Abstract
This review article examined the gender symmetry debate in light of recent research relating to the feminist and family research perspectives on intimate partner violence, providing a context for rethinking perpetrator programs. The concept of coercive control is considered as an explanatory factor in an attempt to integrate the feminist and family research perspectives. The limited effectiveness of perpetrator programs is examined. Research highlighting potential factors that could improve the effectiveness of perpetrator programs is introduced, followed by a discussion of the rejection–abuse cycle, one attempt to incorporate current research into a more inclusive program. The rejection–abuse cycle identifies a pattern of perpetrator behavior, which links rejection, threat to self, defense against threat, and abuse. Finally, suggestions for changing perpetrator programs are elaborated, incorporating past research, which would make them appropriate for both male and female perpetrators. These implications are contextualized within a meta-theory to provide greater clarity for the development of future perpetrator programs.
One of the most controversial aspects of intimate partner violence (IPV) is the role of patriarchy in understanding and treating men who are violent in their intimate relationships. Although feminist researchers (e.g., Dobash & Dobash, 1988, 1995) have clearly identified patriarchy as the major factor that influences men’s violence toward women, family violence researchers (e.g., Straus, 2006, 2008) have conducted research that presents quite a different view. The family research perspective demonstrates that both men and women report equal amounts of violence between themselves and their partners, usually measured on the Conflict Tactic Scale (Straus, Hamby, Boney-McCoy, & Sugarman, 1996). This anomaly has polarized treatment initiatives in working with men who are perpetrators and has created controversy as to whether or not men can be victimized by women in their intimate relationships. Current treatment programs for perpetrators are heavily focused on ideas drawn from patriarchy, making such treatment regimens of little use for female perpetrator groups. Thus, it is time to reexamine the gender symmetry debate in light of current evidence, in the interests of providing male perpetrators access to treatment informed by a wider range of current research. This article will consider the gender symmetry debate within the light of feminist and family research perspectives, current perpetrator programs, and how the current ideas from research on IPV could be used in developing more universal programs appropriate for all perpetrators.
Feminist Perspective: Dealing With Life-Threatening Crises
From this perspective, it is argued that the dominant pattern between couples is male perpetrators directing abuse toward female victims or survivors. This perspective argues that violence against women is accepted in society largely because of the dominant discourse around which society is structured whereby men are seen as dominant, sexism is rampant, patriarchal norms exist, and women are regularly subordinate to men (Rothenberg, 2003). Although this view acknowledges that women may be violent toward men, this is seen as predominantly because of their attempts to defend themselves against the assaults of men (O’Neill, 1998). Furthermore, the relatively few men who are subjected to female abuse are not in fear of their lives and can more easily leave their female partners (Pence & Paymar, 1993). A study identified three types of self-protective reactions from women in response to men’s violence: nonphysical, physical, and a combination of nonphysical and physical (Downs, Rindels, & Atkinson, 2007). Nonphysical was the most common response; however, this was often followed up with physical when the nonphysical response was ineffective. In this study, there were very few women who initiated violence against their male partners. Thus, the context of women being violent in relationships is important in explaining any reciprocal violence between partners. In arguing for the feminist perspective, Kimmel (2000) pointed out the inconsistency of the gender symmetry debate by noting that men are more violent everywhere else in society but in the home, whereas women are shown to be equally or even more violent in the home. He argued that this perspective does not make sense in light of the evidence and points to an inconsistency that cannot easily be explained. Dobash, Dobash, Cavanaugh, and Lewis (2000) acknowledged that women may be violent against men but that the majority of the reported violence is directed against women (Dobash & Dobash, 1992; Dobash, Dobash, Wilson, & Daly, 1992; Nazroo, 1995). The National Violence Against Women Survey reported that men assaulted women 3 times more than women assaulted men (Tjaden & Thoennes, 2000). Furthermore, for women who separated from their male partners, the rates of violence perpetrated against them in the same survey was 8 times higher than for women who continued to be married (Bachman & Saltzman, 1995). It is clear from these data that women are more likely to be victimized, and increasingly so if they have previously separated, than are men.
