Responsive regulation in child welfare: systemic challenges to mainstreaming the family group conference.

The purpose of this article is to examine the challenges inherent in transforming child welfare services. We apply Braithwaite's model of responsive regulation to the restorative practice of family group conferencing in child welfare. Shifting the role of the state away from controller of families in the child protective services system to one of regulatory partner with them is extraordinarily difficult. The paper looks at the complexities of reorienting child welfare services through the use of family group conferences on a large scale.

Key words: responsive regulation, restorative justice, family group conference, child welfare, child protection

**********

In this article, we argue that Family Group Conferencing (FGC) in child welfare has a transformative potential, but one that is hard to realize for systemic reasons. We first examine the application of Braithwaite's (2002) model of responsive regulation to child protective services (CPS) in the United States. We argue that model enables us to understand the relation of two apparently contradictory but essential elements of FGC--empowerment and the context of social control or state coercion. We discuss the difficulties of implementing FGC in the United States and suggest that the range of models and variations currently being practiced across the country may be in part evidence of and a response to these difficulties.

Drawing on the experience of the State of Hawaii in adopting a uniform model of FGC and its application statewide to over 2,000 cases, we conclude with a discussion of the kinds of system change needed to facilitate and support the mainstreaming of FGC. The wide variations in FGC and in other forms of family group decision-making (FGDM) may be decried as representing compromises of core principles that undergird the model. This variety also is making evaluative research difficult since researchers are failing to specify a consistently applied model of the elements of FGC prior to analyzing its effectiveness. On the other hand, the diversity of forms of FGC may be celebrated both as creative adaptations to local conditions and cultures and as providing a natural experiment without prejudging the key, efficacious components of the approach. We sidestep these disputes and take a different path here one of exploring the systemic context in which policy-makers and practitioners are attempting to apply FGC principles and processes.

Conferencing as Regulation

Braithwaite's (2002) work on restorative justice and responsive regulation provides a valuable conceptual framework for this undertaking. It enables us to see both the restorative aspects of FGC--its relation to indigenous practices aiming at solving problems and setting things right, its empowerment of families, and its widening the circle of care and control beyond the professional-client relationship--and FGC's role in the context of responsive regulation of families. Braithwaite's discussion of responsive regulation draws on the field of business regulation. It enables us to see, in the complex field of child welfare, how combining the empowering aspects of FGC with the coercive power of the state is not necessarily a limitation or contradiction. Rather, empowerment and control are different, but necessary and mutually enriching aspects of a dynamic model of state regulation of families to protect children.

The Braithwaite Pyramid

Braithwaite (2002) contends that "restorative justice, deterrence and incapacitation are all limited and flawed theories of compliance" (p. 32). Each needs to be understood and applied in a model that includes all three. In his figure entitled "Toward an Integration of Restorative, Deterrent and Incapacitative Justice," Braithwaite (2002, p. 32) hierarchically orders these concepts and places restorative justice at the base of the pyramid, filling up most of the space.

The pyramid provides a dynamic, non-formalist model of governmental regulation, whether of a nursing home, a nuclear power station, an insurance company, or a family. The formalist approach to regulation seeks to define in advance which problems or failures of compliance require what official responses and mandate them in regulations. In responsive regulation, by contrast, there is a presumption, regardless of the seriousness of the offense or violation, in favor of starting official intervention at the base of the pyramid. Moving up the pyramid to deterrence and, ultimately, incapacitation, is a response not to the seriousness of the harm done but to the failure to elicit reform and repair at the base with restorative justice processes. Of course, as with other violent crimes--a shooting spree in progress, for example--an immediate move to incapacitation (at least temporary) may be necessary in cases of child abuse where there is imminent and continuing danger to the child.

The presumption in favor of starting at the base of the pyramid, Braithwaite (2002) argues, not only favors less coercive and costly state intervention where possible, but also makes more coercive measures more legitimate when escalation up the pyramid is necessary. This is important because "when regulation is seen as more legitimate, more procedurally fair, compliance with the law is more likely" (Braithwaite, 2002, p. 33; see also Neff in this issue).

By analogy, accreditation of a professional school, whether of law, medicine, nursing, or social work, is a process of required self-regulation through a process of self-study and reform. The accrediting body, e.g., the Council on Social Work Education in the case of U.S. baccalaureate and masters of social work programs, has a range of escalating options to identify and encourage schools to address concerns and come into compliance, culminating in the ultimate and rarely used action of withdrawing or denying accredited status. In extreme situations, withdrawal might be immediate rather than a final step in a succession of regulatory actions. Although this option of denying a school accreditation is understood by all to be at the Council's disposal as a last resort, site visitors and accreditation commissioners work in collegial partnership with schools with the shared aim of avoiding escalation up the regulatory pyramid.

