Despite numerous studies that report the preponderance of domestic violence is perpe- trated by men against women, other empirical studies suggest that rates of domestic vio- lence by women and men are equivalent. This article explores these claims of gender sym- metry in intimate partners’ use of violence by reviewing the empirical foundations of the research and critiquing existing sources of data on domestic violence. The author sug- gests methods to reconcile the disparate data and encourages researchers and practi- tioners to acknowledge women’s use of violence while understanding why it tends to be very different from violence by men toward their female partners.

Domestic violence has emerged as one of the world’s most pressing problems. The United Nations estimates that between 20% and 50% of all women worldwide have experienced physical violence at the hands of intimate partners or family members (Leeman, 2000; United Nations Population Fund, 2000). In the United States, more than 1 million cases of intimate partner violence are reported to police each year according to the U.S. Department of Justice (see Goldberg, 1999). One of the major platforms for action adopted at the World Conference on Women in Beijing in 1995 was the prevention and elimination of violence against women and girls.

Efforts to prevent domestic violence and to facilitate the suc- cessful prosecution of batterers have followed research and

1332

AUTHOR’S NOTE: An earlier version of this article was commissioned by the Irish Department of Education and Health. Subsequent debates in Ireland led to two op-ed pieces in The Irish Times (Kimmel, 2001, 2002.). I am grateful to Harry Ferguson for com- ments and support. I am also grateful to two anonymous reviewers for their helpful feed- back and to Sue Osthoff, Andrea Bible, Shamita Das Dasgupta, Aisha Baruni, and Claire Renzetti for their careful readings and judicious editing.

http://vaw.sagepub.com

advocacy on behalf of its victims. New laws, police procedures, and medical and forensic efforts to collect and preserve evidence have all encouraged prosecution; at the same time, refuges and shelters for battered women and education and therapy groups for men who are violent toward their partners have sought to transform the conditions that have traditionally supported and sustained domestic violence (see, for example, McNeely & Jones, 1980; Pence & Paymar, 1993).

In recent years, a serious debate has erupted among activists, activist organizations, and individuals about the nature of domes- tic violence and, especially, the gender of the perpetrators. Decades after first bringing the problem to public awareness, fem- inist activists now confront a growing chorus of researchers and political activists who claim that women and men are victimized by domestic violence in roughly equal numbers (see Pleck, Pleck, Grossman, & Bart, 1978; Schwartz & DeKeseredy, 1993; Steinmetz, 1978; Straton, 1994).

Despite numerous studies that report the preponderance of domestic violence to be perpetrated by men against women, there are also now more than 100 empirical studies or reports that sug- gest that rates of domestic violence are equivalent (see, for exam- ple, Archer, 2000; Fiebert, 1997.) In the United States, numerous studies have found that women and men are equally likely to report to researchers that they have hit their partners during the preceding 12 months. In Great Britain also, 4.2% of women and men said they had been physically assaulted by partners during the previous 12 months (Tendler, 1999).

Thus, activists for “men’s rights” have suggested that policy- oriented efforts for women have been misplaced because they focus entirely on women as the victims of domestic violence. Instead of the picture painted by feminist researchers and activ- ists, these activists argue, as one writer put it, “Men are the victims of domestic violence at least as often as women” (Brott, 1994). Domestic violence, they argue, exhibits gender symmetry; that is, an equal number of women and men are its victims.

Although such activists draw our attention to the often- ignored problem of men as victims of domestic violence, their efforts are also often motivated by a desire to undermine or dis- mantle those initiatives that administer to female victims. To many of these advocates of gender symmetry, compassion is a

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1333

http://vaw.sagepub.com

zero-sum game, and when we show any compassion for women who are the victims of domestic violence, we will never address the male victims.

These apparent discrepancies between claims of gender sym- metry and claims of dramatic asymmetry have led to significant confusion among policy makers and the general public. Is domes- tic violence a “women’s issue,” or do equivalent rates indicate that domestic violence is a problem shared by women and men equally or even not a problem at all? In this article, I examine the claims of gender symmetry in domestic violence. I review existing sources of data on domestic violence and suggest why the rates of domestic violence appear so varied. I offer some ways to under- stand and reconcile these discordant data so that both scholars and policy makers alike may acknowledge the male victims of domestic violence within the larger context of domestic violence. In particular, I argue that claims about gender symmetry exclude a thorough analysis of gender and how gender identity and ideol- ogy—the cultural definitions of masculinity and femininity— may help to clarify these seemingly discordant claims.

THE IDEA OF GENDER SYMMETRY

Reports of gender symmetry have come to play a significant role in public and media discussions of domestic violence. Because these reports run counter to existing stereotypes of male- female relationships, they often have the headline-grabbing value of a “man bites dog” story. One review of the literature (Fiebert, 1997) found 79 empirical studies and 16 reviews of literature that demonstrated gender symmetry among couples. In a more recent meta-analytic review of this literature, Archer (2000) looked at 82 studies that found gender symmetry.

These empirical studies raise troubling questions about what the public is thought to “know” to be true of domestic violence: It is something men overwhelmingly “do” to women and not the other way around, it is among the leading causes of serious injury to women every year; and worldwide, men’s violence against women is one of the world’s most widespread public health issues.

The questions these studies raise are indeed troubling, but the questions they themselves ask are far from clear. For example,

1334 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

does gender symmetry mean that women hit men as often as men hit women? Or does it mean that an equal number of men and women hit each other? Does symmetry refer to men’s and women’s motivations for such violence, or does it mean that the consequences of it are symmetrical? These questions are often lumped together in reviews of literature and meta-analyses, which review existing data sets.

The two large-scale reviews of literature that demonstrate gen- der symmetry are useful indicators of the types of evidence offered and arguments made by their proponents (Archer, 2000; Fiebert, 1997). Of the 79 empirical articles that Fiebert reviewed, 55 used the same empirical measure of family conflict, the Con- flict Tactics Scale (CTS), as the sole measure of domestic violence. This scale was also used in 76 out of the 82 studies that Archer examined. In addition, 28 of those studies noted by Fiebert dis- cussed samples composed entirely of young people—college stu- dents, high school students, or dating couples younger than 30— and not married couples. (These two groups overlap somewhat, as 13 of those studies of young, dating couples also used the CTS.) I will discuss the CTS at some length below and also examine some of the reasons that studies of college-age and dating couples yield different rates of violence and aggression than do studies of somewhat older married couples.

Of the remaining nine studies in Fiebert’s (1997) survey that used neither the CTS nor sampled only young, dating, unmarried couples, two were based on people’s perceptions of violence but offered no data about violence itself, whereas another was based on reports of witnessing violence that contained no useful data (Fiebert, 1997). Another was a study of spousal homicide that did not include homicides by ex-spouses (to which I shall also give some attention). One was a study of young people that had no comparisons by gender (Elliott, 1997), and one was based on vio- lence in American comic strips in 1950 (Saenger, 1963).

Of the three remaining studies, two were based on clinical sam- ples undertaken by my colleagues (O’Leary et al., 1989; Tyree & Malone, 1991). Although these studies suggest that couples that seek clinical therapeutic help have high rates of mutual aggres- sion, O’Leary (1999, 2000) has insisted that the age of the individu- als dramatically changes the data and that clinical samples cannot necessarily be generalized to a national population. Even so, as

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1335

http://vaw.sagepub.com

Fiebert (1997) noted, the study by Tyree and Malone found that women’s violence was a result of a “desire to improve contact with partners” (p. 11), by which they meant that women tended to slap or push their partners to get them to pay attention but not to hurt them.

It would appear, therefore, that Gonzalez’s (1997) unpublished master ’s thesis, written apparently under Fiebert’s supervision, is the only quantitative survey that purports to find gender symme- try without relying on the CTS. Although it may be of interest that most of the women said their violence was a “spontaneous reac- tion to frustration,” Gonzalez did not survey men nor administer the same questionnaire to a sample of men; thus, one can make no inferences whatsoever about gender symmetry.

Fiebert’s (1997) scholarly annotated bibliography thus turns out to be far more of an ideological polemic than a serious schol- arly undertaking. But because it has become a touchstone for those who support a gender symmetry analysis, it is important to consider the studies on which it is based. Despite the vituperative ideological debates, there are serious and credible social science researchers who have used reliable social science and found gen- der symmetry. As follows, I examine (a) the CTS, especially what it measures and what it does not measure; and (b) the effects of age and marital status on domestic violence.

