Are Heterosexual Men Also Victims of Intimate Partner Abuse?
Joanne Belknap and Heather Melton
Perpetration of intimate partner abuse (IPA) by women against men has received widespread attention from both practitioners and researchers. Some research suggests that contrary to popular belief, women are just as likely as men to be perpetrators of IPA (Brush, 1990; Madgol, Moffit, Caspi, Fagan, & Silva, 1997; Moffit & Caspi, 1999; Morse, 1995; O'Leary, Barling, Arias, Rosenbaum, Malone, & Tyree, 1989; Straus & Gelles, 1990). Others argue that IPA continues to be perpetrated largely by males against their female partners and ex-partners ( Dasgupta , 2001; Dobash & Dobash, 1984-1988 in References; Dobash, Dobash, Wilson, & Daly, 1992; McLeod, 1984; Melton & Belknap, 2003; Saunders, 1986; Tjaden & Thoennes, 2000; Vivian & Langhinrichsen-Rohling, 1994). While the debate continues regarding women's use of violence against intimate and former intimate male partners, several issues have emerged regarding research on woman-perpetrated IPA. The purpose of this essay is to critically review the existing research on the question of gender symmetry in IPA. Gender symmetry is the terminology often used to indicate that men and women are equally likely to be IPA offenders.
This paper presents and discusses the varied findings on women's roles as perpetrators of IPA. The reasons for these varied findings are examined and the implications of the research finding gender symmetry in the perpetration of IPA are discussed. This paper documents the importance of the approach taken by the researcher regarding whether IPA is found to be gendered. This overview of scientific research concludes that IPA is indeed gendered, that the perpetrators are more commonly men and the victims are more commonly women. This review also emphasizes the importance of not simply examining types of abuse reported, but the consequences of the abuse . We hope to clarify women's use of violence in IPA as having typically different intentions than men's abuse of intimate partners.
Women and girls who report that they are abused by intimate partners are far more likely than men and boys reporting such abuse to say they are afraid and seriously injured (Hamberger & Guse, 2002; Johnson & Bunge, 2001; Melton & Belknap, 2003; Morse, 1995; O'Keefe & Treister, 1998). Our own work is grounded in a feminist approach, starting with the belief that there is significant gender inequality in the culture, political system, justice system, and the like, which results in the gendered dynamics and rates of IPA. Thus, as we believe that all research is inherently subjective, we begin this review by acknowledging our own viewpoint. However, we have conducted a thorough review of the existing work to explain the nuances of the gendered nature of the problem of IPA as well as the nuances of the varied approaches to measuring IPA and its impact.
The terminology we use in this paper is slightly different than that seen in most other publications on the topic. The terms used to describe the behavior of individuals abusing their current or former intimate partners are lengthy: domestic violence, woman battering, spousal violence, intimate partner violence, and so on. In this paper, we choose the term intimate partner abuse , because as described here, there is a wide range of behaviors including but not limited to physical violence that constitute the larger concept of abuse. Indeed, victims often identify some forms of psychological and emotional abuse (e.g., fatal threats, demeaning statements, threats to abuse children or obtain custody, pet abuse, withholding access to money, etc.) perpetrated by intimate partners as more harmful than their intimate partners' physical violence (e.g., hitting, slapping, shoving, etc.) (See, for example, Hamby & Sugarman, 1999). Therefore, in this paper, we use the term intimate partner abuse to connote a wide range of abuses, some of which are physically violent behaviors and others that are emotionally abusive. As we make apparent in this paper, addressing the nuances of the gendered nature of IPA is more comprehensive when understanding and including the impacts of nonphysical abuse as well as physically violent behaviors.
There is significant conflict over the amount of female perpetration of IPA against male partners and ex-partners in existing research. In a previous paper we discussed the two approaches regarding the gendered nature of IPA perpetration (Melton & Belknap, 2003). First, some researchers tend to view IPA as strongly gendered, with males as the primary offenders and females the primary victims. Others, however, are far more likely to report IPA as a gender-neutral, or gender symmetrical, phenomenon.