In applying this model to a business such as a nursing home, nuclear power station, or insurance company, a regulator would begin to work with the firm's management at the base of the pyramid. Both management and regulator are aware that if the firm proves unable or unwilling to make the changes needed to come into compliance, the next level of regulation will be more coercive. At the first level, the assumption is that the firm's management is a "virtuous actor," with the will and capacity to respond to the regulatory process by taking the steps needed to come into compliance. Regulator and regulated work together to prevent a more coercive regulatory response. At the next level, management may have no wish to cooperate with regulators or to make the necessary changes, but is assumed to be a "rational actor" who--faced with a fine or other penalty and the threat of being put out of business--will calculate that it is better to comply. At the highest level of the regulatory pyramid, management is assumed to be an "incompetent or irrational actor" who is unable or unwilling to comply and who therefore needs to be incapacitated by losing its license to operate.

Applying this model to FGC, we see conferencing as a restorative process at the base of the pyramid. Following Marshall's widely quoted definition of restorative justice as process, FGC may be seen as a "process whereby all the parties with a stake in a particular offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future" (Braithwaite, 2002, p. 11). Conferencing involves core values as well as processes, having to do with healing and setting right, moral learning, community and kin participation, respectful dialogue, responsibility, apology, forgiveness. "So restorative justice is about restoring victims, restoring offenders, and restoring communities ... Stakeholder deliberation determines what restoration means in a specific context" (Braithwaite, 2002, p. 11).

In contrast to youth or criminal justice contexts, conferencing in child welfare is not directly focused on the wrongdoing and the harm done, nor is there a process, in any usual sense, of resolving a conflict between offender and victim and between their respective families. Instead the focus is on the future safety and well-being of the children involved. The extended family partners with the system professionals--the family regulators--in order to develop a plan to achieve this. The family members receive significantly more information than usual from the state about their "case", including the actions of the social worker, the official concern about the abuse or neglect, and any other pertinent facts about resources and constraints that could affect decision-making. The process is concerned not with holding the "offender"--the maltreating parent--passively accountable for past actions, but with engaging the extended family group in taking active responsibility for generating and implementing solutions. Conferencing aspires to form a true partnership of family and state, even though the state retains the ultimate veto power in light of its legal mandate and responsibility to protect children. In this sense, child welfare conferencing shares with other forms of conferencing and restorative justice certain principles--about healing, moral learning, taking active responsibility for resolving the situation--but de-emphasizes others such as apology, making amends, and forgiveness.

These differences of emphasis notwithstanding, FGCs in child welfare fit well at the base of Braithwaite's regulatory pyramid. They offer a decision-making process in which the regulator (the public child welfare agency) works with the family in a respectful mode that assumes that the family group (if not necessarily the parents) has both the capacity and the caring concern to come into compliance with the law and community standards so that children are protected from abuse and neglect. The process taps into and mobilizes the knowledge, wisdom, and caring capacity of the extended family, its culture and community, in order to plan for the children's safety.

Levels of the Pyramid

FGC, from this perspective, is a process of state-enforced family self-regulation, a collaborative regulatory process in which the professionals and any other community representatives help the family design a strategy to come into compliance by providing information, a structure, and access to resources. However, the regulatory context of this empowering, professional-family partnership is one in which all understand and seek to avoid escalation up the pyramid's levels to a more coercive regulatory response in which the state, through its social workers and/or family court, take over the decision-making process. At the apex of the pyramid lies the regulatory option of incapacitation, which in the case of families in the United States could mean the involuntary termination of parental rights and a permanent alternative placement for the child, such as adoption.

Braithwaite (2002) suggests that in the case of business regulation, "Perhaps the most common reason ... for successive failure of restorative justice and deterrence is that noncompliance is neither about a lack of goodwill to comply nor about rational calculation to cheat. It is about management not having the competence to comply" (p. 32).

Much the same may be said about families. That is, restorative, empowering practices like FGC as well as more direct control by CPS may fail, not because rational and competent adults in the family group choose to defy the child welfare regulators (CPS) or fail to appreciate the threat of losing parental rights if they do not comply, but rather because they lack the capacity or competence to make the changes required for the safety and well-being of their children. Escalation up the pyramid results from failure, for whatever reason, to respond at lower levels of coercion. The model also allows for movement back down the pyramid away from coercive interventions when a family demonstrates its will and capacity to plan for and assure the children's safety. The model provides for movement to more or less coercive forms of social control according to the family's behavior.

In Braithwaite's model, deterrence represents an intermediate place in the pyramid between restorative justice and incapacitation. This fits well with the use of FGC's in the youth or adult justice systems, for generally restorative justice occurs in the context of violations of criminal law with the potential for imposition and escalation of formal punishment. Although child abuse and neglect violate the law, criminal penalties are rarely invoked except in the most egregious cases. In those circumstances, criminal sanctions are retributive measures commensurate with the perceived seriousness of the offense, not necessarily a response to the failure of restorative regulation. Deterrence in the form of fines or other penalties analogous to those invoked in the deterrent phase of business regulation is similarly rare.

In child welfare, therefore, deterrence does not seem to fit neatly between restorative justice at the base of the pyramid and incapacitation at the top. Rather, if restorative justice at the base takes the form of the collaborative, respectful, non-coercive decision-making process of FGC, the place of deterrence in the middle is taken by professional- or court-determined disposition of the case, imposed with or without the family's agreement. In practice, that might mean a temporary custody arrangement for the child with a stringent time line in which the family must comply or the court would begin taking steps toward permanently removing the child. The deterrent for the family is not the threat of a punishment such as a fine, as might be imposed in a case of business regulation, but loss of control, of active, decision-making responsibility for their children, to the state.

