In the study of intimate partner violence (IPV), there is perhaps no controversy greater than the question of whether women are as violent as men. According to Pagelow (1984)), the controversy dates back to around 1977 when sociologist Suzanne Steinmetz began writing and talking about what she termed the battered husband syndrome. The controversy has been intense and acrimonious, with charges and counter-charges of selective data reporting (Pagelow 1984), selective citation (Medeiros and Straus 2006) and misrepresentation of the evidence (Straus 2009), ignoring disconfirming evidence due to belief perseverance and groupthink (Dutton and Nicholls 2005), and accusations of those with opposing views providing “comic book caricatures” and “gross misrepresentations” (Johnson 2011, p. 295) of opposing positions. Indeed, the controversy about the use of IPV by both men and women seems so entrenched that in different decades, commentators have suggested that it cannot be resolved (Straus 1999; Winstok 2011).

Archer’s (2000) meta-analysis of over 80 studies of male and female violence in intimate partner relationships reported that women were slightly more likely than men to use physical aggression and to use physical aggression more frequently. Women were also more likely than men to be injured through IPV. The vast majority of studies examined by Archer were of student and community samples; only 13 were from agency or clinical samples. Archer reported some supplementary analyses showing that in shelter and abuse abatement treatment programs, more men than women engaged in assaults on intimate partners. These findings, however, were described as “limited,” “small scale” (Archer 2000, p. 664), and biased. For example, the shelter samples consisted of only women who reported on both their own acts and those of their partners (O’Leary 2000). Thus, an important limitation of Archer’s meta-analysis was the lack of in-depth analysis of male and female IPV among agency or clinical samples. As pointed out by Straus (1999)), studies of large, representative population samples or large, community cohorts are most applicable to the development of primary prevention programming, but they are not generalizable to clinical samples.

Hamberger (2005) reviewed the literature on male and female IPV among clinical samples by using a combination of single-sex studies and 18 direct male–female comparisons. Hamberger concluded that bi-directional violence was the most common pattern of IPV, with sole perpetration by either males or females being relatively rare. Hamberger also evaluated many aspects of the context and impact of IPV and found that compared to women, men initiated more IPV and were more likely to be the first in the relationship to ever use IPV. Men who engaged in IPV were also more likely than women to be motivated by control, whereas women were more motivated by self-defense and retaliation for previous assaults. In addition, women were found to experience more fear of their violent partner and the violence, to sustain injury and to be more severely injured, and to experience clinical levels of depression.

Although the Hamberger (2005) review was the first comprehensive, in-depth review of male and female IPV in clinical samples, it had a number of limitations. First, many of the studies reviewed were preliminary and unpublished. In addition, several studies reviewed presented data for one sex only, relying on self-report and report of partner’s behavior from the same respondent. Other studies investigated couples but collected data from conjoint interviews, and most studies consisted of small samples. Likewise, there has been criticism (e.g., Archer 2000) of other clinical research for relying on samples comparing men and women from disparate populations (e.g., female victims from shelters and male perpetrators in contact with police). In recent years, researchers have increasingly begun to examine samples that include both men and women drawn from similar sources, thus providing more “fair” gender comparisons in clinical samples (Dobash and Dobash 2004).

Another criticism of using clinical samples has focused on small sample sizes and the fact that clinical samples comprise a very small proportion of all IPV, and as a result, are of extremely limited value for developing general theories about IPV (e.g., Medeiros and Straus 2006). Although we would agree that exclusive reliance on clinical samples is not appropriate for general theory development, we also argue that the study of clinical samples is legitimate in its own right. First, clinical samples (and their related populations) are, by definition, of interest to clinicians and policymakers alike. These are the individuals who will be arrested, prosecuted, ordered to treatment, involved in family court actions, and seek shelter and other advocacy services as well as emergency, medical, and mental healthcare assistance. Further, women may have different intervention needs and different risk factors for violence and for recidivism than men. Therefore, in order to develop effective interventions, laws, and policies, it is important to understand the nature of the most directly affected population.

Our second argument regarding the importance of using clinical samples is that this population experiences heavy morbidity and mortality. For example, IPV has a profound impact on health (Black 2011), mental health (Bonomi et al. 2006), and increased healthcare utilization (Rivera et al. 2007). In addition, although females are six times more likely than males to be killed by an intimate partner, both sexes are at risk of homicide from IPV (Cooper and Smith 2011). Recent estimates by the Centers for Disease Control and Prevention place the economic burden from the health impact of IPV at over eight billion dollars per year (Max et al. 2004). Therefore, despite being a small proportion of IPV in the general population, clinical samples comprise a large share of the suffering from IPV, and they are heavy consumers of resources designed to alleviate suffering and to bring justice in efforts to end IPV.

To summarize, there are three reasons for conducting a review of gender differences in IPV among clinical samples. First, major reviews to date have focused primarily on large, representative sample surveys and have not included sufficient numbers of clinical samples, which leads to problems with generalizing such studies to clinical populations. Second, clinical populations are directly relevant to clinicians, law enforcement, the judiciary, and policy makers. Third, clinical samples are large consumers of societal resources due to morbidity and mortality. Thus, the study of clinical samples is important and necessary.

The Current Review

The purpose of the present review is to update and extend the work of Hamberger (2005) by examining literature published since 2002 regarding female and male experiences of IPV among clinical or agency samples. A goal of this review was to examine and evaluate sex differences in violence participation rates and severity. Another goal was to evaluate gender differences in use of various types of violence (e.g., psychological abuse, sexual assault). Finally, the review will report on gender differences in violence motivations, and consequences (including injuries and fear).

Method

Inclusion Criteria

We selected peer-reviewed studies that included respondents recruited from some type of clinical or service setting, such as domestic violence advocacy programs; abuse abatement programs; medical, law enforcement, and criminal justice settings; family court; marital and family therapy; substance abuse programs; and mental health settings. Thus, participants from such settings were not randomly recruited but were seeking services for violence or came into contact with a helping or criminal justice system. Studies of dating violence and community cohort samples that did not include some type of agency utilization or help seeking were excluded. We also restricted our sample to studies of adult heterosexual couples in Western countries, as different issues are likely to be present with homosexual couples or non-Western countries. We did not examine intimate partner homicide or stalking unless this was raised in a study of IPV.

