Women's attempts to exit violent relationships are encouraged by the increased availability of protective orders; the growing social stigmatization of domestic violence; the more frequent issuance and enforcement of protection orders; pro-arrest and mandatory arrest policies; and the incarceration, treatment and incapacitation of batterers. However, exiting violent relationships can be extremely dangerous for battered women and their children. The Violence Against Women Act (VAWA) recognizes this danger and introduces further protections for women and children. In this brief exploratory essay we argue that battered women who flee across state lines are engaging in an acute form of separation that may elicit tremendous rage on the part of abusers. Batterers' rage at the extreme distancing may make it more likely that some will seek lethal retribution against women they may perceive as emphatically deserting them. Such perceptions make full faith and credit guarantees all the more relevant to the future lives of these women, and all the more valuable as a potential life saver.
In what follows, we note the connections between the risk of lethal violence and women's estrangement from their batterers. Using data from a recent study of lethal homicides in Florida, we identify a number of risk factors, correlates with lethal domestic violence, or what we call "red flags." We suggest these red flags might usefully be communicated to authorities in asylum states to warn them of potential retributive acts by enraged batterers. In cases where battered women cross state lines with protective orders in their possession, we already see three key red flags: prior history of violent victimization; attempting to flee a violent relationship; and, possession of a protection order. One way of doing this might be to attach some kind of cautionary information on the risk of lethal violence to the NCIC protective order file. Whatever mechanism is used, we assert the need to somehow link lethality assessments to full faith and credit guarantees, if only for the fact that leaving issuing state does not merely result in a jurisdictional change, but rather signals and entirely new twist in the dynamics of abusive relationships. Finally, we reflect upon the possible policy initiatives that might be pursued to effect a greater awareness of the potential risk of lethal violence that battered women face who move to asylum states and receive the benefits of full faith and credit provisions.
When battered women leave violent relationships their exposure to violence at the hands of their abuser does not necessarily cease. Numerous studies show that women experience an elevated risk of severe or lethal violence in the process of leaving batterers ( Allen, 1983 ; Barnard et al., 1982 ; Browne, 1987 . Wilson and Daly's (1993) analysis of intramarital homicide from summary data in Canada (1974-1990), New South Wales, Australia (1968-1986) and Chicago, USA (1965-1990), shows that wives in all three countries experienced a "substantially elevated" risk of lethal victimization when estranged from their husbands as opposed to co-residing with them. Wilson and Daly (1993) reject the argument that murdered wives leave husbands because of escalating violence, and that homicide may have been the outcome whether women stayed or remained. Rather they point out that batterers themselves often do exactly what they say they'll do: that is, they threaten lethal violence if their wives threaten to leave, and they carry out that violence in the event of the wives attempted departure (1993:10). It is also clear from the research of Wallace (1986) that the decision to leave a relationship usually precedes physical separation. Once the decision to leave is communicated to her partner, women also face an elevated risk of lethal violence while they cohabit with the husband prior to separation.
The qualitative studies of homicide-suicide utilize small sample sizes but tend to scrutinize the case details very closely, revealing useful insights into the relationship dynamics preceding the lethal episode. These studies clearly reveal that one of the key triggering events for homicide-suicide is the woman's decision to leave the relationship, or her act of departure, or both. Buteau et al. (1993) found a history of recent separation in 59 percent of the homicide suicides they studied. In her study of homicide-suicides between adult sexual intimates in Australia, Easteal identifies the "inability to see oneself as a separate entity from one's partner" in combination with the attendant "obsessive possessiveness," as a "principal theme" (1993:102).
The obsessive possessiveness of some abusers means they may do practically anything not to lose their partners. Wilson and Daly cite one case in which an Illinois man told his wife that he would kill her if she ever left him. He commented, "I swear if you ever leave me, I'll follow you to the ends of the earth and kill you" (People v. Wood, 391 N.E. ed 206)( 1992:89). Other authors point to the role of obsessiveness about one's partner as an important factor in propelling an abuser toward the use of lethal violence. Hart (1988) notes that if the partner has threatened homicide or suicide, and has fantasized about killing his partner, then obsessiveness could be a major predictor of lethality. She notes, "A batterer who is obsessive about his battered partner-who either idolizes her and feels he cannot live without her or believes he somehow owns her because she is his wife - is likely to act more dangerously to keep his partner (1988:242).