We may assume that such loss of control is a powerful deterrent for families that may not wish to comply with legal requirements or collaborate with the child welfare authorities, but who nevertheless have the capacity to make and act on a rational calculation of the inevitable consequences of their failure to do so. In Braithwaite's scheme, these are rational actors, in contrast both to the virtuous actors at the base of the pyramid--who seek the best for their children and willingly collaborate in the restorative FGC process--and to the incompetent or irrational actors at the apex who are unable or unwilling to comply. Despite some peculiarities of the child welfare context, FGC's provide both a form of restorative justice and an empowering, non-coercive, respectful form of collaborative decision-making at the base of the regulatory pyramid.

The key to applying the regulatory pyramid to child protection is to see that the levels of the pyramid do not reflect particular outcomes but different decision-making processes with different degrees of coercion or non-domination. For example, a decision to remove children permanently from the care of their parents, through guardianship or other mechanisms, could be an outcome arrived at even at the base of the regulatory pyramid. That is, it could be part of the plan developed by the extended family group in its private family time and endorsed by the other participants in the conference, by the responsible agency, and subsequently by the court. The pyramid in child welfare reflects a continuum from state-managed family self-regulation to an outcome unilaterally imposed by the state, not a continuum of outcomes from reunification with supportive services to termination of parental rights.

Empowerment and Coercion

If a formal regulatory philosophy objects to the inconsistency of punishment implicit in restorative justice--the punishment does not fit the crime, but depends upon the offender's willingness and capacity to set things right--an objection may be made from the opposite direction that truly restorative justice, not to mention empowerment, is impossible in the presence of an implicit threat of escalating coercion.

Braithwaite (2002) makes the point that as a matter of fact, very few criminal offenders would participate in restorative conferences in the absence of some degree of coercion in the form of detection and/or arrest, and perhaps the specter of a trial. A conference or sentencing circle puts the offender's behavior under the scrutiny of family and community and inevitably involves shame. Shame, as Braithwaite (1989) argued in his earlier classic work, Crime, Shame, and Reintegration, may stigmatize or, in a restorative justice process, help to reintegrate the offender. In the case of child abuse and neglect, the FGC process brings secrets, e.g., of violence within the family, out into the open, brings contributing behavior such as drug use to the attention of both maternal and paternal relatives, and focuses the attention of all participants on the harm done to the children and the need to protect their future well-being. It is reasonable to assume official investigation and the prospect of court involvement and ultimately of losing one's children, create a coercive context that provides a strong and often necessary incentive for abusing or neglecting parents to collaborate with the conferencing process.

Braithwaite's (2002) hypothesis is that restorative justice works best in the context of coercion, but where the implicit threat--in this case of loss of control over decisions involving one's children's future--is in the background. Similarly, Burford and Pennell's (Burford & Pennell, 1996, 1998; Pennell & Burford, 2000) research on an FGDM project to reduce family violence in Newfoundland and Labrador suggested that the sanction, support, and resources of government and professionals could facilitate the family's decision to stop violence within it. Vulnerable family members were protected both by widening the circle of responsibility and control to include those who love and care for each other, ending secrecy and isolation and providing for active family responsibility for its members, and by the sanction of the law. The legal context of FGC--the fact that in New Zealand the state has veto power if the family's plan is contrary to the principles of the Children, Young Persons, and Their Families Act--sets up a tension that may make for sharper focus and better decisions.

The legal system, similarly, constrains the "people's justice" of FGC by providing a recourse for the protection of individual rights, at the same time as FGC loosens the dead hand of bureaucratic-professional domination and unleashes the creativity, wisdom, and active responsibility of families and communities. The hypothesis is that the interweaving of formal and informal, state and family, professional and community care and control in this way makes for a fairer process and better decisions than either would achieve if left to itself (Adams, 2002; Braithwaite, 2002; Strang & Braithwaite, 2001).

Adapting Braithwaite's (2002) hypothesis to child welfare conferencing enables us to see, however, even in the social control sphere of child protection, an important difference between the ultimatist social worker who threatens the parent with loss of her child if she does not comply with her instructions, on the one hand, and on the other, the process of collaborating with the parent and others involved to prevent escalation up the regulatory pyramid. This is true even though all understand that escalation is inexorable if they cannot work together to find ways to keep the children safe. The focus of the conference is not on blaming or threatening the maltreating parents, but on meeting the needs of the children themselves. The threat is in the background, not the foreground. This difference makes possible the empowering nature of an FGC, which enables the family group to tap into its own knowledge, wisdom, and resources in collaboration with the others involved, including the responsible agency and professionals. The family group takes active responsibility for mobilizing its collective capacity to care for the children and keep them safe through whatever plan they, rather than the professionals, play the lead role in developing.

