Female intimate partner violence and developmental trajectories of abusive females.

A review of the research literature indicates that female intimate partner violence (IPV) is a frequent as male IPV. It is just as severe and has much the same consequences for males as for females. Despite these findings, criminal justice intervention and custody evaluation operate from the unwarranted assumption that males are the greater risk for IPV perpetration.

Keywords: females as perpetrators, males as victims, intimate partner violence

**********

Research data contradicting the feminist view of intimate partner violence (IPV)--exclusively male perpetrators and female victims--have been available since 1980. Stets and Straus (1992) reported data from the 1985 United States National Survey on the incidence of abuse showing that 45 percent of abuse reports were bilateral (corrected for level of severity) and, of the rest, women were three times as likely as males to use severe violence against a non-violent or minimally violent partner. Dutton (1994) reviewed data patterns inconsistent with feminist theory, including higher rates of IPV in lesbian relationships compared to heterosexual relationships, the absence of a direct relationship between power and violence in couples, and a small number of men who used severe violence unilaterally.

As new incidence data were collected, they became more troubling for feminist theory (Dutton 2006; Dutton & Nicholls, 2005). These included data from the United States National Surveys of 1975 and 1985 showing women to be as violent as males. However, female violence rates have been portrayed as indications of self-defensive behavior, as less serious, or as a result of reporting differences. In fact, they are equivalent or exceed males' rates, they include female violence against non-violent males, and they have serious consequences for males. A large meta-analytic study reviewed below (Archer, 2000) found male and female IPV to be best represented as overlapping skewed distributions. Females were injured more often than males, but by only .083 of a standard deviation from the mean.

New Studies of Dating Aggression

New data from dating violence studies are remarkably consistent with the adult partner abuse literature. Watson and colleagues sampled 475 high school students (266 males and 209 females) from a large, metropolitan area on Long Island, New York (Watson, Cascardi, Avery-Leaf, & O'Leary, 2001). Using a modified Conflict Tactics Scale (CTS), of students with past or current romantic relationships (N = 401), 45.6 percent reported at least one incident of physical aggression by their current or former partners, but just 9 percent reported exclusive victimization (that is, they had been a victim but not a perpetrator of physical aggression). Using a measure they developed, Watson and colleagues also studied gender differences in response to aggression by a dating partner. Female students were significantly more likely than male students to report an aggressive response. Specifically, girls were significantly more likely (42%) to fight back than boys (26%). Male students (24%) were more likely than female students (6%) to do nothing in response to abuse by a partner. There was also a trend for female students (28%) to be more likely to report breaking up with an abusive partner compared to male students (21%).

Follingstad et al. (2002) conducted a particularly well-designed study of dating violence. They developed a structural equation model, which delineates primary and moderator variables, to best relate psychological etiology and dating violence in a sample of 412 college students. Since the authors assessed both members of the couple, individual characteristics could be statistically connected to couple variables. They also used subjects with a history of IPV (30%) and a subsample of those without such a history (about 15%). Psychological variables measured included anxious attachment and angry temperament. The outcome variables included controlling behaviour and physical violence. The study found that anxious attachment related to angry temperament, which was causally correlated with controlling behavior. In turn, controlling behavior generated violence (p. 42). As the authors put it,

   ... the primary path leading to the use of force in dating
   relationships was initially due to the presence of anxious
   attachment influencing the development of an angry temperament
   which then leads to behaviors to control one's partner.
   Controlling behavior is the significant mediator between the angry
   temperament and greater frequency and severity of dating violence.
   (p. 44)

This study was conducted on 233 males and 199 females. No gender differences were reported. In this study attachment seemed to be predictive of control and physical abuse in either gender.

The largest and most comprehensive of all dating violence studies was a cross-cultural study of partner violence in a sample of 6,900 university students from seven teen nations (Douglas & Straus, 2003). The authors found that adolescent girls were 1.15 times more likely to assault male partners than were adolescent boys, regardless of whether overall assault or severe assault rates were considered. In Scotland, severe assault was 5.52 times much more likely to be perpetrated by females than by males. In the Singapore sample, females were 4.57 times more likely than males to assault their partner. In the New Zealand sample, females were 2.96 times more likely than males to assault their partner. In this study, male-perpetrated injury rates were 8.1 percent (serious injury 2.6%), while female-perpetrated injury rates were 6.1 percent (serious injury 1.2%).

Some have argued that dating relationships are different than cohabitating and married relationships. However, while they differ in terms of commitment and ease of exit, all are intimate relationships.

Longitudinal Cohort Studies 

Longitudinal cohort studies are methodologically soundest for studying a variety of problem behaviors. The samples are large and representative of an age cohort and are not self-selected through their presence in a shelter or a court-mandated group or even a college. Hence, their external validity (the degree to which the results can be generalized to an entire population) is higher than in most studies. Also, since the cohort is followed over time, cause-and-effect relationships can be inferred. Unlike "snapshot" studies that freeze behaviour to a one-time sample and rely on correlations at that time, a protracted longitudinal study can examine effects from a measure taken years before. Three major peer-cohort longitudinal studies on our theme have been carried out and will now be examined in detail.

