|Violence against Women (World Bank, 1994, 84 pages)|
Violence against women is an extremely complex phenomenon, deeply rooted in gender-based power relations, sexuality, self-identity, and social institutions. Any strategy to eliminate gender violence must therefore confront the underlying cultural beliefs and social structures that perpetuate it. To be effective, such a strategy would have to draw on a wide range of expertise and resources, both governmental and non-governmental.
Our understanding of the exact causes of gender violence still needs refining. But the results of several recent cross-cultural studies on family violence and rape reinforce the feminist contention that hierarchical gender relations-perpetuated through gender socialization and socioeconomic inequalities-play an integral role in violence against women (box 4). Using complex statistics and coded ethnographic data from 90 societies throughout the world, Levinson (1989) identified four factors that, taken together, are strong predictors of the prevalence of violence against women in a society. These factors are economic inequality between men and women, a pattern of using physical violence to resolve conflict, male authority and control of decision making in the home, and divorce restrictions for women. The study suggests that economic inequality for women is the strongest factor, reinforced by male control in the family and a woman's inability to divorce.
These findings reinforce the feminist view that violence against women is not an inherent part of "maleness" but a function of socially constructed norms of acceptable behavior. As Cheryl Bernard, Director of Austria's Ludwig Boltzmann Institute of Politics, notes:
Box 4 Correlates of gender violence In cross-cultural studies
Predictive of high violence
1. Violent interpersonal conflict resolutions
Predictive of low violence
1. Female power outside the home
a Sanday 1981.
b. Counta, Brawn, and Campbell 1992.
c. Levinson 1989.
Violence against women in the family takes place because the perpetrators feel, and their environment encourages them to feel, that this is an acceptable exercise of male prerogative, a legitimate and appropriate way to relieve their own tension in conditions of stress, to sanction female behavior...or just to enjoy a feeling of supremacy. (Bernard 1986 p. 26)
Indeed, in many societies women are defined as inferior and the right to dominate them is considered an essential aspect of being male. A strategy to prevent violence must therefore begin by dismantling these cultural beliefs and deconstructing notions of masculinity that promote aggressive sexual behavior and domination of women. To the extent that male sexual behavior is "predatory" in certain cultures, it is not because male "sexuality" is aggressive, but because sexuality is used to express power relations based on yonder. Thus any effort to eradicate violence must also address the underlying power dimensions.
Clearly, any systematic effort to root out violence must be multidimensional, drawing on the expertise and resources of many sectors, both governmental and non-governmental. Although the response of the health sector is clearly important, a strategy that seeks to go beyond treating "the symptoms" of abuse must focus on eliminating the attitudes and beliefs that legitimize violence and justify male control of female behavior. And it must improve women's access to power and resources so as to give them realistic alternatives to staying in abusive relationships. As Joanne Leslie, codirector of the Los Angeles-based Pacific Institute for Women's Health, observes: "To the extent that more education, higher incomes, occupations outside of the home, access to credit etc. empower women and enhance their self esteem, these may prove much more effective in reducing the morbidity and mortality associated with domestic violence, than more direct health sector intervention". (1992, p. 26).
A strategy to prevent violence must also promote nonviolent means to resolve conflict (between all members of society-men, boys, family members). Passing laws to criminalize violence within family relationships-in the same way that societies criminalize violence between strangers-is an important way to redefine the frontiers of acceptable behavior. Using violence to resolve conflicts is a learned behavior- children are exposed to violence by their parents' behavior in their homes and through television, film and videos. Levinson's study suggests that violence against women is particularly prevalent in societies in which the use of force to resolve interpersonal conflicts is condoned.
This policy section recommends actions involving a broad range of actors. Appendix A outlines a comprehensive program designed to confront and eliminate violence against women, which includes both long-term actions to prevent future violence and shorter-term responses to victims' needs. The reader may wish to review the appendix before reading further The section that follows explores some of the interventions and approaches that have been tried in different parts of the world, to give a serve of the innovative programming that is beginning to evolve.
Unlike for many health and development issues, the most important step that can be taken to combat violence is fairly clear: support the nascent initiatives already under way. Some of these are at the governmental level (see box 5 for examples), but most represent the untiring efforts of autonomous women's organizations that have pushed this issue forward despite local and national resistance. To be effective, work to combat violence must be site-specific, emerging from the cultural and political realities of each country. A wealth of well organized NGOs are already working throughout the developing world on many of the programs outlined below. A recent directory published by ISIS International (1990) lists 379 separate organizations working on gender violence issues in Latin America alone. These groups, which function with little outside support, could easily be strengthened with a minimal investment of resources.
Justice system reform
The gross inadequacy of most laws in protecting victims or sanctioning violent perpetrators has made legal reform an important priority for many groups working on violence against women. Clearly, amending laws on paper is not enough to ensure change, but strong laws can be a considerable asset in helping women protect themselves from violence. Three critical tasks in legal reform are changing laws that keep women trapped in abusive relationships, removing barriers to prosecution, and eliminating aspects of the law that are prejudicial to women.
A number of laws have worked to trap women in relationships. Article 114 of Guatemala's Civil Code, for example, grants a woman's husband the right to prohibit her from working outside the home; among other things, this drastically limits a woman's ability to gain the financial independence needed to escape an abusive relationship (Garcia 1992). In Ecuador, until a 1989 legal reform, a husband had the right to force his wife to live with him no matter how abusive he may have been ( Fonce, Palan, and Jacome 1992). And in Chile divorce is illegal for any reason, even in cases of extreme violence (Valdez 1992). Such laws put women living in violent relationships at substantial risk.