The family research perspective relies on particular samples, which are unlikely to find the extreme examples of abuse that support the feminist perspective. In his comprehensive review of the gender symmetry literature, Archer (2000) reported that 37 studies were based on data from college students, 27 studies were based on community samples, 5 studies came from data based on couple treatment programs, 2 studies from refuges for battered women, 3 studies from homeless, and 3 studies were on couples referred for IPV. In addition, Archer reported that 33 studies targeted married cohabiting couples, whereas 47 studies targeted noncohabiting respondents. This review by Archer, which is supportive of gender symmetry in relation to IPV, is thus highly skewed in favor of young people and community samples of which the majority are not cohabiting. Thus, these data are not equivalent to the data where women are coercively trapped in marriages with children that make it very difficult and often dangerous to leave, such as those few studies reviewed by Archer involving shelters, homeless, and couples in treatment.
Kimmel (2002) observed that when considering populations in shelters and emergency care facilities, it is clear that women make up the majority of this population. Thus, it seems that the feminist position has been articulated from extreme samples of male abuse where there are few apparent ways of understanding such senseless violence. Yet even the more likely representative of the American populations samples, such as the National Violence Against Women Survey, support the feminist conclusion that men are more abusive toward their partners than women. What is clear is that the statistics for the two views are usually taken from different populations. Although the feminist perspective relies on crime victimization studies of usually married couples, the family research perspective frequently relies on community samples of young, unmarried couples, where rates of aggression are assessed through self-report. Both these sources of data identify men as more likely to be perpetrators than women in many instances. However, in situations where the reported violence is relatively minor, it is more likely that gender symmetry is reported, a conclusion supported by the Archer (2000) review.
In dealing with the IPV problem, the feminist perspective was very influential in motivating women initially, and subsequently health professionals, to obtain resources to respond to the crises that emerged with women and children being terrified of and terrorized by male partners and fathers in situations demanding immediate action in order to avoid life-threatening situations. Thus, shelters were created to protect these women and children, and steps were taken to ensure that domestic violence was viewed as a crime to be addressed by the police and courts rather than being ignored by patriarchal complacency and some would argue complicity. This response included court orders that required men to stay away from their female partners and children and criminal charges being laid to deal with the violence. Initially, the focus was on getting women and children away from the men who were abusing them. It was evident that women often found it difficult to leave these men because of threats, lack of financial and other resources, and their own fears of the unknown. Subsequently, with the knowledge that these men would get involved in other relationships and create similar patterns of violence, programs were created to help deal with them. Thus, the initial response of the feminist perspective to IPV was very important in dealing with domestic violence in offering hope to female and child survivors who were terrorized by their male partners and fathers, as well as providing male perpetrators the opportunity of relinquishing their violent patterns to help them establish egalitarian relationships that were more sustaining. This initial response was extremely valuable in saving many lives. The phenomenon of similar rates of violence between male and female partners was reported by other researchers, which became known as the family research perspective.
Family Research Perspective: Reflecting on Gender Symmetry Research
The family research perspective relies on evidence from some broad community-based surveys indicating that violence between partners is largely equal and reciprocal, leading to the conclusion that gender symmetry exists and that explanations other than patriarchy must be found to understand the total phenomenon of IPV. Abusive behavior between partners has been examined over a period of many years, with the most common definition based on the subscales of the Conflict Tactics Scale (CTS), which includes self-report of personal and partner use of psychological, physical, sexual, and injury within the past 6 months (Straus et al., 1996). There has been extensive research with a range of populations using this questionnaire. Langhinrichsen-Rohling (2010), in highlighting major controversies in IPV, noted the commonality of bidirectional violence, as one of the most contentious issues in IPV. Various studies have noted the frequent research that identifies bidirectional violence as the most usual form of violence between partners in a range of samples, including couples in therapy (Vivian & Langhinrichsen-Rohling, 1994) and dating couples (Straus, 2008), and in a wide range of community samples rather than in clinical samples in a large meta-analysis (Archer, 2000). In a study of dating violence, it was concluded that it was rare for men to be violent in the absence of violence from their female partners (Skuja & Halford, 2004). Consequently, by attending to female violence, it was argued that women could be protected, as men more commonly retaliated if they were the object of female partner abuse (Kimmel, 