Those who insist on gender symmetry must also account for two statistical anomalies. First, there is a dramatic disproportion of women in shelters and hospital emergency care facilities. Why is it that when we begin our analysis at the end point of the domes- tic violence experience—when we examine the serious injuries that often are its consequence—the rates are so dramatically asymmetrical? Second, claims of gender symmetry in marital vio- lence must be squared with the empirical certainty that in every single other arena of social life, men are far more disproportion- ately likely to use violence than are women. Why are women so much more violent in the home that their rates approach, or even exceed, those of men, whereas in every other nondomestic arena, men’s rates of violence are about nine times those of women (on rates of violence generally, see Kimmel, 2000)?

1336 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

HOW DO WE KNOW WHAT WE KNOW: TYPES OF DATA

Our understanding of domestic violence has relied on a wide variety of evidence, from clinical observations to narrative accounts of victims and batterers, the experiences of advocates, and qualitative data gleaned from police and medical sources. Large-scale surveys have fallen into two distinct types (see, for example, Bachman, 1998, 2000; Nazroo, 1995; see also Walby, 1999). These are crime victimization studies, which rely on large- scale aggregate data on crime victimization, and family conflict studies, which measure the prevalence of aggression between married or cohabiting couples. These two sources of data find very different rates of domestic violence, in part because they are measuring two different things.

CRIME VICTIMIZATION STUDIES

Data about crime victimization are gathered from a variety of sources. Some are obtained from household surveys, such as the National Violence Against Women Survey (NVAW), sponsored by the National Institute of Justice and the Centers for Disease Control and Prevention (see Tjaden & Thoennes, 1998, 2000a, 2000b). This nationally representative sample surveyed 8,000 women and 8,000 men representing 16,000 households in the United States. Other crime studies are compiled from police sta- tistics, the National Crime Survey, and the National Crime Victim- ization Study in which 60,000 households are surveyed annually. Police data typically rely on calls to domestic violence hot lines or calls to police departments.

Crime victimization studies have large sample sizes, in part because they are funded by national, state, and local government agencies. Crime victimization studies include a wide range of assaults, including sexual assault, in their samples. They ask not only about assaults by current partners (spouses or cohabiting partners) but also about assaults by ex-spouses or ex-partners. But they ask only about those events that individuals experience—or even report to municipal authorities—as a crime, and therefore

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1337

http://vaw.sagepub.com

miss those events that are neither perceived nor reported as crimes. They also find significantly lower rates of domestic vio- lence than do family conflict studies, ranging from significantly less than 1% to about 1.1% of all couples. Some reasons that they find lower rates of violence are that crime victimization studies include all individuals in a household older than age 12, although rates of domestic assault are far lower for women older than age 65 and between 12 and 18. All family members are interviewed, which also may prevent some respondents from disclosing inci- dents of violence out of fear of retaliation (for a summary of these findings, see DeKeseredy, 2000; Gelles, 2000; Straus, 1999).

These studies uniformly find dramatic gender asymmetry in rates of domestic violence. The National Crime Victimization Sur- vey found women reported six times as many incidents of vio- lence by intimates as men did in 1992 and 1993 (Bachman & Saltzman, 1995; see also Dawson & Langan, 1994). The NVAW found that in 1998, men physically assaulted their partners at three times the rate at which women assaulted their partners (Tjaden & Thoennes, 2000b).

Crime victimization studies further find that domestic violence increases in severity over time, so that earlier, “moderate” vio- lence is likely to be followed by more severe violence (Johnson & Ferraro, 2000). This emerges also in discussions of spousal homi- cide, where significant numbers of women killed by their spouses or ex-spouses were also earlier victims of violence (see Browne, Williams, & Dutton, 1999; Dugan, Nagin, & Rosenfeld, 1999). In sum, crime victimization studies typically find that domestic vio- lence is rare, serious, escalates over time, and is primarily perpe- trated by men.

FAMILY CONFLICT STUDIES

By contrast, family conflict studies are based on smaller scale nationally representative household surveys such as the National Family Violence Survey (Straus & Gelles, 1990) or the National Survey of Families and Households and the British and Canadian national surveys. These surveys interview respondents once and ask only one partner of a cohabiting couple (older than 18) about their experiences with various methods of expressing conflict in the family. Other survey evidence comes from smaller scale

1338 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

surveys of college students or dating couples, and some draw from clinical samples of couples seeking marital therapy. Still other data are drawn from convenience samples of people who responded to advertisements for participants placed in newspa- pers and magazines. According to Fiebert (1997), the total number of respondents for all studies that find gender symmetry is slightly more than 66,000—that is, slightly more than the single annual number of one of the crime victimization studies in any year.

These surveys both expand and contract the types of questions asked respondents compared to crime victimization studies. On one hand, they ask about all the possible experiences of physical violence, including those that are not especially serious or severe and that do not result in injury—that is, those that might not be reported or even considered a crime. On the other hand, they ask questions only about cohabiting couples (and therefore exclude assaults by ex-spouses or ex-partners) and exclude sexual assault, embedding domestic violence within a context of “family con- flict.” So, for example, the CTS asks respondents about what hap- pens “when they disagree, get annoyed with the other person, or just have spats or fights because they’re in a bad mood or tired or for some other reason” (Straus, 1997, p. 217).

Family conflict studies tend to find much higher general rates of domestic violence than do crime victimization studies—typi- cally, about 16% of all couples report some form of domestic vio- lence (Straus, 1990). One summary of 21 of the approximately 120 studies that have explored family conflict found that about one third of men and two fifths of women indicated using violence in their marriages (Sugarman & Hotaling, 1989). As surprising as it may be to see high levels of violence, the most surprising finding has been the gender symmetry in the use of violence to try to resolve family conflicts; as Fiebert (1997) wrote, “Women are as physically aggressive, or more aggressive, than men in their rela- tionships” (p. 273).

These studies also find much lower rates of injury from domes- tic violence, typically about 3% (Stets & Straus, 1990). When “minor” forms of injury (such as slapping, pushing, and grab- bing) are excluded from the data, the yearly incidence falls signifi- cantly from 16% noted previously to around 6% of all couples (Straus & Gelles, 1986). They also find that violence is unlikely to

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1339

http://vaw.sagepub.com

escalate over time (see Johnson & Ferraro, 2000). In sum, then, family conflict studies tend to find high rates of domestic vio- lence, stable levels of severity, and low rates of injury and find it perpetrated equally by women and men.

How are such different conclusions to be reconciled? A first step is to make the sources of data similar and make sure they are asking similar questions and comparing the same sorts of events. Crime victimization studies rely on two types of data: surveys of national probability samples that are representative of the popu- lation at large and “clinical” samples as well as calls to police and shelters and visits to emergency rooms. Family conflict studies are based on three sources of data: nationally representative prob- ability samples, clinical samples, and convenience samples based on responses to advertisements.

Nationally representative probability samples are the only sources of data that are consistently reliable and generalizable. Whereas clinical samples may have important therapeutic utility, especially in treatment modalities, they are relatively easy to dis- miss as inadequate empirical surveys because they do not offer control groups from the nonclinical population and therefore offer no grounds whatsoever for generalizability. Therefore, I shall omit further discussion of both types of clinical data— police, shelter, and emergency room data and data drawn from marital therapy cases.

Recruitment via ads in newspapers and magazines offers related problems of sample representativeness and therefore undermines efforts at generalizability. Often people who respond to such ads respond because they have a “stake” in the issue and feel they want to contribute to it somehow. The representativness of such people to the general population is unclear at best.1

Virtually all the “family conflict” surveys rely on the CTS and CTS2, survey measures developed by New Hampshire sociolo- gist Murray Straus and his collaborators, so we must examine that scale a bit further.

THE CTS

Developed by Straus and his colleagues during the past two decades, the CTS is enormously useful, especially for eliciting the quotidian, commonplace acceptance of violence as a means to

1340 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

“communicate.” Let us begin our discussion where the CTS begins. Here is the opening paragraph to the survey as adminis- tered (Straus, 1990):

No matter how well a couple gets along, there are times when they disagree, get annoyed with the other person, or just have spats or fights because they’re in a bad mood or tired or for some other rea- son. They also use many different ways of trying to settle their dif- ferences. I’m going to read some things that you and your (spouse/ partner) might do when you have an argument. I would like you to tell me how many times . . . in the past 12 months you . . . (p. 33)

Such a framing assumes that domestic violence is the result of an argument, that it has more to do with being tired or in a bad mood than it does with an effort to control another person (for critiques of the CTS and CTS2 generally, see Brush, 1993; Currie, 1998; DeKeseredy & Schwartz, 1998a, 1998b, 1998c; Dobash, Dobash, Wilson, & Daly, 1992; Kurz, 1993; Okun, 1986).