Laws in other countries make it almost impossible to prosecute violence against women, especially violence perpetrated by an intimate partner. In Pakistan, for example, four male Moslim witnesses must testify before a man can be convicted and subjected to the hadd punishment (the most severe) for rape (Human Rights Watch 1992). It is extremely difficult in Pakistan to get any conviction of rape, even for the lesser tazir penalty (public flogging, rigorous imprisonment, or fines), because the Law of Evidence considers women "incompetent" as witnesses in cases of rape and grants their testimony only the status of corroborative evidence. In 1979 Pakistan passed the hudood ordinance, which made all forms of sex outside marriage-including fornication and adultery-crimes against the state. Women who have failed to meet Pakistan's high standard of proof for rape have themselves been thrown in trial for adultery or fornication based on their admission of intercourse. Human rights activists estimate that up to 1,500 Pakistani women are in prison awaiting trial for hudood violations (Human Rights Watch 1992).
Rape laws that are prejudicial against women are not uncommon. The definition of rape is extremely narrow in most countries, and the law and judicial systems often treat rape as a crime against public morality, family honor, or-as in African customary law-property, rather than as a crime against the woman. In cases of sexual violence the justice system is almost universally biased against women who are not virgins. In some Latin American countries-for example, Brazil, Costa Rica, Ecuador, and Guatemala-the law defines certain sexual offenses as crimes only if they are committed against "honest"-that is, virginal-women or girls. Laws in Chile and Guatemala specifically exonerate a man who agrees to marry the girl he has raped; his marriage to the victim is perceived as restoring her honor and that of her family (Garcia 1992; Valdez 1992).
Box 5 Government initiatives against gender-based violence
In February 1991 the government of Canada announced a new four-year Family Violence Initiative, a "call to action" intended to mobilize community action, strengthen Canada's legal framework, establish services on Indian reserves and in Inuit communities, develop resources to help victims and stop offenders, and provide housing for abused women and their children (Government of Canada 1991).
In 1991 Chile's Congress created El Servicio Nacional de la Mujer (SERNAM) to advance the rights and opportunities of Chilean women. SERNAM has proposed a program to prevent family violence by promoting legal reform to criminalize domestic violence, documenting the dimensions of the problem, organizing community awareness campaigns to change public consciousness, and opening crisis centers to provide legal and psychological support (Servicio Nacional de la Mujer 1991).
Brazil's new constitution, enacted in 1988, contains the following provision: The states should assist the family, in the person of each of its members, and should create mechanisms so as to impede violence in the sphere of its relationships (Americas Watch 1991). And Colombia's 1989 constitution states that "any form of violence within the family is considered destructive to its harmony and unity and will be sanctioned by law" (ISIS International 1993).
In March 1990 the prime minister of Australia established the Commonwealth/State National Committee on Violence Against Women (NCVAW) with a three year budget of$1.35 million The NCVAW initiates research, coordinates the development of policy, programs, legislation, and law enforcement on a national level, and conducts and coordinates community education on violence against women. Among the committee's founding principles: "women have a fundamental right to be safe," and "men muse be held completely responsible for their violence. (NCYAW 1991).
In 1992 Bolivia's National Council for Solidarity and Social Development opened an Office of Battered Women, which runs a halfway house where women can seek refuge and receive medical treatment and counseling. "We are absolutely new at dealing with this, " said Maria Luisa Palacios, national director of the Social Welfare Agency. "We're looking for other women's groups to help us understand how to handle this problem." (Nash 1992).
Ecuador's National Development Plan for 1988-92 included, for the first tune, an entire chapter on improving the situation of women. And in 1°,91 the government body in charge of women's affairs (DINAMU) added a new line of action-consciousness-raising and action against violence against women. DINAMU has opened a legal services office and a battered women's shelter, but the effectiveness of these state services for women has so far been limited (Ponce, Palan, and Jacome 1992).
In addition, in the vast majority of countries the law does not recognize marital rape or domestic violence. Although most legal systems have laws against assault, these provisions are often difficult to use to convict an intimate partner. In Latin America the definition of assault often requires a finding of injury sufficient to incapacitate the victim for a se' number of days (frequently a week or more). Normally, such findings must be made by an official forensic doctor, who evaluates whether the incident meets the legal definition of injury, or lesion. But because incapacitation is usually framed in terms of inability to work, and because work is seen as employment outside the home, a woman's injury may not be interpreted as a lesion, regardless of the physical or mental health consequences it may have for her.
Similar distinctions hold in India, where assaults that do not cause "grievous harm" are "noncognizable" offenses that is, the police can take no action without first seeking a warrant from a magistrate. "Grievous harm. includes only certain types of permanent injures, such as emasculation, loss of sight or hearing, or permanent disfigurement the face (Articles 319-26 of the Indian Penal Code; Agnes 1988).
In the legal systems of many countries the burden of proof and the penalties for violence against women are biased against intimate assaults. Article 276 of the Bolivian Penal Code, for example, states that lesions caused by a husband are punishable only if they incapacitate a woman for more than 30 days (Rosenberg 1992). In Peru the stiffest sentence possible for wife abuse is 30 days of community service. Lawyers who represent battered women say that, in practice, even this minimal penalty is rarely enforced (Kirk 1993).
