Mary P. Koss received the Award for Distinguished Contributions to Research in Public Policy. Award winners are invited to deliver an award address at APA's annual convention. This award address was delivered at the 108th annual meeting, held August 4-8, 2000, in Washington, DC. Articles based on award addresses are not peer reviewed, as they are the expression of the winners' reflections on the occasion of receiving an award.

Justice processing for crimes against women is reviewed. The data reveal conviction rates for partner violence and rape by known acquaintances are miniscule; mandatory arrest, protection orders, and diversion programs inadequately deter rebattering; few losses are compensated; and the adversarial justice process is retraumatizing, exacerbating survivor's self-blame. To better address crimes against women, several nations and tribal communities use communitarian approaches, forms of restorative justice. The offense is framed to include the perpetrator, victim, and community. The process forgoes incarceration to have family, peers, and advocates design perpetrator rehabilitation, victim restoration, and social reintegration of both victim and perpetrator. Evaluations suggest communitarian justice may increase victim satisfaction, raise the social costs of offending, multiply social control and support resources, and open a new avenue to targeted prevention.

Violence against women is a significant public health issue globally and domestically. The U.S. Department of Justice and the Centers for Disease Control and Prevention estimated that 15% of U.S. women have been raped and 22% physically assaulted by an intimate in their lifetimes ( Tjaden & Thoennes, 1998 ). Worldwide, at least one woman in every three has been beaten, coerced into sex, or otherwise abused in her lifetime ( Heise, Ellsburg, & Gottemoeller, 1999 ). Beyond external wounds and well-documented long-term chronic mental and physical health consequences (see Koss et al., 1994 ), violence against women is a violation of social equity ( Cosmides & Tooby, 1992 ). Many view a need for justice as a basic human motive. Although the actual social institutions that developed as a response to this need have differed across time and place, many nations use an adversarial justice system based on Anglo-Saxon law to adjudicate violence against women and address victims' needs for protection and retribution. To preserve its monopoly on the right to prosecute and punish and to prevent vigilante justice, the justice system requires the support of the governed. When members of one group see or experience the system as biased or ineffective, they may withdraw their support of that system ( Konradi & Burger, 2000 ). This article is a critical analysis of the application of adversarial justice to crimes of violence against women and includes an assessment of how well it punishes and deters such violence and at what cost to the survivor. Then communitarian justice, a promising new model to address intimate violence, is described and evaluated. Appropriate justice response to battering and rape is critical, because these crimes are most women's only exposure to criminal victimization. In all other categories of crime, men are victimized more often than are women.

Traditional retributive justice aims for punishment of the guilty, whereas restorative justice is a new paradigm seeking victim-oriented criminal justice ( Dignan & Cavadino, 1996 ). Justice is adversarial when two sides of the case square off to uncover truth and affix responsibility by examining evidence through the questioning of witnesses. Procedural justice refers to the fairness of the adjudication process and substantive justice to the appropriateness of the penalty ( Konradi & Burger, 2000 ). Because somewhat different justice processes apply to rape and physical assault, each is reviewed separately.

In 1997, the American Bar Association adopted a resolution recognizing no-drop prosecutions as being helpful in sparing victims the sole responsibility for prosecuting their batterer. However, these policies create the spectacle of victim testimony only on subpoena, victim nonappearance, and victim undermining of the prosecution (for a review, see Hanna, 1996 ). For example, in Milwaukee, 95% of those arrested for domestic violence were prosecuted, but only 1 % were convicted ( Zorza, 1995 .) Women have many reasons for deciding that they do not wish to prosecute their batterer, not the least of which is their vulnerability to continued threats of violence to themselves, their children, their relatives, and their property. Among batterers arrested on victim complaints, 27% rebattered prior to trial ( Ford & Regoli, 1998 ). Nevertheless, women persevered even in the face of mental illness and child care responsibilities ( Goodman, Bennett, & Dutton, 1999 ). The major causes of what is called victim reluctance or noncooperation were lack of tangible support and substance abuse. One third of U.S. jurisdictions report uncooperative victims in more than half the cases of domestic violence ( Rebovich, 1998 ). Actually, it is more often workers in the criminal justice systems that are reluctant to proceed with prosecution, not victims ( Erez & Belknap, 1998 ). However, the most critical evaluation of no-drop policies hinges on their outcomes. In the six months following case settlement, many men rebattered, including 40% of offenders in cases initiated by on-scene arrest and 29% of offenders accused in victim-initiated complaints ( Ford & Regoli, 1998 ).