The extant literature on the events preceding domestic fatalities laid the groundwork for our study of domestic fatalities in Florida. We now outline that study and go on to discuss its implications for the implementation of full faith and credit under the VAWA.
The Florida Governor's Task Force on Domestic and Sexual Violence (1997) recently began a detailed study of domestic fatalities in order to understand, prevent and intervene in these crimes. We examined 317 domestic fatalities that occurred in 1994 with a view to uncovering themes/common characteristics of the cases. While the findings are of a preliminary nature and await statistical confirmation through the addition of further data, they do provide useful insights into what seem to be a number of important red flags. In our search for pragmatic "red flags" or "strong correlates" with domestic fatalities, we have broken down abusive behavior into discrete variables. It could be objected that many of the variables we report upon are inseparable and profoundly interwoven parts of an abusive relationship. As such, to distinguish between threats to kill, prior acts of violence, possessive beliefs, prior police calls to the residence, etc., is to ignore the overall complexity of the abusive relationship. Our response to this criticism is that it is important to breakdown the various sub-components of an abusive relationship so that we can determine which combinations of sub-components are the most likely to be associated with, and therefore predictive of, lethal domestic situations. The strength of a "red flag" type of analysis is not in the relative predictive power of each red flag, but in the way we can present combinations of red flags that seem to be the best predictors of a domestic fatality. It is in this spirit and with the aforementioned provisos, that we present, in order of priority, those red flags most frequently associated with the 1994 domestic fatalities we identified in Florida.
Of all the correlates with domestic fatalities, the strongest is a prior history of domestic violence in the perpetrator-victim relationship. This holds true as much for adult heterosexual couples, homosexual couples, and parent/caretaker-child fatalities in cases where children are killed. This finding also holds true to differing degrees for white, black and Latino subgroups. Taken by itself this red flag is not particularly helpful. Obviously there are millions of domestic assaults per year. Very few people who are assaulted, even multiple times, end up a victim in a domestic fatality. However, a prior history of domestic violence assumes much more predictive power when combined with other important correlates, even though, as we have already pointed out, the other correlates are less likely than "prior history" to be associated with domestic fatalities.
Historically men have regarded women as their property. Many men still treat women as their property. However, this belief about the social place of women vis-a-vis men does not mean that all men with those beliefs end up killing the women they perceive that they own. Nevertheless in the 1994 domestic fatalities, acutely possessive beliefs about women surface in a large number of cases. From our data, women are twice as likely to be killed if the perpetrator had possessive beliefs about her.
After history of domestic violence and possessive beliefs, "threats to kill" is the third most important correlate with domestic fatalities. We need to know more about these threats. In particular we need qualitative data on the nature of threats, whether the abuser has hatched a plan, whether he has killed before, etc. Regardless of the limited nature of the data on this variable, threats to kill are very significant correlates with domestic fatalities. Our case studies clearly demonstrate that perpetrators share their homicidal ideas with victims, immediate family, relatives, friends, work mates, and neighbors. We have to take these threats more seriously.
This potential red flag is similar to the possessiveness variable insofar as betrayal can take the form of the partner leaving for another relationship. However, betrayal can also mean other things such as departing from role prescriptions and for example, getting a job or going to school. In cases where the betrayal factor is evident, our research shows that men have pinned many of their (highly exaggerated, self-centered and often unrealistic) hopes and dreams upon a future with their wives, partners and/or family. Any threat to these projections is a threat against the perpetrator himself. One of the core elements in the thinking pattern of the batterer or abuser that makes what he sees as "betrayal" so threatening is his narcissism. Often insensitive to or totally oblivious to the impact of his controlling behavior upon his female partner's own autonomy, the more she flexes her muscles, the more dangerous he becomes. We must learn more about this particular red flag because it has profound implications for the manner in which we intervene at the level of social policy.
Women who are attempting major life changes as part of a strategy of breaking away are also much more likely to die. Again this does not mean that women should not attempt to break away. Rather it tells us what we already know: that is, breaking away is a very dangerous time for women and children. Consequently those agencies involved with battered women and their children must take extra precautionary steps to help women through this transition. This is likely to be an especially pressing issue for women who take the decision to make a substantial break by moving to another state. These women will be in particular need of full faith and credit guarantees, and be in particular need of sensitive police support.