The forensic medicine system in Latin America and parts of Asia complicates the prospects of rape convictions. In most countries only government-employed forensic doctors are authorized to collect evidence on rape admissible in court. These physicians are generally located only in large cities, and their offices are closed in the evenings and on weekends. (In Lima, Peru, where there are only five forensic doctors, a rural woman raped on a Thursday night might have to travel all day by bus and then wait two days before she could be examined.) Moreover, forensic doctors seldom have specialized training in rape or domestic violence, and they routinely omit from their reports information crucial to establishing the commission of a crime (Kirk 1993).
In recent years some countries have made significant strides toward improving written laws relating to violence against women. In July 1991 Mexico revised its rape law in several important ways. It redefined rape as "a crime against a person's freedom (rather than against morality), expanded the definition of rape to include anal oral, and vaginal penetration, increased the sentence to rapists, eliminated a provision allowing a man who rape' or to avoid prosecution if he agrees to marry- and required judges to hand down a decision regarding access to an abortion within five working days (Women's World 1991-92).
Steps toward reform have also been taken in the Philippines. In 1993 a coalition of 14 women's groups developed and got introduced in Congress a progressive rape law that incorporates an expanded definition of rape, redefines rape as a crime against the person (and not her chastity), makes marital rape illegal, recognizes "physical or verbal resistance in any degree...as prima facie evidence of lack of consent,. disallows information on the sexual history of the women to prejudice her claim, and establishes a woman's right to a closed-door court hearing (Women's Legal Bureau 1992).
A growing number of governments, including some in the developing world (the Bahamas, Barbados, Belize, Malaysia, Puerto Rico), have passed laws or reformed their penal codes to criminalize domestic violence. And a substantial number of countries-including Argentina,
Bolivia, Brazil, Chile, Colombia, Ecuador, Peru, and Venezuela-have bills under consideration (ISIS International 1993). Often such laws criminalize psychological as well as physical violence and provide for orders of protection- legally binding court orders that prohibit one person from abusing another. Under the most progressive statutes judges can require a man to leave his home, establish temporary custody and visitation arrangements for children, make the husband pay financial support, forbid telephone threats and harassment, and order the battered to attend counseling. Many women prefer protection orders to criminal persecution because they do not want their abuser jailed or wish to avoid the trauma and expense of a trial.
The effectiveness of protection orders depends largely on how well they are enforced. Too often, protection orders become meaningless pieces of paper because police and judges refuse to impose the penalties for noncompliance. Where orders are enforced rigorously, however, they can offer a substantial subset of women considerable protection, and make it possible for them and their children to shy at home. Of course, for some men-those who are especially violent, jealous, and obsessive-the orders are essentially useless; the only way to protect their partners is to incarcerate the abuser.
Elsewhere, governments have passed laws against particular types of violence common in their countries. The Indian government has passed a law against "Eve teasing, " the sexual and physical harassment of girls and women in public. Both India and Pakistan have passed laws against dowry harassment, and Bangladesh has outlawed acid throwing. None of these laws is widely enforced, however (Heise 1989). Colombia's congress is considering a law that would make secuestro (the confinement or isolation) of a wife by a husband a crime. The law was proposed in re spouse to the growing trend among Colombian men to lock up their wives to prevent infidelity (Maridos Secuestradores 1992).
As with protection orders, such laws are only as good as their enforcement, and it is in implementation that the legal response to violence most notably fails. Nonetheless, important initiatives have been taken in recent years to improve the response of the justice system to gender-based violence (box 6). Perhaps the best known has been the creation of women-only police stations, an innovation that has spread from Brazil to Colombia, Uruguay, Peru, Costa Rica, and Argentina." Data from Brazil's special police stations show that women only units have greatly facilitated the reporting of abuse. In Sao Paulo, for example, reported rape cases went from 67 in 1985, before the women's police stations were opened, to 841 in 1990. Sao Paulo has 96 of the country's 125 women's police stations; these registered 79,000 of the national total of 205,000 crimes against women reported between July 1991 and August 1992, suggesting that the number of reported cases would be much higher women's police stations were widely available in other states (Dimenstein 1992).
Although the women's police stations are an important innovation, they have also had problems. Many stations have been overrun with women seeking assistance that the stations do not provide: counseling, legal advice, and help with state bureaucracies. The original plans to assign social workers and lawyers to each station have not materialized. The female police officers assigned to these stations become easily demoralized because their male peers do not consider their job "real police work."
Moreover, observers have learned that, without training, female officers are not necessarily more sensitive to women's needs than their male colleagues (although women have proven more open and responsive to training). Finally, the special stations have not increased the rate of prosecution because at higher levels the justice system remains unchanged. Nonetheless, because the stations encourage women to come forward, they have helped deter violence among men who worry about being reported to the police.
In industrial countries the most recent innovation in combating violence against women has been "coordinated community intervention." This strategy brings together the policymakers concerned-from battered women's groups, law enforcement agencies, the justice system, batterer treatment programs, and other relevant groups-in regular meetings to develop a coordinated response to domestic violence. Once policies are developed, lower level representatives are assigned to meet regularly to oversee their implementation. Roughly 75 to 100 communities in the United States have adopted this model. The strategy includes several key elements. There are written policies about how each agency should respond and agreements on coordination and sharing data. A paid coordinator manages the task force and oversees the processing of cases. Victims' advocates are trained to help battered women negotiate the court system and other social service agencies, and training is also given to all relevant staff on the dynamics of abuse and related policies. Local shelters or safe homes are provided, as are batterer treatment programs. And there is active monitoring-preferably by an autonomous women's group-to ensure that each agency carries out its policy and coordinates properly with other actors. Similar community intervention strategies are being implemented in Canadian cities (Heise and Chapman 1992).