2002). Furthermore, there has been some evidence to suggest that gender symmetry is prevalent in countries where there is relative equality between women and men but not in countries where women’s status remains considerably lower than men (Harvey, Garcia-Moreno, & Butchart, 2007). Where there is conformity to sex roles and nonconformity to gender roles, it is more likely for IPV to occur (Reidy, Shirk, Sloan, & Zeichner, 2009; Reidy, Sloan, & Zeichner, 2009). Consequently, the initial understandings relating abuse to patriarchy have been questioned in light of current research, particularly in Western countries where efforts have been made to address gender inequality and also where the abuse that is reported is only minor in nature. In these situations, for example, rather than gender, it has been argued that personality disturbance is a better predictor of IPV (Dutton & Corvo, 2006). However, this research does not mean that injury inflicted is equivalent across gender, as it is clear that men cause more physical injury than women in IPV (Archer, 2000; Stets & Straus, 1990; Warner, 2010). It is also clear that more serious forms of abuse are more likely to be attributed toward men (Kimmel, 2002). Warner (2010) reported that men have higher rates of violence than women in clinical samples or more serious instances where violence has occurred. However, in general population studies of family conflict in the community, Kimmel (2002) concluded that there are relatively high rates of minor forms of domestic violence, which are equally perpetrated by males and females, specifically and particularly where there are low rates of injury. There are those who argue that the CTS is a flawed measure and that its design—(a) (form of abuse) . . . I did this to my partner; (b) (form of abuse) . . . My partner did this to me—is one of the reasons that gender symmetry abuse has been reported. In defense of this argument, Kimmel (2002), in his analysis of Fiebert’s (1997) data review, concluded that only one serious study of IPV found gender symmetry using a measure other than the CTS. Thus, using the CTS, in studies where there is less serious abuse reported, in countries where there is greater amounts of gender equality, and in samples that are more likely to be community based, gender symmetry is commonly reported.
The value of the family research perspective lies in its focus on abuse in the broader community and a consideration of data from a wider range of couples than have been reported from the feminist perspective. These researchers examined community samples and reported surprisingly high rates of IPV in many couples where it was not necessarily expected. This perspective suggests that other factors must be examined in considering abusive patterns of IPV. An exploration of other factors related to abuse does not minimize the violence nor blame the victim, as both partners can be seen as perpetrators as well as victims. The perspective also encourages the development of perpetrator treatment approaches that move beyond gendered explanations of abuse that can be used with males and with females and that may prevent less abusive couples from becoming more serious cases of abuse where women are more frequently the victims. Thus, the family research perspective cannot be seen as an exhaustive approach to IPV, but it certainly offers the potential for a wider range of options than currently exist. For example, dealing with couples’ conflict before it escalates to the extreme violence associated with fear, intimidation, and injury may be a preventative measure. The variable of coercive control has been introduced as a key concept in understanding the integrity of both the feminist and family research perspectives (Johnson, 2006).
Coercive Control: Linking the Feminist and Family Research Perspectives
Although the feminist and family research perspectives have led to different responses in relation to IPV, there have been attempts to find some common ground between them. Bidirectional violence in couples is more complex than the data suggest and has been explained in light of coercive control. Johnson and Ferraro (2000) identified various patterns when they considered both partners who reported bidirectional violence and linked this to controlling behaviors. Control accounted for the most variance in predicting violent behavior. They noted that there were different types of violence in these dyads. Common couple violence was the most frequently occurring type of violence identified and consisted of couples where the conflict got out of control on occasion and led to minor forms of violence. This type of violence was relabeled situational couple violence (Kelly & Johnson, 2008).
Intimate terrorism, on the other hand, was violence perpetrated by one partner, which was usually the male partner, which was repeated and often led to injury (Johnson, 1995). This type of violence has been recently relabeled as coercive controlling violence (Kelly & Johnson, 2008). This violence appeared to be linked to domination or the motive of control over another, which involved manipulative control, and was less frequently occurring than situational couple violence. Finally, this type of violence was usually associated with cases where female partners were involved with women’s shelters, police, and the legal system: samples that were more frequently associated with the feminist perspective. Although this violence is usually perpetrated by men, there are instances where women have been identified in same sex violence (Renzetti, 1992) and in heterosexual relationships (Hines, Brown, & Dunning, 2007; Migliaccio, 2002). Hines et al. (2007) reported that 95% of men stated that they were controlled by their female partners in ways that were identified on the power and control wheel as described by Pence and Paymar (1993).