Gender symmetry is often reported in studies employing large, random, and national or community samples and using the Conflict Tactics Scale (CTS), a questionnaire that asks about recent use of specific tactics by an intimate partner against the respondent to measure IPA (Brush, 1990; Cantos, Neidig, & O'Leary, 1994; Johnson, 1995; Morse, 1995). Morse (1995) suggests that the goal is to reconceptualize the social problem of ""woman battering"" to one of ""family violence."" From this perspective, some scholars tend to conclude that men and women are equally likely to be both perpetrators and victims of IPA. Most notably, the 1975 and 1985 National Family Violence Survey by Straus and his colleagues (Gelles, 1979; Straus & Gelles, 1990) found ""gender symmetry"" in spousal assault (see Dobash et al., 1992; Grandin & Lupri, 1997, for overviews). Some studies report that women are more often the perpetrators of IPA, while men are more often the victims of IPA, especially at less serious levels. For example, the National Youth Survey (NYS) reports for the year 1983 that slightly over one-third of the self-reported intimate partner abuse was male-to-female (male offender, female victim), and almost one-half was female-to-male (female offender, male victim), and the remainder (15%) were ""mutually combative,"" or cases where both members of the couple were abusive to each other (Morse, 1995). Following subsequent NYS surveys up to 1992 indicate that these rates declined at all time periods; however, female-to-male partner violence remained higher than that of male-to-female partner violence from 1985 to 1992 (Morse, 1995). Moffit and her colleagues compared the NYS findings to two family violence surveys in New Zealand and reported that across these three studies, 36.4 to 51.3 percent of all women reported some form of IPA, while 21.8 to 43.0 percent of all men reported perpetrating some form of IPA (Madgol et al., 1997; Moffit & Caspi, 1999).
Despite these findings, other researchers often identifying themselves as feminists (as well as most shelter workers) continue to maintain that there is not gender symmetry in IPA. Contrary to the national samples mentioned above, studies of samples from shelters, hospitals, and police reports find that as many as 90 to 95 percent of IPA involves a male perpetrator against his female partner or ex-partner (Dobash et al., 1992; Johnson, 1995; McLeod, 1984; Melton & Belknap, 2003; Saunders, 1986). In addition, reviews of some of the criminal victimization surveys using national probability samples report that even using national samples, women are far more likely than men to be victimized by their intimate partners. Indeed, analyses of the U.S. National Crime Survey, the National Crime Victimization Survey (NCVS), the 1981 Canadian Urban Victimization Survey, the 1987 General Social Survey, and the National Violence Against Women Survey have all resulted in findings where women are far more likely than men to be abused by intimate partners (Bachman, 1998; Dobash et al., 1992; Rennison & Welchans, 2000; Tjaden & Thoennes, 2000).
Several arguments have been made to explain the huge discrepancies in scholars' interpretations of findings regarding women's use of violence against intimate partners. These arguments include criticisms of the CTS as a measure of IPA, concerns over gender differences in reporting of IPA and its impact on abuse rates, the differences due to settings in which the data have been collected and the samples studied and, finally, issues related to studying victimization only versus victimization and perpetration. Each of these will be discussed in the following paragraphs.
First, and at the heart of much of this criticism, is CTS: a set of scales designed by Strauss and Gelles, leaders in the ""family violence"" approach, to measure IPA (Strauss & Gelles, 1990). Some scholars have harshly criticized the CTS for ignoring the context, motivations, meanings, and consequences of IPA (see, for example, Bachman, 1998; Belknap, 2001; Berk, Berk, Loseke, & Rauma, 1983; DeKeseredy, 1995; Dasgupta , 2001; DesKeseredy & Schwartz, 1998; Dobash & Dobash, 1988; Kurz, 1993; Schwartz, 1987; Stark & Flitcraft, 1983; Yllo, 1983, for critiques). As stated previously, the studies reporting gender symmetry and the higher rates of female-perpetrated IPA predominantly use national sampling plans. These are also the studies most likely to employ the CTS to measure IPA. Thus, it is likely not the ""national"" sample that results in the different rates of female-perpetrated IPA, but rather the strong correlation between a study's likelihood of using the CTS and using a national sampling plan that leads to the reporting of gender symmetry. The use of the CTS clearly explains some of the discrepancies between the two approaches.
Also, some researchers are concerned that the studies using the Conflict Tactics Scale (CTS) interpret the results incorrectly, and that this may account for the differences in the studies. Dobash and colleagues report that in many of the studies finding higher rates of female-perpetrated IPA, a respondent who reports that he or she has ever, ""pushed,"" ""grabbed,"" ""shoved,"" ""slapped,"" or ""hit or tried to hit"" another person, is regarded as a perpetrator of IPA (Dobash et al., 1992). This may include only one instance not taken in context. Thus, when victims resist abuse in any way, including defending themselves or their children, they will mistakenly be portrayed as intimate partner abusers. For an excellent review of the history of the CTS, its past, revised, and current uses and limitations, read DeKeseredy and Schwartz (1998).
