Partner Homicide in Context: A Population-Based Perspective

Interviews to develop profiles of all partner homicides that occurred in North Carolina in a single year. This methodology allowed us to investigate gender differences that might shape the context for male and female homicide perpetration and victimization. Five themes emerged: (a) The context for partner homicides is often chronic women battering, (b) leaving an abusive partner and remaining are both dangerous options, (c) protective measures for battered women are inadequate, (d) domestic violence is not necessarily private violence, and (e) alcohol and firearms often accompany homicide. These themes suggested: Partner homicides emanated almost uniformly from a history of male-perpetrated aggression; analysis of partner homicide should not be detached from the daily life created and sustained by battering; and a gender analysis of partner homicide focuses on the context of gender-based power imbalances rather than on frequency or severity of injury. Article: In 1994, 28.4% of the 4,739 women homicide victims were killed by their husbands or boyfriends. In fact, American women are more likely to be killed by their male partners than by anybody else (Kellermann & Mercy, 1992; Koss et al., 1994; McGuire & Pastore, 1996). Not all victims of partner homicide, however, are women: 3.3% of the 17,337 men killed in 1994 were killed by a wife or girlfriend (McGuire & Pastore, 1996). Efforts to prevent partner homicide, the murder of men and women by their current or former intimate partners, are challenged in part by the lack of sufficient knowledge about its incidence as well as about the causes and circumstances surrounding it. Much research has been based on large data sets that provide population-based quantitative data but is limited in contextual or circumstantial data. At the other extreme is case-study research, which provides qualitative contextual data but lacks population representativeness. Although both methodologies are useful because they provide us with different snapshots of partner homicide, they leave many questions unanswered. Research based on the FBI's Supplemental Homicide Reports provides important information on high-risk groups and certain risk factors. It indicates that African Americans are at higher risk than Whites (Mercy & Saltzman, 1989; Plass, 1993; Stout, 1991) and both men and women are victims of partner homicide (Mercy & Saltzman, 1989; Wilson & Daly, 1993). However, women's and men's risk of victimization varies considerably by race and across type of relationship (Browne & Williams, 1993). Among Whites, women are at higher risk for spousal homicide, whereas among African Americans, men and women are at similar risk (Block & Christakos, 1995; Mercy & Saltzman, 1989; Plass, 1993). Additionally, women are more likely to be killed by a spouse than by a common-law or dating partner whereas men are just as likely to be killed by a spouse as by a common-law partner (Browne & Williams, 1993; Rosenfeld, 1997; Wilson & Daly, 1993). Further, women are more likely than men to be killed by a former partner (Mercy & Saltzman, 1989). Population-based research has also provided information on weapons used in the commission of these homicides. Firearms and knives are 1 AUTHORS' NOTE: An earlier version of this article was presented at the 1992 annual American Public Health Association conference. Partial funding for the research described in this article was provided by the Society for the Psychological Study of Social Issues and the Injury Prevention Research Center of the University of North Carolina at Chapel Hill through a student research award. We would like to thank Carol Runyan, director of the Injury Prevention Research Center, a community advisory committee, and University of North Carolina—Greensboro graduate assistants Anne Green and Gloria Edwards for their help with this article, and anonymous reviewers for their helpful comments. typically the most common weapons, although the percentage of homicides committed with each varies across studies (Block and Christakos, 1995; Mercy & Saltzman, 1989). Although population-based research on partner homicide suggests that both women and men are at risk, it provides less information on whether and how the circumstances that lead men and women to kill may or may not differ. It has been suggested that female-perpetrated partner homicide may frequently be a response to the male partner's actual or threatened physical assault, whereas male-perpetrated partner homicide occurs most often in the context of women attempting to leave their partners (Browne & Williams, 1993; Wilson & Daly, 1993; Wolfgang, 1958). However, Browne and Williams (1993) note that gender is often omitted from analysis in much of this research. They have argued that exclusion of gender may lead to the erroneous conclusions that "The conditions of women's lives are essentially the same as those of men" or that "Although the conditions of women's lives may sharply differ from those of men, those differences are not germane to general theories on homicide" (p. 79). Case-study research, in contrast, has often explicitly considered the role gender plays in the lives of the men and women involved in partner homicide. This research suggests that, regardless of the sex of the perpetrator, partner homicide is rarely an isolated incident; rather, it often occurs as a result of a chronic pattern of abusive and threatening behavior by the man against his female partner (Browne, 1987; Chimbos, 1978; Jurik & Winn, 1990; Trotman, 1978). It also suggests that women frequently kill in response to threatened or actual physical assault by their partners (Browne, 1987; Goetting, 1991). However, case-study research has tended to focus on selected subgroups such as battered women charged with murdering or seriously wounding their partners (Browne, 1987, Walker, 