There was a final type of violence called mutual violent control, where both partners were violent and controlling, which was very rare (Johnson, 2006). In this situation, both partners were seen as perpetrators. Graham-Kevan and Archer (2003) concluded that situational couple violence and coercive controlling violence represent two very separate types of violence in relationships where there are differences on physical aggression, controlling behaviors, injuries, and fear of injury and that the sampling procedures would result in very different populations being accessed to obtain these results. Where there is coercive control involved in IPV, there is more likely to be injury (Felson & Outlaw, 2007). Thus, IPV appears to divide into different types of violence that encompasses both the feminist and the family research perspectives, with coercive control and the seriousness of the violence being the variables that place the two views in perspective.
A further type of violence has been identified by Kelly and Johnson (2008), called separation-instigated violence. This violence was associated with those cases where there had not been previous violence, control, or intimidation but was instigated when separation or divorce was imminent. This violence appeared to be triggered by the traumatic thoughts of separation and thus could lead to be either situational couple violence or coercive control, depending on the tactics used by partners and the responses to those tactics.
Finally, violent resistance was the type of violence identified where coercive control was resisted. It has been argued by those advocating the feminist perspective that this is the only type of violence in which women are engaged (Dobash & Dobash, 1988, Walker, 1989). However, others note research that suggests that the majority of women do not report self-defense as a motive for their violence (Felson & Messner, 2000; Follingstad, Wright, Lloyd, & Sebastian, 1991). Consequently, there is lack of conclusive evidence for the universality of this type of violence.
Although there are a range of categories of violence, they are not necessary distinct types, as the last two, separation-instigated violence and violent resistance, reflect motives for violence and, depending on the motives, could be either labeled as situational couple violence or coercive control. When discussing categories, it is important to remember that categories have inherent problems associated with labeling, which results in difficult questions. For example, when does a person begin to demonstrate coercive control? Is this after the first instance, or is a pattern of this type of control required to achieve the status of the label? Furthermore, if coercive control is used once by a partner and a fear response is evoked that results in compliance so that coercive control never again needs to be used, what type of violence is this now called? However, the ideas behind this categorization system are helpful in understanding some of the confusion around this topic, and it is the idea of coercion that adds insight here. Therefore, mindful of the pitfalls, and with these categories as a guide, the topic of victims should now be revisited. Ultimately, acknowledging that men can be victims does not detract from the fact that there are many more women who are victims and the object of very serious forms of violence. Kimmel (2002) noted that compassion is not in limited supply and may be shown to both women and men who are victims. It is the controversy related to who is the victim that is fundamental to the key perspectives. The feminist perspective identified coercive control by focusing on patriarchy, which they noted in the cases with which they were concerned, leading to important treatment initiatives that clearly saved lives of women and children. On the other hand, the family research perspective added important other factors associated with abuse that could lead to more effective ways of managing IPV that may prevent the destructive dynamic of coercive control in relationships. By recognizing that violence must be contextualized beyond patriarchy, researchers have identified variables that could lead to the creation of more effective treatment programs for all perpetrators.
Perpetrator Treatment Programs
Treatment of perpetrators has concentrated mainly on the goal of changing men’s cognitions, emotions, and behaviors that are associated with interpersonal conflict through two types of groups: (a) those that focus on stimulating new attitudes toward violence and the power that they as men hold over women, inspired by the feminist perspective and often referred to as the Duluth model; and (b) those that focus on the cognitive-behavioral skills related to anger management, including managing internal dialogue, monitoring signs that often precede violence, and practicing skills to control violence inspired by cognitive-behavioral therapy (CBT; Shamai & Buchbinder, 2010).
The Duluth model, developed by Pence and Paymar (1993), emerged from extensive work with female victims, many of whom went through refuges and shelters to avoid the horrific violence perpetrated on them by their male partners. Thus, issues of control were of fundamental importance. The program highlights the power and control wheel that elaborates the many ways that men use to control women, including coercion and threats, intimidation, emotional abuse, isolation, financial control, male privilege, threatening children, and minimizing and denying any responsibility for the violence. The program emphasizes the creation of egalitarian relationships by embracing the following themes: nonviolence, nonthreatening behavior, respect, trust and support, honesty and accountability, sexual respect, partnership and negotiation, and fairness. These messages are reinforced by discussion, role-play, and worksheets that the men complete, which are centered on their own experiences.