None of the alternative treatments for battering, which include counseling diversion, adjudicated guilt with counseling, probation, or traditional sentencing, have been shown to have a unique preventative effect on the prevalence, severity, or frequency of battering ( Ford & Regoli,1998 ; McCord, 1992 ). Comparisons of treated and untreated offenders showed that that those offenders who were not treated were significantly less likely to continue physical aggression ( Harrell, 1998 ). Further, contrary to prevailing beliefs among those who emphasize retributive punishment, the highest rebattering rate (44%) was among men who received noncounseling sentencing ( Ford & Regoli, 1998 ).

Women seem to understand that the credibility of rape survivors is examined more closely than that of other crime victims. Just 16% of rapes are reported to the police, according to the Rape in America study ( Kilpatrick, Edmunds, & Seymour, 1992 ). Official justice department estimates put the reporting rate at 36% ( Bureau of Justice Statistics, U.S. Department of Justice, 1997 ). Surveys of U.S. rape crisis centers revealed that in 17 states, adult rape complainants had been required to take a polygraph exam before their charges would be accepted, and in 11 states, child victims were polygraphed ( Sloan, 1995 ). Many survivors faced with these unsupportive early warning signs withdrew their charges, which were then documented by police as false rape allegations or recantations ( Kanin, 1994 ). Although police training has improved, case processing is still influenced by officers' private stereotypes ( Campbell & Johnson, 1997 ).

Among reported rapes, approximately half or more, depending on jurisdiction, are rejected for charging by prosecutors ( Frazier & Haney, 1996 ; Frohmann, 1991 , 1997 , 1998 ). Many studies have shown that rape attributions are affected by the race, age, and occupation of the perpetrator and the victim; their relationship; the severity of the violence; and the victim's risk-taking behavior, drug use, reputation, or moral character (for meta-analysis, see Whately, 1996 ). Although prosecutors may personally reject the appropriateness of these grounds, they feel themselves positioned downstream of jurors, so they nevertheless incorporate these factors into decision making. The process of charging an offender confronts a survivor with a prosecutor whose job is to look for holes that could make her story unconvincing. A victim's report may be discredited if her behavior conflicts with prosecutors' "knowledge" about the characteristics of rape and the behavior of victims ( Frohmann, 1991 , 1997 , 1998 ). Prosecutors may also use victims' living circumstances, relationship with the suspect, and behavior to construct a hypothetical scenario in which the forced sex was consensual or to impute motives for false allegations, such as to cover up infidelity, pregnancy, or sexual disease. If victims reside in racially mixed, lower-class neighborhoods, prosecutors often conclude that misinterpretation of them by a White middle class jury will lead to a not-guilty verdict ( Frohmann, 1997 ). Both women of color and victims of acquaintance rape were less likely to have their cases pursued by members of the criminal justice system ( Campbell, 1998 ; Razack, 1998 ).If allowed to proceed, survivors of color must contend with tension between their need for justice and felt obligations to buffer racism in the criminal justice system: African American men constitute 35% of the rapists currently incarcerated and 48% of the total prisoners in the United States ( Greenfield et al., 1999 ).