One of the problems with this red flag is that police officers do not always document the nature of their involvement with prior domestic violence incidents. There were several cases among the 1994 cases in which women were involved with police departments immediately before being killed, or killing their abuser in self defense, where the department did not even log the nature of the complaint. We must ask ourselves whether we can do more to discern the distress of women at the critical junctures in their abusive relationships. Given that most women who are battered do not call the police, we must do more to recognize the acute distress women are in when they do call. We must recognize this distress whether or not a woman's affect corresponds with what helping agencies deem to be the appropriate affect for a distressed person. It seems that among those cases where the police were called to the residence, fewer contacts correlated with a higher likelihood of lethality. This does not mean that police departments should take their "chronic cases" less seriously.
The use of drugs and/or alcohol immediately prior to the fatal incident correlates with death. However the correlation is not a particularly strong one and we do not in any way suggest it is a causal link. Many people use drugs and alcohol and do not become violent. We need more research into this topic.
Our data are clear: among the women victims, those who had injunctions were twice as likely to be killed as those that did not. Of course this may reflect the fact that those women who have sought injunctions are far enough along that continuum of violence that they are closer to losing their lives. Findings like these do not mean that women should not seek injunctions. Rather they tell us that we have to do much more to assist women who chose to exercise these particular resistive options and take out injunctions.
It is also evident from the case studies that some very violent abusers had extensive criminal histories including crimes of violence. However, our criminal history data is incomplete. This area of the data gathering process needs much work. We need to know precisely how many victims and perpetrators have criminal histories. Then we need to develop some means of quantifying the degree of violent behavior evident in those histories. Clearly an abuser who has convictions for marijuana possession may not be as dangerous as the offender with aggravated assault, armed robbery and attempted homicide in his criminal history.
According to the FBI's Crime in the U.S. Report, an estimated 1,432 females were killed by intimates in 1992. In that year female victims represented 70 percent of the intimate partner homicide victims. These numbers, however, tell us nothing of the circumstances leading up to these tragedies. Homicide data shed little light on how many women, for example, were killed in the process of either terminating or fleeing from an abusive relationship. Nor does the narrative in a typical homicide report necessarily describe the history of abuse prior to the incident.
Indeed many of the large data sets (e.g. Mercy and Saltzman, 1989 ) on domestic fatalities draw upon supplemental homicide reports which are rather poor sources of data on prior histories of domestic violence.
Child fatality review or death review panels currently exist in most states and are comprised of representatives from multiple agencies who investigate child homicides. These death review panels tell us many things about how these children died and the lessons we can learn from these tragedies. Further, these assessments raise many questions about what we do not know and at the same time yield extremely useful information about how these deaths can be prevented. Drawing data from hospitals, health departments, schools, social service agencies, law enforcement, and other community organizations, child death review panels have served to help better understand the dynamics of child homicides and subsequently have helped agencies to intervene more effectively to prevent violence directed toward children. These teams have also improved interagency coordination in the protection of survivors and in the prosecution of perpetrators. Child death review teams are providing critical multi-agency input into identifying factors that precipitate lethal violence. These teams also engineer strategies to prevent these occurrences. In short, the merit of having child death review teams has been firmly established.
The existence of adult fatality review teams, however, remains woefully rare nationally. One exception is the state of California which authorized its counties to set up Domestic Violence Death Review Teams pursuant to California Penal Code ss.11163.3-11163.5 ( see Bowman, 1997 ). Several states and local jurisdictions have adopted either formal or informal adult fatality review panels or teams in the last several years ( see Stone, 1995 ). We firmly believe that the findings of child fatality review will similarly benefit and protect women, the elderly, and others who find themselves at risk from potentially lethal violence. Our own discussions with law enforcement officers, prosecutors, judges, social service workers, and advocates for battered women, indicates that many are still unaware of this innovation, but when fully explained, there is considerable interest in establishing adult fatality review teams. For example, we have had judges tell us they would love to have objective assessments of the potential for lethal violence in a given relationship. Such information would be extremely useful in the judicial handling of domestic cases and, for example, the nature and number of conditions attached to protective orders. The introduction of domestic fatality reviews and the attendant interagency liaison to produce preventive strategies, requires a major change in public policy. Such a change is reminiscent of the life-saving introduction of seat belts and the placement of children in the back seats of cars.
Most states and tribal lands have no centralized mechanism (either through the police, prosecutors or courts) for monitoring and handling domestic violence, domestic fatalities or both. Fatality review teams might be seen as a first step in the direction of establishing a centralized mechanism for monitoring domestic violence and domestic fatalities.
