«BRIEF REPORT Understanding Domestic Violence Against Women: Using Evolutionary Psychology to Extend the Feminist Functional Analysis Jay Peters ...»
Violence and Victims, Volume 17, Number 2, April 2002
Understanding Domestic Violence
Against Women: Using Evolutionary
Psychology to Extend the Feminist
University of Maine
Todd K. Shackelford
Florida Atlantic University
David M. Buss
University of Texas at Austin
Evolutionary psychologists such as Wilson and Daly (1993b) hypothesize that one goal of
male-perpetrated domestic violence is control over female sexuality, including the deterrence of infidelity. According to this hypothesis, domestic violence varies with women’s reproductive value or expected future reproduction, declining steeply as women age. We tested this hypothesis with a sample of 3,969 cases of male-perpetrated partner-abuse reported to a single police precinct in a large urban area over a 14-year period. Results show that (a) rates of domestic violence decrease as women age, (b) younger men are at greatest risk for perpetrating domestic violence, (c) younger, reproductive age women incur nearly 10 times the risk of domestic violence as do older, post-reproductive age women, and (d) the greater risk of domestic violence incurred by reproductive age women is not attributable solely to mateship to younger, more violent men. Discussion addresses theoretical implications of these findings and suggests a refinement of the feminist hypothesis of domestic violence against women.
Keywords: domestic violence; intimate partner violence; abuse D omestic violence affects many people in North America. Each year, women in the United States report to law enforcement about 572,000 violent attacks by spouses, ex-spouses, boyfriends, and ex-boyfriends, with young women who are divorced or separated incurring the greatest risk of assault (Bureau of Justice Statistics, 1994; Rennison & Welchans, 2000; Wilson & Daly, 1993a, 1993b; Wilson, Daly, & Wright, 1993). Rates of domestic violence victimization range from 13.9% (Kennedy, Forde, Smith, & Dutton, 1991) to 16% (Williams & Hawkins, 1989) annually in the United States. Within a lifetime, 30% percent of women in the United States (Williams & Hawkins, 1989) and 27% of women in Canada (Randall & Haskell, 1995) are battered by a partner or ex-partner.
© Springer Publishing Company 255 256 Peters et al.
The substantial impact of domestic violence results not just from the large number of women who are victimized, but also from the associatedpersonal and social costs. One fifth of the women reporting domestic violence have been attacked repeatedly in the previous six months (Bureau of Justice Statistics, 1994). One in two victimized women receive physical injuries, with 40% of victimized women seeking medical treatment for their injuries (Rennison & Welchans, 2000). Although most domestic violence involves nonfatal injuries, domestic violence is a key precursor of many homicides. During 1999, for example, 48% of all homicides (partner and non-partner) in Maine were preceded by domestic violence (Maine Department of Public Safety, 2000).
Given the number of women victimized by intimates and the seriousness of the sequelae, understanding the etiology and occurrence of male-perpetrated intimate-partner violence is a crucial first step toward successful intervention and eventual reduction of violence against women. Over the past 30 years, at least four theoretical approaches have generated hypotheses about the occurrence of male-perpetrated domestic violence.
Psychoanalytic theorists propose deficits in the ego structure of the violent man, which cause him to lose control of anger that is unconsciously linked to a conflictual relationship with his maternal figures (see, e.g., Adams, 1990). Many family systems theorists also argue that a man’s loss of control over angry feelings is the cause of domestic violence (see, e.g., Geller, 1992). A third explanation of domestic violence proposes that dysfunctional interactional patterns between partners and maladaptive cognitive processes in the individuals result in escalating conflict and eventual violence (Deschner, McNeil, & Moore, 1986; Neidig & Freidman, 1984).
These first three theoretical approaches to domestic violence view the violence as caused by specific deficits: Deficits in ego skills or coping mechanisms; deficits in anger control; and deficits in communication and cognitive skills. In contrast, a fourth hypothesis, derived from feminist theory, states that the violence inflicted on the victim is not a byproduct of underlying deficits, but instead is inflicted strategically and intentionally.