Of 1,198 reported rapes in Philadelphia, only 158 (7.5%) led to a verdict of guilty, a finding that is explained both by the high attrition of rape cases from the system prior to trial and by the outcomes of trials ( McCahill, Meyer, & Fischman, 1979 ; also see Frazier & Haney, 1996 ; Horney & Spohn, 1991 ). Only 25% of consent-defense rape cases were reported to result in convictions ( Weninger, 1978 ). In Washington, DC, only 9% of defendants who were ex-spouses, boyfriends, or cohabiting partners of the victim were convicted ( Williams, 1981 ). There are significant differences in the courts' response to intimate and nonintimate crime ( Ferraro & Boychuk, 1992 ). The net jury leniency for simple rape (i.e., acts among acquaintances where force was limited to that necessary to complete intercourse against consent) is tied for the highest of any crime and is much higher than any other crime against one's person of equivalent severity. In contrast, the net jury leniency for aggravated rape is near the bottom of the list ( Bryden & Lengnick, 1997 ).

Because an attorney retained by the survivor pursues civil justice, the survivor is more in control of decision making and is better informed of case progress, but the process is nevertheless adversarial. How successful is civil justice in compensating rape victims? The U.S. Department of Justice estimated that the out-of-pocket expenses of rape are $7.5 billion per year ( Miller et al., 1996 ). When pain, suffering, and lost quality of life were monetized, they reached $127 billion. Few of these costs were ever compensated. Surveys of victims of sexual violence who pursued a civil remedy revealed that they wanted more than money. They were seeking to be heard and searching for validation of the wrong that they suffered ( Des Rosiers, Feldthusen, & Hankivsky, 1998 ).

Despite some advantages over criminal processing, tort procedures introduce additional antivictim biases unique to civil proceedings and have antitherapeutic effects ( Bublick, 1999 ). Comparative fault is a doctrine that partially blames a victim for rape. Most states do not allow the rapist to use comparative fault defenses, but third party tortfeasors have unlimited access to the defense of "victim fault." An example is the case of Morris v Yogi Bear's Jellystone Park Camp Resort (1989) involving the gang rape of a 13 year-old by three 17-year-old youths with whom she was drinking on camp property. The three defendants were found 78% responsible, the park 10%, and the victim 12%. Another rape victim was held 30% negligent for being in streets that "were dangerous for a young lady at 3:00 AM" (as quoted in Bublick, 1999, p. 1460 ). A woman raped by a man with whom she had gone to a bar was found 51 % responsible for being raped.

Among the actions that have been viewed as unreasonable and to which comparative fault was assigned were going outside alone at night to hail a cab, walking to a car in a hotel parking lot, taking five steps inside a door before closing it, and failing to double check door or window locks (cases documented in Bublick, 1999 ). In contrast, many states do not allow the failure to wear a seatbelt to be considered comparative fault in torts involving vehicular injury. Comparative fault doctrine perpetrates victim blame and shifts blame from perpetrator to victim. "If a woman does not consent to sex, the fact that she did consent to a drink or the like should not make the rapist's act 'her fault"' ( Bublick, 1999, p. 1476 ).

Women whose rapes and assaults are adjudicated learn that even successful convictions or civil damage awards exact a psychic price. Partner violence victims whose cases are prosecuted face stressors not present in other crimes, including being forced to testify about intimate details of their relationships, fear of losing their children and vice versa, and fear of having to raise the children alone without child support ( Goodman et al., 1999 ). Rape survivors may be dismayed that their identity is a matter of public record, that they are expected to testify about graphic details of sexual assault in open court, and that even rape shield laws fail to protect them from questions about their social and sexual history as they pertain to consent issues. Victims may be further shocked when a plea bargain to which they were not party to and do not agree with ends their recourse for justice, precluding any face-to-face encounter with the perpetrator.