The Concordia Longitudinal Risk Project. One of these studies, the Concordia Longitudinal Risk Project, in Montreal, used data collected in a longitudinal study of 4,109 French-speaking school children in 1976 (Serbin et al., 2004). The children were categorized into Aggression and Withdrawal categories using a French version of the Pupil Evaluation Inventory. Extremes in Aggression and Withdrawal were indicated when children scored above the ninety-fifth percentile on Aggression and below the 75th percentile on Withdrawn. The resulting sub-sample yielded 101 girls and 97 boys. Reverse criteria yielded a Withdrawn group of 129 girls and 108 boys. Age-matched comparisons were developed by taking children who were average (between the 25th and 75th percentiles) on both the Aggression and Withdrawal categories. The researchers describe their sample a being "community based" and therefore "avoiding biases inherent in clinic-referred samples and more representative of the (general) population" (p. 266). Aggressive children of both sexes had lower IQs and academic achievement than comparison controls. Both were more physically aggressive during play.

Girls' aggression was associated with a preference for male partners who were also aggressive. As they approached adolescence, these girls had elevated rates of smoking, alcohol, and illicit drug use and "continue[d] to seek out behaviourally compatible peer groups, probably comprised of boys and girls with similar aggressive or 'pre-delinquent' behavioural styles" (p. 268). They had elevated rates of gynaecological problems, were more likely to use birth control methods sooner, had higher rates of sexually transmitted diseases between the age of 14 and 20, and became pregnant sooner and more frequently. Girls who scored above the 75th percentile on both Aggression and Withdrawal had a teen pregnancy rate of 48 percent.

The aggressive group had elevated levels of depression and anxiety disorder by late teens. When they married, their children had higher health risks, and the aggressive girls had become aggressive mothers, exhibiting maternal aggression and having children who had more visits to hospital emergency rooms for treatment of injuries. These risk factors for women are completely overlooked in the advice given by domestic violence experts to custody evaluators (Jaffe et al., 2003) who concentrate solely on male intimate partner violence and warn evaluators of this risk. The Concordia study and the Dunedin study (see below) present clear examples of the development and independent expression in women of aggression toward others. This is not, then, aggression in reaction to male aggression but aggression that develops independently of the behaviours of the intimate male partner. They show the developmental trajectory and "trait" character of aggression. Both studies indicate that these women will select aggressive men and contribute to ongoing intra-couple aggression.

The Dunedin Study. Magdol and her colleagues followed a birth cohort of 1,037 subjects in Dunedin, New Zealand (Magdol et al., 1997). The original cohort was

   ... a complete cohort of births between April 1972 and March 31,
   1973, in Dunedin, New Zealand (population 120,000). The group has
   been studied every two years since for a variety of health,
   development, and behavioural measures. The sample is predominantly
   of European ancestry and is representative of the social class
   distribution of New Zealand. (p. 69)

This study of partner violence took place when the members of the group were 21 years of age. It was based on items embedded in a larger group of questions relating to mental health issues. Of the original cohort, 941 provided data. Respondents had to have had "a relationship with a romantic partner during the last twelve months that lasted at least one month." Of the 941 in the sample, 861 qualified. Partner violence was assessed using the CTS, each respondent reporting for both self and partner. Measures were also reported of the following: socioeconomic status, social ties, substance abuse, criminality, and mental health (anxiety, depression, mania, and psychosis). Criminality focused on self-reports of crimes against strangers, using the DSM-IV criteria of antisocial personality disorder.

The 425 women and 436 men from the cohort who were in intimate relationships indicated that both minor and severe physical violence rates were higher for women, whether self-reported or partner-reported. The female severe physical violence rate was more than triple that of males (18.6% vs. 5.7%). Based on this same sample, Moffitt et al. (2001) reported that pre-existing characteristics of the women (at age 15) predicted their (1) later choice of an abusive male panner and (2) use of violence with that partner apart from the male's violence (see below). As Magdol et al. (1997) put it, "Early studies of partner violence assumed that men's perpetration rates exceeded those of women, in part because these studies relied almost exclusively on clinical samples of women who sought assistance or of men in court-mandated counseling programs" (p. 69).

Moffitt, Moffitt, Caspi, Rutter, and Silva (2001) completed a comprehensive analysis of the Dunedin data. These authors reported that the following characteristics predicted intimate violence in females: approval of the use of aggression, excessive jealousy and suspiciousness, a tendency to experience intense and rapid emotions, and poor self-control. These are the same characteristics found in male batterers. The antisocial female sample also had earlier puberty, earlier initiation of intercourse, more friends who were older than the subject, and more delinquent friends. The researchers found that antisocial traits measured in females at age fifteen made them more likely to be involved in a relationship with an abusive man at age twenty-one, and even after controlling for their partners' physical abuse, "women with a juvenile history of conduct problems were still more likely to commit violence against their partners" (p. 64). A similarly designed study in the United States found the same gender equality of violence (Morse, 1995).

The abusive personality in males (Dutton, 2002) has the same profile as Moffitt et al. found in females and termed "negative emotionality": jealousy, impulsivity, rapidly fluctuating emotions, and poor self-control. With the men, these features were related to independently assessed borderline traits and borderline personality disorder, which unfortunately were not assessed in the Dunedin women. From the descriptors given by Moffitt et al., however, it seems that an identical "abusive personality" exists for male and female intimate abuse perpetrators. In sum, a woman's conduct problems correlated +.44 with her later use of violence against her partner (while his violence partialled out). It also correlated +.36 with his use of violence toward her. Essentially, the pattern of correlations between early conduct problems and later intimate violence and partner's use of violence was similar for both sexes. The correlations were not significantly different.