As indicated above, women fleeing abusive relationships are likely to be at increased risk of lethal violence. Indeed, these are the women that full faith and credit guarantees are designed to protect. But how do we assess the potential risk of lethal violence? This is a question that many agencies and social service providers presently cannot answer and we do not yet have measures that reliably identify or predict risks. However, fatality review and lethality assessments can begin to make this predictive information available to multiple agencies. Death review teams not only review domestic fatalities, they will also, whenever appropriate, come up with recommendations for systemic reform that, with the benefit of hindsight, may be instrumental in helping to prevent similar tragedies in the future.
Since these teams collect a wealth of data, we believe it is advantageous to construct a "classification system" or an "index of lethality" that quantifies the risk of lethal domestic violence. Such a system or index would have direct practical application and could be used as a guide for decision-making by those agencies intervening in potentially lethal domestic scenarios. For example, with the profile of a high risk abuser available, agencies could better strategize about arrest, prosecution, incarceration, diversion, etc. At the same time, agencies could begin to prioritize their services to victims of pre-lethal violence (safe independent housing and relocation services).
In chapter 2 of this collection, Sims and Zaorski discuss the critical role of technology in the effective implementation of full faith and credit. We envisage a system, in the not-too-distant future, where data from death review panels in Miami will be made available through a centralized statewide tracking system to multiple agencies and jurisdictions throughout the state of Florida. Consequently, a woman (and her children) fleeing to Jacksonville, Florida, from Miami will not only have her Miami protective order enforced in Jacksonville, but police and other relevant agencies could also have access to the lethality assessment attached to the order.
Criminal justice agencies may soon be able to obtain electronic fingerprint data on abusers as well as check criminal histories and Brady firearm prohibitions. Law enforcement will be able to decipher whether the abuser has violated a protection order or a history of violations.
Further, the NCIC protective order file has the potential to make available lethality risk assessments to criminal justice agencies checking the validity of out-of-state protective orders. Such a system could alert criminal justice professionals in asylum states and tribal lands to the potential risk of lethal violence simply by tapping into the NCIC protective order file. This would not only protect battered women and their children, but also enhance the enforcement of the law, including the full faith and credit provisions of the VAWA. We believe fatality review and its obvious relevance to the enforcement of full faith and credit guarantees will contribute toward preventing future domestic tragedies.
Public policy makers often view domestic violence homicides like powerful forces of nature; forces we can do nothing to stop. What this study from Florida shows is that these deaths are far from inevitable aspects of the human condition. Unlike earthquakes or hurricanes, we can intervene in and prevent these domestic fatalities.
Before any changes in the behavior of judges, law enforcement, victims advocates, prosecutors, health care professionals and others who deal with cases of abuse, these professional groups must receive comprehensive education about the signs and, routinely incorporate them into their daily practices. For instance, a physician should ask the victim/patient about acts of coercion, intimidation and violence utilized by her partner. It is important to ask if the perpetrator ever threatened to kill them, since perpetrators often have disclosed intent to kill to the victim and third parties before they make good on their deadly threat. That physician who identifies threats, obsessive possessive conduct and jealousy or suicidal fantasies by the perpetrator can then make active referrals to emergency shelter and strongly encourage the patient to talk to a trained counselor who could work on a safety plan with her.
A judge should/could ask a petitioner seeking an order of protection: "Have you changed anything about your life, like gone back to school or gotten a new job," as these steps toward independence are danger triggers for abusers. A "yes," for instance, should make that judge automatically suggest a ban of restrictions on, visitation and exchange of children. Early and effective use of this information can, then, limit the escalation of violence. Specifically, judges should ascertain whether the woman has told her abuser of her intent to leave. The judge must also probe for more information about his threats to kill her, his fears and phobias about her impending breakaway, his feeling of being betrayed by her, and any history of other red flags. With this information in hand, the judge can then place restrictions upon the abuser in accordance with the judge's informed assessment of risk of lethal violence.
Relevant to dangerousness assessment is a victim's movement across state or tribal lines. Therefore, in addition to specific behavioral questions and investigations, professionals must routinely ask if the victim has changed residences or recently come from another city. This line of questioning, in and of itself, can open the door to a full and frank discussion with that victim of safety risks and options for security. Relocation may be a "risk-marker" for stalking and escalated violence. Prosecutors for example, should note this possibility and assess risks accordingly. Red flags for lethal domestic violence should become integral parts of the training and education of all professionals who intervene, support and advocate for battered women.