The ecological model, commonly called the CBT approach, described by Edleson and Tolman (1992), focuses on changing men’s attitudes toward violence, increasing their self-awareness, and learning nonabusive alternative behaviors. The group examines male socialization and how this has led to unacceptable ideas about abuse and encourages men to develop empathy for women and a broader context for understanding abusive behavior. Men are taught to identify and monitor emotional, cognitive, and situational cues that often lead to violence so that they can develop alternate patterns of behavior. This paves the way for men to learn strategies to develop their own safety plans for potential abusive situations and to manage their arousal in nonviolent ways that include cognitive restructuring around their new understanding of violence and its unacceptability. Thus, the model includes CBT and is informed by the feminist perspective.
These men’s groups have been running for many years, which have resulted in a substantial number of effectiveness studies. In one study, 40% of perpetrators were nonviolent following treatment, whereas in a control group, 35% were nonviolent without treatment (Babcock, Costa, Green, Eckhardt, 2004). Although dropout rates are high, of those who complete the treatment, approximately one-third do not subsequently engage in violence (Scott & Wolfe, 2003). In a major meta-analytic study of the effectiveness of 22 perpetrator group programs, Babcock, Green, and Robie (2004) reported small effects sizes with the Duluth model (d = .35) and with CBT programs (d = .29). In terms of therapeutic outcome, these results are disappointing, in that most psychotherapy effectiveness studies report much larger effect sizes (d = .85), regardless of the therapy being evaluated (Smith, Glass, & Miller, 1980; Wampold, 2000). When perpetrators are in treatment, it has been suggested that their partners are less likely to report reabuse to police (Cattaneo & Goodman, 2005), implying that the treatment may be even less effective than is currently being reported. The low effects sizes could be because of the ineffectiveness of the actual treatment techniques or could simply reflect the difficulties in dealing effectively with perpetrators who have very entrenched abusive behavior patterns that are resistant to change.
Another way of reflecting on the success rates of these programs for perpetrators is a consideration of the dropout rates, which vary consistently between 30% and 60%, even in programs that are mandated by the court (Daly & Pelowski, 2000). This suggests that either the programs are frequently not engaging participants or they are not seen as relevant or effective enough to continue participation. There has been some interesting evidence to suggest that participants demonstrating different types of violence have different results. In a recent study, perpetrators who were categorized as demonstrating situational couple violence were most likely to complete their program (77%), whereas those who were categorized as coercive controlling, were much less likely to complete their programs (38% and 9%) (Eckhardt, Holtzworth-Munroe, Norlander, Sibley, & Cahill, 2008). This outcome suggests that different types of violence result in different responses to the treatment process and may imply that coercive controlling violence is most resistant to change. For results to improve, the actual programs may need to be tailored more specifically for the individual participants. Thus, it appears that IPV is considerably more complex and may include a range of factors that the two current treatment models are unable to adequately manage.
With low overall effectiveness rates for treatment outcome, it is perhaps time to begin considering aspects of the therapeutic process that may make a difference in treatment (Scott, 2004). An issue that these programs fail to address is the common factors of therapy, which include the importance of the therapeutic relationship (Wampold et al., 1997). There have been a number of initiatives to explore some of these other factors. Engagement factors, including the therapeutic alliance, and group cohesion predicted less subsequent violence (Taft, Murphy, Kings, Musser, & DeDeyn, 2003). Support for retention in treatment by focusing on attendance and the creation of supportive and caring environments led to increased attendance, tended to decrease the dropout rates, and resulted in lower subsequent violence (Taft, Murphy, Elliott, & Morrel, 2001). These factors suggest relatively simple initiatives that should be easy to implement to enhance treatment effectiveness.