Moffitt et al. emphasize the importance of puberty as a developmental crossroads for these girls. They also make a provocative argument, based on their data set that males later form two types of antisocial behaviour, one against strangers, which may be neurologically based, and another against intimate females. Females form only one type of antisocial behaviour--against intimate males.

The use by these authors of sophisticated path analysis, which differentiates independent, moderator, and dependent variables, on this huge and representative sample gives added weight to their findings. The reader is reminded that other studies (for example, Serbin et al., 2004) found a broader array of antisocial actions by women, including physical abuse of children.

Ehrensaft, Moffitt, and Caspi (2004) also studied the Dunedin birth cohort and found that nine percent were in "clinically abusive relationships," defined as those that required intervention by a physician, the police, or a lawyer. More such help exists for women than for men, and they are more likely to use it, so the results may be skewed. However, the authors found comparable rates of violence in both sexes, with 68 percent of women and 60 percent of men self-reporting injury. Both male and female perpetrators evidenced signs of personality disturbance. The authors noted, for instance, that the women had "aggressive personalities and/or adolescent conduct disorder" (p. 267). As the authors put it, "These findings counter the assumption that if clinical abuse was ascertained in epidemiological samples, it would be primarily man-to-woman, explained by patriarchy rather than psychopathology" (p. 258).

The U.S. National Youth Survey. This preponderance of female stranger aggression was replicated in the United States Youth Survey (Elliott, Huizinga, & Morse, 1985). This survey used a national probability sample of 1,725 respondents. Begun in 1976, it provided nine waves of data over 17 years. Respondents were interviewed annually using structured, face-to-face, confidential interviews. Violence was measured using the eight-item intimate violence subscale of the CTS. Injury was also assessed. For the years 1983, 1986, 1989, and 1992, female-to-male violence and severe violence was about double the rate of male-to-female violence and severe violence.

On the intimate violence scale, females scored higher than males on the following acts: kicked, bit or hit with fist, hit with object, threatened with a knife or gun, and used knife or gun. Identical victimization rates were reported by men and women, but higher perpetration rates were reported for women. Men scored higher than women on the following acts: beating up and choking or strangling. Morse (1995) found little or no change in the ratio of female-to-male vs. male-to-female violence over the years, with slightly more than twice as much female-initiated as male-initiated violence. This pattern remained unchanged in all years that data were collected. When the most serious fight had involved one-sided violence that was not reciprocated, both men and women were twice as likely to report that the perpetrator was female (see also Stets & Straus, 1992).

As with so many studies, the reader has to examine the data tables, not the conclusion or abstract. In the Morse report, for example, "gender differences" are described, while, in fact, there were tiny minorities of each gender being compared to keep the "gender paradigm" alive.

The New York Children in the Community (CIC) Study. Ehrensaft et al. (2004) followed a randomly selected cohort sample of 543 children over 20 years to test the effects of parenting, exposure to domestic violence between parents (ETDV), parental maltreatment, adolescent disruptive conduct disorders, and substance abuse disorders on the risk of violence to and from an adult intimate partner. Conduct Disorder (CD) was the strongest predictor of perpetration for both sexes, followed by ETDV. Essentially, the CD in some individuals failed to disappear but developed into one or more of a variety of adult personality disorders. Ehrensaft et al. call these "personality disorder trajectories." A failure of personality disorders to diminish from adolescence to adulthood predicted intimate violence in both sexes. Women with a pattern of distrust, interpersonal avoidance, unusual beliefs, and constricted affect were more likely to assault intimate male partners. It was personality disorder, not gender, which predicted violence. As the authors put it, "It was personality functioning measured prospectively from adolescence to early adulthood [that] can distinguish individuals who will go on to perpetrate partner violence" (p. 263; italics author's own).

The U.S. National Comorbidity Survey (NCS). The NCS was a nationally representative household survey completed between September 1990 and March 1992 to assess mental illness (Kessler, Molnar, Feurer, & Appelbaum, 2001). The NCS involved face-to-face interviews with 8,098 people between the ages of 15 and 44. A probability subsample was then generated to assess the social consequences of certain mental disorders. Of that sample, 3,537 participants (1,738 men, 1,799 women) were cohabitating or married when they completed the CTS.

Kessler et al. (2001) reported there was no significant gender difference in the prevalence of minor physical victimization reported by women (17.4%) and men (18.4%), respectively. Similarly, for minor violence, victimization exceeds perpetration in reports of both men and women (18.4% of men report victimization and 17.7% report perpetration; 17.4% of women report victimization and 15.4% report perpetration). For severe violence, reports of victimization did not differ significantly by gender (6.5% of female respondents and 5.5% of male respondents), but reports of perpetration did differ, with men reporting significantly less perpetration (2.7%) than women (6.5%).

Kessler and his colleagues also examined reciprocal aggression--that is, couple aggression in which both partners perpetrate aggressive acts and are victimized by their partner. As in prior studies (for example, Stets & Straus [1992]), reciprocity was the norm. Similar proportions of women (10.5%) and men (11.7%) reported that both partners engaged in the same amount of minor violence. Of the participants reporting reciprocal minor aggression, "70.7 percent of the women who reported minor violence victimization and exactly the same percent of comparable men [70.7%] said that they reciprocated minor violence" (p. 492). Of those reporting reciprocal minor violence, by far, most women (85.4%) and men (90.0%) reporting it described the frequency by each partner as the same.