Another inclusion criterion was date of publication. We assumed a three-year lag time between completing a manuscript and publication. Therefore, to avoid duplication of articles reviewed by Hamberger (2005), we included articles that were published between 2002 and June 2013 (so long as they were not included in the earlier review).

A third inclusion criterion was gender comparison. Studies that directly compared male and female respondents on their use and/or experience of IPV were included. Studies that investigated only males or only females were excluded.

Procedure

Three strategies were used to identify articles for review. First, we conducted a comprehensive PsycInfo search using the following key words: [“sex differences” OR “gender differences” OR “gender”] AND [“intimate partner violence” OR “domestic violence” OR “spouse abuse” OR “intimate partner abuse” OR “partner abuse” OR “family violence”]. We also searched for “female offenders” and each violence term. Further, we conducted a Google Scholar search using the key words “men’s and women’s use of intimate partner violence”; “sex differences in IPV perpetration and victimization”; “gender and intimate partner violence”; “women as perpetrators of IPV”; and “men as victims of IPV.”

The second search strategy used was inspection of the table of contents of six specialty journals that frequently publish articles on gender differences in IPV. We searched the tables of contents for each journal from the year 2002 to June 2013. The journals searched were Journal of Interpersonal Violence, Journal of Family Violence, Journal of Aggression, Maltreatment & Trauma, Violence & Victims, Psychology of Women Quarterly, and Violence Against Women. The third strategy used to identify articles was to scan the reference section of each selected article for other articles not identified by the other two approaches. Our search yielded 64 articles that compared male and female use of intimate partner violence.

Results

Following Tolleson and Gross (2009; citing Belknap 2001), we organized the review into methods (the how of the violence), motivation (the why of the violence), context and risk factors (e.g., environmental and situational variables relevant to understanding IPV), and consequences (the actual physical and psychological consequences of the violence). In this Part I article, we focus on those areas that were covered in Hamberger (2005)), so as to provide direct comparison to that earlier review: methods, motivation, consequences of injury and fear, and overall victimization.

Methods (The How of the Violence)

The question of gender symmetry has commonly been addressed through participation rate analysis—that is, analyzing gender differences in acknowledgement of violent behaviors, typically using the Conflict Tactics Scale (Straus 1979). Hamberger (2005)) concluded that the predominant pattern of intimate partner violence within clinical sample couples appears to be bi-directional, suggesting roughly equal participation rates between men and women.

Participation Rates in Intimate Partner Violence.

Several studies compared males and females on rates of physical violence perpetration. In general, the research has suggested few gender differences. Among emergency room (ER) patients in a Level 1 trauma setting, Houry et al. (2008)) found that males and females reported very similar rates of IPV perpetration (females, 7 % vs. males, 5 %). Among couples seeking relationship therapy at a Veterans Administration setting, Teten, Sherman, and Han (2009) found that the prevalence of one-sided violence was slightly higher for males (56 %) than for females (44 %), but was not statistically significant. Robertson and Murachver (2007) reported no significant gender differences for IPV perpetration among incarcerated men and women.

Other studies, however, have found gender differences, particularly when surveying clinical samples on the experience, as opposed to perpetration, of IPV. Selic, Pesjak, and Kersnik (2011) noted that among men and women in primary care medical practices, women reported experiencing significantly more physical violence in the past five years than did men (77.6 vs. 22.4 %, respectively). Walton et al. (2009)) found that women reported higher rates of both IPV victimization (8.2 vs. 6.1 %) and perpetration (6 vs. 2.3 %) in response to a single question for each. Schneider et al. (2009) asked a single question about lifetime physical IPV victimization of men and women entering substance abuse treatment and found that women reported higher victimization rates (46.7 %) than men (9.5 %). Dixon et al. (2007) studied women and men who underwent forensic psychological evaluations following allegations of child abuse. Of the 66 parents (40.7 % of the total sample) found to have perpetrated violence against both an intimate partner and a child, 43 were fathers (65.2 %) compared to 23 mothers (34.8 %); among parents determined to be victims of IPV, 92 % were mothers (Dixon et al. 2007). See Table 1 for a summary.

Table 1 Gender differences in IPV participation rate
Full size table

Severity or Level of Physical IPV

Another approach to assessing gender differences in IPV is to evaluate severity of physical violence used. This section has been divided into studies that found no gender differences, those reporting women use more severe violence, and those reporting men use more severe violence. See Table 2 for a summary.

Table 2 Gender differences in IPV physical violence severity
Full size table

No Gender Differences

Kernsmith (2005a) reported no gender differences between women and men participating in court-ordered batterer intervention programs. Feder and Henning (2005) found no gender differences among dually arrested men and women in weapon use during the index offense and in prior use of both minor and severe IPV. In divorcing couples in mediation, Tanha et al. (2010) observed no gender differences in use of physical violence.

Women Use More Severe Violence than Men

One study found that women perpetrated more low-level violence than men (Taft et al. 2010). Several studies observed that women engaged in more severe levels of IPV than men, particularly weapon use (Babcock et al. 2005; Busch and Rosenberg 2004; Melton and Belknap 2003; Melton and Sillito 2012). In a study of couples who volunteered for alcohol abuse treatment, Drapkin et al. (2005) found that females were more likely than male partners to endorse engaging in more hitting, biting, kicking, beating up, or use of severe violence from the CTS (23 vs. 11 %). Hester (2013)) reported similar (though non-significant) male–female percentage differences showing that women were more likely to use a weapon than men who were arrested for IPV, particularly in dual perpetrator situations. Wupperman et al. (2009)) studied female partners of men in treatment for IPV, and they reported different and even contradictory results, depending on the mode of inquiry. Using self-report of perpetration, women were found to use more mild and severe violence than men. However, using partner report, men were observed to use more mild and severe violence than women. It should be noted that the sample size studied by Wupperman et al. was quite small (n = 22) and further hampered by a low female study participation rate (30 %).