Box 6 Innovative justice system reform projects
In the United States feminist lawyers organized the Judicial Education Program to Promote Equality for Women and Men in the Courts (Heise and Chapman 1992). This program has succeeded in getting more than half the states to form "gender-bias" task forces to detect and attack sexism in the courts. The task forces, made up of judges and community representatives, have uncovered devastating testimony by victims of abuse about their mistreatment in the courts and have prompted the recall of some judges and increased training for judges and prosecutors. The Asia Pacific Forum on Women, Law, and Development is undertaking a similar project that is analyzing the laws in seven Asian countries (Fernando 1993).
In Harare, Zimbabwe, the Musassa Project works with local police and prosecutors to sensitize them to issues of domestic violence and rape. Commenting on their work, the organizers observe that " the specifics of an educational strategy aimed to justice system professionals must be very carefully devised. In many cases, credibility must be ensured by involving a legal professional in the education process, and the content must be highly dependable and informed. Another effective technique is to facilitate a workshop with one part of the legal system acting as host to another (police hosting prosecutors, for example)" (Stewart 1992).
In Costa Rica, El Instituto Legal de los Naciones Unidas y Desarollo (ILANUD) offers gender sensitivity training, emphasizing violence against women, to prosecutors, judges, lawyers, and other professionals. In 1992 the project conducted 32 workshops throughout Latin America (Facto 1993).
In Malaysia five organizations joined forces at the end of 1984 to form the Joint Action Group Against Violence Against Women (JAG). JAG organized a major media campaign against rape, initiated dialogue with the police and the medical profession, and successfully lobbied for the creation of women-only rape teams on the police force. The Health Ministry agreed in 1987 to set up one-stop crisis centers in all hospitals, staffed by medical personnel and trained volunteers from local women's organizations. The centers have not yet been established, however (APDC 1989).
Health care system reform
The health care system is well placed to identify and refer victims of violence. It is the only public institution likely to interact with all women at some point in their lives-as they seek contraception, give birth, or seek care for their children. Experience has shown that this access is important.
Even in countries with a strong movement against violence, many abused women never choose to call the police or a crisis hot line, the two most widely developed sources for referral. Advocates in Connecticut, for example, estimate that only 10 percent of battered women living in that state ever come in contact with its extensive network of legal advocates, shelters, and crisis centers. This may be in large part because the system relies primarily on the police and crisis hot lines to inform victims about the services available (Heise and Chapman 1992). In politically repressive countries, the likelihood of the police serving as an adequate referral system is oven leas realistic."
Women who are unable or unwilling to seek help from the police or other government authorities may nonetheless admit abuse when questioned gently and in private by a supportive health care provider. Providers have found that, contrary to their expectations, women are willing to admit abuse when questioned directly and non-judgmentally. For example, when Planned Parenthood of Houston and Southeast Texas added four abuse assessment questions to its standard intake form, 8.2 percent of women identified themselves as physically abused. When a provider asked the same questions in person, 29 percent of women reported abuse (Bullock and others 1989). Researchers have found that three to four simple questions are generally enough to screen for physical and sexual abuse (see box 7 for examples). Questions should be asked in person and in private, and the questioner should make sure that the potential abuser is not present to avoid putting the woman at additional risk.
Some who have implemented programs to screen for abuse at prenatal care clinics and emergency rooms note that asking itself can be an important intervention "It is my impression that some women have been waiting their whole lives for someone to ask." notes Dr. Ana Flavia d'Oliveira (1993), a Brazilian public health physician who initiated an abuse screening program among her prenatal care patients. Providers can emphasize to a woman that no one deserves to be beaten or raped, and help her think through options for protecting herself (for example, seeking safety at a friend's house). In urban areas providers can refer women to a growing number of services for legal or psychological support (see the section below on assisting victims). Even where no external support exists, having a sympathetic individual acknowledge and denounce the violence in a woman's life offers relief from isolation and self-blame.
Providers can also help a woman better assess the degree of danger that her abuser poses to her and her children. In the United States, for example, researchers have developed a list of warning signs that indicate that a woman is at substantially increased risk of homicide or serious injury. The list includes the following (see appendix D for a complete danger assessment):
1. Has the violence escalated over the past year?
2. Is there a gun in the house?
3. Has he threatened to kill you?
4. Has your partner forced you into sex when you did not want it?
5. Is your partner violent outside the home?
6. Is your partner drunk every day or almost every day?
Box 7 Abuse assessment screen
1. Have you ever been emotionally or physically abused by your
partner or someone important to you?
Statistical analysis shows that women who answer yes to a cluster of these questions are at higher risk of being fatally wounded by their partner. By administering this simple danger assessment checklist, providers can help women evaluate their situation and take precautionary measures. Although a checklist along precisely these lines may not be appropriate in settings outside the United States, the analysis required to derive such indicators is straightforward and a list appropriate to any area could easily be developed.
Despite the potentially critical role of health care professionals, evidence indicates that few providers identify and respond appropriately to victims of abuse (Warshaw 1989). In evaluating 481 medical records of women seeking aid for injuries at a major U.S. urban emergency room, physicians identified only 2.8 percent as battered. But closer examination of the records showed that 16 percent of the women had injuries considered probably or highly suggestive of abuse and almost 10 percent could be positively identified as battered. Another 15 percent had trauma histories suggesting battering (Stark and others 1981). This means that emergency room staff identified only about one in eight battered women who passed through their service.