Among those who reported committing severe aggression, each gender was more likely to report their partner as reciprocating than was the partner likely to report this. Of those reporting reciprocal severe aggression, again most men (96%) and women (80.4%) stated that the abuse frequency was the same for both partners. In this study, Kessler et al. replicated Stets and Straus's findings on this issue of reciprocity.

Williams and Frieze (2005) recently analyzed the National Comorbidity Survey data. They found several different violence patterns based on a sample of 3,519, concluding that the most common was "mutual and mild violence" followed by "mutual severe violence." In terms of reactions to the violence, the similarities by gender out-weighed the differences. More women than men reported perpetrating violence, and slightly more men than women reported being the victim of severe violence. The authors conclude that this "may challenge assumptions about women's victimization in relationships."

The Oregon Youth Study. Capaldi and her colleagues (Capaldi, Kim, & Shortt, 2004) examined data from an ongoing community-based longitudinal study of youth in Oregon (the Oregon Youth Study [OYS]). By young adulthood, nine percent of men and 13 percent of women were identified as engaging in frequent partner violence. Consistent with prior findings, frequent violence was most common in relationships with bi-directional abuse. Regarding injuries, 13 percent of the young men and nine percent of the young women indicated they had been hurt at least once by partner violence, and injury was also likely to be mutual (p. 233). No gender differences were found in fear of partner's abusive behaviour (p. 232). Women's prior antisocial behaviour and depressive symptoms predicted both their own abusive partner behaviour, as well as their male partner's abuse. Notably, the women's characteristics were predictive over and above the contribution of the male partner's antisocial characteristics. These findings suggest assortative mating for antisocial behaviour, as well as the independent contribution of women's risk factors to the development of violent relationships. Stability of aggression was found for both genders. As Capaldi et al. concluded, "aggression thus appears to be predominantly bidirectional" (p. 235) and based on "assortative mating," in other words, that aggressive people chose each other.

The Finnish Study. Kokko and Pulkinnen (2005)assessed aggression in a sample 300 Finnish men and women based on teacher and peer ratings. They evaluated the stability of aggression from ages eight to 14 and then to adulthood. Aggression was found to be stable essentially from eight to adulthood, for both men and women.

The results of these longitudinal developmental studies of men and women, which were all based on large community samples, suggest the same conclusions: female violence (1) is common, (2) occurs at about the same rate as male violence, (3) is generated independently of the actions of the "current boyfriend" or husband, and (4) has a long developmental history preceding the current adult relationship and so cannot be dismissed as self-defense. When violence does occur, it is most commonly mutual, then followed by more severe female violence, and, finally, more severe male violence. As Moffitt et al. (2001) wrote: "The argument that women's abuse perpetration in the community is too trivial to research could prove to be tantamount to arguing that smoking in the community is too trivial to research and scientists should focus on cases of lung cancer" (p. 69; italics author's own).

The Archer Meta-analytic Study

The most comprehensive study of gender differences in intimate violence was conducted by Archer (2000). This meta-analytic study examined combined results from eighty-two independent studies (including the United States Violence Against Women Survey) where data were available for comparing gender rates of abuse perpetration. Based on combined data across studies (a total sample of 64, 487), women were slightly more likely than men to use physical abuse (defined using the CTS) against an intimate partner (effect size or d' = -0.05). Regarding use of violence, an effect size of -0.05 means that women are one-twentieth of a standard deviation from the mean more often violent than men. Similarly, women are injured about one-twelfth of a standard deviation from the mean more often than men. In both cases, therefore, the rates by gender are virtually identical. This was true whether or not outliers were removed and whether or not studies with a ceiling N of 800 were considered to offset swamping of the outcome by studies with huge samples. Samples from shelters were unrepresentative of community samples, since by definition they were male-violent/female-victim samples. This was not true of community samples. As with the dating samples reported above, the younger the sample, the higher the level of female violence relative to male violence.

Medical treatment for injuries across studies revealed an effect size of +0.08, with women being slightly more likely than men to seek treatment. Neither the use of violence nor medical treatment resulted in a large effect size. An effect size of d' = 0.08 is a difference between genders of less than one-tenth of one standard deviation from the mean. Given the methodology employed by Archer, his work has to be considered the "gold standard" of studies in gender usage of violence.

Archer cites the norms ordinarily used regarding use of violence: (1) the so-called acceptance norm repeatedly cited by Dobash and Dobash (1979), for which no evidence has been reported and (2) the contrary norm against males using physical aggression toward women. A subsequent analysis of the data that disaggregated the violence into discrete acts found that women were more likely than men to slap, bite, kick, punch, or hit with an object. Men were more likely to beat up or choke and strangle their partner, although "a substantial minority of endorsements of 'beat up' and 'choke or strangle' involved women perpetrators" (Morse 1995, p. 339). Gender differences ranged from "very small to medium." Samples selected for marital problems showed large effects in the male-perpetrator direction. Student samples showed effects more in the female-perpetrator direction than did community samples. Patterns of findings did not differ depending on whether the perpetrator or the victim was reporting.

