’ supporters. The only victim who did not believe the process served justice was the 10-year-old whose case was incorrectly conferenced.

Conference participants’ satisfaction in Honolulu was compared with satisfaction levels of juvenile conferencing programs in Bethlehem, Pennsylvania during 1995 - 1997; four sites in Virginia during 1998 - 1999; Indianapolis, Indiana during 1997 - 2000; and Canberra, Australia during 1997 - 1999. Analysis of these programs shows Honolulu participants had a similar level of satisfaction with the conferencing process as the other projects (McCold, 1998; McCold, 1999; McGarrell, Olivares, Crawford, & Kroovand, 2000; Sherman, Strange, Barnes, Braithwaite, Inkpen, and The M.M., 1998 & 1999). Table 2 shows conference participant satisfaction rates.

Result: Offender Compliance with Agreements

All of the 85 conferences resulted in agreements. Approximately six months after the conferences, a telephone survey was conducted to determine whether offenders fulfilled the agreements. Victims were contacted in 47 of the cases (representing 61 juveniles). When the victim was not available, the victim’s supporters were contacted. Victim supporters were contacted in 14 of the cases (representing 15 juveniles) and when they were unavailable, the offender’s supporters were contacted. Offender’s supporters were used to determine whether the agreements were fulfilled in 15 cases (representing 17 juveniles), and as a last resort, in three cases offenders were relied on. One of these offenders said that he did not fulfill the agreement while the other two said they had.

The telephone survey showed that there was a high rate of offender compliance with the agreements. Out of the 102 juvenile offenders, at least 90 of them complied with the terms of the agreements as shown in Table 3. Even the 16-year old whose case was incorrectly conferenced complied with his agreement to “stay away from [the victim’s] home,” although the victim’s mother said the offender was seen near the victim’s street. Six juveniles did not fulfill their agreements. In six other cases, the outcome is unknown because no one was contacted.

Out of the 102 juvenile offenders, only eight were required to provide monetary restitution. Seven of these juveniles fulfilled the agreements. Only one offender did not pay the money back to the victim as agreed. In this case, he agreed to repay $250.00 damage to a candle display in a store. The store representative had no authority to lower the restitution amount, which its head out-of-state office had established without any negotiation possibility. Even with this juvenile’s failure to pay the restitution, the overall 87 percent restitution payment rate is significantly higher the current system’s.

First, there is no requirement of other Honolulu Police Department diversion programs that offenders pay any restitution. While courts order restitution, there is no data on the percent of collected court-ordered restitution in Honolulu’s juvenile cases. A long-time juvenile probation officer of the court believes, however, that less than 10 percent of all restitution orders for juveniles are completely paid.

Result: Recidivism Rates

Recidivism in this study looked at re-arrest rates six months after the conference. As shown in Table 4, the overall recidivism rate for the juveniles who had conferences was 28 percent six months following the last conference. The recidivism rate was only 11 percent in September 2000 when the last conference was held. This increase from 11 percent to 28 percent six months later confirms the likelihood of repeat offenses over time.

As shown in Table 5, a statistically significant recidivism difference was found for juveniles arrested for non-violent violations, i.e. theft, status offenses, and drugs. A matched group for time of offense, type of offense (non-violent) and gender was used to compare the difference between diversion programs. Although the overall recidivism rates between the two groups was not different, the juveniles who had conferences for non-violent offenses were less likely to escalate to violent crimes, compared to juveniles without conferences. In the group of 102 conferenced juveniles, 59 were arrested for non-violent offenses. Of those, only one was rearrested within the following six months for a violent crime. In the matched group of 82 juveniles, 75 of them were arrested for non-violent offenses. Of those, six were arrested for violent crimes within the following six months.

Keeping juveniles from escalating to violent crimes is important for reducing recidivism. The latest national study analyzing recidivism differences found that juveniles charged with assault are 44 percent more likely to repeat future offenses, while juveniles charged with theft are only 34 percent more likely to repeat offenses (Snyder, 1988).