In Alexandra Township, South Africa, a rapidly urbanizing community near the heart of Johannesburg, a similar retrospective study reviewed the charts of 398 women presenting with a history of assault to the Casualty Department of Alexandra Health Clinic during October and November 1991. The survey found that providers failed to record the identity of the perpetrator in 78 percent of cases; charts included such agent-less descriptions as "chopped with an axe" or "stabbed with a knife." Charts providing more complete information include the following cases:
· A 35-year-old woman was kicked and stabbed with a screwdriver by her boyfriend. On physical examination, she had suprapubic tenderness, laceration of the labia minora, and swelling on her forehead.
· A 32-year-old woman was assaulted and raped by her husband. When she reported the incident to the local traditional healer, he also raped her.
· A 32-year-old woman who was being assaulted by her husband left the house to seek help from her relatives. On her way, she was raped by three men.
· A 15-year-old, married and six months pregnant, was raped by a neighbor. Her husband was away, working on a contract job in one of the homelands.
Studies show that with proper training and protocols, however, health care facilities can greatly improve their staff s sensitivity to gender-based abuse (see appendix D for information on protocols). After the emergency department of the Medical College of Pennsylvania introduced training and protocols, the share of female trauma patients found to be battered increased more than fivefold, from 5.6 percent to 30 percent (McCleer and Anwar 1989). Similarly, prenatal care providers who received training through a program funded by the March of Dimes significantly increased their rate of screening for abuse. After six months 75 percent of program directors reported that they had implemented abuse screening, up from zero before training (Helton, McFarlane, and Anderson 1987).
Despite their usefulness, protocols and training arc still rare in
the United States. Only 20 percent of emergency departments in Masschusetts-one
of the better organized states-had a written protocol for domestic violence in
1991 (Isaac and Sanchez 1992). That same year, however, two important
initiatives were launched that should improve the health sector's involvement in
the issue of violence. First, the American Medical Association (AMA) initiated a
major campaign to educate physicians and the public about family violence and
devoted an entire issue of its prestigious Journal of the American Medical
Association to the theme.
Second, the U.S. Joint Commission on Hospital Accreditation included emergency mom protocols and training on family violence among the criteria used to evaluate hospitals for accreditation (Heise and Chapman 1992). This policy change should encourage more active screening and referral of abuse victims.
A new project sponsored by the Family Violence Prevention Fund, a private, nonprofit group in San Francisco, and the Pennsylvania Coalition Against Domestic Violence seeks to help institutionalize the new hospital accreditation standards by developing model protocols, training programs, and dissemination strategies that can tee applied throughout the country (Family Violence Prevention Fund 1993).
Training providers is essential not only to increase referral rates, but to ensure that victims are not revictimized by the health care system. Victims of rape and domestic assault frequently report being humiliated and degraded by the very providers who are supposed to help them (Kirk 1993). Providers who fail to collect and record evidence properly in rape and assault cases can jeopardize any legal cases that a victim might bring. And providers ignorant of violence and its sequelae can exacerbate the consequences for women by labeling them hypochondriacs or by treating them for nonexistent mental illnesses. Research from the United S tales shows that emergency room doctors are more likely to prescribe tranquilizers and pain medication to battered women than to trauma victims who are not battered (Stark, Flitcraft, and Frazier 1979). By deadening the pun "d clouding judgment, tranquilizers can prolong the battering relationship and make it more difficult far women to assess their options or take action to protect themselves. It is widely acknowledged that Valium and other tranquilizers are over-prescribed by the medical profession in the developing world as well (Busto 1991).
The issue of violence can and should be incorporated into the training of community health workers as well as professional staff. Project workers report that such issues as domestic violence and men's alcoholism arise spontaneously during health promoter meetings, especially in all - female groups. Increasingly, NGO-sponsored projects are incorporating themes on gender violence and women's status into training materials for health promoters. The Women's Program of Uraco Pueblo in Honduras, for example, includes socio-dramas, discussions, and role playing on domestic violence and sexual harassment in its health promoter training; promoters regularly hold community meetings on domestic violence, inviting lawyers to offer women legal advice and holding joint meetings with husbands and other men from the village (Maher, personal communication, 1993). And female health workers in the SARTHI project in Gujurat, India, consider offering individual and community support to victims of violence an integral part of their job. Health promoters have accompanied women to the police station to register complaints and worked with family members to marshal support for women's decisions to take action against an abusive husband. Project organizer Renu Khanna notes that the women themselves defined violence as a priority; SARTHI merely supported their leadership on the issue (Khanna, personal communication, 1992).
On the international front tentative progress has also been made toward recognizing violence as an obstacle to women's health and development. In 1991 the Pan American Health Organization (PAHO) sponsored a conference in Managua enticed "Violence against Women: A Problem of Public Health. (OPS 1°,92). Colombia's Ministry of Health issued an action agenda for women's health which included a program on "Prevention of Abuse and Attention to Victims of Violence.. And the United Nations Fund for Women (UNIFEM) published "Battered Dreams: Violence against Women as an Obstacle to Development" (Carrillo 1992). But the World Health Organization has no program or policy related to genderbased violence. In fact, in planning World Health Day 1993-whose theme was Injury and Violence Prevention- officials included no mention of violence against women until women's health advocates persuaded them to do so.
Although violence is in theory largely preventable, few preventive programs have been undertaken on a wide scale. Among those programs that do exist, many focus on helping adolescents and young children learn nonviolent ways to resolve conflict (box 8). Some concentrate on developing self-esteem and the ability to express emotions in constructive, nonviolent ways. Others work to challenge the gender stereotypes and notions of male prerogative perpetuated in the media and in the culture at large. Programs in some schools encourage children to disclose to an adult unwanted touching by strangers or family members. And on college campuses consciousness-raising programs are being developed to combat acquaintance rape.