However, there are more broad-based changes that could also be implemented. The concept of readiness for treatment seems like a fundamental idea that is appropriate to any type of treatment. Results from intervention studies suggest that male perpetrators have different levels of readiness to change, which is influenced by factors such as cognitive, affective, and behavioral differences (Eckhardt et al., 2008). The transtheoretical model of change (TTM), used for the promotion of health-related behaviors has been adapted to deal with abusive behavior (Murphy & Baxter, 1997; Prochaska et al., 1994; Scott & Wolfe, 2003). TTM identifies four stages of change, based on attitudes and behaviors. Stage 1, precontemplation, is the stage where there is a denial of any need for change. Stage 2, contemplation, focuses on those who have thought that change would be good but have not taken any efforts in that direction. Stage 3, action, describes those who have begun to attempt to change their behavior or aspects of the environment in which the problem occurs. Finally, Stage 4, maintenance, identifies those who have made changes and are trying to retain their achievements and also prevent relapse. Hellman, Johnson, and Dobson (2010) examined a group of men in perpetrator treatment programs and reported that the action stage was predicted only by contemplation. They concluded that energy should be directed at moving perpetrators from precontemplation to the contemplation stage as a necessary initial step in promoting change in abusive behavior. This concept is not part of the curriculum in the two current treatment models. Thus, there are a range of factors that have been explored that clearly show some promise in increasing the effectiveness of treatment of perpetrators. However, as well as treatment design changes, there are other program content changes that could also be made, based on other research.
Theoretical Treatment Perspectives: Moving Beyond Gender
From the research evidence, it is clear that where there is IPV in the general population, that partners appear to report similar amounts of less serious violence in their relationships and that much of the violence is probably situational couple violence. Langhinrichsen-Rohling (2010) suggested that once gender is eliminated as an explanation for significant amounts of IVP, there are many new levels of explanation, including environments, culture, and individual characteristics reflecting the backgrounds of partners.
In a qualitative study of men who were abused by their female partners, it was reported that their female partners were easily able to get restraining orders against their partners, rendering their own abusiveness as invisible (Hines et al., 2007). Men in this study reported that they experienced life-threatening violence, including choking and stabbing. Thus, removing gender from the debate may be a very important step in not only planning better treatment programs for perpetrators who are frequently male, but also providing effective treatment options for males who are victims. It is also possible to consider a range of factors that can become the focus of treatments that may create more effective models than have been evident so far.
There are a number of models that have advocated classifying perpetrators, including pit bulls and cobras (Jacobson & Gottman, 1998), exploders and tyrants (James, Seddon, & Brown, 2002), family only, dysphonic/ borderline, and generally violent/antisocial (Holtzworth-Munroe, & Stuart, 1994), as well as the categories discussed above, based on the presence or absence of coercive control (Johnson, 2006). These classifications appear to be attempts to refine the gendered view that because of patriarchy, only men are perpetrators. Thus, psychological categories, rather than gender, are used to explain the most horrific cases of abusive behavior by men that were previously explained simply by gender or patriarchy. However, all classification systems suffer from difficulties associated with labeling, which creates discrete entities that are inaccurate descriptions of real-life situations. Widiger and Trull (2007), in discussing changes being suggested for the Diagnostic and Statistical Manual of Mental Disorders classification system, advocated for the shift to a dimensional model. This has also been recommended in the area of IPV, as the classification of perpetrators has not lead to significant breakthroughs in their treatment (Ross & Babcock, 2010). Focusing on relevant dimensions, new directions in treatment have to be provided that would enable movement beyond the current ineffective ways of dealing with perpetrators, be they male or female. A number of dimensions have already been identified and researched by family violence researchers. However, much of the research has only been conducted on males at present, leaving many questions with regard to the factors that may be related to female perpetration.
Self-esteem has been related to IPV. Although the results are inconsistent, generally lower self-esteem is associated with violent behavior (Murphy, Meyer, & O’Leary, 1994; Nunn & Thomas, 1999). Baumeister, Smart, and Boden (1996) found the opposite result, with high self-esteem related to violent behavior in their review of the literature. These results may be related to other intervening factors, such as narcissism (particularly exploitativeness, exhibitionism, and entitlement), which has been associated with psychological aggression (Blanchard, 2001; Bushman & Baumeister, 2002; Russell, 2004). Where there is low self-concept clarity, some studies found higher levels of aggression (Stucke & Sporer, 2002), providing evidence of narcissism as an intervening variable between self-esteem and abusive behavior.