The re-arrest rate for all youth in Honolulu was not significantly different from the conferenced juveniles. During the project, a total 3376 youth, excluding the 102 in this project, were arrested in Honolulu. After six months, 863 juveniles or 25 percent were rearrested. It is unknown what crimes the juveniles were arrested for.


Conferencing is a welcome juvenile justice process in Honolulu. As one assault victim’s parent said: “We were able to share our feelings and felt that the children understood how their actions affect others. I felt a healing when my thoughts and feelings were heard. I also felt good to show the children that we support them and care about their well being. We are very glad we could solve this problem in this manner.” The mother of the 10-year victim in the incorrectly conferenced case said, “Talking with the [offender’s] dad was very helpful.” She also added that the “conference would have been better if [the offender] would have admitted assaulting my son. He cannot get better until he can face that he assaulted my son. Until he admits it nothing can be repaired.” This statement illustrates why conferencing criminal cases should only be used when offenders accept full responsibility for their behavior. When offenders take responsibility, conferencing offers a process where healing can occur compared to traditional processes that focus primarily on retribution.

Second, while the Real Justice conference model relies on the offender admitting guilt, and may fail to address the often-complex problems that are relevant for status offenses, it is an effective intervention for many non-violent offenses. Conferenced non-violent offenders are less likely to escalate to more serious crimes six months later. A conferencing model needs to be developed to address relevant social issues for status offenders.

Finally, this research confirms what South Africa’s Archbishop Desmond Tutu says: "justice, restorative justice, is being served when efforts are being made to work for healing, for forgiving, and for reconciliation." Most victims of juvenile crime in Honolulu agree with Archbishop Tutu because most that attended conferences wanted an apology. As one victim said a "verbal apology was all I needed." And, as the 10-year-old victim’s father from the incorrectly held conference said, “Until the [offender] can be remorseful and admit what happened nothing can be accomplished!” When those most affected by crime participate in a process focused on addressing their needs, healing can begin, and restorative justice happens.


1. The case of the 16-year-old boy who held the 10-year-old upside down should not have been conferenced. Although the conference had value because the offender’s father assured the victim he would not be touched again, the case did not meet the criterion of the offender taking responsibility for his actions.


Cameron, L. and Thorsborne, M., (1999). "Restorative Justice and School Discipline: Mutually Exclusive?" [online]. Available:

Daly. K. (2001). “Conferencing in Australia and New Zealand: Variations, Research Findings, and Prospects.” In A. Morris & G. Maxwell (eds.), Restoring Justice for Juveniles: Conferencing, Mediations and Circles. Oxford: Hart Publishing.

Graber, L., Keys, T., & White, J., (1996). “Family Group Decision-Making in the United States: The Case of Oregon.” In Hudson, J., et al. (eds.), Family Group Conferences: Perspectives on Policy and Practice. New York: Criminal Justice Press.

Hall, D. (1996). Criminal and Procedure. Albany N.Y: Delmar Publishers.

Hassall, I. (1996). “Origin and Development of Family Group Conferences.” In Hudson, J., et al. (eds.), Family Group Conferences: Perspectives on Policy and Practice. New York: Criminal Justice Press.

Hyndman, M., Thorsborne, M., and Wood., S. (1996). "Community Accountability Conferencing." Trial Report. Queensland Department of Education.

Lewin, K. (1997). “Experiments in social space (1939)” In: G. Lewin (ed.), Resolving Social Conflicts & Field Theory in Social Science. Washington D.C.: American Psychological Association. Maxwell, G. & Morris, (1993). Family, Victims and Culture: Youth Justice in New Zealand. Wellington, NZ: Victoria University.