Inherent features of adversarial justice within the courtroom also shape survivors' experiences, such as the environment of formality, the sequestering of witnesses who may also be the family and supporters of the victim, attorney questioning that exacerbates self-blame, and the perpetrator's unmovable stance that he is not guilty of a crime ( Holmstrom & Burgess, 1975 , 1978 ; Madigan & Gamble, 1989 ; Martin & Powell, 1994 ; Matoesian, 1993 ; McCahill et al., 1979 ; Sanday, 1996 ).Testifying is one of four significant predictors of post-traumatic stress disorder symptoms among adult survivors of child rape ( Epstein, Saunders, & Kilpatrick, 1997 ). Data from 990 criminal trials reveal most victims believed that rapists had more rights, the system was unfair, victims' rights were not protected, and they weren't given enough information about or control over the handling their case ( Frazier & Haney, 1996 ). For some survivors, the trial did fulfill their desire for retribution ( McCahill et al., 1979 ). However, this conclusion is flawed, because investigators failed to examine differences by relationship to perpetrator. Those assaulted by a stranger, as opposed to those assaulted by an acquaintance, would be predicted to find the courtroom experience more constructive on the basis of stark differences in jury leniency and conviction rates between these two types of crime.

Increased occurrence of nightmares, decreased participation in social activities, more dissatisfaction with heterosexual relationships, loss of appetite, recurrence of phobias, and greater psychological distress have been documented among victims whose cases went to trial ( Holmstrom & Burgess, 1975 ). Insensitive attempts to obtain the testimony of Bosnian rape survivors resulted in severe traumatization of the victims, leading to feelings of shame, lack of trust, fear of reliving bad memories, fear of reprisals, and suicide attempts ( Allen, 1996 ).

Critics assert that the adjudication of intimate crimes against women by retributive justice represents state "theft of interpersonal conflicts" ( Christie, 1977 ), robbing many women of an important ritual encounter with the perpetrator and contributing to the segmentation of industrial society by undermining citizen participation and regular public discussion of the norms of conduct. The role of the victim within this justice system has been described as "evidentiary fodder for the defense attorney" ( Braithwaite & Daly, 1998 p. 154 ). Furthermore, the introduction of personnel such as police psychiatrists and psychologists into the presentencing investigation to evaluate perpetrators has shifted attention away from evaluating the damage done to the victim or establishing the defendant's obligation to repair it. In response to criticism, significant reforms of retributive justice response to victims have been implemented.

More than 20 years of legal reform have focused on removing spousal exclusions, legal requirements in rape cases that victims resist, and corroboration requirements that intimate crimes must have witnesses and on adding partial shields against revealing victim's sexual and social history. Evaluations of these initiatives in terms of changes in rates of reporting, charging, prosecuting, and convicting have revealed limited effects ( Horney & Spohn, 1991 ; Matosian, 1993 ). Likewise, legislation to educate or alter the behavior of criminal justice and judicial personnel generally has proved ineffective in terms of affecting case outcomes. For example, indigenous women in Canada were denied justice to a greater degree after judicial education than they were before ( Razack, 1998 ). Further, victim-witness advocates, whose role within the criminal justice system is to support victims, may align themselves with the prosecutor instead. Even at its best, victim support throughout adversarial processes has a flavor of hand-holding on a walk through hell ( Weisz, 1999 ).

A final procedural reform has been that of establishing victim participation statutes in all states of the United States, Canada, New Zealand, and Australia, to give survivors the right to communicate with their case's judge in writing and to speak at or attend the sentencing hearing ( Dignan & Cavadino, 1996 ; Konradi & Burger, 2000 ). However, victim participation statutes can do nothing to alter decisions or processes that have gone beforehand, including traumatizing investigatory, charging, and trial procedures, and apply only to the fraction of reported crimes that survive the attrition process. Furthermore, to exercise their rights, survivors must adopt a victim status, which works against overcoming the helplessness induced by crime victimization. At best, victim participation statutes have been labeled a "placebo" (Walsh, as quoted in Konradi & Burger, 2000, p. 387 ). At worst, there is concern over their potential to increase sentence severity at the expense of minority men, creating "an instrumentally empty opportunity [for women] to raise their voices at the end of a process so that the White men who control the criminal justice system can use individual women's testimony to further their classed, raced, and gendered agenda" ( Snider, 1998, p. 359 ).