The Promise of FGC

The child welfare FGC brings together those affected by a situation of harm to children--parents, children, extended family members, other community supports or fictive kin, and professionals (such as therapists, school officials, the CPS social worker, and the conference coordinator/facilitator)--to provide the family group with all the pertinent information and engage them in a planning or decision-making process. It widens the circle of care and responsibility, shares information with the family that under professional-dominated practice would be kept confidential to the professionals alone, and potentially places the social worker's behavior under scrutiny as well as that of the parents.

FGC in child welfare differs from some other kinds of conferencing, mediation, or alternative dispute resolution. It is not a third-party mediated negotiation of a dispute between offender and victim, or parents and child. It does, however, deal with the aftermath of the (normally undisputed) harm done to the victim and involves a process of taking responsibility for resolving the concerns about the child's safety and well-being. Like other forms of restorative justice, it draws on a cultural treasure of pre-state societies and ancient civilizations that has persisted in many indigenous people's practices around the world. Grounded in traditions of justice that have roots and resonance in many, perhaps all cultures, as Braithwaite (2002) argues, restorative justice was eclipsed with the emergence of the modern state, which "stole conflicts" and their resolution from those affected (Christie, 1977).

The resulting homogenization and impoverishment of justice traditions has, for our purposes, at least two important aspects. One is the domination of the processing of both conflicts and disputes, both within and between families, by professionals licensed by and/or acting on behalf of the state--police, lawyers, social workers, probation officers, among others. The other is that in certain fields like child protection where workers face a high level of public skepticism about their professionalism and competence, their capacity to negotiate solutions with those directly involved is increasingly circumscribed by formal rules, regulations, and procedures (Adams & Krauth, 1995; Adams & Nelson, 1997)

FGC offers a paradigm shift in this context. It fundamentally alters the relationship of professionals and the families and communities they serve. It de-centers the professional-client relationship and widens the circle of responsibility and decision-making to include those whose relationship to the children at risk is based not on professionalism but on caring and kinship. It rests on the assumption, as Burford and Hudson (2000) put it, that "lasting solutions to problems are ones that grow out of, or can fit with, the knowledge, experiences, and desires of the people most affected" (p. xxiii). The key shift in practice and in such legislation as mandates it (above all the 1989 Children, Young Persons, and Their Families Act of New Zealand) is from professional-dominated practice resting on a model of regulatory formalism to a process for decision-making and planning that mobilizes and empowers the children's kin and community. It does not replace the formal processes of the court system, but-and this is surely the potential of restorative justice in modern societies--it enables both formal and informal care and control to enrich and constrain each other.

Threats to Implementation in the United States

So much for theory, principles, and promise, but what of the challenge of translating these into policy and practice in the United States? Bringing about such a fundamental shift in the balance of power between child welfare professionals and families, a shift from domination to partnership, is both at the heart of FGC and a daunting task. Even the New Zealand model, the gold standard for those who emphasize this aspect of FGC, in practice has the potential to be subverted by social workers' reasserting their control over the process and outcome (Lupton & Nixon, 1999, Walker, 2003).

In the United States, there are many variations on FGC, most of which increase this potential to weaken FGC's radical core. These variations both reflect and highlight the political and bureaucratic difficulties faced by those who wish to adopt this innovation in their child welfare systems.

In the remainder of this article, we explore the variations and the threats to implementation in an American context, and discuss possible reasons for them. We discuss the case of Hawaii, which provided a particularly favorable environment for the adoption of its own version of FGC. The cultural relationship and contacts of Native Hawaiians with the Maori of New Zealand who provided the early impetus for FGC, the small size of the state, and its pattern of political-bureaucratic centralization allowing for statewide adoption of a uniform program, all made for statewide adoptability of this innovation. All this is considered in order to arrive at a realistic and sober assessment of the difficulties involved in adoption of FGC as an empowering, paradigm-shifting practice, and to suggest what systemic changes are needed to support the full realization of FGC's potential.

Model Variations

Pennell (2003) stresses the importance of model fidelity among the variety of conferencing programs across the country and articulates several minimum components required for conferencing. Without legislative mandates or judicial guidance in the United States, as is the case in New Zealand and England, the practice models of FGC in the U.S. vary widely. Of particular concern is the fact that even before the basic model of FGC has been implemented widely and empirically tested, criticisms are beginning that the model is too expensive; should be limited to certain types of case; and does not need specially trained staff or extensive preparation time. We lack empirical research with a robust design that compares outcome variables of family conferencing with matched control cases. Given this paucity of empirically tested outcome research on conferences, there is little to guide policy or social work practice. However, how do we get to reform if we do not vigorously implement and test new models of practice?

Pennell (2003) identified nine key principles and practices for assessing family conferencing model fidelity. Using the North Carolina Family Group Conferencing Project, she developed a 25-item questionnaire to evaluate if the family participants, coordinators, and service providers understood the conferencing process, actually experienced it during the conference and agreed that the listed conferencing practices had been conducted and its purposes achieved. Participants were asked questions like "[Was the] FGC coordinator respectful of the family group?" and "[Did] the family group have private time to make their plan?" (Pennell, 2003, p. 19). Her findings are encouraging for two reasons. One, the participants generally agreed that the core principles of FGC were indeed being implemented and experienced. Secondly, and perhaps most importantly, Pennell is asking the correct question: "How widespread is FGC and is there model fidelity?" Merkel-Holguin (2003) reports that in 2003, more than 150 communities in 35 states and approximately 20 countries are implementing FGDM initiatives. This suggests that a revolution in child welfare practice is under way. However, given the challenges in implementing such a model, we are concerned about the ability of many states to put this reform in place in ways that are true to FGC's core principles.