Anger, and the ability to control it, has been related to violent behavior (Norlander & Eckhardt, 2005). Perpetrators have higher levels of anger than do nonperpetrators (Maiuro, Cahn, Vitaliano, Wagner, & Zegree, 1988; Margolin, John, & Gliberman, 1989). However, there are differences in the types of anger that have ramifications for IPV. Three types of anger have been identified from a group of male perpetrators; pathological anger, low anger control, and normal anger (Murphy, Taft, & Eckhardt, 2007). Those men categorized as having pathological anger had higher levels of abuse before and after completing perpetrators groups, as well as lower rates of attendance at the groups.
There is a link between abuse and depression (Brown, James, & Seddon, 2002; Coker et al., 2002). This may be surprising, as depression is often associated with withdrawal rather than an active engagement. However, it is another indication that mental health issues may be involved in IPV and require treatment that addresses the underlying issues, rather than focusing only on the symptoms.
Attachment patterns have been linked to violent or abusive behavior in relationships. Hazan and Shaver (1987) reported a continuation of past attachment patterns in their romantic relationships, where the anxiously attached tended to fear abandonment and the avoidantly attached seemed to fear closeness in their relationships.This included the continuation of the anger and rage of childhood into adult relationships where there was the threat of rejection and abandonment (Bartholomew, 1990; Dutton, 1999; Dutton, Saunders, Starzomski, & Bartholomew, 1994). To some who are being rejected, violence or abuse seems a justifiable response (Downey, Feldman, & Ayduk, 2000), particularly those who are anxiously attached and are overwhelmed by thoughts and feelings of rejection (Brown et al., 2002; Henderson, Bartholomew, Trinke, & Kwong, 2005; Mikulincer & Nachshon, 1991). However, for the securely attached, their reactions to rejection are much less intense (Mikulincer & Nachshon, 1991). The link between attachment, abandonment, and intense anger is thus well established in intimate relationships.
Shame, often called the master affect, is a variable that has been linked with abusive behavior. Lewis (1987) hypothesized that shame produced humiliation and that anger was a way of defending against the feelings of poor self-esteem or the sense of feeling diminished. This resulted in attacking the other who was seen as the source of the shame, demonstrating very powerful negative feelings (Leary, Koch, & Hechenbleikner, 2001; Leary, Tambor, Terdal, & Downs, 1995), including “threats to self” (Bourgeois & Leary, 2001). Shame has been linked with abusive behavior (Brown, 2004; Brown, James, & Taylor, 2010; Harmon, 2002). The development and management of affect is central to the development of a strong sense of self (Stolorow, Brandchaft, & Atwood, 1987). Excessive amounts of shame or the inability to manage this affect reflects a poor sense of self, suggesting a threat to self, which is linked to abuse.
Alexithymia is the inability to identify and thus discriminate between personal feelings (Nemiah, Freyberger, & Sifeneos, 1976). It is common for all emotional states to be related to a general state of arousal that may or may not be recognized by the individual. This general state of arousal is often associated with anger in men and may lead to abusive behavior. Brown et al. (2010) noted the link between psychologically abusive behavior and alexithymia. Alexithymia was also used as a variable to discriminate between men who reported abusive behavior and those who did not (Yelsma, 1996). The lack of engagement with feelings has ramifications with a lack of self–awareness or sense of self generally, as well as a lack of empathy with a partner (Scott & Wolfe, 2000). A more emotionally based type of psychodynamic psychotherapy fits well with the exploration of emotion and the subsequent development of a stronger sense of self.