Law enforcement domestic violence training, must incorporates the "red flags" for lethality. In particular, breaking away from a relationship and moving to another state, tribal land, or territory, must be seen as a risk factor for lethal violence. Indeed, agencies must include "red flag" items in standardized forms and criminal investigation materials. Particular training should be designed for airport and transit police and persons who live close to state lines.
Education for health care professionals, attorneys, judges or any other groups of people who come in contact with domestic violence cases should likewise utilize "red flags" in assessing danger to victims.
Judges, law enforcement agencies and others in the justice system charged with full faith and credit enforcement of protection orders from other states should be trained to treat those orders with great seriousness as risks for violence is very high since the petitioner has crossed state lines.
States should have coordinated data systems to access histories of violence, including civil and criminal history records. States should make sure that their injunction registries or protection order files are maintained so that histories of past protection orders are not purged once the protection order expires.
Data systems should include "red flags" for law enforcement so that anyone accessing these records from within the jurisdiction or out-of the jurisdiction can be advised of the potential for danger.
Victim advocates and advisors, lawyers, social service and health care workers, should be specifically trained on recognizing "red flags" and how to safely bring these to the attention of law enforcement, judges and any other persons who could provide for the safety of victims, both resident or transient.
Public education campaigns can positively focus on the "red flags" by showing how dangerous leaving a relationship can be for a victim of abuse. Such a campaign can tell people not to minimize threats to kill, can say that making the victim leave town and flee to another jurisdiction or even get an injunction is insufficient alone and must be accompanied by the development of an active safety plan.
The federal government, tribes and states should step up their efforts to share data across state lines -- both in criminal and civil information systems and through national conferences and organizations. When a victim crosses state lines, information such as the number of prior police calls to a residence, the existence of an order of protection and the elements of a jurisdiction's laws can be readily obtained by a foreign state.
Local and state law enforcement, prosecutors and judges and those on reservations should become familiar with one anothers' laws governing tribal nations so as to know how to enforce foreign protection orders.
Jurisdictions should establish regular and standardized fatality reviews. These reviews should be designed to prevent future homicides by assessing a recent killing -- not be a media event or finger pointing sessions. States may look to already established child fatality review teams for guidance.
State and local government should fund the above practices with additional resources as greater emphasis is placed on assessing risks, conducting safety planning and community education and establishing fatality review teams.
Domestic homicides are, in fact, some of the most preventable of crimes. Further, the wide dissemination and publication of the "red flags" can reduce domestic violence, provided that those who come in contact with battered women and children pay attention to them and incorporate them into their thinking and actions.
Allen, N., H. 1983. "Homicide followed by suicide: Los Angeles, 1970-1979." Suicide and Life Threatening Behavior, 13, 3: 155-165.
Barnard, G. W, Vera, H, Vera, M, and Newman G, 1982. "Till Death Do Us Part: A Study of Spouse Murder," 10 Bulletin of the American Academy of Psychiatry and Law 271.
Bowman, Alana. 1997. "Establishing Domestic Violence Death Review Teams." Domestic Violence Report, August/September, 1997, 83, 93-94.
Buteau, Jacques, Alain Lesage, and Margaret Kiely. 1993. "Homicide followed by suicide: A Quebec case series, 1988-1990." Canadian Journal of Psychiatry, 38:552-556.
Easteal, Patricia W. 1993. Killing the Beloved: Homicide between adult sexual intimates. Australian Institute of Criminology.
Hart, B. (1988). "Beyond the Duty to Warn: A Therapist's Duty to Protect Battered Women and Children." Pp. 234-247 in Yllo, K. & Bograd, M. (Eds.), Feminist Perspectives on Wife Abuse. Newbury Park, California. Sage.
Mercy, James A. and Linda E. Saltzman. 1989. "Fatal Violence Among Spouses in the United States, 1976-85." American Journal of Public Health 79: 595-9.
Stone, Maria. 1995. "Domestic Violence Fatality Reviews: One step closer to a solution." Boalt School of Law. Graduation writing requirement.
Wallace, Alison. 1986. Homicide: The Social Reality. Sydney, Australia: New South Wales Bureau of Crime Statistics and Research.