Good Intentions

While the conferencing model uses the language of family empowerment, partnerships, and child welfare reform, if it is to be truly implemented, the role of the state in child welfare has to be dramatically altered. Recent evaluations underscore the dangers of basing policy decisions on evaluations of programs that may depart substantially from the principles of the innovation it seeks to replicate. For example, in the study of the family decision meeting in Oregon, only a little more than half the family members reported knowing they could invite others to the conference (Rodgers, 2000). In Miami, 38% of the parents and family members interviewed reported that the private family time was "not useful at all" (NCJFCJ, 2002). It is hard to believe, however that many families who were offered the opportunity to design a safety plan for their child and believed that their input would be seriously considered by the child welfare agency could not find this useful. We wonder if these families were truly offered this opportunity or if the family members believed that the plan had been made and agreed to in advance by the professionals. In that case, private time does indeed become superfluous. We wonder if private time is really being implemented as a period in which the family has, and understands it has, the opportunity for taking responsibility for developing a safety plan. Private family time is not a part of Oregon Family Unity Meetings, and in Michigan it is optional. Many communities do not have independent or trained facilitators and to date, there is no required training for conference facilitators. What exactly is happening inside the conference circle in many communities and how radically it differs from traditional case conferences with the family present is as yet unclear.

Reform is Hard to Implement

FGC uses an exchange-based model of assessment and issue definition that includes the family in all stages of the decision-making about child placement. This arrangement forms a new collaboration based on shared information and trust between the state and the family. This is a significant change from the traditional child protection practice in which the public agency and the court manage most of the information unfettered by the family's questions or inputs. We have previously described the core elements of conferencing as: widening the circle, taking/sharing responsibility for solutions, culturally competent practice, family leadership and empowerment, family driven solutions, community partnerships and private family decision making time (Adams & Chandler, 2003). Pennell's (2003) key principles and practices include these and add a few more. She includes creativity in planning; ensuring the conference belongs to the family group and "chang[ing] policies, procedures and resources to sustain partnerships among family groups, community organizations and public agencies" (p. 17). Creative social workers may be able to apply an empowering model of practice and participate in conferences that meet many of these principles. However, changing policies, procedures, and resources to promote such practice is a profound administrative challenge. It is likely that it will be difficult, in the metaphor of diffusion of innovation as contagion (Smale, 1996, 1998) to infect the public child welfare agency with the conferencing bug and infuse it throughout the system.

Possible Reasons for the Variations

Merkel-Holguin (2003), after reviewing the FGC literature in the United States, concludes that the model is not well diffused across the nation and is at best a marginalized practice in most communities. When one considers the significant change in values, practice, behavior, and policies of FGC if implemented on a large scale, it is not surprising that this system reform has not taken root in many places. The reluctance of professionals to fully enable consumers to describe their own needs has been reported (Lupton & Nixon, 1999). This has been seen in the struggles toward the development of a person-centered practice in the fields of mental health and developmental disabilities. In these areas, federal and state laws were required in order to compel state agencies to implement initiatives that involved consumers and their family members in case planning as well as policy making (Rothman, 1990; Stroman, 2003).

Implementation Barriers (Real and Perceived)

Lupton and Nixon (1999) suggest an intriguing reason for social workers' reluctance to partner with their clients--that professional autonomy was declining in public child welfare agencies due to frequent litigation, class action suits, and court consent decrees. Their theory suggests that whenever a state lost a lawsuit or settled a consent decree, the public agency would hunker down further under their rules, make new rules, and constrict and constrain workers' discretion in an effort to demonstrate and document that the state was complying with the lawsuit (Rothman, 1990). Often the court's or plaintiff's monitoring process resulted in the agency's becoming less responsive and less innovative. An unintended consequence was more bureaucracy and paperwork and less communication and openness with consumers. Rarely is this a time of creativity and reform even though that may be what is required and often what the suit is demanding.

Even professional, masters level social workers lose discretion in these circumstances and see themselves subject to surveillance by even more levels of supervisors who oversee multiple elements of each case, presumably concerned about the case holding up under attorney scrutiny in court. When this siege mentality sets in, individual social workers believe that they will be held personally accountable for any negative case outcome such as re-abuse of a child, whether in foster care or in the biological home. Fear of making a mistake leads to high turnover, job burnout, and dissatisfaction (US-GAO, 2003). While there is logic to the idea that social workers should embrace reform efforts that result in the sharing of some of the responsibility with family members and community partners, since this is deeply embedded in their core values (see NASW Code of Ethics), it is difficult to discern this level of empowerment in many child welfare agencies.