Much of the public education and media work by the women's movement can also be loosely classed as prevention work, although more effort has been directed at reaching potential victims than at changing men's attitudes. Women's groups have held hundreds of workshops and produced thousands of pamphlets, comic books, and other consciousness-raising materials to give women basic information about their rights. Although few of these materials have been evaluated, they have clearly been useful in initiating dialogue on this often taboo subject (Zurutuza 1993). Several developing countries, among them Ecuador and Peru, have sponsored national media campaigns to sensitize the public to issues concerning rape and domestic violence (Zurutuza 1993). The Family Violence Prevention Fund in San Francisco is trying to take media-based prevention a step further by doing sophisticated market research to craft messages aimed at changing public attitudes toward violence. This effort represents the first time that the media techniques successfully used to change drinking and smoking behavior in the United States will be applied to domestic violence.
In recent years justice system intervention (arrest, prosecution) has been advanced in the United States and Canada as a tool to prevent future violence among already violent men. Until the late 1970s the traditional response by police to domestic calls (when they responded at all) was to walk the abuser around the block, using arrest only as a last resort This changed dramatically in the United States during the 1980s as police departments-responding to pressure from advocates for battered women, fear of liability suits, and a flurry of new mandatory arrest laws-began to arrest offenders for intimate assaults. Also prompting the shift in social policy were research results from Minneapolis, published in 1984, suggesting that, compared with separating couples or advising them to get help, arrest cut in half the risk of future assaults over a six-month follow-up period (Sherman and Berk 1984). These findings were widely publicized by advocates seeking to criminalize wife assault and thus end the double standard of policing for private and public violence.
But recent studies have called into question the results of the Minneapolis experiment (Schmidt and Sherman 1993). Of five studies, only two (Colorado Springs and Miami) found even weak support for the greater efficacy of arrest compared with other police interventions. Detailed analysis reveals that the effect of arrest varies with characteristics of the perpetrator. When the perpetrator is married or employed, or both, arrest reduces recidivism, but for unemployed and unattached perpetrators, arrest actually increased abuse in some cities. Some have interpreted employment and marriage as measures of "social embeddedness," arguing that arrest deters men who have more to lose (Sherman and Smith 1992). But it is equally plausible that employment is a surrogate for other factors not measured, such as education, self-esteem, and socioeconomic standing. Teasing out the exact nature and causes of the differential effect will require further analysis and research. .
Box 8 Initiatives to prevent gender-based violence
In Kingston, Jamaica, three groups use popular theater for prevention education on Bender violence. The artistic collective System uses interactive workshops and street theater to prompt discussions on issues of domestic violence arid rape. The Women's Media Watch protests violence and objectionable portrayals of women in the media and uses theater work with young people to help them grapple with complex questions relating to sexuality and sexual violence. Teens in Action, a community group formed after the brutal rape and murder of a young girl, performs drama to encourage critical reflection in their neighborhood on issues of sexuality, male-female relationships, and rape (Popular Education Research Group 1992).
In Ontario, Canada, the Ministry of Education's Violence Prevention Initiative schools children in the three "R's" plus one: relationships. The program includes a school-based family violence prevention curriculum tested in the schools in 1991-92, a Handbook for the Prevention of Family Violence developed with the input of more than 60 professionals, and a family violence training program for school personnel. The project has published and distributed thousands of pamphlets and storybooks, sponsored theater groups, organized parent-teacher days, and worked with schools, the police, medical staff, shelter workers, and social service agencies to make them more aware of the issue (Etue 1991).
In Brooklyn, New York, the Anti-Violence Education Project uses self-defense training as an entree for discussing violence prevention with children in the public schools. The project holds weekly sessions to teach children self-defense and nonviolent ways to resolve conflict. It draws analogies between relationship strategies and the philosophy of karate; it teaches that the martial arts do not condone violence, but instead that the true master is the one who can use the least force to achieve his or her ends. It also teaches children to look critically at how the media misrepresent the martial arts through the depiction of such pseudo-heros as Bruce Lee (Ellman 1993).
Education Wife Assault (EWA) in Toronto, Ontario, works with immigrant and refugee women to help them develop culturally appropriate violence prevention campaigns for their community. The EWA holds "skill shops" that give women leaders the skills they need to develop their own culturally specific programs against domestic violence. It then provides technical support to the women carrying out the campaigns. The EWA's staff also lend emotional support to women organizers to help them overcome the isolation and the backlash often directed at women working against domestic violence because of perceptions that they are threatening community and cultural cohesiveness (Center for Women's Global Leadership 1992).
This is not to say that arrest serves no useful purpose in domestic violence cases, but only that it may not, on average, reduce recidivism more effectively than other possible police interventions. In fact, advocates' original objective in promoting arrest was not to deter future violence, but to interrupt current abuse and to ensure women's equal protection under the law (Stark 1993). The questionable ability of arrest to deter intimate assaults makes it no less effective than for other crimes: the literature on juvenile delinquency and on general criminology offers little empirical or theoretical support for arrest deterring juvenile or adult offenders (Gellees 1993). It may also be that arrest alone is not enough. A study by Steinman (1989) found that arrest in isolation from other criminal justice sanctions produced greater subsequent violence. But v/hen part of an integrated justice system response that included prosecution and court-mandated treatment, arrests offered significant protection from further abuse.