Men Use More Severe Violence than Women

Melton and Belknap’s (2003) qualitative analysis of police report content found that men showed a pattern of increasing severity of violent acts and used more (and more varied) weapons than women, such as a machete and a box fan. Melton and Sillito (Melton and Sillito 2012) found that men eclipsed women in grabbing their partner. Feder and Henning (2005) reported that men (vs. women) arrested for IPV had higher total physical abuse scores on a physical abuse scale. In addition, compared to women, men reported more serious prior instances of IPV, including strangling and weapon use. Basile (2004) observed that male respondents in restraining order actions were more likely than female respondents to slam their partner against a wall or use an object against them. Busch and Rosenberg (2004) observed that men used more severe violent tactics during the index offense than women, including punching, kicking, strangling, and head butting; on the other hand, women and men did not differ in having used at least one severe violent tactic during the index offense. Phelan et al. (2005) found that women emergency department patients were over three times as likely as men to report severe violence victimization (100 vs. 29 %) and were more likely to report being beaten up, strangled, or having a knife or gun used against them.

Three studies of gender differences in IPV severity were conducted with military personnel, with sample sizes ranging from 2,991 to over 20,000 cases (Brewster et al. 2002; McCarroll et al. 2004; Taylor and Pittman 2005). In these military studies, violence severity for each case was determined by multidisciplinary case review committees. Each study reported that males perpetrated significantly more severe physical violence than females. McCarroll et al. (2004) also found that females experienced more severe violence in both mutual violence cases and non-mutual violence cases. Hester (2013) studied IPV offenders entering the criminal justice system in the North East of England over a three-year period (92 % of perpetrators were male and 91 % of victims were female). Males perpetrated more physical violence than females, though the severity or type of violence was not reported; however, men in Hester’s study who were identified as the sole perpetrator were more likely than women to use a weapon (60 vs. 40 %, respectively).

Two studies examined couples in which the males came in contact with the judicial system due to their violence. In a study of 95 couples, Dobash and Dobash (2004) reported that men were significantly more likely to choke and threaten violence, whereas women were significantly more likely than men to perpetrate no violence in the past year. Gondolf (2012) reported that men were significantly more likely than women to beat, burn, or choke their partner (38 vs. 3 %, respectively) and to use a weapon (20 vs. 8 %, respectively); more women than men, however, reported kicking or hitting their partner (32 vs. 14 %, respectively). In contrast to the other studies reviewed in this section, the research by Gondolf and Dobash and Dobash compared women with male partners who had been arrested and referred to the criminal justice system for their IPV. Thus, by definition, the women would be classified as victims. Comparing victims’ use of violence with that of perpetrators may yield a bias toward the conclusion that men use more severe violence than women.

Initiation of and Response to Partner’s Violence

Few studies have investigated actual initiation of and response to violence in a violent episode. One small study by Phelan et al. (2005) examined violence initiation and response among 34 male and female emergency room patients who were in a current violent relationship. Respondents indicated the percent of physical fights in which they or their partners initiated physical force (with significant differences by gender). Via self-report, few women (9 %) initiated violence all the time, and 55 % reported they never initiated violence. All men self-reported initiating violence at least half the time. Via partner report, nearly all women (91 %) reported that their male partners initiated physical violence between 80 and 100 % of the time. Men reported either that their female partners initiated violence half the time (60 % of men) or less than half the time (40 % of men). The researchers also assessed responses to violence, finding that more women (40 % of women vs. none of the men) “almost always” or “always” responded with physical violence to partner-initiated violence, and 30 % of women (vs. 74 % of men) “never” responded to partner-initiated violence with physical violence (p = .003, Phelan et al. 2005). Further, women were significantly more likely to call the police (50 % of women almost/always did, and 95 % of men almost/never did).

Although Kernsmith (2005b) found no gender differences were observed in overall perpetration of physical aggression, 88 % of the women in that study reported using violence in response to the ongoing physical violence of their male partners (i.e., revenge, retaliation, or self-defense). In contrast, only 15 % of male perpetrators reported using physical violence in response to their female partners’ ongoing physical violence. Further, Muftic et al. (2007) found that female partners of male arrestees had significantly fewer prior domestic violence arrests (4.8 %) compared to male partners of female arrestees (19.4 %), suggesting that females arrested for IPV are more likely responding to a pattern of IPV set by the male partners.

Sexual Abuse

The six studies that investigated gender differences in sexual IPV all showed that males perpetrated more sexual abuse than females. Dobash and Dobash (2004) reported that none of the women in their study had used any form of sexual coercion or abuse compared to their male partners. Specific prevalence of forced sex and coerced sex differed by mode of report. Female reports on their partners’ sexual coercion showed that 20 % experienced forced sex and 40 % reported that their partners “demanded sex; male self-reports showed that 3 % perpetrated forced sex and 15 % demanded sex (Dobash and Dobash 2004). Kernsmith (2005a) showed that nearly 60 % of women IPV perpetrators (vs. 29 % of male perpetrators) reported having been sexually abused by their partner. Studying petitioners for temporary restraining orders, Basile (2004)) reported that only males committed sexual coercion. Feder and Henning (2005) reported that arrested men acknowledged more extensive histories of sexual abuse perpetration than women. Taft et al. (2010) found that among chronic pain patients, males were significantly more likely to report perpetrating sexual coercion than females. Tanha et al. (2010) reported that women were significantly more likely than men to report experiencing sexual assault, intimidation, and coercion. Although a limited number of studies, the research on sex differences for sexual IPV is consistent across clinical samples: while some men are sexually victimized by their female intimate partners, women are much more likely to experience such victimization and less likely than men to sexually aggress against their heterosexual intimate partners (see Table 3).

Table 3 Gender differences in sexual abuse (SA)
Full size table

Emotional Abuse

Some studies of emotional abuse have shown few differences between men and women. For instance, in a small, incarcerated sample, Robertson and Murachver (2007) found no gender differences in the perpetration of or victimization from psychological abuse. Studying men and women involved in substantiated mutual and non-mutual abuse in the U.S. Army, McCarroll et al. (2004) reported that the severity of emotional abuse in mutual abuse cases was the same in 90 % of cases. Females, however, experienced more severe emotional abuse in 8.5 % of cases compared to 1.7 % for men.