The CTS asks about frequency, although only for 1 year. Asking how often in the past year either spouse or partner hit the other may capture some version of reality but does not capture an ongo- ing systematic pattern of abuse and violence over many years. This is akin to the difference between watching a single frame of a movie and the movie itself.

Context

The CTS simply counts acts of violence but takes no account of the circumstances under which these acts occur. Who initiates the violence, the relative size and strength of the people involved, and the nature of the relationship all will surely shape the experience of the violence but not the scores on the CTS. Thus, if she pushes him back after being severely beaten, it would be scored 1 conflict tactic for each. And if she punches him to get him to stop beating their children or pushes him away after he has sexually assaulted her, it would count as 1 for her and none for him.

In response, Straus and his various colleagues acknowledged that the context is important but believed that it is preferable to explore the context separately from the incidence. This response is unpersuasive. Imagine simply observing that death rates soared for men between ages 19 and 30 during a period of a few years

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1341

http://vaw.sagepub.com

without explaining that a country has declared war. Context matters.

Initiation

Some critics (see, for example, Currie, 1998; DeKeseredy & Schwartz, 1998a, 1998c) have argued that simply asking how many times a person or his or her spouse used a series of conflict tactics is inadequate to measure the initiation of the violence. Straus (1993) argued that using the CTS, initiation is about even and that self-defense is not the motivation for most women. Straus found that women initiated in 53% of the cases; in 42%, they reported their husbands initiated the aggression, and about 3% said they could not tell who initiated it. Data from other stud- ies, however, indicated that women were far more likely to use violence defensively against the aggression of their partners (DeKeseredy, 1997). With such discordant findings, the CTS’s value is limited unless there are a variety of measures incorpo- rated to adequately ascertain the motivation for violence.

Intention and Motivation

Asking people how often they used various conflict tactics dur- ing an argument assumes that people use violence expressively— that is, in the heat of anger, as a way to settle an argument, to get one’s point across, or to get a spouse or partner to listen or pay attention. It misses the way violence might be used instrumen- tally—to control or subdue and to reproduce subordination. Such an absence would be analogous to discussing rape and only focusing on those date and acquaintance rapes in which there had been some sexual foreplay and the boundaries were less than fully clear while ignoring, for example, rapes that ended in homicides, rape as a systematic policy of militarily subduing a population, rape in prison, and rape of strangers that has nothing to do with sexual ardor. In short, motivation for violence matters.

Does Location Matter? The Public-Private Split

In general, men are more aggressive than are women (for a summary of this research, see Kimmel, 2000). In fact, violence is the only behavioral variable for which there are intractable and

1342 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

overwhelmingly skewed results showing gender differences. Although gender differences on a host of other variables—such as spatial orientation and visual perception and academic achieve- ment and ability—have been demonstrated, these differences are typically quite small. Rates of violence based on gender, however, are large and consistent. In their path-breaking work, The Psychol- ogy of Sex Differences, Maccoby and Jacklin (1974) found that vio- lence exhibits the greatest gender variation; 20 years later, an anal- ysis by Baron and Richardson (1994) found the same thing. So we would have to ask why women would hit men inside the house in roughly equal numbers but almost never commit violence toward men—or women—outside the home.

Studies that propose gender symmetry must explain this apparent paradox. Some argue, for example, that women assume that their violence toward their male partners is harmless (see Fiebert & Gonzalez, 1997). Straus (1999) believed that slapping a man might actually be considered appropriately feminine behav- ior. It is likely that each of these has some validity, but neither addresses the motivation of women’s violence nor the context in which it occurs. Actually, most empirical research on female aggression points in a very different direction. For example, Bjorkqvist and Niemela (1992) found that women are as aggres- sive as men but only when they are in no danger of being recog- nized, that is, when the target is not a family member, and there is no danger of retaliation. When parties know each other, women’s violence tends to be defensive and men take the initiative (Adams, 1992). Obviously, domestic violence cannot fit the pat- tern of women retaining their anonymity.

Two final criticisms of the CTS, one methodological and one substantive, deserve somewhat fuller elaboration: the method- ological problem of memory and retrospective analysis and reporting bias.

Methodological problems invariably skew substantive results. The CTS relies on retrospection, asking people to accurately remember what happened during the past year. (It shares this method with crime victimization studies, and these biases may well extend to those studies as well.) Retrospection may not be completely reliable because memory often serves our current interests but is unlikely to provide an accurate rendition of what actually happened. There is some evidence that the gender

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1343

http://vaw.sagepub.com

symmetry of domestic violence breaks down when retrospective studies are used alone. Why?

One argument commonly made (see, for example, Brott, 1994) is that men would be likely to underestimate how often they were victimized because being hit by a woman is so emasculating that they would be too ashamed to admit it, wheras women would tend to overestimate how often they were hit because it might serve their interests to make false allegations of domestic assault in divorce or custody proceedings. Both of these assumptions turn out to be empirically groundless; in fact, the evidence points decidedly in the other direction.

Bringing Gender Into the Equation

What is missing, oddly, from these claims of gender symmetry is an analysis of gender. By this, I mean more than simply a tally- ing up of which biological sex is more likely to be perpetrator or victim. I mean an analysis that explicitly underscores the ways in which gender identities and gender ideologies are embodied and enacted by women and men. Examining domestic violence through a gender lens helps clarify several issues.

For example, both women and men tend to see their use of vio- lence as gender nonconforming, but the consequences of this non- conformity might lead women and men to estimate their use of violence and their victimization quite differently. Women are socialized not to use violence, and as a result, they tend to remem- ber every transgression. As Dobash, Dobash, Cavanagh, and Lewis (1998) wrote,

Women may be more likely to remember their own aggression because it is deemed less appropriate and less acceptable for women than for men and thus takes on the more memorable qual- ity of a forbidden act or one that is out of character. (p. 405)

Men, however, might find it emasculating to reveal that their assumed control over “their women” is so tenuous that they are forced to use violence to keep them “in line.” They may find it dif- ficult to admit that they cannot “handle” their wives. Thus, men might underestimate their violence, and women might tend to overestimate theirs.

1344 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

Furthermore, in addition to overestimating their own violence, women may also tend to underestimate their partners’ violence given the norms of domestic life, which frequently find women discounting, downplaying, normalizing, or even excusing their partners’ violent behavior because they (the women) “deserved” it. By the same token, in addition to underestimating their own violence, men may overestimate their partners’ violence for the same norms of masculinity. American men, at least, believe vio- lence is legitimate if used as retaliation for violence already com- mitted (see, for example, Kimmel, 1996; Mead, 1950). The expres- sion “having a chip on one’s shoulder” actually has its literal origin among young, southern White boys after the Civil War, who would place a piece of wood on their shoulder and dare someone to knock it off so they might legitimately fight and prove their manhood. Initiating violence is never legitimate according to the norms of traditional masculinity in America; retaliating against a perceived injustice with violence is always legitimate. As a result, men will tend to overestimate their victimization and women will tend to underestimate theirs (see also Archer, 1994; Bograd, 1990; Bowker, 1998).

In response to the notion that men would be too ashamed or humiliated to call the police or go to the hospital if they were beaten by their wives, available empirical evidence suggests a very different picture: Men who are assaulted by intimates are actually more likely to call the police, more likely to press charges, and less likely to drop them (Ferrante, Morgan, Indermaur, & Harding, 1996; Rouse, Breen, & Howell, 1988; Schwartz, 1987). This makes sense in the terms outlined previously, as women would be more likely to forgive being hit and normalize it with statements about how he really does love her. Another study found that men underreport the violence they perpetrate against women by 50% (Edleson & Brygger, 1986; see also Browning & Dutton, 1986; Brush, 1993; and especially Dobash et al., 1998). Dobash et al. found a useful measure of the gender asymmetry in reporting: Women’s narrative descriptions of the events of their experiences are far longer and more richly detailed, entering the narrative at a much earlier point in the unfolding drama and e x t e n d i n g t h e n a r r a t i v e t o i n c l u d e i n j u r i e s a n d o t h e r consequences.