Archer concluded that concern with the (female) victims in such cases is certainly not misplaced but regarding them as the only victims of partner violence is too narrow a view of the problem according to the present findings.

Finally, I want to mention Fiebert (1997), who has amassed a bibliography of 159 peer-reviewed publications that find equal or greater aggression by females than males. The total collected sample is greater than 109,000.

In sum, across several large sample studies, with varying demographic compositions, dating violence is more likely perpetrated by females than males, especially when they are young. These findings clearly run counter to the common assertion that female aggression in intimate relationships is uncommon and inconsequential. They also lead us to query if the rate of intimate abuse is at risk of increasing rather than decreasing with the next generation.

Effects More Severe for Women?

It is widely supposed that women's aggression against male partners is less likely to result in serious physical or psychological harm than male aggression against female partners. The Archer study cited above revealed a much smaller effect size for injuries by gender (one-sixth of one standard deviation from the mean) and hospitalization (one-eleventh of one standard deviation from the mean) than had been claimed by prior studies based on feminist theory.

Other studies also have supported this finding. Coker et al. (2002) reanalyzed data from the National Violence Against Women Survey (NVAWS) (6,790 women and 7,122 men) to assess associations between physical, sexual, and psychological abuse, and current and long-term physical and psychological effects in men and women. Results indicated that psychological and physical abuse was associated with much the same outcomes and had similar effects for men and women. The authors cautioned that it is possible male victims were also perpetrators and their mental health status resulted from inflicting abuse rather than from being victimized. Interestingly, they did not present this hypothesis for women.

A reanalysis of the Canadian General Social Survey (GSS) data by LaRoche (2005) (a sample of 25,876) strongly refutes the idea that males do not suffer ill effects from intimate partner violence. It is of interest that, although not all "victim" data in that survey were available for men, what was available indicated great similarity in male and female victimization. LaRoche reports that 83 percent of men who "feared for their life" did so because they were unilaterally terrorized by their female partner compared to the 77 percent of women who were unilaterally terrorized by a male partner. Of the terrorized men, 80 percent reported having their everyday activities disrupted (compared to 74% for terrorized women), 84 percent received medical care (the same rate as for terrorized women), and 62 percent sought psychological counseling (63% for women). Hence, in a huge and nationally representative Canadian sample, victim reactions for abused men were virtually identical to those of abused women. It was simply that earlier research was driven by a paradigm that avoided asking the right questions of men. When these questions are asked, the results are surprising.

An emergency clinic in Philadelphia, Pennsylvania, found that 12.6 percent of all male patients over a thirteen-week period (N = 866) were victims of domestic violence (Mechem, Shofer, Reinhard, Hornig, & Datner, 1999). These patients reported having been kicked, bitten, punched, or choked by female intimate partners in 47 percent of cases, and in 37 percent of cases the men reported a weapon being used against them. The authors observe the numbers would have been higher except they had to stop counting after midnight, and also screened out "major trauma" cases, which could have upped the proportion of males injured by female partners. It is noteworthy that many emergency clinics ask women but not men about potential domestic violence origins for injuries.

An emergency clinic study in Ohio found that 72 percent of men admitted with injuries from spousal violence had been stabbed (Vasquez & Falcone, 1997). The most frequent cause of admission for women victims was assault (53%). The authors report that burns obtained in intimate violence were as frequent for male victims as for female victims. As this study demonstrates, community samples, unless they require subjects to self-report as crime victims, show a different and more equivalent pattern of violence by gender than that alleged by the radical feminist perspective.

Domestic Violence Helpline Calls

Rarely have men had their fear of female violence assessed. The LaRoche analysis of the Canadian National GSS data is the one exception. Hines, Brown, and Dunning (2003) examined calls from men to the American national domestic violence helpline for men. As the authors pointed out, it would be unlikely for male perpetrators or co-perpetrators to use this line. When the line opened, it received one call a day. When it was advertised in state telephone directories, it began to receive 250 calls a day. Given that 2.6 million men in the United States are victims of severe violence (Straus & Gelles, 1992), further usage increase can be expected. All but a few callers experienced physical abuse from their female partners (only 4% were gay), and a substantial minority feared their partners' violence and were stalked. Over 90 percent experienced controlling behaviors, and several men reported frustrating experiences with the domestic violence system. More than half (52.4%) of males who were currently in an abusive relationship indicated they were fearful that their female partner would cause a serious injury if she found out that they had called the helpline. Hines et al. (2003) state that "according to qualitative accounts, several physical attacks were reported to have occurred to the groin area" (p. 17). Callers reported forms of violence that are not measured in surveys, such as having their partner try to run them down with a car. Twenty-nine percent reported being stalked by their female partners. Callers' reports indicated that their female abusers had a history of trauma, alcohol and/or drug abuse problems, mental illness, or homicidal and suicidal ideation. Hines et al. concluded that the "system in place to prevent IPV re-victimizes these men and hence, no help is available for half the population" (p. 18).

Differential Fear

Jacobson et al. (1994) have reported that only women become fearful during marital conflicts. However, as I have pointed out (Dutton, 2006), women use fear scales differently than men and report fear during exposure to conflict which poses no danger to them.