Other studies found that women used more emotional abuse. Using a novel approach, Basile (2004) translated descriptive statements from restraining order affidavits to CTS-2 scores. Basile reported that females were significantly more likely than males to make harassing phone calls and to threaten to make false IPV allegations. Babcock et al. (2005) also reported that women (vs. men) entering batterer intervention programs were significantly more likely to use Engulfment and Denigration forms of emotional abuse. The sexes did not differ in Dominance or Withdrawal.

On the other hand, several studies found that men used more emotional abuse. Ross (2012) reported that in a sample of men and women referred for batterer intervention, women reported more psychological abuse from their partners on the CTS2. Phelan et al. (2005) administered a gender-neutral version of the Psychological Maltreatment of Women Inventory (PMWI; Tolman 1999 to emergency department patients who reported IPV in the past year. Results showed that men were significantly more likely than women to report experiencing none or only one form of emotional abuse from their female partner in the past year (30 vs. 9 %, respectively), whereas women reported experiencing more forms or types of emotional abuse (82 vs. 22 % respectively) and more forms or types of domination/isolation (64 vs. 22 %, respectively) than males. Among primary care medical patients (Selic et al. 2011), significantly more women than men reported receiving psychological abuse (80.8 vs. 19.2 %, respectively) in the past five years. Further, in an assessment of police reports of IPV calls, Hester (2013) found that men were significantly more likely than women to make threats and engage in harassing behaviors. Tanha et al. (2010) observed that among divorcing couples, women were more likely than men to experience psychological abuse as well as threatened (and escalated) violence.

Finally, three studies point to the importance of measurement in determining gender differences in emotional abuse. For instance, in a sample of men in an IPV abatement program and their partners, self-report showed that women reported perpetrating more emotional abuse than men (91 vs. 62 %, respectively; Wupperman et al. 2009. However, when using partner report, men showed higher levels than women (96 vs. 57 %, respectively). Feder and Henning (2005) found no overall gender differences in emotional abuse among dually arrested couples, though an item analysis did reveal differences: whereas female arrestees were more likely than males to raise their voices and shout, male arrestees were more likely to prevent their partners’ independent activities. Males also made significantly more lethal threats against their partners and children.

Finally, Melton and Belknap (2003) studied both quantitative and qualitative aspects of police reports among men and women arrested for domestic violence. Quantitative analysis showed no gender differences in threats to kill the partner but did find that males made more nonlethal threats than females. Qualitative content analysis, however, revealed that males made threats that were more hostile and detailed than females, both generally and specifically if the victim called police. Melton and Sillito (2012) also reported that the sexes differed in type, but not prevalence, of threats. In particular, males made more lethal threats, as well as threats to harm their victims, and engaged in more name-calling and the use of expletives. Females made threats to male partners if the male partner did something to harm a child, whereas no arrested males made such threats. Melton and Sillito further found that male offenders were significantly more likely than female offenders to engage in stalking behavior. However, the sexes did not differ with respect to other types of intrusive behaviors, such as breaking into a car or house; leaving unwanted messages, phone calls, or gifts; or threatening or causing harm to a new partner. Table 4 summarizes research on gender differences in emotional abuse.

Table 4 Gender differences in emotional abuse (EA)
Full size table

Motivation (The Why of Violence)

Motivation and Situational Contexts of Violence

In a sample of men convicted for their violence, 75 % of their female partners reported using violence in self-defense, whereas 6.3 % of men did so (Dobash and Dobash 2004). Using a checklist developed by Follingstad et al. (1991), Kernsmith (2005a) studied the situational context as well as motivation for IPV among court-ordered perpetrators. Males reported using IPV in the following situations: partner nagging (40 %) or started an argument (32 %), and being under stress (28 %). In contrast, female perpetrators reported using violence when they felt their partners disrespected them (48 %), were trying to control them (37 %), and were not listening (30 %). No gender differences were found for self-defense or for disciplining the partner or exerting power. Females did, however, eclipse males on using violence to get back at their partners and to punish their partners (i.e., “striking back for abuse”; Kernsmith 2005a).

Henning et al. (2005) studied attributions of blame for IPV (including a self-defense motivation) in 1,267 men and 159 women arrested for domestic violence. Results showed that 65.4 % of females endorsed self-defense compared to 50 % of males (p < .01). Ross (2011) studied self-reported motivations for IPV by 30 women and 56 men in court-ordered IPV abatement counseling and found that men and women reported a similar “top three” list of motivations, but in different order. The top three motivations for men were retaliation, emotional dysregulation, and self-defense; for women, the order was self-defense, retaliation, and emotional dysregulation. Women were significantly more likely than men to report being motivated by self-defense. There were no gender differences for the other two top-rated motivations. While there were no gender differences for dominate/punish, it was not a highly rated motivation for either sex. Ross also observed that females reported higher rates of controlling behaviors for themselves and for their partners than did men. Further, women showed greater variability than men, suggesting that women may be a more diverse group or that the smaller standard deviation for men could reflect a systematic under-reporting bias.

Tanha et al. (2010) found that women experienced significantly more coercive control from their male partners than men reported experiencing from their female partners, as measured by the (partner-reported) Relationship Behavior Rating Scale (Attala et al. 1994). However, study results also showed that the pattern of relationship between coercive control and victimization was similar for both men and women. Tanha et al. interpreted the findings as indicating that while men are generally more coercive than women, when women do use coercive control, they use similar tactics as men. See Table 5 for a summary.

Table 5 Gender differences in motivation and situational contexts for IPV
Full size table

Consequences of IPV

Injuries Inflicted

Research with clinical samples has elucidated few gender differences regarding infliction of injury by male and female IPV perpetrators, either generally (Basile 2004) or during the index offense for which the perpetrator was arrested (Busch and Rosenberg 2004; Feder and Henning 2005; Henning and Feder 2004; Melton and Sillito 2012). Busch and Rosenberg (2004) noted that females were more likely to inflict injury using a weapon or object, whereas men were more likely to inflict injury with their hands and feet or through head butt. Almost all of the studies that did not find sex differences in injury infliction based their findings on review of criminal justice records (e.g., police report, arrest affidavit). One study that reported no gender differences from criminal justice data did find from interview that male dual arrestees were more likely to acknowledge inflicting injury than females (Feder and Henning 2005).