A second explanation for the discrepancies among scholars' findings is that research indicates gender differences in honesty in reporting use of force by members of intimate couples . At the same time that many women victims of IPA minimize the abuse perpetrated against them, many of the male IPA abusers also minimize or underreport their own use of abuse (see Berns, 2000; Campbell, 1995; Dobash, Dobash, Cavanagh, & Lewis , 1998; Goodrum, Umberson, & Anderson, 2001; Heckert & Gondolf, 2000; Morse, 1995; Stets & Straus, 1990; Walker, 1979), particularly sexual abuse (Meyer, Vivian, & O'Leary, 1998). Moreover, studies of male intimate partner abusers find they typically use excuses and justifications to minimize their use of violence (Dutton, 1988; Ptacek, 1988). These gendered practices would obviously lead to data misrepresenting something closer to gender symmetry in IPA. Thus, who the researcher interviews is likely to impact the direction of the data. In national surveys, often the only person interviewed is the first person to answer the phone. Furthermore, women victims of IPA have many reasons to underreport IPA victimization, even to researchers (see Smith, 1994) because of:
ï Danger and the fear of reprisals or retaliation
ï Shame or embarrassment
ï The view that IPA is too personal to report to anyone during a survey
ï A lack of trust of the interviewer
A third plausible explanation for the higher rates of female-perpetrated IPA has to do with the data collection setting . More specifically, the clinical types of settings (e.g., shelters, hospitals, etc.) are likely and legitimately perceived as safer contexts for answering questions than on the phone in her home or to an interviewer in her home. Research on IPA consistently reports highly controlling and stalking behaviors by abusers, thus, the home is often one of the least safe places for the victims. Thus, victims are likely more able and willing to disclose abuse victimization in the settings more typically used in smaller scale studies.
A fourth explanation for the varied findings concerns the difference in samples used. Perhaps this is best understood through Johnson's (1995) distinction between common couple violence and patriarchal terrorism. Common couple violence refers to those cases where both members of a couple use occasional outbursts of abuse toward their partner in response to everyday stimuli. According to Johnson, this is not gendered, rarely escalates into serious, life-threatening violence, and is not a product of patriarchy, but of a culture that accepts violence in general. Patriarchal terrorism is a form of ""terroristic control"" of women by their male partners and ex-partners, and is a direct result of patriarchy: the historical tradition of men's legal and social ""right"" to control women. Patriarchal terrorism tends to be more serious, occur over a longer period of time, and be more likely to escalate into life-threatening violence than common couple violence. In addition, patriarchal terrorism includes not only violence, but also other terrorist tactics such as economic control, threats, and isolation (Johnson, 1995). Thus, Johnson (1995) argues that studies using community samples are more likely measuring common couple violence, while studies using clinical, self-help, or official (e.g., police report) samples are examining patriarchal terrorism. Stated alternatively, community samples are biased toward common couple violence and clinical or official samples are biased toward patriarchal terrorism. Men who commit patriarchal terrorism and women who experience it would be unlikely to directly answer survey questions (or at least, do so as honestly), and victims of patriarchal terrorism are more likely to need or seek out assistance in combating violence (Johnson, 1995). However, it should be noted that Johnson's argument has been criticized--some national samples have found higher levels of male-to-female violence that is terroristic in nature (see Dobash et al., 1992; Rennison & Welchans, 2000; & Tjaden and Thoennes, 2000). In the same vein, abused women in the clinical and official samples are likely more willing and able to address and admit the abuse they receive from their intimate partners. Ferraro and Johnson (1983) report that being in a battered women's shelter was a pivotal point for many women to start to define themselves as ""battered"" and plan their exit from their abusers.
Notably, Bachman (1998) compares two national samples: the National Crime Victimization Survey (NCVS) and the National Family Violence Survey (NFVS), reporting that the latter study's rates are more consistent with findings of gender differences in IPA perpetration. They attribute this to improved methodology of the NFVS over the NCVS: The NFVS designed the survey items to screen better for intimate partner abuses. This study found that women were eight times as likely as men to be victims of IPA.
Finally, Tjaden and Thoennes (2000) identify one more methodological difference that may account for the difference found between studies' findings. Studies reporting gender symmetry tend to focus on both victimization and perpetration, whereas studies that show gender asymmetry tend to focus only on victimization. When asking study respondents to report how much violence or abuse they use (perpetrate) against their intimate partners, it is not unusual for men to minimize these abuses while women tend to be more honest or forthcoming. The next section details this gendered phenomenon.
Unfortunately, given the problems with measurement, it is difficult to know how many cases of IPA are truly female-perpetrated. Our perusal of the scientific studies on IPA indicates that if we include a careful screening of what constitutes IPA, females are the perpetrators in five percent or fewer of the cases (see Berk et al., 1983; Browne, 1987; Dobash et al., 1992).