Pimlott-Kubiak and Cortina (2003) carried out a large-scale (N = 16,000) study of gender differences in traumatic reactions to intimate violence, stalking, and emotional abuse. Arguing that earlier studies had focused too specifically on Post-Traumatic Stress Disorder (PTSD), the authors broadened their assessment of traumatic reactions and found eight distinct combinations of exposure profile, gender, and mental and physical health. No meaningful interactive effects of gender and interpersonal aggression were found once lifetime exposure to aggressive events was adequately taken into account. The only thing that predicted the severity of psychological consequences was the degree of exposure--not gender. The authors concluded that their findings argue against theories of greater female vulnerability to traumatic outcomes.

Regardless of the variation in the studies, two conclusions seem reasonable: (1) women are injured more than men, but (2) men are injured as well and are not immune to being seriously injured. Simply because the injury rates are lower, men should not be denied protection.

The argument is sometimes made that men use threats more frequently than women, but Straus and his colleagues (1996) found that women reported using psychological aggression somewhat more than did men. The "threatened to hit or throw something at partner" item correlated .52 with this scale, of which it is a part. Also, Giordano et al. (1999) in a study of 721 young adults found that women were more likely than men to threaten to use a knife or gun.

Underestimating Male Victimization from Criminal Justice Statistics

Male victimization is not taken seriously because of (1) the "gender paradigm" described above, (2) a cultural belief that men should be able to defend themselves, and (3) a disbelief in female violence. The item "burned him with something" is rated less serious by males (.83, where the upper limit is 1.0) than the item "burned her with something" is rated by females (.96) (Marshall, 1992a & b).

Law enforcement is lax when men are injured in domestic violence disputes (Brown, 2004). When John Wayne Bobbit's penis was cut off by his wife in 1993, it became material for late night comedy routines. His wife was found not guilty by reason of temporary insanity. This reaction would have been unthinkable with the genders reversed.

Buzawa and her colleagues (1992), in a study of the police arrest policy in Detroit, found that "male victims reported three times the rate of serious injury as their female counterparts" (38% compared to 14%). The police rarely arrested a female perpetrator. As the authors put it, "Not one male victim was pleased with the police response. They stated that their preferences were not respected by the officers, nor was their victimization taken seriously" (p. 268). The lack of police responsiveness occurred regardless of the degree of injury. For example, one male reported requiring hospitalization for being stabbed in the back, with a wound that just missed puncturing his lungs.

Brown (2004) studied differences in responses by the criminal justice system to assaults committed by males and females equated for severity. Using police and prosecutorial case files, Brown examined 2,044 cases where the man was charged, 155 where the woman was charged, 118 where both were charged, and 617 where a complaint was filed but neither was charged. Injuries made little difference to charging rates for female offenders (52.5% vs. 60.2% for males). Arrested women were significantly more likely to have used knives or blunt instruments than were arrested men. Brown found that women were more likely to have used weapons and caused injuries, and also to have received more serious charges. They were more than twice as likely to have been charged with aggravated assault or assault with a weapon. Self-defense arguments by men tended to be disbelieved by police. Women who were prosecuted tended to have inflicted higher levels of injury against their victim than prosecuted men and, as with arrested women, were more likely than men to have used weapons. In severe injury cases, 71.4 percent of men and 22.2 percent of women were found guilty. The low percentage of women found guilty was due to "witness problems" (few men being willing to testify). More than half the male victims refused to testify, and female perpetrators of severe injuries had charges withdrawn 77.8 percent of the time. This pattern was reversed for woman victims; the more seriously injured, the more likely they were to testify. In a similar study, Henning and Renauer (2005) found the same thing: almost one-half (47%) of the cases involving women arrested for domestic violence against a heterosexual intimate partner were rejected by prosecutors, and another 16 percent were dismissed by a judge. Legal factors, such as a defendant's prior criminal arrests, use of a weapon, victim injury, and, most important, the type of arrest (dual vs. single arrest), all affected prosecutors' decisions to take these cases. Female defendants arrested for offending against a male intimate partner were treated more leniently than either male defendants or women arrested for domestic offences involving other types of relationship (for example, familial or homosexual).

When we examine the Buzawa et al. (1992) study or the Brown (2004) study, it becomes apparent that the criminal justice system under-responds to male victimization. The situation for males is now much the same as it was for women in 1965. Hence, government "victim" surveys of intimate violence based on crime report data (for example, police arrest) underestimate the incidence of male victimization. For this reason, the studies using the CTS, which focuses on conflict rather than on crime victimization, provide a fuller estimate of actual violence.

How We See Abuse

What constitutes "abuse" varies considerably according to the gender of both the perpetrator and the victim. Sorenson and Taylor (2005) implemented a random-digit dialed survey in four languages of 3,769 adults in the Los Angeles area. Respondents were presented with five vignettes in which characteristics of the victim, assailant, and incident were experimentally manipulated. The vignette variables (assailant's motive, type or intensity of abuse, whether alcohol was involved, presence of weapons, presence of children, and frequency of abuse) and respondent characteristics were examined using a multivariate log regression. Respondent characteristics turned out to be unrelated to their judgments. Judgments about women's violence against male intimates compared to the opposite (men's violence against female intimates) were less harsh and took contextual factors into account. The type of violence and the presence of a weapon played a central role in respondent judgments. Across vignettes, male violence was seen as more likely to be illegal, suggested that the police should be called, indicated that the assailant should be arrested and should serve jail or prison time, and suggested that a restraining order should be issued. It is important to note that while some of the abuse types portrayed were physical, others were psychological, involving control or humiliation. Physical abuse (slap, forced sex) was more likely to be perceived as illegal by women when the assailant was a man. In general, the genders agreed more than they disagreed, whatever the event. To test the hypothesis that social norms about female abuse are less clear, Sorenson and Taylor (2005) examined the standard deviation value of the residuals (a measure of response variability). It was 36 percent greater when the assailant was a woman, suggesting either much greater diversity of opinion or lack of a clear cut norm (or both).