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1345

http://vaw.sagepub.com

If men underestimate their own violence and overestimate their victimization while women overestimate their own violence and underestimate their victimization, this would have enor- mous consequences in a survey that asks only one partner to recall accurately how much they and their spouses used various conflict-resolution techniques.

The Causes and Consequences of Violence: Severity and Injury

A final substantive critique of the CTS is that it does not mea- sure the consequences of physical assault (such as physical or emotional injury) or the causes of the assault (such as the desire to dominate). Straus (1997) responded that assessing causes and consequences may be interesting, but it is not a necessary part of the picture. He scolded his critics saying that to fault his research on this question “is akin to thinking that a spelling test is inade- quate because it does not measure why a child spells badly, or does not measure possible consequences of poor spelling, such as low self-esteem or low evaluations by employers” (Straus, 1997, p. 218).

Were Straus not a credible social scientist, one might suspect the reply to be disingenuous. As such, it is simply inadequate. It is more akin to a teacher who does not look at how far off the spell- ing mistakes are or whether there is a pattern in the mistakes that might point to a physiological problem such as dyslexia or some other learning disability rather than academic laziness and thus leaves the learning problems untouched and misdirects funds away from remediation toward punitive after-school programs for lazy students. And even that analogy is imperfect because, unlike spelling, domestic violence is not about what happens to the perpetrator (the poor speller) but to someone else. Can one imagine any other issue in which causes and consequences are thought to be irrelevant?

The consequences of violence raise perhaps the most telling criticism of the CTS—a criticism, not incidentally, that Straus and his more thoughtful collaborators share. The CTS lumps together many different forms of violence so that a single slap may be equated with a more intensive assault. In the NVAW, for example, lifetime percentages of persons physically assaulted by intimate partners found dramatic differences in some types of assault but

1346 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

not others. For example, just less than 1% of men and women (0.9% of women and 0.8% of men) said their attackers used a knife in the attack, but 3.5% of women and only 0.4% of men said their partners threatened to use a gun; 0.7% of women and 0.1% of men said their spouses actually did use a gun (Tjaden & Thoennes, 1998). (It is interesting to note that these differences inside the home are actually slightly smaller than the differences outside the home, where men are overwhelmingly more likely to use weap- ons in an attack.)

Even more telling are the gender disparities in serious physical injuries without weapons. For example, in a British study that found equal rates of reporting victimization of violence, there were no injuries at all reported in 59% of incidents that involved pushing, shoving, and grabbing (these are the behaviors more typically reported being committed by women than by men). In the NVAW Survey (a crime victimization–type study), half the number of men than women (4.4% of men and 8.1% of women) said their partners threw something at them, and three times as many women (18.1% of women and 5.4% of men) said their part- ners pushed, grabbed, or shoved them or that their partners slapped or hit them (16.0% of women and 5.5% of men). But more than 10 times as many women (8.5% of women and 0.6% of men) reported their partner “beat them up” (Tjaden & Thoennes, 1998).

The consequences of violence range from minor to fatal, and these are significant in understanding domestic violence in gen- eral and its gendered patterns. Far more men than women kill their spouses (and, of course, “couples” in which one spouse killed the other could not participate in the CTS studies because both partners must be cohabiting at the time of the study). And rates of homicides of ex-spouses are even more gender asymmet- rical. According to the FBI, female victims represent about 70% of all intimate homicide victims (see Bachman, 2000). About one third of all female homicide victims in the United States were killed by intimate partners compared with 4% of male homicide victims (see, for example, Bachman & Saltzman, 1995; Kellerman & Mercy, 1992). (What this suggests, of course, is that both women and men are more likely to be killed by men; efforts to end all types of violence should properly focus on the association of mas- culinity and violence, the legitimacy of violence to men, and men’s sense of entitlement to use violence.) In the United States,

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1347

http://vaw.sagepub.com

the number of men killed by intimate partners has dropped by 69% since 1976. The number of women killed by intimates was rel- atively stable until 1993 when it too began to drop, but only by about 15% (Fox & Zawitz, 2001).

Gender symmetry tends to be clustered entirely at the lower end of violence (Dobash et al., 1998). According to some data, women are six times more likely to require medical care for inju- ries sustained by family violence (Kaufman Kantor & Straus, 1987; Stets & Straus, 1990). Straus (1997) also reports that in family conflict studies, the injury rate for assaults by men is about seven times greater than the injury rate for assaults by women (Stets & Straus, 1990). This dramatic difference in rates of injury, found in both types of studies, leads Straus, the creator of the CTS and the researcher who is most often cited by those claiming gender sym- metry, to write that

although women may assault their partners at approximately the same rate as men, because of the greater physical, financial, and emotional injury suffered by women, they are the predominant victims. Consequently, the first priority in services for victims and in prevention and control must continue to be directed toward assaults by husbands. (p. 219)

Straus also understands that women, on average, suffer much more frequent and more severe injury (physical, economic, and psychological) than do men (see also Stets & Straus, 1990; Straus, Gelles, & Steinmetz, 1980).

These different rates of injury are so pronounced that when injury data have been obtained in studies using the CTS, the rate of violence drops to that predicted by crime victimization studies and the gender asymmetry of such studies is also revealed (see Straus, 1997). This led another researcher to conclude that both husbands and wives may be said to be “aggressive,” but many more husbands are “violent” (Frude, 1994, p. 153).

Age and Aggression

The CTS measures family conflict in intact partnerships between either cohabiting or married couples. However, as I have previously mentioned, more than one third of the studies noted

1348 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

by Fiebert (1997) that found gender symmetry were surveys of college age, dating couples who were not cohabiting. About one half of Archer ’s (2000) samples in his meta-analytic review involved high school or college students. Therefore, it is impor- tant to examine the way age exerts an effect on domestic violence.

According to all available research, age, especially being youn- ger than 30, is a strong predictor of partner violence (see Suitor, Pillemer, & Straus, 1990). O’Leary et al. (1989) have consistently found that age is a significant variable in the distribution of part- ner violence. Rates of violence rise significantly between ages 10 (less than 2% violent) and 25 when levels peak at 35% of all cou- ples. But after 25, rates begin to drop and keep dropping to return to about 5% by age 75. This suggests that younger couples are most likely to have the highest rates of violence (O’Leary, 1999). The National Survey of Adolescents in the United States found that of 22.3 million adolescents (ages 12 to 17), 1.8 million had been victims of what researchers label serious sexual assault and 3.9 million had been victims of serious physical assault. Women were four times more likely to have been sexually assaulted (13% com- pared with 3.4% of men), and young men were significantly more likely to have been physically assaulted (21.3% compared with 13.4% of young women; Kilpatrick & Saunders, 1997, 2000). This is because violence means different things to younger dating cou- ples than to married couples at midlife, when violence is usually associated with significant marital discord (O’Leary, 1999, 2000). The two populations—young, unmarried dating couples and older married couples at midlife—are so dissimilar that results from one population cannot be generalized to the other.

Younger people also report using only a few of the various forms of conflict: pushing and slapping. These are not typically associated with injury or with fear of partner (O’Leary, 2000). Stets and Pirog-Good’s (1990) work on the centrality of control in dat- ing violence also helps explain the relationship of age and gender on nonspousal violence. It is possible that men’s rates of violence drop after marriage because they establish their (financial, physi- cal, emotional) control over the relationship, and therefore, overt acts of violence are less necessary as long as the threat of violence is present (Stark & Flitcraft, 1988).

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1349

http://vaw.sagepub.com

What the CTS Leaves Out

It is important not only to understand what the CTS measures but also to make explicit what it does not measure. First, the CTS does not include sexual assault in its definition of family conflict. This is crucial because a significant number of spousal assaults are sexual assaults. The NVAW found that 7.7% of all female respon- dents had been raped by intimate partners at some point in their lifetimes; this translates into approximately 201,394 U.S. women who are raped by intimate partners each year (Tjaden & Thoennes, 2000c). Yet, Straus and Gelles (1990) did not include rape as a category in the index.