Feminist theorists argue that domestic violence is used by men to exert power and control over their female partners (Adams, 1988, 1990; Walker, 1979, 1994).
In recent years, some evolutionary psychologists have suggested a refinement of the feminist analysis of domestic violence. In agreement with certain aspects of feminist analyses, evolutionary psychologists argue that domestic violence is about controlling women, but that it is specifically about controlling women’s sexuality (see, e.g., Buss & Malamuth, 1996;
Daly & Wilson, 1988). Because ovulation in women is concealed and fertilization occurs internally, men have recurrently faced the adaptive problem of uncertainty of paternity in offspring. Men who failed to solve this problem risked investing resources in children to whom they were genetically unrelated. In addition, cuckolded men incur opportunity costs by forgoing other mating opportunities, and the reputational damage a man incurs by being cuckolded can jeopardize his future mating opportunities (for reviews of the relevant theoretical and empirical literatures, see Buss, 2000, and Daly & Wilson, 1988).
According to one evolutionary psychological perspective, a “solution” to male paternity uncertainty is an evolved male psychology that motivates feelings of proprietary jealousy (Daly, Wilson, & Weghorst, 1982; Wilson & Daly, 1992). These feelings motivate men to guard their mates to prevent them from having sex with other men or to prevent them from leaving the relationship. Such sexual proprietariness is found worldwide (Buss, 1996; Daly et al., 1982). In men, it is responsive to threats (real or imagined) of sexual infidelity or defection by a mate (Buss, Larsen, Westen, & Semmelroth, 1992; Daly & Wilson, 1993; Shackelford, Buss, & Peters, 2000). In one study of married couples (Buss Domestic Violence 257 & Shackelford, 1997), for example, men who anticipated partner infidelity in the future engaged in greater concealment of their partners, exacted greater punishment for a known, suspected, or threatened infidelity, and derogated rivals more than men who did not anticipate future infidelities.
Male sexual proprietariness motivates a range of abusive behaviors (Daly et al., 1982;
Malamuth, 1996) that are commonly labeled domestic violence and that serve to circumvent women’s choices, especially choices regarding sexual behavior (Buss, 1996; Geary, Rumsey, Bow-Thomas, & Hoard 1995; Smuts, 1996). Although one evolutionary psychological perspective on domestic violence focuses on the use of violence to deter infidelity or relationship break-up, evolutionary perspectives also recognize that many contextual factors contribute to the likelihood that men will perpetrate violence against an intimate partner. These factors include lack of protection of the woman by her relatives (Figueredo & McClosky, 1993; Smuts, 1996), economic resources (Figueredo & McClosky, 1993), social tolerance of domestic violence behavior and, for the abusive man, childhood socialization, neurobiological predisposition, and loss of control of angry feelings (Wilson & Daly, 1993b).
Wilson and Daly (1993b; see also Wilson, Daly, & Scheib, 1997) propose that, en route to inflicting violence on an intimate partner, men solve information-processing problems that generate risk assessments of a partner’s sexual infidelity or desertion. This information processing includes assessments of the presence of potential rival males, the social costs of inflicting violence, and the woman’s reproductive value or expected future reproduction (Trivers, 1972). Because reproductive value peaks for women soon after puberty and then declines with a woman’s age, Wilson and Daly (1993b) hypothesize that attempts to control her reproductive choices through violence will follow a similar trajectory.
Specifically, Wilson and Daly (1993b) hypothesize that younger women, with higher reproductive value, are at greater risk of domestic violence than are older women, with lower reproductive value.
Support for Wilson and Daly’s (1993b) hypothesis comes from several sources, including their 1988 study in which they found that the risk of uxoricide or wife killing—which they argue is an extreme form of domestic violence—is highest for younger wives in North America (Daly & Wilson, 1988; see also Mercy & Saltzman (1989) for similar findings for United States national homicide data). The United States National Crime Victimization Survey (NCVS) for 1987-1998 revealed that young women had the highest rates of violent victimization by an intimate male partner (Bureau of Justice Statistics, 1994; Rennison & Welchans, 2000). Similar findings were reported for a Canadian sample by Wilson, Johnson, and Daly (1995). Buss and Shackelford (1997) found that men married to younger women engaged in more forms of mate guarding, including violence, than did men married to older women.