Not only the general public but also professional psychologists rate aggression as less serious when it is performed by females and even when it is psychological aggression (Follingstad et al., 2004). Two scenarios describing a context and psychologically abusive behaviors but with the genders reversed were given to 449 clinicians (56% male) whose median age was fifty-two. Psychologists rated male-perpetrated behaviour as more abusive and severe than the wife's use of the same actions. Contextual factors (frequency, intent, and perception of recipient) did not affect this tendency. The authors concluded that "the stereotypical association between physical aggression and males appears to extend to an association of psychological abuse and males" (p. 446). Unfortunately, this sometimes leads to serious problems. Coontz, Lidz, and Mulvey (1994) found that clinical predictions of dangerousness made in psychiatric emergency rooms consistently underestimated female dangerousness. Predictions that a male would not be violent were correct 70 percent of the time, but for females, they were correct only 55 percent of the time.

Conclusion

In the best studies, that is, with the largest and most representative (community or epidemiological) samples, when presented without a "crime victim" filter for the data, female violence rates were higher than male rates. Also, in the largest study done on effects of abuse, no gender differences were found. The best predictor of intimate partner violence is not gender but personality disorder.

We have given reasons to believe that female terrorism of males exists if it is defined by unilateral severe violence and instrumental uses of violence. Nor is it the case that men do not suffer from intimate abuse victimization. The LaRoche data analysis clearly refutes this idea. However, police treatment differs in the extreme by gender. Judgments about violence and abuse differ according to the gender of the assailant and the victim. While this is easily understood for acts such as being punched or beaten up, where the consequences may be greater for women victims, the results also apply to actions where the consequences are the same (for example, being stabbed or burned). The view that women are more vulnerable is shared not only by the general public but also by psychologists.

A question raised by the data presented above is this: What happens to these men? No shelter houses exist, and the problem is neither recognized nor sympathized with. A man who presents to the police with an injury is likely to be viewed as having brought it on himself. Ironically, the situation for these men is identical to that of abused women during the Age of Denial.

References

Archer, J. (2000). Sex differences in aggression between heterosexual partners: A meta-analytic review. Psychological Bulletin, 126, 651-680.

Brown, G. A. (2004). Gender as a factor in the response of the law-enforcement system to violence against partners. Sexuality and Culture, 8, 3-139.

Buzawa, E. S., Austin, T. L., Bannon, J., & Jackson, J. (1992). Role of victim preference in determining police response to victims of domestic violence. In E. S. Buzawa & C. G. Buzawa (Eds.), Domestic violence: The changing criminal justice response (pp. 255-269). Westport, CT: Auburn House.

Capaldi, D. M., Kim, H. K., & Shortt, J. W. (2004). Women's involvement in aggression in young adult romantic relationships. In K. L Putallaz & K. L. Bierman (Eds.), Aggression, antisocial behavior, and violence among girls (pp. 223-241). New York: Guilford.

Coker, A. L., Davis, K. E., Arias, I., Desai, S., Sanderson, M., Brandt, H. M., et al. (2002). Physical and mental health effects of intimate partner violence for men and women. American Journal of Preventive Medicine, 23,260-268.

Coontz, P. D., Lidz, C. W., & Mulvey, E. P. (1994). Gender and the assessment of dangerousness in the psychiatric emergency room. International Journal of Law and Psychiatry, 17, 369-376.

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

Douglas, E. M., & Straus, M. A. (2003, August, 2003). Corporal punishment experienced by university students in 17 countries and it's relation to assault and injury of dating partners. Paper presented at the European Society of Criminology, Helsinki, Finland.

Dutton, D. G. (1994). Patriarchy and wife assault: The ecological fallacy. Violence and Victims, 9, 125-140. Dutton, D. G. (2002). The abusive personality: Violence and control in intimate relationships. (Revised Paperback Edition). New York: Guilford Press.

Dutton, D. G. (2006). Rethinking domestic violence. Vancouver: University of British Columbia Press.

Dutton, D. G., & Nicholls, T. L. (2005). The gender paradigm in domestic violence research and theory: Part 1--The conflict of theory and data. Aggression and Violent Behavior, 10, 680-714.

Ehrensaft, M. K., Moffitt, T. E., & Caspi, A. (2004). Clinically abusive relationships in an unselected birth cohort: Men's and women's participation and developmental antecedents. Journal of Abnormal Psychology, 113, 258-270.

Elliott, F., Huizinga, D., & Morse, B. J. (1985). The dynamics of delinquent behaviour: A national survey progress report. Boulder: Institute of Behavioral Sciences.

Felson, R. B. (2002). Violence & gender re-examined. Washington, DC: American Psychological Association.