Elsewhere in the world women have begun to organize to prevent genital mutilation and other traditional practices harmful to women and girls. In 1984, at a World Health Organization-sponsored conference in Dakar, Senegal, 22 national committees working to eradicate genital mutilation joined to form the Inter-African Committee on Traditional Practices Affecting the Health of Women and Children. In the words of Berhanc Ras Work, the committee's president, "female circumcision is a clear example of social violence which women have to bear in silence as a price for marriage and social identity. (Heise 1993, p. 180).
In 1990, at a conference in Addis Ababa, Ethiopia, on Traditional Practices Affecting the Health of Women and Children: How Far Forward?," African delegates voted to support laws to ban female genital mutilation and punish those who carry out the practice. The first United Nations Human Rights Seminar on Traditional Practices- held in Ouagadougou, Burkina Faso, in May 1991- recommended using legislation, education. and such means as street theater and community organizing to end harmful practices (Dorkenoo and Scilla 1992). African women living in Europe, Canada, and the United States have also begun to organize to combat the practice among immigrant and refugee populations (Hedley and Dorkenoo 1992).
Programs to assist victims
In the industrial world rape crisis centers and battered women's shelters have been the cornerstone of programs to assist abuse victims. A typical rape crisis center supports a 24-hour hot line, community education programs, staff or volunteers who accompany raped women to the hospital and police station, training for health and justice professionals, and ongoing counseling or support groups for victims. Shelters provide emergency housing for women and children for up to several months. They also provide hot lines, support groups for residents, community education programs, basic child care, and women's advocates who help residents negotiate the legal system and social service bureaucracies. Some better funded shelters provide legal assistance and job training, and employ staff who carry out training and institutional reform efforts to reduce violence against women (Heise and Chapman 1992).
Since the early 1980s shelters and women's crisis centers have sprung up in many developing countries too. At least 40 developing countries now have at least a handful of NGOs that assist victims of violence. Some countries have both specialized rape crisis centers and separate services for battered women; others have one or more all-purpose crisis centers to which female victims of any kind of gender violence can go for services. Because of the expense of maintaining shelters, many developing country NGOs have decided not to open any, and instead to concentrate their limited resources on self help support groups, legal services, counseling, and institutional reform. Some women's organizations have also rejected the shelter model because they see it as reinforcing dependency rather than promoting networking and problem solving (Carcedo, personal communication, 1993).
A few governments, especially at the municipal level, have also begun to provide services for battered women and rape victims. In 1989 the Mexican government, for example, created specialized agencies under the District Attorney's office for the reporting of sexual violence. These agencies aim to provide legal, medical, and psychological care to survivors in a sensitive and confidential atmosphere. New legislation has been introduced to provide long-term support for the five specialized agencies in Mexico City and for agencies in all 31 state capitals (Women's World 199192). Likewise, Peru's Ministry of Health is seeking low cost ways to shelter female victims of violence. The Director of Mental Health, Mariano Querol, recently visited the United States to investigate models for assisting victims in sparsely populated rural areas (Querol, personal communication, 1992).
In both the industrial and the developing world, support groups play an integral role in healing and consciousness raising. As Lucrecia Oller, coordinator of the Violence Prevention Program at Lugar de Mujer in Buenos Aires, Argentina, observes:
The group becomes a space where the abused woman can begin to visualize her recuperation. She has ongoing communication with women who have been working at this for a long time, and who share what they've gone through with her. Under these circumstances, the first tangible tool she receives from the group is hope; she sees that others have been able to free themselves, and begins to think that perhaps she can too. (Popular Education Research Group 1992, p. 11)
Support groups help abused women begin to live fuller, more productive lives. They play a critical role even for women unable or unwilling to leave an abusive partner, by helping them to stop blaming themselves, by reducing their isolation, and by providing a springboard for social activism. The role of social support in reducing; stress reactions has also been well documented (van der Kolk 1988; Rachley 1990). As Lucrecia Oller observes, "to feel heard and understood is to begin to feel human. (Popular Education Research Group 1992, p. 11).
Support can also help women escape future violence. A recent evaluation by the Cash Rican women's CEFEMINA found that 60 percent of the abused women who attended its support groups were able to achieve a violence-free life within six months-most by leaving the relationship. Requiring less than $20,000 a year to reach more than 400 women, or roughly $50 per woman served, the CEFEMINA program is highly cost-effective (Carcedo, personal communication, 1993).
Support groups are also used to assist victims of childhood and adolescent sexual abuse. As part of its community prevention project, for exemple, the Costa Rican group Ser y Crecer trains professionals in three communities-teachers, therapists, social workers-to run self-help support groups for victims of sexual abuse. The project provides three months of training and guidelines on conducting a 16-session support group. Ser y Crecer also teaches local leaders from these same professional communities to identify and refer victims of abuse, and trains other professionals as "prevention experts." to conduct prevention education in their communities. A nurse, for example, might be trained to give talks to parents waiting in a doctor's office (Babes, personal communication, 1993).