Failure to comply with a male partner’s demands for a sexually exclusive relationship can be costly for the woman (Smuts, 1996) and can lead to escalation of the seriousness or frequency of “autonomy limiting techniques” (Wilson et al., 1995). These techniques are directly linked to the use of violence, according to Wilson and colleagues (1995), who found that the number and degree of men’s controlling behaviors were positively correlated with the seriousness of the violence they inflicted on their female partners. That sexual jealousy is a key factor in domestic violence is supported by numerous studies which show that the majority of domestic violence femicides are precipitated by a man’s suspicion of his partner’s infidelity or her intention to end the relationship (Wilson & Daly, 1992; see Buss, 2000, for a review of the literature).
258 Peters et al.
An alternative explanation for the steady decrease in violence against women as they age is that young women tend to be mated to young men and young men are the most violent age-sex subset of the population. Men aged 16 to 24 years, for example, commit the majority of violent acts, including homicide (Wilson & Daly, 1985). The decrease in violence toward women as women age therefore may be attributable to the aging of men and not attributable to the aging of women. Wilson and colleagues (1995) show that uxoricides per million wives per annum decrease slightly with husband’s age, but begin decreasing about 10 years earlier than do uxoricides as a function of wife’s age. If the decrease in domestic violence were solely a function of men’s decreasing use of violence, that decrease should occur when women are in their late 20s and early 30s, not late 30s and early 40s. Shackelford, Buss, and Peters (2000) found that young women are at greatest risk for uxoricide when they are married to men aged 45 to 54 years. In both of these studies, the “young male syndrome” hypothesis was not supported.
We conducted a new test of Wilson and Daly’s (1993b) hypothesis that the risk of domestic violence varies with the woman’s reproductive value. Specifically, we tested the hypothesis that domestic violence rates will decrease with the increasing age of the woman. We further hypothesized that this risk distribution is not solely attributable to the age of the male perpetrator. To test these hypotheses, we secured data for 3,969 cases of male-perpetrated partner-abuse reported to a single police precinct in a large urban area over a 14-year period.
New York City police officers are required to complete an incident report form for each reported case of lethal and non-lethal domestic disturbance. These reports, which track emergency (“911”) telephone calls from alleged victims, family members, neighbors, and witnesses, contain some information about the victim and perpetrator. Included in the data are the nature of the relationship (for example, intimate relationship, strangers) and the ages of the victim and the offender. Reports concerning intimate partner violence are furnished to the Domestic Violence Prevention Project (DVPP) of Safe Horizon, a non-profit social service agency serving crime victims. In the data set provided by Safe Horizon, lethal and non-lethal reports were not differentiated.
From 1986 to 1999, DVPP received 5,298 domestic violence reports from the 103rd police precinct in Queens, New York. For the current study, we excluded 908 cases because of missing data for the age or sex of the victim or perpetrator. In addition, we excluded 421 cases in which the male was the alleged victim or the female was the alleged perpetrator. The remaining 3,969 cases involve one female victim and one male perpetrator who were (a) married or cohabiting (marital status was not differentiated in the dataset), or who have a child in common and are (b) both 16 years of age or older.
Domestic violence rates were calculated according to relevant population estimates for married and cohabiting individuals provided by the United States Census for the population served by the 103rd precinct (available from the second author upon request).
and corresponding domestic dispute rate (number of domestic disputes per 1,000 persons in the relevant population) for each of five age groups, by sex. These age groupings were used because these are the age groupings according to which the census data were provided. Table 1 shows that women aged 15 to 24 years are at greatest risk for domestic violence, a risk that is more than twice as great as the risk for women aged 25 to 34 years. Rates of domestic violence decrease throughout women’s reproductive years, to a low of 0.6 domestic disputes per annum per 1,000 women aged 65 or older.