McCold, P. (1999). "Restorative Justice Practice: State of the Field." Paper presented at Building Strong Partnerships for Restorative Practices, Burlington, Vt. August 5 - 7, 1999. Retrieved December 13, 2001 from the World Wide Web:

McCold, P. (1999). "Virginia Conference Project A Four-Site Evaluation." Retrieved August 27, 2001 from the World Wide Web:

McCold, P. & Wachtel, B. (1998). "Restorative Policing Experiment: The Bethlehem Pennsylvania Police Family Group Conferencing Project." Retrieved March 15, 2001 from the World Wide Web:

McCold, P. & Wachtel, T. (2000). "Restorative Justice Theory Validation." Paper presented at Fourth International Conference on Restorative Practices for Juveniles, Tubingen, Germany, October 1 - 4, 2000. Executive Summary Restorative Justice Theory Validation. Retrieved August 27, 2001 from the World Wide Web:

McGarrell, E., Olivares, K., Crawford, K., & Kroovand, N., (2000). "Returning Justice to the Community: The Indianapolis Juvenile Restorative Justice Experiment." Hudson Institute, Crime Control Policy Center. Retrieved June 12, 2001 from the World Wide Web:

Moore, D. B. & Forsythe, L. (1995). "A New Approach to Juvenile Justice: An Evaluation of Family Group Conferencing in Wagga Wagga." A report to the Criminology Research Council. Wagga Wagga, Australia: Centre for Rural Social Research, Charles Stuart University-Riverina.

O’Connell (1998). "From Wagga Wagga to Minnesota." Paper presented at Conferencing: A New Approach to Wroingdoing, Minneapolis, MN, Aug. 7, 1998. Retrieved February 5, 2001 from the World Wide Web:

O’Connell, T., Wachtel. B. and Wachtel, T. (1999). Conferencing Handbook. Pipersville, PA: The Piper’s Press.

Pranis, K. (1996). “A State Initiative toward Restorative Justice: The Minnesota Experience.” In: B. Galaway and J. Hudson (eds.) Restorative Justice: International Perspectives. New York: Monsey Criminal Justice press, pp. 493-504.

Schiff, M. (1998). "Restorative Justice Interventions for Juvenile Offenders: A Research Agenda for the Next Decade." Western Criminology Review 1(1). Retrieved October 14, 2000 from the World Wide Web:

Sherman, L., Strang, H., Barnes, G., Braithwaite, J., Inkpen, N., and The M.M. (1998). "Experiments in Restorative Policing. A progress report on the Canberra Reintegrative Shaming Experiments (RISE) to the National Police Research Unit." Canberra ACT: Australian National University, June.

Sherman, L., Strang, H., Barnes, G., Braithwaite, J., Inkpen, N., and The M.M. (1999). "Experiments in Restorative Policing. A progress report on the Canberra Reintegrative Shaming Experiments (RISE) to the National Police Research Unit." Canberra ACT: Australian National University, July. Retrieved August 29, 2001 from the World Wide Web: rjustice/rise/progress/1999.html

Shook, E. V. (1985). Ho’oponopono: Contemporary uses of a Hawaiian problem-solving process. Honolulu: University of Hawaii Press.

Snyder, H., (1988). "Court Careers of Juvenile Offenders." National Center for Juvenile Justice, Research Division for the National Council of Juvenile and Family Court Judges.

Some, M. P., (1998). The Healing Wisdom of Africa. New York:Tarcher/Putnam.Schiff, 1998

Tutu, D., (1999). No Future Without Forgiveness. New York: Image.

Umbreit, M. & Fercello, C. (1998). "Family Group Conferencing Program Results in Client Satisfaction." Juvenile Justice Update, 12.

Umbreit, M. & Fercello, C. (1999). "Client Evaluation of Family Group Conferencing in 12 Sites in the 1st Judicial District of Minnesota." Center for Restorative Justice and Mediation, School of Social Work, University of Minnesota.

Walker, L. (2001). "Conferencing: Western Application of Indigenous People’s Conflict Resolution Practices." Paper presented at the Fifth National Conference on Family and Community Violence Prevention, Los Angeles, April 2001.

Zehr, H., (1990). Changing Lenses. Scottdale, PA: Herald Press.

Zion, J. W., (1998). "The Dynamics of Navajo Peacemaking." Journal of Contemporary Criminal Justice, 14 (1), 58-74. Retrieved February 5, 2001 from the World Wide Web: citing Palk, G., H. Hayes and T. Prenzler. (1998). "Restorative justice and community conferencing: Summary of findings form a pilot study." Current Issues in Criminal Justice. 10(2):138-155.