In addition to addressing whether the frequency of IPA is gendered, it is necessary to ask whether the nature of IPA is gendered. Some argue that men and women are likely to use different types of violence. Milardo (1998) found that while women are more likely to perpetrate more varied forms of violence, men are more likely to perpetrate more serious forms of violence. Melton and Belknap (2003), in their comparison of male and female perpetrators of IPA in police reports, also found that men were more likely to perpetrate more serious forms of IPA.
Many scholars and victim advocates report that women have different motivations for using force against their current or former intimate partners. More specifically, women are far more likely than men to employ force with their intimate partners in the context of self-defense (Barnett, Lee, & Thelan, 1997; Cascardi & Vivian, 1995; Dasgupta , 2001; DeKeseredy & Schwartz, 1998; Hamberger, Lohr, Bonge, & Tolin, 1997; Hamberger & Potente, 1994; Molidor & Tolman, 1998; Saunders, 1986; Schwartz, 1987). Indeed, Anderson and Umberson's (2001) study of male IPA offenders concluded that these men were effective in twisting their less serious (female) partners' behaviors into the major violence, while they excused their own abusive behaviors as rational, capable, and nonviolent. While women are more likely than men to use force to resist violence initiated by their intimate partners, men are more likely than women to use force in order to control and exercise power over their partners (Barnett et al., 1997; Hamberger et al., 1997; Hamberger & Guse, 2002; Hamberger & Potente, 1994). Indeed, Worcester (2002) emphasizes that any analysis of women and girls' use of force in intimate relationships must be through a ""framework that keeps power and control central to the definitions of domestic violence""(p. 1390). This important distinction between women's primary motivation as self-defense and men's primary motivation as control has major gender implications for practitioners responding to those charged with IPA--many charged women may actually be victims of IPA acting in self-defense, rather than the offenders.
This raises the issue in national surveys regarding who is reporting the abuse to the research investigator ? Typically, the respondent to these national studies is whoever answers the phone first. If an intimate partner abuser is monitoring all incoming calls, it is likely that he will be the respondent. At the same time it is not unusual for victims to minimize their time on phone calls when jealous and controlling batterers ""check up"" on them by calling to make sure they are not talking to anyone on the phone. A long survey would be something such a victim would want to avoid. Moreover, if a victim answers the survey phone call, but the batterer is home, it is likely that s/he would minimize the abuse or choose not to take part in the study.
Das Dasgupta (2001) has developed a very complete report on women's use of nonlethal violence in heterosexual relationships. This report carefully reviews the research on this topic, including how women's violence in intimate partner relationships is defined and researched, and the implications for reshaping the existing societal responses to IPA.
Research findings are consistent regarding extreme gender differences in the consequences of men and women's violence in their intimate relationships. A significant amount of research reports that women suffer more negative consequences as a result of violence from a current or former male partner than men do from a current or former female partner (Brush, 1990; Cantos et al., 1994; Cascardi & Vivian, 1995; Cook & Harris, 1995; Dobash et al., 1992; Holtzworth-Munroe, Smutzler, & Bates, 1997; Milardo, 1998; O'Leary et al., 1989; Stets & Straus, 1990; Tjaden & Thoennes, 2000; Vivian & Langhinrichsen-Rohling, 1994). Women involved in IPA are more likely than their male counterparts to suffer from injuries, require medical treatment, lose time from work, and experience bedridden days than are men (Archer, 2000; Berk et al., 1983; Brush, 1990; Cantos et al., 1994; Cascardi & Vivian, 1995; Cook & Harris, 1995; Dobash et al., 1992; Holtzworth-Monroe et al., 1997; Johnson & Bunge, 2001; Morse, 1995; O'Leary et al., 1989; Rand, 1997; Stets & Straus, 1990; Tjaden & Thoennes, 2000; Vivian & Langhinrichesen-Rohling, 1994). More negative consequences for females (relative to males) have also been reported in dating violence (Molidor & Tolman, 1998; O'Keefe & Treister, 1998). Moreover, men's violence against women is far more likely than women's violence against men to produce fear in the victim (Hamberger & Guse, 2002; Johnson & Bunge, 2001; Melton & Belknap, 2003; Morse, 1995; O'Keefe & Treister, 1998). Notably, two studies found that men and boys are more likely than women and girls to report that they think the abuse or force perpetrated by an intimate partner of the opposite sex was funny or humorous (Hamberger & Guse, 2002; O'Keefe & Treister, 1998). Clearly, there are significant gender differences regarding men and women's use of abuse and force against their current and former intimate partners.