Injuries Sustained

A few studies have investigated gender differences in injuries sustained from IPV. In general, this group of studies has shown that women are more likely to be injured than men. For instance, Ross (2012) found that in a batterer intervention sample, women reported having sustained significantly more injuries. Among IPV arrestees, women also reported experiencing more injuries than men (24 vs. 7 %, respectively; Busch and Rosenberg 2004). Using a national database of 468,451 nonfatal gunshot wounds and 97,697 nonpenetrating gun injuries, Wiebe (2003) found that women sustained more gunshot injuries and nonpenetrating gun injuries than men. In addition, women were 3.6 times more likely to be shot by a spouse or ex-spouse than by a stranger. When shot by a spouse or ex-spouse, women were twice as likely as men to suffer a head or neck injury. In their emergency department study, Phelan et al. (2005) found that women were more likely than men to sustain injuries due to IPV, and they reported higher rates of lifetime and past-year injury than men (100 vs. 39 % and 100 vs. 30 %, respectively).

Using a subsample of men and women who participated in the National Violence Against Women Survey (NVAWS, Tjaden and Thoennes 2000) who sought clinical intervention for their IPV-related injuries, Arias and Corso (2005) compared injuries sustained by males and females and the type of medical intervention they sought. Although the NVAWS consisted primarily of a national probability sample, the Arias and Corso study investigated a subsample of respondents who sought clinical intervention for their IPV-related injuries. It was determined that more women sustained injuries than men (21.9 vs. 7.3 %, respectively). Further, women suffered more of the following types of injuries than men: central nervous system and internal injuries, broken bones, broken teeth, burns, scratches, bruises, and welts. Men suffered more lacerations and cuts than women.

Fear and Intimidation

Hamberger (2005) argued that fear is an important dynamic in IPV because of its aversive nature. Fear constitutes aversive arousal that the victim is motivated to reduce through a variety of tactics, including compliance and acquiescence. Melton and Belknap (2003) and Melton and Sillito (2012) studied police report narratives and found that victims of male offenders reported more fear for their safety than victims of female offenders. Basile (2004) found no gender differences among restraining order petitioners who endorsed the check box “fear of the offending partner.” In contrast to Basile,, Phelan et al. (2005) reported that 70 % of women (vs. 4.3 % of men) reported feeling strongly afraid when their male partners initiated violence, while 85% of males stated they experienced no fear when their female partners initiated physical violence. Phelan et al. also reported that significantly more females felt intimidated by their partners’ size than did males (36 % vs. 0, respectively).

In samples of men and women in batterer intervention programs, Ross (2012) and Kernsmith (2005a) found that female perpetrators were significantly more likely to report being afraid of their partners’ initiated violence. Kernsmith also found that females reported feeling more scared, powerless, and weak, and experienced a threat to their personal liberty. This latter finding is similar to research by Dobash and Dobash (2004) who reported that 79 % of female respondents felt frightened when their partner used violence. In addition to feeling fear, 60 % of the women reported feeling helpless, 65 % reported feeling alone, 57 % felt trapped, and 65 % felt abused (Dobash and Dobash 2004). In contrast, the modal male response was that they were not bothered by their partners’ use of violence (26 %), followed by feeling that their partners were justified in use of violence (20 %). Another 17 % of the men reported they ridiculed their partners for their use of violence. Although percentages were not provided, Hester (2013) reported from qualitative analysis of police reports that men were far more likely than women to be described as creating a context of fear and control by their abusive acts; only one woman was described as having created such an environment.

Kernsmith (Kernsmith 2006) also found that female offenders reported significantly more fear of their partners than did male offenders. However, when both gender and prior abuse variables were used to predict fear, gender no longer predicted generalized fear of one’s partner. Rather, such fear was predicted by prior domestic violence victimization and a history of sexual abuse. Further, a significant Sexual Abuse x Gender interaction showed that men who experienced sexual abuse as a children reported experiencing more fear at the time of the abusive incident than women with a history of sexual abuse. Table 6 summarizes the research on gender differences in consequences of IPV.

Table 6 Gender differences in consequences of IPV
Full size table

Overall Victimization

Some studies of the impact of IPV incorporated participation rate while also speaking to a bigger picture of how violence affects men and women. The concept of overall victimization goes beyond participation rate counts and examines patterns of violence, impact of violence, and history of intimate partner violence in a person’s life to present a picture of victimization. For example, some studies examined prior IPV, whether in the current relationship or previously. Kernsmith (2006) observed that women participants in IPV offender treatment reported significantly more physical IPV victimization in prior relationships than men (58 vs. 24 %, respectively) and higher levels of sexual abuse as a child or as an adult (59 vs. 29 %, respectively). In addition, men were more likely than women to have stalked their intimate partners/victims. Other studies (Melton and Sillito 2012; Phelan et al. 2005; Tolleson and Gross 2009) found that significantly more women reported prior IPV in the relationship than men. Phelan et al. (2005) also observed that women experienced more past-year and lifetime IPV-related injuries. Additional researchers have reported that compared to women arrested for IPV, men have longer records of prior IPV arrests (Busch and Rosenberg 2004; Henning and Feder 2004).

As previously noted, Houry et al. (2008), who found no gender differences in IPV perpetration, victimization, or both (as noted above), also compared men and women with the Women’s Experience with Battering (WEB) Scale, which measures degree of victimization from IPV and uses gender-neutral language. Results showed that significantly more women than men scored above the cut-off for severe victimization. Houry et al. concluded that although men and women did not differ with respect to IPV perpetration and victimization, the experience of victimization was greater for women, suggesting gender-based differences in emotional impact and reaction to IPV, loss of power and autonomy, and increased danger and vulnerability. Further, Tanha et al. (2010) reported in their sample of divorcing couples that, despite men and women exhibiting similar patterns of abusive behaviors and use of coercive control, overall, women experienced much greater victimization than men as evidenced by higher levels of psychological abuse and coercive control, threats of and escalated physical violence, sexual assault, intimidation and coercion, and total intimate partner violence. Please see Table 7 for a summary.