Crime compensation schemes are a vehicle for victims to receive money to pay for material losses or medical treatment. However, these plans have been viewed as welfare schemes, because crime victims must come begging for help, the eligibility criteria are restricted, and they are vulnerable to political expediency ( Dignan & Cavadino, 1996 ). Specifically, most plans apply only to victims whose incidents involved people other than family or friends and were reported to police. Because domestic violence is defined as assault among those who are married, separated, or cohabiting, and because rape is primarily an acquaintance phenomenon, these restrictions effectively eliminate from compensation most violence against women. These schemes also fail when evaluated from the perspective of restoring control to those who have been made helpless, because they do not involve victims in the decision-making process.

Other substantive reforms have focused on intensifying the retributive punishment for crimes against women by increasing sentence length, creating new categories of aggravated crimes such as drug-assisted sexual assaults, and introducing public shaming as a punishment. For example, Judge Thomas Brownfield ordered a resident who had been convicted of aggravated battery to post a sign in front of his house reading, "Warning: A Violent Felon Lives Here" (as quoted in Massaro, 1997, p. 690 ). The convicted man's wife divorced him, saying she could not live with the humiliation. Others have recommended that college date rapists be ordered to wear bright orange armbands broadcasting their offenses and listing all affiliations, including sports teams and fraternities, as well as to make public apologies in classrooms ( Baker, 1999 ). Massaro's ( 1997 ) carefully reasoned critique of retributive shaming concluded that its effects are unpredictable, and it could make some perpetrators more violent. Moreover, shaming expresses no genuine concern for the condemned, seeks humiliation, lacks a dues-paid endpoint from which the perpetrator is reintegrated into society, departs from social norms of privacy, and erodes standards of decency and respect for dignity.

Restorative justice focuses on involving victims in the justice process and addressing the harm they sustained. It includes (a) civil proceedings, (b) victim-offender reparation through mediation, and (c) communitarian approaches. Experts consider communitarian justice the most promising of the restorative justice models to challenge the dominant retributive model ( Dignan & Cavadino, 1996 ). By giving up the overriding concern with incarceration, communitarian justice appears capable of producing a better balance between parties. In contrast, civil justice suffers from the problems with victim blaming and secondary traumatization documented earlier. Further, mediation presents a different but equally serious set of concerns when applied to rape and battering. Its conceptual foundation fails to acknowledge the structural inequalities between parties (for a procedural critique, see Brown, 1994 ; for a contrasting view, see Ellis, 2000 ). Women may find themselves negotiating for their safety instead of for their autonomy. Mediation also frames the issues as a conflict between victim and perpetrator, leaving unaddressed the harm violence against women inflicts on their loved ones as well as on the broader society. Many advocates view programs that focus on apology and forgiveness as cheap justice, because sincere offender apologies can be part of the ongoing cycle of violence ( Walker, 1992 ). Finally, mediation lacks a vehicle to reintegrate perpetrator and victim into the community.