Second, the CTS only includes violence by current spouses or cohabiting partners. It does not include violence by ex-spouses or partners. Crime victimization studies do include these. This is important because crimes by former spouses comprise a signifi- cant number of domestic assaults. It may be that when women exit relationships, they have no “need” for violence, whereas men tend to continue or even escalate their use of violence when women leave. The National Crime Victimization Survey found that rates of intimate-perpetrated violence for separated women are more than eight times higher than rates for married women (Bachman & Saltzman, 1995; see also Greenfeld et al., 1998). It may be true that these might be somewhat overrepresented in crime victimization studies because people who are assaulted by former spouses would be more likely to report that as a crime because for- mer spouses clearly had no “right” to aggress against victims; therefore, it would clearly be seen as a crime and more likely to be reported. But to ignore these data would so skew any study as to make it unreliable. For example, in one Australian study, only 1% of all violent victimizations of men involved ex-spouses or ex- partners, but ex-spouses or ex-partners involved fully one third of all female incidents (Ferrante et al., 1996). Failure to include ex- spouses may fail to capture up to one third of all cases.

Failure to include sexual assault and assaults by ex-spouses or ex-partners compounds the problem that the CTS does not ade- quately measure rates of serious injury from domestic violence.2

The NVAW Survey (Tjaden & Thoennes, 1998) found that 72.6% of rape victims and 66.6% of physical assault victims sustained inju- ries such as scratches, bruises, or welts and that 14.1% of rape vic- tims and 12.2% of physical assault victims sustained broken bones

1350 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

or dislocated joints. Rape victims were far more likely to sustain internal injuries (5.8% to 0.8%) or chipped or broken teeth (3.3% to 1.8%). On the other hand, physical assault victims were more likely to sustain lacerations or knife wounds (16.9% to 6.2%), head or spinal cord injuries (10.1% to 6.6%), and burns and bullet wounds (0.7% and 1.8%, respectively; rape rates too low to esti- mate; Tjaden & Thoennes, 1998).

Violence by ex-husbands also tends to be more serious. For example, the risk of spousal homicide goes up by about 50% for women who leave abusive husbands. (This may also help explain the “rationality” in the decision by women to stay in abusive rela- tionships.) Men may kill their ex-wives because their ex-wives left them; women may kill their ex-husbands because they believe their ex-husbands will otherwise kill them for leaving. In both cases, then, the larger context for both women’s and men’s vio- lence is men’s violence. One study of spousal homicide (Barnard, Vera, Vera, & Newman, 1982) found that more than half of all defendants were separated from their victims at the time they were accused of committing the murder. (For more on relation- ship status and violence, see DeKeseredy, 1997.)

In sum, the gender symmetry found by CTS-based studies result from the omission of severity of injury, sexual assault, and assaults by former spouses. Some fail to adequately account for marital status and age. Including these would certainly make the gender asymmetry of domestic violence more clear.

HOW CAN WE UNDERSTAND THE USE OF AGGRESSION IN DOMESTIC LIFE?

These two different types of studies—crime victimization stud- ies and family conflict studies—rely on two different theoretical perspectives and two different sources of data. They measure two different phenomena based on different conceptualizations of aggression in families. But they can be reconciled conceptually and methodologically.

If one is interested in the level of aggression in family conflict— that is, the likelihood of any type of aggression occurring when a couple has an argument—then the CTS may be somewhat useful. I say “somewhat” because among other problems that I have pre- viously outlined, the utility of the CTS is limited by the fact that it

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1351

http://vaw.sagepub.com

fails to take into account sexual assault and also assault by ex- spouses. But it does enable us to see the overall amount of a partic- ular kind of violence in families, what we might call expressive violence—the way individuals might express anger, frustration, or loss of control. If, however, one were interested in the ways in which partners use violence not expressively but instrumentally to achieve some end of control, injury, or terror, then the CTS would be a poor measure. Then, crime victimization surveys will be more valuable because these measure serious injury and include sexual assault and assaults by ex-spouses in their pur- view. These surveys may capture those family conflicts in which the level of violence escalates beyond a mere “conflict tactic” to something far more ominous and perhaps lethal.

Some violence by men against women is motivated not by the desire to express anger, frustration, or some other immediate emotion during a family conflict but may be more instrumentally motivated by the desire to control. However, the use of violence may indicate not the experience of control but the experience of loss of control. “Violence is a part of a system of domination,” writes R. W. Connell (1995), “but it is at the same time a measure of its imperfection” (p. 84).

In that sense, we might say that many men who assault their partners or ex-partners are using violence when they fear their control is breaking down, their ability to control their partners by the implicit threat of violence is compromised, and they feel com- pelled to use explicit violence to “restore” their control. Thus, men see their violence as restorative and retaliatory.3 For example, in an earlier study, Dobash and Dobash (1979) found three anteced- ents of men’s use of violence: their sexual jealousy, their percep- tion that the women failed to perform a household task such as cleaning or preparing a hot meal, and women’s challenging the men’s authority on financial matters. All of these are indicators of a breakdown of men’s expected dominance and control.

This understanding of control-motivated, instrumental vio- lence is particularly important in our understanding of claims of gender symmetry. For one thing, men’s control over women has clearly broken down when their spouses have left them; thus, measures of physical assault that do not include assaults by ex- spouses will entirely miss these events. Second, breakdowns of men’s control over women may be revealed not by physical

1352 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

assault but by women’s withholding or refusing of sexual inti- macy. They may exert what limited power they may have by attempting to refuse their partners’ sexual advances. Thus, mea- sures that do not include sexual assaults among acts of aggression will be equally inadequate to measure the problem.

Control-motivated instrumental violence is experienced by men not as an expression of their power but as an instance of its collapse. The men may feel entitled to experience that control over women, but they become violent when they do not feel that con- trol. Masculinity, in that sense, has already been compromised; violence is a method to restore manhood and domestic inequality at the same time (see, for example, Kimmel, 1994, 1996, 2000). Such control-motivated, instrumental violence is more likely to escalate over time, less likely to be mutual, and more likely to involve serious injury. This difference between expressive and instrumental violence is a difference not only in purpose but also in frequency, severity, and initiation. It addresses whether the vio- lence is part of a systematic pattern of control and fear or an iso- lated expression of frustration or anger. These two types of vio- lence are so different that Johnson (1995; Johnson & Ferraro, 2000) has come to call instrumental violence “intimate terrorism” and the types of expressive violence measured by the CTS as “com- mon couple violence.”

Social control–motivated abuse can be illustrated in another form of domestic violence: stalking. Control-motivated abuse refers to intentionally inflicted physically or psychologically painful or hurtful acts (or threats) by individuals as a means of compelling or constraining the conduct, dress, or demeanor of their partners (Ellis & Stuckless, 1996). Rates of stalking by inti- mates, more prevalent than previously thought, can best be understood as an effort to restore control or dominance after part- ners have left. Stalking exhibits dramatic gender asymmetry: Nearly 5% of American women and about one half of 1% (0.6%) of men report being stalked by current or former intimate partners at some time in their lives (Tjaden & Thoennes, 2000a).

Claims about the gender symmetry of conflict-motivated expressive violence must be complemented with claims about the dramatic gender asymmetry in control-motivated instrumental violence. When these two are factored together, it is clear that women and men may express their anger or frustration during an

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1353

http://vaw.sagepub.com

argument more equally than we earlier thought. This, however, is by no means fully symmetrical, because even the CTS leaves out two of the dominant forms of expressive, conflict-motivated aggression—sexual assault and assault by ex-spouses. And when control-motivated instrumental violence is added—the violence that more typically results in serious injury, is more systematic, and is independent of specific “conflict” situations—the gender asymmetry is clear (see Johnson, 2000).

WHY WE SHOULD BE CONCERNED ABOUT WOMEN’S VIOLENCE TOWARD MEN

Despite the evidence that gender symmetry is largely a myth, we should nonetheless be concerned about women’s violence for a variety of reasons. For one thing, compassion for victims of vio- lence is not a zero-sum game. Reasonable people would naturally want to extend compassion, support, and interventions to all vic- tims of violence. It is an indication of the political intentions of those who argue for gender symmetry that they never question the levels of violence against women, only that the level of vio- lence against men is equivalent. Their solution, however, is not more funding for domestic violence research and intervention but to decrease the amount of funding that women receive, although they never challenge the levels of violence against women.

Second, acknowledging women’s capacity for intimate vio- lence will illuminate the gender symmetry in intimate violence among gay men and lesbian couples. According to the NVAW Survey, slightly more than 11% of women living with same-sex partners report being raped, physically assaulted, or stalked by a female cohabitant (compared with 30.4% of women with a live-in male partner). About 15% of men living with male live-in partners report having experienced violence (compared with 7.7% of men with female live-in partners; Tjaden & Thoennes, 2000b).