Threats to the implementation of FGC abound. Implementation may be seen as requiring too much time and overtime pay for workers, as too expensive or just unnecessary since there is little evidence to support its superiority. With its creativity and variety of elements, a family's safety plan is likely to look very different from one drawn up by a public child welfare agency and a traditional agency may be unenthusiastic about embracing such differences. Child welfare workers in Hawaii estimate that over 80% of the compliance plans in child protection cases include some or all of the same components: substance abuse treatment, anger management, and parenting classes (Hawaii Department of Human Services, 2002). This seems to be what the CPS workers believe is necessary and the courts like to see. There is a strong belief that these address the deficits that almost all families in CPS have.

If professionals in child welfare are experiencing a continuing loss of discretion and autonomy (Lupton & Nixon, 1999), it is to be expected that they will not readily embrace a concept of sharing responsibility that appears to put them professionally at risk (Adams & Krauth, 1994). NCJFCJ (2002) reported that workers often do not refer cases to FGC, believing that families in the child welfare system do not have the capacity to be decision-making partners. Social workers may have been trained to do comprehensive assessments of family deficits rather than focusing on the strengths of larger family systems as a base for building solutions. These are just some of the barriers that are likely to prevent FGC from moving into the mainstream.

We suggest that the wide variety of models now being practiced, while at best creative and responsive to local conditions and stakeholders, at worst compromises FGC principles in an accommodation to prevailing legal, bureaucratic, and professional norms. Nonetheless, the experience of FGC in Hawaii indicates the potential as well as the challenges of implementing a statewide model, of taking a conferencing pilot project to scale (Schorr, 1997). It indicates that FGC can be the practice of choice and not a marginalized or fringe project of child welfare reform.

The Case of Family Group Conferencing in Hawaii

The experience of the State of Hawaii illustrates the kind of collaborations and political and community support needed to implement FGC as a model of child welfare reform. FGC in Hawaii (called 'Ohana Conferencing; 'Ohana is the Hawaiian word for extended family and/or supportive networks) was implemented as a pilot project in November 1996 as the result of collaboration between the Family Court in Honolulu and the Department of Human Services (DHS). (In discussing this experience, we draw on the first-hand knowledge of one of the authors, Susan Chandler, who was the state director of human services from 1995 to 2002.) Hawaii's conferencing program was originally one of four demonstration court diversion programs funded by the National Council of Juvenile and Family Court Judges (NCJFCJ, 2003) to facilitate systems change in the processing of child abuse and neglect cases (NCJFCJ, 2000).

After several staff trainings conducted by experts from New Zealand, Honolulu was selected as the pilot site for a conferencing project. The state agency chose to contract out all of the needed conferencing services, including the initial planning and organizing of the conferences, as well as the facilitation of the conference. The state contracted with the private agency, Effective Planning and Innovative Communication (EPIC) for all staff training of community-based facilitators as well as the CPS workers. The decision to privatize this reform effort, which had support from many community organizations, social services agencies, and State legislators, is believed to have been an essential factor in the successful statewide implementation of conferencing. EPIC was a newly formed, non-profit agency with the single goal of conducting family group conferences. It immediately established high standards of professional practice for its entire program staff. Conference facilitators are required to have at least a bachelor's degree and experience in working with children as well as demonstrated multi-cultural practice experience. In each conference, there must be a professionally trained group facilitator as well as a neighborhood-based community facilitator to insure family comfort and participation.

The pilot project began as a court diversion effort using conferencing to divert CPS cases away from the judicial system. This diversion goal, rather than family empowerment or child welfare reform, perhaps is one reason why conferencing in these first demonstration projects has been implemented successfully. The judges in Honolulu wanted to test strategies that included parents and other family members early in the case processing and have them actively participate in the decision-making related to their children. The court wanted to open up its proceedings and make the whole experience more humane. The project also aimed to increase family participation as a strategy to move the cases through the system more quickly.

The child welfare agency director in Hawaii was a strong supporter of this practice reform. Improving communication with family members and providers, and diverting cases out of the system completely or having families spend less time in the system was a primary agency goal. After one year, the court stopped funding the project and since 1997 the Department of Human Services has been the primary advocate for conferencing and now finances the entire program with state and federal funds.

Elements of a Successful Innovation

"A key ingredient for organizational change is the commitment and involvement of leaders throughout the agency," according to Sahonchik (2003, p. 1). This is classic management advice, however leadership in a child welfare agency is often quite decentralized with local branches, or county structures. Daft and Marcic (2001) suggest that political appointees at the top of a public human service agency rarely can bring about lasting change, since the permanent civil service staff may decide to just wait out their latest reform idea and wait for the next political appointee to arrive with yet another agenda for change. To inoculate against this known resistance to change strategy, seven-year contracts were written in Hawaii with EPIC to ensure that FGC would endure past the changes of the political administration.