Potter (in press) emphasizes the need to account for racial and ethnic differences in examining any of the instruments measuring IPA in terms of their applications to People of Color. She notes that the measures of IPA have been criticized for being designed to ""tap"" the White population's rates and dynamics, and then applied these same measures to Communities of Color without appropriate testing (see Crowell & Burgess, 1996; Sorenson, 1996). Early research employing the CTS and studies such as the NFVS reported higher rates of Black than White women's use of ""severe violence"" toward their husbands (Straus, Gelles, & Steinmetz, 1980), but these findings must be considered cautiously given the problems with instruments designed for primarily for White populations (Potter, in press).
Dasgupata (2001) describes how the CTS could completely miss the gendered nature of IPA in defining the victim and offender regarding an immigrant woman:
""For instance, consider a situation where an immigrant woman has thrown a pot at her husband who has just destroyed her passport and conditional residency status card. On the CTS, the magnitude of the woman's violence would be considered much greater than that of her husband. On levels of severity also, the tearing up of papers would compute much lower than the physical violence that has just occurred. Yet, the consequences of the trashing of papers that lend this woman legitimate residency are extremely devastating. She may lose her job, be deported, and lose custody of her children because of her abuser's behavior. Thus, the woman may view such an act as intensely violent"" (p. 3).
This research overview stresses that IPA is gendered: Men and boys are more likely (than women and girls) to be the perpetrators, and women and girls are more likely (than men and boys) to be the victims of IPA. At the same time, it is important to recognize that there are some women and girls who are abusive and violent to their intimate male partners . In our minds, they are not what Johnson (1995) would envision as ""common couple"" abusers, but rather, are the primary aggressors in their relationships.
One of the few studies that has focused on these women and girls, reports three components of a model attempting to explain female-perpetrated IPA: learning, opportunity, and choice (Perilla, Frndak, Lillard, & East, 2003). Learning is a means by which the girl or woman learns to be abusive through experiencing or witnessing IPA or other violence. Indeed, other research claims that women charged with domestic violence have disproportionately high experiences of childhood abuse (see Swan & Snow 2003, for a review). In a study of women victims of IPA, women who responded with more aggression and violence toward their abusers were significantly more likely to have a history of childhood abuse than women IPA victims without such childhood histories (Swan & Snow, 2003). Perilla et al. (2003) cite examples of instances in which the woman had a previous intimate relationship in which she was severely abused or watched her mother's serious IPA, and ""learned"" to be violent in her subsequent relationships.
In addition to ""learning,"" Perilla et al. (2003) identify opportunity and choice as components of a model to explain female-perpetrated IPA. Opportunity, as described by Perilla and her colleagues, closely indicates retaliation opportunity, but could be seen as self-defense by some. Their example of an ""opportunity"" is where the power shifts for a period of time, for example, when the male abuser is passed out from alcohol, and the woman/victim chooses to use violence against him in this vulnerable state. In this case, Perilla et al. (2003) report that the victim who became the offender ""believed that she could die at the hands of her abuser and saw her use of violence as a way to defend herself from further battering"" (p. 31). Thus, to some, the opportunity component may be more self-defense than retaliation. This victim-come-offender had repeatedly used the police to no avail. This is consistent with Miller's (2001) findings on women arrested for domestic violence. She identifies a paradox for many women arrested for domestic violence, that many used violence in self-defense.
Finally, Perilla and her colleagues (2003) emphasize that just as those in the domestic violence movement have the mantra ""violence is a choice "" when talking about male batterers, this approach must also be taken for women who are violent toward intimate partners. They report cases in which women chose violence in self-defense, but also in retaliation: a case in which a man quit abusing his wife after many years of marriage, but she used violence once he stopped, as a way to make up for the years of abuse she had experienced at his hands. Swan and Snow (2003) portray some women's violence toward their intimate partners as a ""coping strategy"" for the stress of being in the abusive relationship. They also found that the women in heterosexual relationships who used the most aggression and violence (compared to other women IPA victims in the sample) were also the ones most likely to initiate IPA.
Perilla and her colleagues' effectively showed how the three components (learning, opportunity, and choice) in their model for women's use of violence in intimate partner relationships do in fact often converge: The woman has learned abusive behavior over the course of her relationship and/or life, she has the opportunity where she has more power, and she makes the choice to abuse. Further research needs to more accurately identify women using abuse in their intimate relationships and understand the dynamics behind their behavior.
This Applied Research paper and In Brief may be reprinted in its entirety or excerpted with proper acknowledgement to the author(s) and VAWnet, a project of the National Resource Center on Domestic Violence, but may not be altered or sold for profit.