Third, perhaps ironically, examining women’s violence can better illuminate the dynamics of men’s aggression against women. Because women’s violence is often retaliatory or commit- ted in self-defense, it may help to expose some of the ways men use violence to control women and women’s perceived lack of options except “fighting back.”

1354 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

Fourth, acknowledging assaults by women is important, Straus (1997) wrote, because they “put women in danger of much more severe retaliation by men” (p. 210). In a recent interview, Straus elaborated that because women generally suffer greater fear and more injuries, “when she slaps, she sets the stage for him to hit her. The safety of women alone demands we make a big deal of women hitting men” (Slobodian, 2000, p. 1).

Finally, men actually benefit from efforts to reduce men’s vio- lence against women. It turns out that efforts to protect women in the United States have had the effect of reducing the homicide rate of men by their partners by almost 70% during the past 24 years. According to James Alan Fox, a professor of criminal justice at Northeastern University, homicides by women of their spouses, ex-spouses, or boyfriends have steadily declined from 1,357 in 1976 to 424 in 1999 (Elsner, 2001). Fox attributes this decline to the availability of alternatives for battered women. “We have given women alternatives, including hotlines, shelters, counseling and restraining orders. Because more battered women have escape routes, fewer wife batterers are being killed,” Fox told reporters (as cited in Elsner, 2001). A 1999 study by the National Consor- tium on Violence Research found that the greater availability of hotlines and other resources for battered women, the greater was the decline in homicide of their male partners. (The study found that 80% of these male domestic homicide victims had abused their partners and that nearly two thirds of female homicide vic- tims had been abused before they were killed.) It turns out that those very initiatives that have greatly benefited women—shel- ters, hotlines, and the like—save men’s lives as well.

TOWARD AN INCLUSIVE EXPLANATION OF DOMESTIC VIOLENCE

It is certainly possible and politically necessary to acknowledge that some women use violence as a tactic in family conflict while also understanding that men tend to use violence more instru- mentally to control women’s lives. Furthermore, these two types of aggression must also be embedded within the larger frame- work of gender inequality. Women’s violence toward male part- ners certainly does exist, but it tends to be very different from that of men toward their female partners: It is far less injurious and less

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1355

http://vaw.sagepub.com

likely to be motivated by attempts to dominate or terrorize their partners (see, for example, Kaufman Kantor & Jasinski, 1998).

The different types of data sources, family violence studies and crime victimization studies, each point to different problems and each is useful to develop intervention strategies. As Straus (1999) wrote, “Research using a broad definition [of violence] and emphasizing injury may be most useful for informing programs designed to treat offenders or help victims of repeated severe assault” (p. 39). On the other hand, “research focusing on the act of assault, most of which does not involve injury but does involve millions of couples, may be most useful in informing programs of ‘primary prevention’ i.e., steps that will prevent physical assaults from ever happening” (p. 39). He concluded,

I believe humanity needs research inspired by the moral agenda and perspective of those who focus on the oppression of women, regardless of whether the oppression is physical, sexual, psycho- logical, or economic; and also research inspired by the moral agenda of those who focus on physical assault, regardless of whether the assault is by a man, woman or child. (p. 40)

Coupled with studies of parental violence toward children— which routinely find that more than 90% of parents aggress against their children—family conflict studies are useful in point- ing out the ubiquity and the casualness with which violence struc- tures our daily lives. Coupled with data about intimate partner homicide, rape, and other forms of sexual assault, crime victim- ization data are useful in pointing out the ways in which men’s domination over women requires the implicit threat and often the explicit instrumental use of violence to maintain that power.

Claims of gender symmetry are often made by those who do not understand the data: what the various studies measure and what they omit. Others make claims of gender symmetry based on disingenuous political motives, attempting to discredit women’s suffering by offering abstract statistical equivalences that turn out to be chimerical. Gelles and Straus (1999) themselves understood the political misuses to which their work has been put and strongly disavow those political efforts. In a summary of their work, they wrote,

1356 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

Perhaps the most controversial finding from our 1975 National Family Violence Survey was the report that a substantial number of women hit and beat their husbands. Since 1975 at least ten addi- tional investigations have confirmed the fact that women hit and beat their husbands. Unfortunately the data on wife-to-husband vio- lence has been misreported, misinterpreted, and misunderstood [italics added]. Research uniformly shows that about as many women hit men as men hit women. However, those who report that husband abuse is as common as wife abuse overlook two important facts. First, the greater average size and strength of men and their greater aggressiveness means that a man’s punch will probably produce more pain, injury and harm than a punch by a woman. Second, nearly three-fourths of the violence committed by women is done in self-defense. While violence by women should not be dismissed, neither should it be overlooked or hidden. On occasion, legislators and spokespersons . . . have used the data on violence by wives to minimize the need for services for battered women. Such arguments do a great injustice to the victimization of women [italics added]. (p. 424)

Gelles (2000) underscored this disingenuous political use of their work with this clear and unequivocal statement: “It is categori- cally false to imply that there are the same number of ‘battered’ men as battered women.” (Note how he even puts the word bat- tered in quotations when describing men.) It is not surprising that credible researchers disavow the political ends to which their work is often put.

Despite the dramatic differences in frequency, severity, and purpose of the violence, we should be compassionate toward all victims of domestic violence. There are some men who are bat- tered by their female partners, and these men are no less deserv- ing of compassion, understanding, and intervention than are women who are battered. And male domestic violence victims deserve access to services and funding, just as do female domestic violence victims. They do not need to be half of all victims to deserve either sympathy or services.

But just as surely, compassion and adequate intervention strat- egies must explore the full range of domestic violence: the differ- ent rates of injury, the different types of violence including sexual assault, and the likelihood of violence by ex-spouses. Such strate- gies must also understand the differences between violence that is an expression of family conflict and violence that is instrumental to the control of one partner over the other.

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1357

http://vaw.sagepub.com

With all the caveats and modifications I have suggested to the family conflict model and, especially, the CTS as the standard of measurement, I would therefore argue that violence as an expres- sion of family conflict is somewhat less than symmetrical but would include a significant percentage of women. I would hypothesize that including assaults and homicides by ex-spouses, spousal homicide, and sexual assault, the gendered ratio of male- perpetrated violence to female-perpetrated violence would be closer to 4:1.4 On the other hand, violence that is instrumental in the maintenance of control—the more systematic, persistent, and injurious type of violence—is overwhelmingly perpetuated by men, with rates captured best by crime victimization studies. More than 90% of this violence is perpetrated by men.

When sexual violence and violence by ex-spouses are consid- ered, the evidence is overwhelming that gender asymmetry in domestic violence remains in full effect. Men are more violent than women, both inside the home and in the public sphere. The home is not a refuge from violence, and it is not a site where gen- der differences in the public sphere are somehow magically reversed. As citizens, we need to be concerned about all victims of violence. We must also be aware that the perpetrators of that vio- lence—both in public and in private, at home or on the street, and whether the victim is male or female—are overwhelmingly men.

NOTES

1. In the best of these studies, O’Leary et al. (1989) found that about 31% of the men and 44% of the women indicated they had engaged in some aggression against their partners in the year before they were married. A year after marriage, rates dropped for both groups, and 27% of the men and 36% of the women indicated they had aggressed; 30 months into marriage, the rates for the previous year were 25% of the men and 32% of the women.

2. The Conflict Tactics Scale–2 does include a measure of sexual coercion, which seems to me a pretty cogent acknowledgement that it must be included in all understandings of gender symmetry.

3. It must be noted, of course, that the “retaliation” is more often for a perceived injury or slight than any real injury (see, for example, Beneke, 1982).

4. As this is a conjecture based on estimates, it remains an empirical question to coordi- nate the synthesis of these two approaches.

1358 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

REFERENCES

Adams, D. (1992). Biology does not make men more aggressive than women. In K. Bjorkvist & P. Niemela (Eds.), Of mice and women: Aspects of female aggression. San Diego: Academic Press.