Where conferences have taken hold, there has been a confluence of forces, from inside the public child welfare agency and in the external environment that joined to make conferencing possible (see special FGC issue of American Humane's Protecting Children (2003)). Smale (1996) wrote that "changes in practice come about through 'convergent' thinking and 'contagious processes'" (p. 20). The first ideas about conferencing came from visits from Maori leaders meeting with Hawaiian leaders concerned about the high rates of incarcerated Hawaiian and part-Hawaiian youths and an overrepresentation of Pacific Islander children in the child welfare system. This dialog, along with broad legislative and community-wide support for a blueprint for child welfare reform initiative as well as a new Director of Human Services with a social work background, focused attention on conferencing as a strategy for improving child protective services. However, like many reform efforts, many inside CPS did not see the need for any innovation or reform, because they did not see a problem. To introduce change into a system that does not see a problem, rarely succeeds. However, when discussions focus on finding a consensus on a common vision, values, mission, and principles, agreements can often be found. Trainings that were titled, "How to reduce your caseload using family group conferencing" were popular.

One strategy used in Hawaii's DHS to infuse the idea of conferencing throughout the agency was to make trainings and workshop attendance competitive rather than mandatory, which had been the usual custom. Staff needed to submit in writing why they wanted to attend and the number of participants was limited. This made the training seem like a scarce commodity and seemed to draw people to it.

After the first few trainings, a core group of innovators affectionately called the "bungee jumpers" emerged. These were the staff that immediately liked the idea of conferences as a philosophy of practice and volunteered to try it out. One rural unit became the first test site. Subsequently, the model of expansion and diffusion was to move unit by unit geographically as the staff got trained. The hope was that conferencing would become contagious. The use of a private, not-for profit agency to conduct the conferences allowed the workers who first adopted the reform to get help with their cases, which was a strong incentive. As conferencing spread throughout the units, it became clear that some supervisors supported the reform and some did not. When supervisors did not support conferencing, their units made little use of it. To better track the level of infusion of the innovation, the agency established a goal of two conferences in every unit each quarter, monitored each unit's progress, and publicized the data. Supervisors who did not have conferences were reminded about the goal and were encouraged to send staff (or go themselves) for more training.

Resistance to adopting the model came in many forms. Mostly, workers explained their resistance by saying that the families in the caseload did not have the capacity to attend a conference or there was no extended family or supports to draw upon. Of course, if a worker does not look for such support, he or she will rarely find it. Some workers complained that certain cases were just not appropriate. Although never explicit, this became a belief among the non-reformers. Agency-wide policies and procedures needed to be developed to provide incentives for workers to participate. Flex-time and glide time working hours were implemented so workers could attend conferences in the evening or on the weekends as well as use comp-time or overtime to encourage worker participation. To insure that family agreements were honored, a new policy established in 2002 that no agreement made within a conference could be overruled by a supervisor who had not attended the conference. There was concern among the workers who had never attended a conference that families would design safety plans that were not safe and social workers would not be able to veto a family's plan. While there is a written policy that permits such a veto, to date it has never been needed or used.

While there have been over two thousand conferences convened throughout the state of Hawaii, the reform is still not consistently implemented throughout the state. Some CPS units simply do not refer cases to conferences. The strategy remains to continue training in each unit and work to show social workers the benefits of conferencing. A state law was passed encouraging the use of conferencing and requiring CPS workers to explain why a conference should not be held for all voluntary cases. Families do not yet have the right to a conference, but perhaps this is the next essential step to insure its implementation.

Conclusion

In synthesizing his two major areas of work, on restorative justice and responsive regulation, Braithwaite (2002) provides a framework for understanding the tension at the heart of social work in general, and child welfare in particular, between empowerment and control. Family group conferencing brings to the project of child protection a form of restorative justice within a framework of responsive regulation that suggests the potential for improving decision-making, procedural justice, and compliance. It has radical implications for practice and policy that generate serious challenges to implementation in an American legal and professional context, even in a state with many forces favorable to such an innovation.

Shifting the role of the state away from controller of families in the child protective services system to one of regulatory partner with them is extraordinarily difficult. Hawaii's experience suggests that adoption of FGC in the mainstream of child welfare and implementation of its core values, principles, and practices require a reorientation of professional practice and bureaucratic functioning. The regulatory pyramid provides a framework for understanding child protection both as a form of state regulation and as a shared responsibility of families and communities. Far from abdicating state responsibility to protect children, it situates FGC within an inexorable process of adapting the level of state coercion to the response of the family to regulatory intervention. Far from offering only the appearance of empowerment to mask bureaucratic and professional domination and appropriation of informal family and community decision-making control and capacity, the regulatory pyramid shows the potential and the value--though not the inevitability in practice--of active family responsibility and empowerment at its base.

We have argued that when restorative practices work well, formal and informal care and control constrain and enrich each other. Perhaps the same may be true of the large-scale adoption and diffusion of an innovation like FGC, which both depend upon administrative support and legislative mandate--or at least a favorable legislative and judicial environment--and also require creative adaptation by local stakeholders and practitioners. Given the pressures to compromise on core principles we have discussed, it may be that only adoption on a large enough scale of FGC as a paradigm shift, a fusion of restorative justice and responsive regulation in child welfare, can provide the empirical information we need to improve both theory and practice.