Table 7 Gender differences in overall victimization
Full size table

Discussion of Part I

The purpose of the present paper was to review the literature on gender differences in use and impact of IPV among clinical samples and to update and extend the work of Hamberger (2005). Hamberger (2005) had concluded that, although both men and women in clinical samples used physical IPV, based on a deeper contextual analysis, women were more likely to be injured and more severely injured, to suffer more negative emotional consequences from IPV, and to experience more fear. Hamberger also showed that in clinical samples, men were more likely than women to initiate violent acts and to have committed more violent assaults in the history of the relationship. Further, compared to women, men were more likely to be motivated by control in their use of violence, whereas women were more likely to be motivated by self-defense or retaliation.

We now discuss findings from the present review as they relate to conclusions drawn by Hamberger (2005). Findings of Part I also have research and clinical implications. Clinical and research implications of Part I will be combined with those of Part II and included in a General Discussion at the end of Part II.

Research Development Since 2005

Participation Rates

Overall, findings regarding rate of violence perpetration seem to show few gender differences across a number of clinical settings. These findings are generally consistent with the observations of Hamberger (2005), though few studies in the present review categorized perpetration in terms of bi-directional violence, male-only perpetration, or female-only perpetration. Many of the studies, however, have limitations that qualify conclusions.

First, some studies consisted of small samples (e.g., Robertson and Murachver 2007). Also, a finding of gender symmetry in participation rates seemed to depend on whether a study was asking about perpetration or victimization. Studies of perpetration rates tended to demonstrate nonsignificant differences (e.g., Teten et al. 2009), whereas studies that found asymmetry tended to ask about violence victimization (e.g., Selic et al. 2011). Additionally, how violence is measured may also be important. While several studies used standardized measures of IPV (e.g., Teten et al. 2009), others used adaptations of violence measures (e.g., Dixon et al. 2007) or even a single question about IPV (e.g., Walton et al. 2009).

Physical Violence Severity

The research as to whether women or men in clinical samples use more severe forms of physical IPV is mixed. As with the findings related to participation rate, the data were clear that women and men both engage in violence and do so at fairly high severity. Several studies showed either no gender differences (e.g., Feder and Henning 2005), or that women used more severe violence than men (Drapkin et al. 2005). However, several studies also showed that men use more severe violence than women (e.g., Melton and Sillito 2012). Reasons for this discrepancy are not entirely clear. Some of the observed asymmetry may be due to comparing female victims with male perpetrators. On the other hand, Hamberger and Guse (2002) did not find such bias in a study comparing men and women arrested for IPV and a group of shelter-based women. In the Hamberger and Guse study, women in both groups reported using severe violence at rates similar to and not significantly different than the men. Study sample size could have influenced findings related to gender differences in violence severity Studies that contained the largest samples, such as those from the Army and Air Force settings, showed that, on average, men perpetrate more severe violence than women. Method of determining IPV severity may also be important. In the military studies, severity was determined by a panel of expert investigators. Other studies utilized check boxes on police report forms (e.g., Melton and Belknap 2003). In contrast to quantitative check box analyses, qualitative studies of the same police reports yielded different results (e.g., Melton and Belknap 2003). Partner report and self-report seemed to yield different results (e.g., Wupperman et al. 2009). Thus, as this research area progresses, it will be important to specify method of violence assessment. In addition, it may be advantageous to combine quantitative and qualitative methods to provide a more complete picture of violence in relationships.

Initiation and Response to Partner’s Violence

Very few studies have examined who initiates or responds to violence. The three available studies (e.g., Kernsmith 2005b; Phelan et al. 2005; Muftic et al. 2007) suggested that women are more likely responding to partner-initiated violence with physical aggression than are males. In addition, several other studies (e.g., Kernsmith 2005a) concluded that women appear to use physical violence more in response to their partner’s initiated violence, even if self-defense is not the primary motivation. However, small sample sizes (Phelan et al. 2005) and relatively indirect measures of violence initiation (Muftic et al. 2007) make conclusions based on this research difficult at present. More work in this area is warranted.

Emotional Abuse

Overall, results of gender comparisons for use and/or experience of emotional abuse among samples of men and women showed mixed results. Some studies reported no gender differences at all (Basile 2004) or that women use more emotional abuse than men (Babcock et al. 2005). Although several studies showed no gender differences in overall prevalence of emotional abuse, when they examined use of different types of emotional abuse, many gender differences were revealed (e.g., Feder and Henning 2005; Melton and Belknap 2003; Melton and Sillito 2012). For example, it was found that males are more likely to make nonlethal, hostile, and specific threats (Melton and Belknap 2003); make lethal threats (Feder and Henning 2005); commit serious emotional abuse (McCarroll et al. 2004); and use more emotional abuse that is controlling of another’s autonomy (Feder and Henning 2005; Phelan et al. 2005). Continued work in this area will require analysis of specific types of emotional abuse that go beyond overall participation rates and scores on emotional abuse measurement scales.

Despite the observed trends in gender differences in emotional abuse among clinical samples, a number of methodological issues prompt caution. These include small sample sizes (Babcock et al. 2005; Phelan et al. 2005; Robertson and Murachver 2007) and nonstandardized measures or use of adapted measures in ways other than their originally developed purpose (Basile 2004; Feder and Henning 2005; Phelan et al. 2005). In addition, the construct of emotional abuse is not as well-established as other forms of IPV (e.g., controlling behavior as a form of emotional abuse versus a motivation for violence). Thus, there may be difficulty generalizing from one study to another or from one study population to another. This is an area that needs more work in the development of gender-specific, standardized, and validated measures.

Summary of Participation Rate Findings

To summarize findings related to forms of violence, there may be a kind of symmetry between males and females in use and severity of physical violence as well as overall emotional abuse. Hamberger (2005) did not review gender differences in use of psychological abuse or sexual abuse. In the present review, the (small) literature on sexual abuse consistently showed that women are more victimized than men. In addition, men appear to perpetrate types of emotional abuse that are different than that of women, with greater impact on autonomy and fear. Women’s psychological abuse, at least within clinical samples, appears to be more expressive (i.e., communicating unhappiness and distress) rather than instrumental (i.e., attempts to control partner’s autonomy or induce fear). Thus, although there is a certain degree of symmetry when the focus is only on physical violence, consideration of sexual IPV, together with what appear to be different forms of psychological IPV characteristic of men and women, leads to the conclusion that asymmetry is noted. When considered in total, then, these results clearly indicate gender differences in IPV and show that women are more victimized and more highly victimized, in general, than men. This suggests that, methodologically, research should focus on the totality of forms of violence that occurs in relationships rather than on a single form only, such as physical abuse tactics. Of course, given the small size of the sexual IPV literature, it will be important to continue to conduct research on this important issue. In addition, while there is some indication that men and women tend to use different forms of psychological IPV, that is an issue that will also need more research going forward.