1989); men and women charged with killing their partners undergoing pretrial psychiatric evaluation (Barnard, Vera, Vera, & Newman, 1982; Daniel & Harris, 1982); men and women convicted of killing their partners (Jurik & Winn, 1990); and women imprisoned for killing their partners (Foster, Veale, & Fogel, 1989). Case-study research thus limits generalizability and may reinforce the view that partner homicide is idiosyncratic rather than representative of an identifiable epidemiological pattern of male violence against women (Saunders and Browne, 1991). This study was undertaken in order to explore the circumstances and context surrounding all cases of heterosexual partner homicide that occurred in 1 year in one state and thus broaden the usual case-study approach. The major purpose was to determine the proportion of all homicides that were preceded by a history of domestic violence and the circumstances that surrounded them. We were particularly interested in exploring how the prehomicide conditions of the men's and women's lives might shape the context for homicide and the role that gender might play in the use of fatal violence by men and women. CONCEPTUAL FRAMEWORKS We made use of two different conceptual frameworks to help us analyze and describe partner homicide. The first, that of victim precipitation, has often been used to analyze sex differences in the perpetration of partner homicide (Browne & Williams, 1993; Campbell, 1992; Chimbos, 1978; Saunders & Browne, 1991). Suggested by Von Hentig (1941, 1948) in the 1940s, the term was coined and used by Marvin Wolfgang in his pioneering research on homicide in Philadelphia in 1958. As defined by Wolfgang (1958), Victim-precipitation is applied to those criminal homicides in which the victim is a direct, positive precipitator in the crime. The victim-precipitated cases are those in which the victim was the first to show and use a deadly weapon, to strike a blow in an altercation—in short, the first to commence the interplay of resort to physical violence. (p. 252) Since then, victim precipitation and the idea that homicide victims sometimes participate in the actions leading to their deaths have played a central role in both homicide research and criminology textbooks (Block, 1993; Mann, 1996; Polk, 1997; Rasche, 1993; Savitz, Kumar, & Turner, 1993; Wolfgang, 1993). Studies have reported varying proportions of victim-precipitated homicides, ranging from 0% to more than 50%, depending on the study population, location, and operational definitions used (Goetting, 1988, 1991; Polk, 1997; Savitz et al., 1993; Wolfgang, 1958). The concept has also been used as a framework to analyze partner homicides, particularly in cases where battered women kill their abusers (Campbell, 1992; Rasche, 1993). Because this concept directs attention to the circumstances surrounding the homicide, it is potentially useful as a framework to examine the similarities and/or differences in the context immediately surrounding men's and women's perpetration of partner homicide. Agudelo (1992) provides the second framework we used. He defines violence as "the exercise of physical, psychological, or moral force directly or indirectly by a person or group of persons in the exercise of power" (p. 367). He argues that a precondition for violence is an "imbalance between heterogeneous entities having unequal power." This gradation of power which, according to Agudelo (1992), "is the delta through which violence flows" (p. 367), could be conceptualized as the conduit for both maleand female-perpetrated violence. The violence that emerges from this power imbalance is of two types: antiaction and proaction. The first, antiaction, is violence that is "aimed against an existing or potential power. It is an attempt to disrupt an order that has been imposed or agreed to. It is an act of aggression against rights that are socioculturally established and regulated" (Agudelo, 1992, p. 367). In contrast, violence that is proaction is that which is "exercised to affirm and defend a right or build another order or legal system" (p. 367). It emerges, according to Latin American epidemiologist Dr. Hector Abad Gomez, from conditions of oppression, injustice, and grievous economic inequality in which violence is not a disease but rather a necessary reaction on the part of the body social—somewhat like the biological organism's response to infection. It is like fever, a mechanism for combating infection, which is the real disease. (quoted in Agudelo, 1992, p. 367-368) Agudelo further argues that understanding violence requires analysis of the "direction of the forces that produce the effects observed, the patterns in question, and the powers in opposition" (Agudelo, 1992, p. 367). Although neither men nor women are free from the use of violence, this framework argues that men's and women's violence might have a different directionality and serve a different purpose. For example, conceptualizing batterers' violence as antiaction would suggest that the purpose behind their use of force is to contain women's potential power and/ or to destroy their legal or socially culturally agreed upon rights. The direction of the force is against those who have lesser social power. In this context, men's use of force helps to consolidate and maintain this power structure. Similarly, conceptualizing battered women's violence as proaction would suggest that their use of force emerges from their oppression as women and as victimized persons and has the purpose of trying to affirm their own rights as humans; furthermore, it may be an unavoidable if not necessary reaction. In contrast to the victim-precipitation framework that directs attention toward the circumstances incident to the homicide, Agudelo's perspective directs our attention toward differences in experience and behavior between battered women and battering men that result from unequal power.