Archer, J. (Ed.). (1994). Male violence. Boston: Routledge Kegan Paul. Archer, J. (2000). Sex differences in aggression between heterosexual partners: A meta-

analytic review. Psychological Bulletin, 126, 651-680. Bachman, R. (1998). Incidence rates of violence against women: A comparison of the redesigned

National Crime Victimization Survey and the 1985 National Family Violence Survey. Retrieved May 7, 2002, from VAWnet Web site: http://www.vawnet.org/vnl/library/ general/AR_incidenc.html

Bachman, R. (2000). A comparison of annual incidence rates and contextual characteristics of intimate-partner violence against women from the National Crime Victimization Survey (NCVS) and the National Violence Against Women Survey (NVAWS). Violence Against Women, 6, 839-867.

Bachman, R., & Saltzman, L. (1995). Violence against women: Estimates from the redesigned National Crime Victimization Survey. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.

Barnard, G. W., Vera, H., Vera, M., & Newman, G. (1982). Till death do us part: A study of spousal murder. Bulletin of the American Academy of Psychiatry and the Law, 10, 271-280.

Baron, R. A., & Richardson, D. (1994). Human aggression. New York: Plenum. Beneke, T. (1982). Men on rape. New York: St. Martin’s. Bjorkqvist, K., & Niemela, P. (Eds.). (1992). Of mice and women: Aspects of female aggression.

San Diego: Academic Press. Bograd, M. (1990). Why we need gender to understand human violence. Journal of Interper-

sonal Violence, 5, 132-135. Bowker, L. (Ed.). (1998). Masculinities and violence. Thousand Oaks, CA: Sage. Brott, A. (1994). Men: The secret victims of domestic violence. Retrieved January 15, 2002, from

http://www.vix.com/pub/men/battery/commentary/brott-hidden.html Browne, A., Williams, K., & Dutton, D. (1999). Homicide between intimate partners: A 20-

year review. In M. D. Smith & M. A. Zahn (Eds.), Homicide: A sourcebook of social research (pp. 149-164). Thousand Oaks, CA: Sage.

Browning, J., & Dutton, D. (1986). Assessment of wife assault with the Conflict Tactics Scale: Using couple data to quantify the differential reporting effect. Journal of Marriage and the Family, 48, 375-379.

Brush, L. D. (1993). Violent acts and injurious outcomes in married couples: Methodologi- cal issues in the National Survey of Families and Households. In P. B. Bart (Ed.), Violence against women: The bloody footprints (pp. 240-251). Newbury Park, CA: Sage.

Connell, R. W. (1995). Masculinities. Berkeley: University of California Press. Currie, D. H. (1998). Violent men or violent women? Whose definition counts? In R. K.

Bergen (Ed.), Issues in intimate violence (pp. 97-111). Thousand Oaks, CA: Sage. Dawson, J. M., & Langan, P. (1994). Murder in families. Washington, DC: Department of Jus-

tice, Bureau of Justice Statistics. DeKeseredy, W. S. (Ed.). (1997). Intimate relationship status variations in woman abuse

[Special issue]. Violence Against Women, 3(6). DeKeseredy, W. S. (2000). Current controversies on defining nonlethal violence against

women in intimate heterosexual relationships: Empirical implications. Violence Against Women, 6, 728-746.

DeKeseredy, W. S., & Schwartz, M. D. (1998a). Measuring the extent of woman abuse in inti- mate heterosexual relationships: A critique of the Conflict Tactics Scales. Retrieved May 7,

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1359

http://vaw.sagepub.com

2002, from the VAWnet Web site: http://www.vawnet.org/vnl/library/general/AR_ ctscrit.html

DeKeseredy, W. S., & Schwartz, M. (1998b). Woman abuse on campus: Results from the Cana- dian National Survey. Thousand Oaks, CA: Sage.

DeKeseredy, W. S., & Schwartz, M. (Eds.). (1998c). Rethinking violence against women. Thou- sand Oaks, CA: Sage.

Dobash, R. E., & Dobash, R. (1979). Violence against wives: A case against the patriarchy. New York: Free Press.

Dobash, R. P., Dobash, R. E., Cavanagh, K., & Lewis, R. (1998). Separate and intersecting realities: A comparison of men’s and women’s accounts of violence against women. Violence Against Women, 4, 382-414.

Dobash, R. P., Dobash, R. E., Wilson, M., & Daly, M. (1992). The myth of sexual symmetry in marital violence. Social Problems, 39, 71-91.

Dugan, L., Nagin, D. S., & Rosenfeld, R. (1999). Explaining the decline in intimate partner homicide: The effects of changing domesticity, women’s status, and domestic violence resources. Homicide Studies, 3, 187-214.

Edleson, J., & Brygger, M. (1986). Gender differences in reporting of battering incidences. Family Relations, 35, 377-382.

Ellis, D., & Stuckless, N. (1996). Mediating and negotiating marital conflicts. Thousand Oaks, CA: Sage.

Elsner, A. (2001, January 4). Protecting battered U.S. women saves the lives of men. Reuters Newswire. Retrieved January 5, 2001, from http://www.eurowrc.org/01.eurowrc/ 04.eurowrc_en/43.en_ewrc.htm

Ferrante, A., Morgan, F., Indermaur, D., & Harding, R. (1996). Measuring the extent of domes- tic violence. Perth, Australia: Hawkins Press.

Fiebert, M. (1997). Annotated bibliography: References examining assaults by women on their spouses/partners. Sexuality and Culture, 1, 273-286.

Fiebert, M., & Gonzalez, D. M. (1997). College women who initiate assaults on their male partners and the reasons offered for such behavior. Psychological Reports, 80, 583-590.

Fox, J. A., & Zawitz, M. W. (2001). Homicide trends in the United States: Intimate homicide. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics. Retrieved May 29, 2002, from http://www.ojp.usdoj.gov/bjs/homicide/intimates.htm

Frude, N. (1994). Marital violence: An interactional perspective. In J. Archer (Ed.), Male vio- lence (pp. 153-169). Boston: Routledge Kegan Paul.

Gelles, R. (2000, October 12). Domestic violence: Not an even playing field. The safety zone. Retrieved October 12, 2000, from http://www.serve.com/zone/everyone/gelles/ html

Gelles, R., & Straus, M. (1999). Profiling violent families. In A. Skolnick & J. Skolnick (Eds.), The family in transition (pp. 414-431). White Plains, NY: Longman.

Goldberg, C. (1999, November 23). Spouse abuse crackdown, surprisingly, nets many women. New York Times, p. A16.

Gonzalez, D. M. (1997). Why females initiate violence: A study examining the reasons behind assaults on men. Unpublished master’s thesis, California State University, Long Beach.

Greenfeld, L. A., Rand, M. R., Craven, D., Klaus, P. A., Perkins, C. A., Ringel, C., et al. (1998). Violence by intimates: Analysis of data on crimes by current or former spouses, boyfriends, or girlfriends. Washington, DC: U.S. Department of Justice.

Johnson, M. P. (1995). Patriarchal terrorism and common couple violence: Two forms of violence against women. Journal of Marriage and the Family, 57, 283-294.

Johnson, M. P. (2000). Conflict and control: Images of symmetry and asymmetry in domes- tic violence. In A. Booth, A. Crouter, & M. Clements (Eds.), Couples in conflict (pp. 178- 204). Hillsdale, NJ: Lawrence Erlbaum.

1360 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

Johnson, M. P., & Ferraro, K. (2000). Research on domestic violence in the 1990s: Making distinctions. Journal of Marriage and the Family, 62, 948-963.

Kaufman Kantor, G., & Jasinski, J. (1998). Dynamics of partner violence and types of abuse and abusers. United States Air Force, NNFR research report. Retrieved May 12, 2002, from http://www.nnfr.org/research/pv/pv_ch1.html

Kaufman Kantor, G., & Straus, M. (1987). The “drunken bum” theory of wife beating. Social Problems, 34, 213-230.

Kellerman, A. L., & Mercy, J. A. (1992). Men, women, and murder: Gender specific differ- ences in rates of fatal violence and victimization. Journal of Trauma, 33, 1-5.

Kilpatrick, D., & Saunders, B. (1997). The prevalence and consequences of child victimization. Washington, DC: U.S. Department of Justice. Retrieved May 7, 2002, from http:// www.ncjrs.org/pdffiles/fs000179.pdf

Kilpatrick, D., & Saunders, B. (2000). The national survey of adolescents in the United States. Ann Arbor, MI: Inter-University Consortium for Political and Social Research.