In clinical samples, it would appear that when investigation ventures beyond the current or index offense and includes assessment of other forms of violence as well as victimization level, the result has been a picture of greater victimization for women than men. This was true whether research was based on instruments designed to assess victimization level (Houry et al. 2008), amount of prior IPV perpetration in the relationship (Feder and Henning 2005), or number of prior IPV arrests in the relationship (Muftic et al. 2007). Further, consistent with Hamberger (2005), the present review found that women experienced more injuries from IPV than men. It is possible that this conclusion is affected by women’s greater likelihood of reporting victimization and men’s greater stigma about reporting their own victimization. However, Hines et al. (2007) found that when given the opportunity, men will report on their victimization. At least some of the present findings are also likely due to women using violence less—and less severely—than men. This summary does not imply that women cannot be classified as perpetrators; researchers have found subsamples of women who appear to use violence in roughly equivalent ways to men (e.g., Tanha et al. 2010). Rather, we would assert that the majority of studies in this area find that men more often initiate and use violence in ways that significantly and adversely affect their partners.

Fear and Intimidation

Hamberger (Hamberger 2005) concluded that women were more fearful of their opposite-sex partners than were men. In the present review, sex differences in fear were found whether inquiry was about fear of one’s partner generally (e.g., Kernsmith 2005b), or fearfulness in the actual violent situation (e.g., Phelan et al. 2005). That women are generally more fearful than men has been found across sample domains, such as individual arrest (Hester 2013; Kernsmith 2005b; Ross 2012), dual arrest (Feder and Henning 2005), victim partners (Dobash and Dobash 2004; Melton and Belknap 2003; Melton and Sillito 2012), and in health care (Phelan et al. 2005).

Research has shown that women are more likely to be fearful of their partners whether or not they are identified as the aggressors in a particular incident. Additionally, the finding that women are more intimidated by their partners’ size is supported by the work of Cullinane et al. (2009)), who argued that size differential (which generally favors males) is important to understanding injury potential. Larger body mass is related to greater force delivered. Thus, on average, it is not surprising that women would experience more fear and intimidation at the prospect of their partner using physical violence against them.

In a review of gender differences in fearfulness, Campbell (2006) noted that biological sex is the most important predictor of fear, with women consistently scoring higher on measures of physical fear than men. Further, Campbell concluded that gender differences in fear account for a considerable proportion of the differences between the sexes in aggression. Of course, these findings are not universal. Although the reported gender differences were statistically significant, the findings also showed that many women were not fearful of their partners, and some men were fearful of their female partners. The findings of Kernsmith (2006) also called for caution in assuming a clear gender difference in fearfulness. Specifically, certain trauma experiences, especially a history of sexual abuse, may be a more important predictor of fear of one’s partner. In addition, Basile (2004) reported no sex differences regarding fear. However, the Basile study is different than the others in that he used a decontextualized measurement of fear (i.e., check boxes completed by respondents to describe their reasons for applying for a restraining order). Such a decontextualized construction of fear may yield different results than a narrative description.

Motivation (The Why of IPV)

Studies have shown conflicting results for gender differences in endorsing the self-defense motive and for the use of the control motive and controlling behaviors. Methodological differences may explain some of the inconsistency. For example, Kernsmith (2005a) measured motivations using a scale that was originally developed for studying dating violence; thus, it is unclear whether the scale was valid for use with clinical samples. Dobash and Dobash (2004) measured self-defense, but did not assess any other potential motivations. Despite these issues, two conclusions have been advanced regarding motivation: (a) women are generally, though not universally, more likely than men to be motivated to use IPV to defend or respond to prior violence initiated against them, and (b), findings related to the control motive among clinical samples are less clear than previously thought. These conclusions will be elaborated below.

First, consistent with the prior body of research on the topic (Hamberger 2005, motivations appear to differ for males and females, with females more likely to report using violence in self-defense or in response to their male partners’ use of violence. However, it is also clear from the literature that women in clinical samples do not universally endorse self-defense as a motivation, and that some men also endorse self-defense. The finding that women also report “striking back for abuse” (Kernsmith 2005a, p. 179) was consistent with other findings reported by Dobash and Dobash (2004), Kernsmith (2005b), and Muftic et al. (2007) that women in clinical samples appear to use violence in response to violence used against them. These findings were also consistent with the conclusion of Saunders (1986) that women in violent relationships do not always easily distinguish between fighting back and self-defense, as well as of Hamberger and Potente (1994) that women in clinical samples who use IPV are often caught up in the dynamics of a violent relationship which they did not initiate and do not control. Thus, their violence is more reactive. It is noted that findings regarding the female self-defense motive were not universal, with some studies showing as many as 50 % of men reporting using self-defense to counter IPV (Henning et al. 2005).

Second, Hamberger (2005) also concluded that men were more likely to be motivated by control and domination. Although Tanha et al. (2010) reported that men used more coercive control than women, women were also found to use coercive control within relationships. Furthermore, none of the studies that assessed self-report showed gender differences in control as a motivation for using IPV. Thus, the concepts of coercive control and self-defense do not appear to be as predictive in understanding sex differences in IPV as previously thought. The current studies do not yet provide enough evidence for the relationship of gender to control; further, it is still unclear what role control plays in violence for men and women.

Though there is an assumption that much domestic violence is in the service of controlling one’s partner, this issue requires more research using validated measurement strategies, as well as qualitative studies, to aid in refining our understanding of women’s and men’s motivations for using IPV. Measurement strategies to assess coercive control in a relationship, such as that used by Tanha et al. (2010), represent a good beginning. Another approach that could be fruitful would be functional analysis of the relationship between actual violent behavior and outcomes. Such an approach would allow examination of violent behaviors and outcomes that reinforce and maintain the behaviors.