Felson, R. B., Ackerman, J., & Yeon, S. J. (2003). The infrequency of family violence. Journal of Marriage and the Family, 65, 622-634.

Fiebert, M. S. (1997). Annotated bibilography: References examining assaults by women on their spouses/partners. In B. M. Dank & R. Refinette (Eds.), Sexual harassment and sexual consent (Vol. 1, pp. 273-286). New Brunswick: Transaction Publisher.

Follingstad, D. R., Bradley, R. G., Helff, C. M., & Laughlin, J. E. (2002). A model for predicting dating violence: Anxious attachment, angry temperament and need for relationship control. Violence and Victims, 17, 35-47.

Follingstad, D. R., DeHart, D. D., & Green, E. P. (2004). Psychologists' judgments of psychologically aggressive actions when perpetrated by a husband versus a wife. Violence and Victims, 19, 435-452.

Giordano, P. C., Millhonin, T. J., Cernokovich, S. A., Pugh, M. D., & Rudolph, J. L. (1999). Delinquency, identity and womens' involvement in relationship violence. Criminology, 37, 17-40.

Henning, K., & Renauer, B. (2005). Prosecution of women arrested for intimate partner abuse. Violence and Victims, 20, 171-189.

Hines, D. A., Brown, J., & Dunning, E. (2003). Characteristics of callers to the domestic abuse helpline for men. Family Violence Lab: University of New Hampshire.

Jacobson, N. S., Gottman, J. M., Waltz, J., Rushe, R., Babcock, J., & Holtzworth-Munroe, A. (1994). Affect, verbal content, and psychophysiology in the arguments of couples with a violent husband. Journal of Consulting and Clinical Psychology, 62, 982-988.

Jaffe, P., Lemon, N., & Poisson, S. E. (2003). Child custody & domestic violence: A call for safety and accountability. Thousand Oaks: Sage.

Kessler, R. C., Molnar, B. E., Feurer, I. D., & Appelbaum, M. (2001). Patterns and mental health predictors of domestic violence in the United States: Results from the National Comorbidity Survey. International Journal of Law and Psychiatry, 24, 487-508.

Kokko, K., & Pulkinnen, L. (2005). Stability of aggressive behavior from childhood to middle age in women and men. Aggressive Behavior, 31,485- 497.

Laroche, D. (2005). Aspects of the context and consequences of domestic violence. Situational couple violence and intimate terrorism in Canada in 1999. Quebec City: Government of Quebec.

Magdol, L., Moffitt, T. E., Caspi, A., Newman, D. L., Fagan, J., & Silva, P. A. (1997). Gender differences in partner violence in a birth cohort of 21-year-olds: Bridging the gap between clinical and epidemiological approaches. Journal of Consulting and Clinical Psychology, 65, 68-78.

Marshall, L. L. (1992a). Development of the Severity of Violence Against Women Scales. Journal of Family Violence, 7, 103-121.

Marshall, L. L. (1992b). The severity of violence against men scales. Journal of Family Violence, 7, 189-203.

Mechem, C. C., Shofer, F., Reinhard, S. S., Hornig, S., & Datner, E. (1999). History of domestic violence among male patients presenting to an urban emergency department. Academic Emergency Medicine, 6,786-791.

Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in antisocial behavior. Cambridge: Cambridge University Press.

Morse, B. J. (1995). Beyond the Conflict Tactics Scale: Assessing gender differences in partner violence. Violence and Victims, 4, 251-271.

Pimlott-Kubiak, S., & Cortina, L. M. (2003). Gender, victimization, and outcomes: Reconceptualizing risk. Journal of Consulting and Clinical Psychology, 71, 528-539.

Serbin, L., Stack, D. M., De Genna, N., Grunzeweig, N., Temmcheff, C. E., Schwartzman, A. E., et al. (2004). When aggressive girls become mothers. In K. L Putallaz & K. L. Bierman (Eds.), Aggression, antisocial behavior and violence among girls (pp. 262-285). New York: The Guilford Press.

Sorenson, S. B., & Taylor, C. A. (2005). Female aggression toward male intimate partners: An examination of social norms in a community-based sample. Psychology of Women Quarterly, 29, 79-96.

Stets, J., & Straus, M. A. (1992). The marriage license as a hitting license. Physical violence in American Families. New Brunswick, NJ: Transaction Publishers.

Straus, M. A. (1980). Victims and aggressors in marital violence. American Behavioral Scientist, 23, 681-704.

Straus, M. A., & Gelles, R. J. (1992). How violent are American families? Estimates from the National Family Violence Resurvey and other surveys. In M. A. Straus & R. J. Gelles (Eds.), Physical violence in American families (pp. 95-108). New Brunswick, N J: Transaction Press.

Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996). The revised Conflict Tactics Scale (CTS-2). Journal of Family Issues, 17, 283-317.

Vasquez, D., & Falcone, R. (1997). Cross gender violence. Annals of Emergency Medicine, 29, 427-429.

Watson, J. M., Cascardi, M., Avery-Leaf, S., & O'Leary, K. D. (2001). High school students' responses to dating aggression. Violence and Victims, 16, 339-348.

Williams, S. L., & Frieze, I. H. (2005). Patterns of violent relationships, psychological distress, and marital satisfaction in a National sample of men and women. Sex Roles, 52, 771-785.