References

Adams, P. (2002). Interweaving formal and informal care: The case of family group conferencing. Social Development Issues, 23(3), 36-42.

Adams, P. & Chandler, S. (2002). Building partnerships to protect children: A blended model of Family Group Conferencing. Family Court Review 40(4).

Adams, P. & Krauth, K. (1994). Empowering workers for empowerment-based practice. In L. Gutierrez & P. Nurius (Eds.), Education and research for empowerment practice (pp. 183-194). Seattle, WA: Center for Policy and Practice Research, School of Social Work, University of Washington.

Adams, P. & Nelson, K.E. (1997). Reclaimimg community: An integrative approach to human services. Administration in Social Work, 21(3/4), 67-81.

Braithwaite, J. (1989). Crime, shame and reintegration. Cambridge: Cambridge University Press.

Braithwaite, J. (2002). Rebuilding a compassionate Republic: The Family Group Conference as a democratic experiment. Address at the 2002 Family Group Decision Making Roundtable, Monterey, California.

Burford, G. (2001). Advancing innovations: Family Group Decision Making as community-centered child and family work. Protecting Children, 16(3), 420.

Burford, G. & Hudson, J (Eds.) (2000) Family Group Conferencing : New directions in community-centered child and family practice. Hawthorne, NY: Aldine de Gruyter.

Burford, G., & Pennell, J. (1996). Family group decision making: Generating indigenous structures for resolving family violence. Protecting Children, 12(3), 17-21.

Burford, G., & Pennell, J. (1998). Family Group Decision Making Project: Outcome report, volume I. St. John's, Newfoundland: Memorial University.

Christie, N. (1977). Conflicts as property. British Journal of Criminology, 17, 1-26.

Daft, R. & Marcic, D. (2001) Understanding management. Fort Worth: Harcourt College Publishers.

Hawaii Department of Human Services. (2001). FY 2000-2004 children and family services plan. Honolulu, HI: State of Hawaii Department of Human Services.

Lupton, C. & Nixon, P. (1999). Empowering practice? A critical appraisal of the Family Group Conference approach. Bristol, U.K.: The Policy Press.

Mazmanian, D. & Sabatier, P. (1989) Implementation and public policy. Lanham, MD: University Press of America.

Merkel-Holguin, L. (1998). Putting families back into the child protection partnership: Family Group Decision Making. Protecting Children 12(3),4-7.

Merkel-Holguin, L. (2000) Practice diversions and philosophical departures in the implementation of family group conferencing. In G. Burford & J. Hudson (Eds.) Family group conferencing: Perspectives on policy, practice and research. Hawthorne, N.Y.: Aldine de Gruyter.

Merkel-Holguin, L. (2003) Promising results, potential new directions: International FGDM research and evaluation in child welfare. Protecting Children, 18 (1&2), 1

National Council of Juvenile and Family Court Judges (2000). Adoption and permanency guidelines: Improving court practice in child abuse and neglect cases. Reno, NV: Author.

National Council of Juvenile and Family Court Judges (2002). Status report 2001: A snapshot of Child Victims Act model court projects. Technical Assistance Bulletin 6(1). Reno, NV: Author.

National Council of Juvenile and Family Court Judges (2003). Empowering families in child protective cases: An implementation evaluation of Hawaii's 'Ohana Conferencing program. Technical Assistance Bulletin, 7(1). Permanency Planning for Children. Reno, NV: Author.

Nixon, P., Merkel-Holguin, L., Sivak, P and Gunderson, K (2001). "How can Family Group Conferences become family-driven? Some dilemmas and possibilities" Protecting Children 16(3), 22-33.

Pennell, J., & Burford, G. (2000). Family group decision making: Protecting children and women. Child Welfare, (79)2, 131-58.

Pennell, J. (2003). Are we following key FGC practices? Views of conference participants." Protecting Children, 18(1&2), 16-21

Rodgers, A. (2002) Family decision meetings: A profile of average use in Oregon's child welfare agency, Final Report. Portland, OR: Child Welfare Partnership. Rothman, D. (1990). The discovery of the asylum: Social order and disorder in the new republic. Boston: Little, Brown.

Sahonchik, K. (2003). Effective strategies for leaders. In Managing Care for Children and Families. National Child Welfare Resource Center for Organizational Improvement. Portland, ME. University of Southern Maine 

Schorr, L. (1997) Common purpose: Strong families and neighborhoods to rebuild America. New York: Anchor Books

Smale G. (1996). Mapping change & innovation. London: HMSO Publications Centre.

Smale, G. (1998). Managing change through innovation. London: The Stationery Office.

Strang, H., & Braithwaite, J. (Eds.). (2001). Restorative justice and civil society. Cambridge: Cambridge University Press.

Stroman, D. (2003). The disability rights movement: From deinstitutionalization to self-determination. Lanham, MD: University Press of America.

United States General Accounting Office (2003). HHS could play a greater role in helping child welfare agencies recruit and retain staff. A report to Congress. Washington, D.C.: GAO Report #03-357

Walker, H. (2003). Personal communication.

PAUL ADAMS AND SUSAN M. CHANDLER

University of Hawaii at Manoa