It will also be important to distinguish control as a motivation from control as a functional outcome. The construct of control as a motivation implies that the person experiences some type of internal state (conscious or unconscious) that he or she strives to satisfy (i.e., they hit their partner in order to satisfy a need to control them. In contrast, control can also be conceptualized as a functional outcome. That is, a person may hit his or her partner and then the partner does something the assailant desires (e.g., they acquiesce). Thus, the hitting behavior is reinforced, but the victim’s acquiescent behavior is also reinforced by removal of hitting. Hence, the behavior of both the perpetrator and the victim is “controlled” by the contingency. The latter is not a mental or internal process, but a product of the contingencies that have been set up. It will be important for researchers to operationally define their use of the concept of control.

Consequences of Violence

Results of gender differences in inflicting and sustaining injury appear to be contradictory. Men and women are frequently found to inflict injury at roughly comparable rates, while women are more likely than men to sustain more severe and more frequent injury from IPV episodes. There are a number of possible explanations for this finding, as noted below.

Most of the studies of injury infliction involved arrested women and men and focused primarily on injuries that occurred during the incident for which the actors were arrested, giving rise to three factors which may be relevant. First, women who injure their partners are at high risk of being arrested (Hamilton and Worthen 2011). Further, several studies have found that women use more severe violence than men, and they may be more likely than men to inflict injury using a weapon (e.g., Melton and Sillito 2012). Weapon use has also been shown to more strongly predict arrest for women than men (Hamilton and Worthen 2011). Thus, the finding of few to no gender differences in injury infliction could be an artifact of how men and women are “selected” for arrest, adjudication, and subsequent study recruitment. Secondly, assessing injury for a single event, such as the index offence, provides only a limited view of injury in the lives of arrested men and women. In contrast, questions about sustaining injury have typically been asked of help-seeking men and women (e.g., emergency medical treatment) and often use a longer timeframe, such as lifetime or past-year injury. This longer view can capture more incidents in which injury may have occurred. Third, it may be easier and more accurate to identify injuries to oneself rather than injuries that have been inflicted. Given these methodological caveats, it appears that injury rates are similar or higher in women. Future research should examine a broader range of injury infliction as well as injuries sustained for both men and women involved in IPV. In addition, future research should explore a broader history of injury, going beyond index offense analysis.

Overall Victimization

We found considerable gender differences in the experience of IPV, as well as in degree of victimization. Compared to men, women show less prior IPV perpetration, suggesting that women may be less likely to use repeated violence. Compared to men, women’s violence may also be more situational (Ross 2011) and less likely to create a pattern of fear or control (Hester 2013). In addition, women report greater histories of previous injury, as well as greater negative impact, than men (e.g., Houry et al. 2008; Tanha et al. 2010). Thus, it appears that in clinical samples, as Feder and Henning (2005) eloquently summarized, “[W]hen violence was used, women got the worst of it” (p. 166).

Limitations

The primary limitation of the present review is that it is based on research conducted with clinical samples. As such, we limit generalizations about the role of gender and IPV to clinical populations and avoid generalization to the general population. Still, study of clinical populations is important to guide development of interventions and policies to assist those who find themselves in the broad system of care for IPV, even though broad theoretical statements and conclusions about the ultimate etiology of IPV cannot be made.

A number of the weaknesses in the clinical sample literature noted by Hamberger (2005) have been ameliorated in the newer research. Samples were much larger, and there were more direct gender comparison studies available. Still, the lack of measurement tools designed and validated for use with clinical samples of males and females is another limitation hampering firm conclusions regarding gender and IPV.

One weakness of the present review may lie in the methods used to identify articles. Specifically, it is possible that more articles could have been identified had we used additional search engines and that our findings, therefore, may represent an underestimate of the literature in this area. This criticism is mitigated by the fact that we used two search engines and hand-searched the entire tables of contents across a 10-year span of the leading interpersonal violence journals. Further, we searched the Reference sections of all articles read in order to ascertain additional articles. According to Akobeng (2005), a systematic review should identify a systematic search strategy that consists of a number of approaches to identify relevant articles; in this review, we used four different search strategies.

Another limitation of the present review is that we did not conduct a meta-analysis. Although we did take a systematic approach to reviewing the literature (Akobeng 2005, a meta-analysis would have provided a more quantitative analysis of gender differences in the experience of IPV. There are several reasons we did not attempt a meta-analysis. First, although we reviewed many studies (35 for Part I), several of the variables we assessed were comprised of only three or four studies, which would not have been enough to generate reliable findings (e.g., Field 2003 indicated that reliable results are generated in a meta-analysis with over 20 studies). In reviewing this literature, we wanted to take a broad survey of research on gender differences beyond participation rate analyses, which would be most amenable to meta-analysis. Further, we were interested in conducting a more in-depth analysis of methods and findings. For example, an overall test of gender differences in emotional abuse would have led to the conclusion of no sex differences. However, when we evaluated item analyses on measures of emotional abuse, we observed gender differences in the kinds of emotional abuse that were used. Because these findings were based on only two studies, it would not have been feasible to conduct a meta-analysis, and an important finding would have been lost. Further, such in-depth analysis based on small numbers of studies can lead to additional hypotheses to be tested in subsequent research. Of course, as the research in this area continues to evolve, the next step would be to conduct meta-analyses.

Summary and Conclusions

The present review of literature on gender differences in the perpetration, motivation, and impact of IPV in clinical samples published between 2002 and 2013 updates and extends a previous review by Hamberger (2005). Results showed that although both women and men are active participants in acts of physical IPV, women’s physical violence appears to be more in response to violence initiated against them. While both men and women participate in emotional abuse tactics, the type and quality appears to differ between the sexes; men tend to use tactics that threaten life and inhibit partner autonomy whereas women primarily use tactics that consist of yelling and shouting. Additionally, men are the predominant perpetrators of sexual abuse. Analysis of patterns of violence and abuse suggests that women in clinical samples are more highly victimized, more injured, and more fearful of their partners than men. Thus, understanding of IPV among clinical samples will continue to benefit from a gender-sensitive approach to conceptualizing the problem, formulating appropriate interventions, and developing policy. Clinical, policy and training implications will be further explicated in the General Discussion accompanying Part II (Larsen and Hamberger 2015) of this review.