Communitarian approaches have many names, reflecting the compromises that have been made so that communitarian justice can coexist with the powerful institutions of retributive justice. These include informal justice, peacemaking, positive justice, relational justice, family group conferencing, effective cautioning, and community accountability conferencing. These approaches spring from the recognition that Anglo-Saxon legalism has stamped out less adversarial methods for restoring social equity or harmony that cfiaracterized societies such as those of the Celts, Maori, Aboriginal Australians, Inuit, Native Hawaiians, Navajo and other American Indian tribes, and Asian ethnic groups ( Braithwaite & Daly, 1998 ; Coker, 1999 ; Nader,1990 ). Other ideological roots are philosophical arguments for a less interventionist criminal justice system ( Braithwaite, 1989 ). Communitarian methods have been applied to juvenile justice and drunk driving worldwide, to juvenile justice in some U.S. jurisdictions, and to domestic violence within North American Indian tribes. The basic principles are simple. "In the wake of an offense, and where guilt is admitted, victims, offenders, and their supporters are given an opportunity to meet in the presence of a coordinator or facilitator . . . encouraged to discuss the direct or indirect effects of the incident on them . . . [and to] negotiate plans for repairing the damage and minimizing further harm" (Moore, as quoted in Stubbs, 1997, p. 110 ; also see Umbreit, 2000 ).

The material reviewed earlier depicted existing responses to physical and sexual assault as a multipronged system of processes that have the effect of drastically reducing the number of women for whom promises of justice were fulfilled. Even though perpetrators of interpersonal violence are very sensitive to its consequences ( Lore & Schultz, 1993 ), the instances where sanctions are applied are too few, too inconsistent, and too separate from the initiating inci- dents to prove potent as preventative measures. The result of these documented shortcomings in justice response is that U.S. society lacks a credible deterrent to crimes against women. These data constitute a strong rationale for seeking new avenues of societal response. Communitarian justice models for domestic violence are designed to operate alongside of, not replace, retributive justice. They represent an additional avenue requiring no rollback of existing legal gains or safety instruments that advocates have fought to establish. The communitarian justice conference is based on a foundational value. For example, Navajo peacemaking is based on traditional teachings about the role of women, incorporating stories such as "Changing Woman" (to read this story, see Coker, 1999 ). Using these values would be inappropriate outside their specific context ( Goldberg, 1997 ). In nonsecular application, the underlying value is that violence against women is unacceptable and unlawful. This value simply applies existing legal codes on sexual crimes, stalking, and domestic assaults that define behaviors that are wrong in that they are prohibited by law. References to the wrongfulness of behavior may be made throughout the conference to counter victim blaming or excuse making by any of the parties. Admission of guilt is required to select conferencing as the method for dealing with the crime, but the process carries no threat of incarceration. Participation is voluntary by both victim and perpetrator. Opting not to participate leaves existing criminal and civil options in place (for a more detailed procedural description, see Umbreit, 2000 ). Because of the threat to safety posed by men who have abused women, it is recommended that compliance with and completion of the commitments a perpetrator has made be monitored by a justice official such as a probation officer, just as such officials currently document the fulfillment of mandatory community service and court-ordered treatment. Should the offender fail to cooperate with the conference, lack follow-through on his commitments, or reoffend, the option remains to activate his arrest warrant and process the case through retributive justice procedures.

Communitarian justice capitalizes on empirical demonstrations that interventions such as mandatory arrest work best on offenders who are socially connected ( Schmidt & Sherman, 1998 ). It builds on the sanctions abusive men said they fear most. Only a minority of batterers feared criminal punishment or job loss (36% and 27%, respectively). Instead, they believed that the major cost of a domestic violence arrest would be self-stigma, family stigma, and broad social disapproval ( Willians & Hawkins, 1989 ). Communitarian methods also respond to data indicating that survivors want to retain choice and be treated as autonomous individuals ( Erez & Belknap, 1998 ).

Appropriate cases for processing by conference include situations where offenders do not have extensive records of past involvement with law enforcement, particularly the youngest offenders facing their first documented incident, and where rape offenders have targeted acquaintances (which is the case in 8 out of 10 rapes). Communitarian methods have been recommended for colleges and universities, because federal law requires them to draft sexual assault policies and conduct disciplinary hearings (for the Higher Education Amendments of 1992, see Baker, 1999 ). In military settings, conferences would formalize communication channels that are already being used. Both in transient environments - such as the military, colleges, and universities-and in larger settings, communitarian methods offer a process to create or augment community ties where they may otherwise have been lost through relocation, transience, or isolation.