Elsewhere in the world, communities are developing their own ways to sanction perpetrators and assist abuse victims, often by increasing the "social cost. of violent behavior through public confrontation and humiliation. Among the Garifuna, an Afro-Indian population inhabiting the Caribbean coast of Belize, for example, women surround the house of a man who is beating his wife, calling out publicly to shame him. In cases of severe abuse, they help the woman to escape by providing her sanctuary in neighborhood homes until the conflict is resolved or she can permanently relocate (Kerns 1992). In Mira de las Flores, a shantytown in Lima, Peru, women have organized themselves into a neighborhood watch committee and they wear whistles to summon other women if attacked (Heise 1989). Levinson, in his cross-cultural study cited earlier, observes that "the presence of kin or neighbors who will intervene in violent or potentially violent situations is a characteristic of societies with low rates of wife beating" (1988, p. 452)
In Canada aboriginal communities are addressing domestic violence by drawing on traditional models of conflict resolution. In Manitoba women from the Hollow Water Reserve have begun to use the tribe's "circle of healing" to address domestic violence. The community members confront the abuser, requiring him to acknowledge his crime publicly, and offer their support for healing both the victim and the perpetrator. The abuser is given a "healing contract" setting out the punishment- usually community work-and arrangements are made to protect the victim. When the contract expires, a cleansing ceremony takes place to symbolize the return of balance to the abuser, the family, and the community. At this point healing is considered complete and the crime is to be forgotten. Healing can take years (MATCH International 1990).
Treatment programs for perpetrators
There has been increasing interest in recent years in rehabilitating offenders through treatment and reeducation. Treatment programs have proliferated in industrial countries as courts have sought a social solution to wife abuse that does not lead to further crowding of jails. Many women, too, have Savored their partner getting help-rather than punishment-as a first step. In the United States most court mandated programs meet weekly for 12 weeks; others range from as few as six sessions to more than eight months (Heise and Chapman 1992). A few treatment programs patterned after the U.S. and Canadian models have sprung up in developing countries as well (for example, in Mexico).
Most treatment programs are based on the idea that violent behavior is a learned and socially reinforced behavior. Counseling styles vary from structured classes that educate men on the power dynamics of battering to free-flowing self-help groups run by experienced facilitators. Some groups concentrate on improving men's ability to handle emotions; others go further, challenging men's perceived right to control women through violence. Many programs include training in relaxation, education on male socialization, and such behavioral techniques as "time out" to interrupt escalation of anger (Stordeur and Stille 1989).
Only recently have researchers attempted to evaluate the effectiveness of programs to treat batterers. Treatment appears to reduce the physical violence of some men, but other variables, such as interaction with the justice system, probably also play a role (Harrel 1991). A recent review by Tolman and Bennett (1990) reports that 53 to 85 percent of men who complete treatment are not violent six to 18 months later, with lower rates for longer follow-up periods and for studies based on victim reports of violence. But these same studies show that even "nonviolent" men continue to use verbal threats and psychological abuse to control their wives. As Edleson and Grusznski speculate, "many men who end their violence may resort to the use of threats as a 'legal' but hardly less terrorizing form of control" (1988, p. 20). Moreover, the success rates in such studies apply only to those who complete the programs. Because more than half of the men participating in such programs drop out before completing their treatment, it remains unclear how successful programs are across a wide range of men (Saunders and Azar 1989).
Perhaps the most definitive evaluation to date found that court-ordered treatment had no positive effect on rates of physical violence or threats of violence (as measured by self-reports, victim reports. and police records), or on key beliefs about wife beating and personal control of violence (Harrel 1991). Unlike other studies, this study compared men ordered by the court to obtain treatment with a group of men not ordered to attend treatment-a true control group. In addition, it compared men's rate of violence before court intervention with their rate of violence after court involvement but before treatment, thus allowing researchers to evaluate whether the measured were due solely to treatment or due in part to the justice system's involvement. The study found that rates of violence were lower after arrest and court hearing, but that treatment itself added no additional benefit. These findings suggest that most of the positive benefits attributed to treatment may actually come from the justice system's intervention rather than from treatment.
It would be premature, however, to conclude that treatment holds no promise for reducing future domestic violence. Treatment and evaluation methods are still in their infancy, and new treatment approaches may yet prove effective. And investigators note that failure to impose harsh penalties for abuse or for failure to attend treatment may undermine the efficacy of treatment. Harrel's (1991) results indicated that none of the men in the treatment or the non-treatment groups believed that they would suffer significant legal or social consequences from committing violence in the future.
In addition to batterer treatment, there is a long tradition in North America and Europe of attempting to treat sex offenders. Techniques have ranged from psychosurgery and pharmacologic interventions (for example, treatment with medrosyprogesterone acetate, or MPA, an androgen antagonist) to cognitive and behavioral therapy. ID the United States roughly three quarters of all sex offender treatment programs are community-based, although 40 states were treating adult sex offenders in prison in 1990 (down from 48 in the mid 1980s; Sapp and Vaughn 1991). In recent years increasing emphasis has been placed on identifying and treating adolescent sex offenders to stop aggressive sexual behavior before it escalates. Research on 411 adult sex offenders treated as outpatients shows that 58 percent began their deviant sexual behavior during early adolescence and that the average adolescent male sex offender has 380 victims during his lifetime (Abel and others 1985, as cited in Stops and Mays 1991). Definitive evaluations of the outcomes of treatment for adult offenders, although scarce, show that at least some programs appear to reduce recidivism (based on rear restrates) among pedophiles and exhibitionists, but none has proved effective with hardcore rapists (Marshall and others 1990).
Treatment programs and evaluation methods for sex offenders-like those for batterers-are still in their infancy, so the possibility that they can provide effective intervention should not be dismissed. Nonetheless, addressing perpetrators one by one after their patterns of abusive behavior have been forged (and reinforced by social norms) is not a particularly promising approach to addressing violence against women. Although treatment may help prevent the future abuse of one or more women, true prevention requires creating a generation of individuals who see violent behavior as inappropriate.