Kimmel, M. (1994). Masculinity as homophobia: Fear, shame and silence in the construc- tion of gender identity. In H. Brod & M. Kaufman (Eds.), Theorizing masculinities (pp. 119-141). Thousand Oaks, CA: Sage.

Kimmel, M. (1996). Manhood in America: A cultural history. New York: Free Press. Kimmel, M. (2000). The gendered society. New York: Oxford University Press. Kimmel, M. (2001, December 5). The myth of gender symmetry in domestic violence. The

Irish Times, p. 16. Kimmel, M. (2002, January 26). Beyond the myths on domestic violence. The Irish Times, p.

16. Kurz, D. (1993). Physical assaults by husbands: A major social problem. In R. J. Gelles &

D. R. Loseke (Eds.), Current controversies on family violence (pp. 88-103). Thousand Oaks, CA: Sage.

Leeman, S. (2000, September 20). U.N. Report: Millions of women suffering. Associated Press.

Maccoby, E., & Jacklin, C. (1974). The psychology of sex differences. Stanford, CA: Stanford University Press.

McNeely, R. L., & Jones, J. M. (1980). Refuge from violence: Designing shelter services for battered women. Administration in Social Work, 4, 71-81.

Mead, M. (1950). And keep your powder dry. New York: William Morrow. Mihalic, S. W., & Elliott, D. (1997). A social learning theory model of marital violence. Jour-

nal of Family Violence, 12, 21-47. Nazroo, J. (1995). Uncovering gender differences in the use of marital violence: The effect

of methodology. Sociology, 29, 475-495. Okun, L. (1986). Women abuse: Facts replacing myths. Albany: State University of New York

Press. O’ Leary, K. D. (1999). Developmental and affective issues in assessing and treating partner

aggression. Clinical Psychology: Science and Practice, 6, 400-414. O’ Leary, K. D. (2000). Are women really more aggressive than men in intimate relation-

ships? Psychological Bulletin, 126, 685-689. O’ Leary, K. D., Barling, J., Arias, I., Rosenblum, A., Malone, J., & Tyree, A. (1989). Preva-

lence and stability of physical aggression between spouses: A longitudinal analysis. Journal of Consulting and Clinical Psychology, 57, 263-268.

Pence, E., & Paymar, M. (1993). Education groups for men who batter: The Duluth model. New York: Springer.

Pleck, E., Pleck, J., Grossman, M., & Bart, P. (1978). The battered data syndrome: A com- ment on Steinmetz’s article. Victimology, 2, 680-684.

Rouse, L., Breen, R., & Howell, M. (1988). Abuse in intimate relationships: A comparison of married and dating college students. Journal of Interpersonal Violence, 3, 414-419.

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1361

http://vaw.sagepub.com

Saenger, G. (1963). Male and female relations in the American comic strip. In D. M. White & R. H. Abel (Eds.), The funnies, an American idiom (pp. 219-231). New York: Free Press.

Schwartz, M. D. (1987). Gender and injury in spousal assault. Sociological Focus, 20, 61-75. Schwartz, M. D., & DeKeseredy, W. S. (1993). The return of the “battered husband syn-

drome” through the typification of women as violent. Crime, Law, and Social Change, 20, 249-265.

Slobodian, L. (2000, November 6). Researcher claims men, women, equal abusers. Calgary Herald, City Section, p. 1.

Stark, E., & Flitcraft, A. (1988). Violence among intimates: An epidemiological review. In V. B. Van Hasselt, R. L. Morrison, A. S. Bellack, & M. Herson (Eds.), Handbook of family violence (pp. 293-317). New York: Plenum.

Steinmetz, S. (1977/1978). The battered husband syndrome. Victimology, 2, 499-509. Stets, J. E., & Pirog-Good, M. A. (1990). Interpersonal control and courtship aggression.

Journal of Social and Personal Relationships, 7, 371-394. Stets, J. E., & Straus, M. A. (1990). Gender differences in reporting marital violence and its

medical and psychological consequences. In M. A. Straus & R. J. Gelles (Eds.), Physical violence in American families (pp. 151-165). New Brunswick, NJ: Transaction Publishing.

Straton, J. (1994). The myth of the “battered husband syndrome.” Masculinities, 2, 79-82. Straus, M. A. (1990). Measuring intrafamily conflict and violence: The Conflict Tactic (CT)

Scales. In M. A. Straus & R. J. Gelles (Eds.), Physical violence in American families (pp. 29- 47). New Brunswick, NJ: Transaction Books.

Straus, M. A. (1993). Identifying offenders in criminal justice research on domestic assault. American Behavioral Scientist, 36, 587-600.

Straus, M. (1997). Physical assaults by women partners: A major social problem. In M. R. Walsh (Ed.), Gender: Ongoing debates (pp. 204-221). New Haven, CT: Yale University Press.

Straus, M. (1999). The controversy over domestic violence by women: A methodological, theoretical and sociology of science analysis. In X. B. Arriaga & S. Oskamp (Eds.), Vio- lence in intimate relationships (pp. 17-44). Thousand Oaks, CA: Sage.

Straus, M., & Gelles, R. (1986). Societal change and family violence from 1975 to 1985 as revealed by two national surveys. Journal of Marriage and the Family, 48, 445-479.

Straus, M., & Gelles, R. (Eds.). (1990). Physical violence in American families. New Brunswick, NJ: Transaction Publishing.

Straus, M., Gelles, R., & Steinmetz, S. (1980). Behind closed doors: Violence in the American fam- ily. Newbury Park, CA: Sage.

Sugarman, D. B., & Hotaling, G. T. (1989). Dating violence: Prevalence, context, and risk markers. In M. A. Pirog-Good & J. E. Stets (Eds.), Violence in dating relationships: Emerging social issues (pp. 91-107). New York: Praeger.

Suitor, J. J., Pillemer, K., & Straus, M. (1990). Marital violence in a life-cycle perspective. In M. Straus & R. Gelles (Eds.), Physical violence in American families. New Brunswick, NJ: Transaction Books.

Tendler, S. (1999, January 22). Men suffer equally from violence in the home. London Times, p. 23

Tjaden, P., & Thoennes, N. (1998). Prevalence, incidence, and consequences of violence against women: Findings from the National Violence Against Women Survey. Washington, DC: U.S. Department of Justice, National Institute of Justice and Centers for Disease Control and Prevention. Retrieved May 7, 2002, from http://www.ncjrs.org/pdffiles/172837.html

Tjaden, P., & Thoennes, N. (2000a). Extent, nature, and consequences of intimate partner vio- lence. Washington, DC: U.S. Department of Justice, National Institute of Justice and Centers for Disease Control and Prevention.

1362 VIOLENCE AGAINST WOMEN / November 2002

http://vaw.sagepub.com

Tjaden, P., & Thoennes, N. (2000b). Prevalence and consequences of male-to-female and female-to-male intimate partner violence as measured by the National Violence Against Women Survey. Violence Against Women, 6, 142-161.

Tjaden, P., & Thoennes, N. (2000c). Full report of the prevalence, incidence, and consequences of violence against women: Findings from the National Violence Against Women Survey. Wash- ington, DC: U.S. Department of Justice, National Institute of Justice and Centers for Disease Control and Prevention.

Tyree, A., & Malone, J. (1991, August). How can it be that wives hit husbands as much as hus- bands hit wives and none of us knew it? Paper presented at the annual meeting of the American Sociological Association, Cincinnati, Ohio.

United Nations Population Fund. (2000). The state of world population: Lives together, worlds apart-Men and women in a time of change. New York: United Nations.

Walby, S. (1999, October). Comparing methodologies used to study violence against women. Men and violence against women. Proceedings from a seminar at the Council of Europe, Strasbourg, France.

Michael S. Kimmel is a professor of sociology at the State University of New York at Stony Brook. His books include Changing Men (1987), Men Confront Por- nography (1990), Men’s Lives (5th ed., 2000), Against the Tide: Profeminist Men in the United States, 1776-1990 (1992), The Politics of Manhood (1996), Manhood: A Cultural History (1996), and The Gendered Society (2000). He edits Men and Masculinities, an interdisciplinary scholarly journal, a book series on men and masculinity at the University of California Press, and the Sage Series on men and masculinities. He is the spokesperson for the National Organization for Men Against Sexism and lectures extensively on campuses in the United States and abroad.

Kimmel / GENDER SYMMETRY IN DOMESTIC VIOLENCE 1363

http://vaw.sagepub.com