Advocates have expressed reservations that any implementation of communitarian justice must address and monitor ( Stubbs, 1997 ). Among them are that the offender may abuse the system. He might use the conference as yet another occasion to abuse the victim. Or the conference could provide him with an expanded forum for the denial and trivialization of his violence. Estranged perpetrators could potentially initiate a conference as a strategy to locate victims who have gone into hiding. Without appropriate security, conferencing raises safety issues, as women could be vulnerable to revictimization in their vehicle or in the parking lot where the meeting is held.

A second area of concern is that the victim could be ill-served. Women who have been physically or sexually abused may not be well equipped to speak in their own best interests or those of their children, and the communication resources could be unequal. Some women have already formed an intention to leave their relationship, but advocates fear the communitarian approach is antidivorce. (However, note that existing responses have been criticized as prodivorce, casting women who stay in abusive relationships as failures.) There is also concern that too much weight can be put on apologies and that victims might be coerced into participation when they would prefer to take their chances with retributive justice.

A third area of concern addresses potential flaws in the process. The communitarian model assumes the community can be relied on to confront the offender, is committed to chastising him for his actions, and has something to contribute to a solution. Critics fear that families will support their offending sons and that the community will simply reinforce and support traditional patriarchal power, misogyny, and victim blaming ( Coker, 1999 ; Dobash & Dobash as quoted in Stubbs, 1997 ).Some women would reject communitarian processes because their family and friends are unsupportive. Others may fear disclosure of personal information.

A final group of concerns is about application. A communitarian model that operated alongside the retributive justice system could create a discriminatory situation if those in the most favored groups were sent to conferencing when the less favored received standard adjudication.

Coker ( 1999 ) has reviewed communitarian approaches specifically in application to partner violence. She emphasized that abusive men do not act alone. They are aided by family, peers, and larger cultural forces and take advantage of victim's vulnerabilities, including deficits in material resources and social power. She identified multiple potential advantages of communitarian approaches over retributive justice for crimes against women. Communitarian approaches address systemic and personal aspects of male violence by searching for underlying problems and transferring material resources so that future vulnerability will be lessened. If a battered women has been controlled by her lack of access to family income, transportation, communication, or child care, an individual bank account can be ordered and other members of the community identified to provide services, thus undermining the offender's power and decreasing the victim's vulnerability. The conference can also attend to the psychological abuse and control that frequently accompany violence but fall outside narrow crime definitions that recognize and directly address physical and sexual assault over other forms of maltreatment. Communitarian methods counter the helplessness induced by victimization. A conference offers the victim a ritual face-to-face contact with her perpetrator under conditions attuned to protecting her safety, providing space for her story to be heard, and involving her as an active participant. No crime victim should be forced to confront her perpetrator, but neither should she be denied the opportunity if she desires it. The support provided the victim in a conference will be of a better quality than in criminal or civil justice forums, because the process is nonadversarial and advocates have a more active role in speaking out on behalf of women's rights to be free from violence and to counter myths and violence-supportive attitudes. Unlike the criminal justice system, the medical care system, and some battered women shelters, where women who stay in violent relationships have been pathologized, communitarian approaches avoid cultural and legal focus on separation and can be adapted to accommodate either women who wish use the conference to safely end the relationship or women who want to work toward a less violent existence with the abusive partner ( Peled, Eisikovitz, Enosh, & Winstok, 2000 ).

The conference creates a setting where peers and family of the perpetrator, who may share attitudes and model behavior that constitute a violence-supportive environment, are exposed to an antiviolence message, thereby creating a new avenue for primary prevention targeted at a high-risk group. Involving advocates or counselors in the process and expanding their ranks to include men who are explicitly trained to respond to the perpetrator ensures that antiviolence messages will be reinforced. There will never be enough police to guard every vulnerable woman. Conferencing offers a methodology to multiply social control resources using the natural community, while also directly aiming to eliminate societal supports for gender-linked abuse. Unlike with retributive justice, where the perpetrator often maintains his innocence even when incarcerated, the communitarian model diminishes the perpetrator's need to deny or minimize his crime or shift responsibility to the victim. The victim has the opportunity to ask for an apology if she desires one and is more likely to receive it. The procedures allow a forum for the offender to air issues of racism, adverse childhood experiences, substance abuse, racial oppression, and economic disadvantage without fram- ing these issues as exculpatory. It invites the members of community to express their solidarity while also repudiating violence against women. Coupled with its nonincarceration focus, and being a process that is shaped by the participation of other members of the victim's and offender's cultural groups, communitarian justice has the potential to mitigate the racism that is perceived to permeate criminal justice. The communitarian justice process attends to the importance of reintegrating both victims and offenders into the community.

Under the status quo, victims often suffer silent shame, afraid to disclose their abuse, and those perpetrators who are incarcerated accrue a lifetime stigma with the negative consequences that labeling theorists have described. Communitarian methods themselves may promote reintegration, but it is also consistent with their roots and philosophy if culturally specific rituals aimed at healing and restoring harmony follow them.

The community conference model has been studied in its application to juvenile justice and drunk driving in Australia and New Zealand (i.e., Prenzler & Hayes-Hennessey, 1999 ; Moore & O'Connell, 1993 ) and to partner violence through Navajo peacemaking ( Coker, 1999 ). Coker's study was based on review of files, conference observations, and interviews with participants in Peacemaking Court in Window Rock, Navajo Nation (Arizona), and Shiprock, Navajo Nation (New Mexico). Coker ( 1999 ) concluded that peacemaking "offers possibilities for women that are largely unavailable in other intervention strategies" (p. 107). Contrary to fears, she did not find that men used the conference to locate their victims. Further, there was no evidence that women were intimidated into reconciliation. The system was used by divorcing couples. Some of the caveats presented earlier about the appropriate application of this methodology were reaffirmed, however. Specifically, Coker stressed that all existing remedies must remain in place, including protection orders, stay-away provisions, child custody rulings, visitation supervision, and child support. She concluded that the conference agreement for those staying together should explicitly stipulate (a) prohibitions against violence, stalking, harassment, and continuance of abuse-supportive associations; (b) arrangements to share housework and child-care responsibilities; and (c) agreements to seek batterer, sex offender, or alcohol treatment. Finally, safeguards must be in place to protect participants' physical safety before, during, and after the conference.

Currently, a large amount of money is spent on justice responses to violence against women that are not rooted in historical approaches to problem solving among many of the groups that constitute U.S. society, ill-serve victims, and are not very effective in reducing the prevalence of violence by perpetrators. Many of the same personnel who are currently paid to implement these strategies could be assigned communitarian justice tasks with better results. Antiviolence advocates are well aware of the shortcomings of the current retributive justice system, and many say privately that they would not use it if they were personally battered or raped by someone they knew. Yet, commentators have expressed surprise that victim advocates have not lobbied harder for new approaches to crimes against women. Instead, they have poured energy into advocating for incremental reforms in retributive justice processing that, if the past 20 years are any indication; will fail to substantially influence either the processes or the outcomes of criminal and civil justice. We, as practitioners of psychological science and practice, can no longer passively support justice responses that the tools of our profession have revealed to be psychologically damaging and ineffective. Further, we cannot expect the law to compete with norms that encourage and condone violence against women ( Baker, 1999 ). We must disseminate information on alternatives like communitarian justice that are better grounded in psychological theory, advocate for demonstration projects using new methodologies such as community conferences, and sustain primary prevention initiatives aimed at decreasing the social policy and cultural supports for violence against women. Without such reforms, women will continue to live with fear, and we will remain in a society where victims blame themselves too much and perpetrators too little.