Thus, there is a range of dimensions that have been related to IPV, particularly for men. Clearly, much more research is needed to relate these dimensional variables to female as well as to male victims. With the current state of knowledge, it is only possible to speculate about changing treatment models that relate to male perpetrators. The dimensions are helpful in providing new directions as suggested by others (e.g., Ross & Babcock, 2010). However, as the number of dimensions increase, it will be increasingly difficult to develop treatment models and programs that focus on all of the dimensions. Other models will be needed to sort out the relationships between variables, and this research is just beginning to appear. Brown et al. (2010) conducted a study using men from perpetrators treatment groups focusing on the relationship between a number of dimensions associated with IPV. They reported a significant relationships between psychological abuse and anxious attachment style, shame, and alexithymia. In their research, they identified and reported evidence for what they called a rejection–abuse cycle, which is diagrammatically represented in Figure 1. In their view, the cycle was triggered by experiences of rejection by the female partner of the male perpetrator, which was hypothesized as linking to previous childhood rejection (as suggested by attachment theory), which was linked to an experience of a depletion of self, commonly associated with internalized shame. The shame was theorized as triggering a reaction with the objective of managing these feelings of shame because of the overwhelming level of distress created. Thus, it was theorized that perpetrators attempted to find ways to manage the intolerable shame. Alexithymia was one strategy that respondents in the research reported. Alexithymia is a complex concept, and there is much speculation regarding the level of consciousness of this process of not recognizing feelings. In support of the idea that alexithymia may serve a protective function from strong overwhelming feelings, research has linked alexithymia with dissociation (Evren et al., 2008). It seems that these attempts to deal with the shame at times may be insufficient, which may lead to an escalation of the abuse. Furthermore, the numbing of distinct feelings that is associated with alexithymia may even encourage abusive responses, as the lack of signals for discrete emotions in the perpetrator does not provide warning of imminent strong emotional reactions. Of course, abuse, rather than having the desired effect of decreasing the shame from a depleted sense of self, may simply lead to further experiences of rejection, completing the cycle. Thus, the cycle documents a circular process. Once the rejection–abuse cycle is established, it is very resistant to change as we know from the difficulties in dealing with IPV. However, the cycle offers new ways of thinking about perpetrator treatment. There is a range of implications for treatment that emerges from research, including the rejection–abuse cycle, which can be applied to perpetrator programs for men as well as for women.
Focusing on coercive control is very important in deciding how to manage a treatment regime (Johnson, 2006). Research cited above indicates that coercion in IPV is not managed well in group programs, as these perpetrators often leave the programs early, suggesting that they have not been adequately engaged in treatment (Campbell, Neil, Jaffe, & Kelly, 2010; Taft et al., 2001). Perhaps it is time to focus on individual treatment where perpetrators are clearly identified by their use of coercive control. These participants could be viewed through a lens of readiness for treatment and engaged appropriately in ways that would not be possible in the group context. Saunders (1996) advocated taking personality into consideration to form treatment goals, but perhaps the dimension of coercive control may provide a more effective assessment when making such a decision.
As currently takes place in many treatment programs, perpetrators need to gain awareness of the violence that they exhibit in all its forms and also need some strategies for controlling it (Shamai & Buchbinder, 2010). However, given the role of the emotion-laden content suggested by the rejection–abuse cycle (Brown et al., 2010), these current strategies need to be augmented by forms of treatment that move to deeper levels than psychoeducation may provide in order to effect lasting change.
If rejection is so fundamental to perpetrators and a part of the problem, then attempts to deal with rejection are important and clearly related to the sensitivity with which group members are treated when they participate. Thus, it is not surprising that attendance improves in programs where there are supportive and caring environments (Taft et al., 2001), and results are improved where there are empathic professionals (Campbell et al., 2010).
The rejection–abuse cycle documents a process that may be understood by perpetrators. Thus, explaining the cycle, finding examples in abusers’ experiences, and talking about the painful experiences associated with each stage of this cycle may help liberate the abuser from these resistant abusive patterns. Insight can be helpful in motivating perpetrators to engage in the often painful process of change.
Programs should focus on insecure attachment patterns and how these dynamics play out in day-to-day interactions. The work on rejection sensitivity (Downey et al., 2000) is helpful in identifying ways of dealing with this issue. For example, rather than defining ambiguous situations as further instances of rejection, encouraging abusers to consider alternative explanations that do not involve rejection could help them deal with the tendency to exaggerate the meaning of experiences of potential rejection.
Internalized shame is a very powerful affect that has ramifications related to violence (Brown et al., 2010). Dealing with shame as a mediating variable may help decrease abuse. Babcock et al. (2007) suggested that the focus on emotion should be central to treatment, including the link between anger and shame. Some group programs have already begun to incorporate a focus on shame (e.g., Wallace & Nosko, 2003). Finding ways for perpetrators to tolerate shame by simply speaking about these feelings and the events that provoke shame is a way of teaching tolerance for these strong emotions, subsequently weakening the link between shame and abuse.
Finding ways of managing all feelings, rather than simply using dissociation, alexithymia, and simple avoidance, may be useful, as suggested by the rejection–abuse cycle (Brown et al., 2010). Much of the treatment in perpetrator groups has focused on changing cognitions (Shamai & Buchbinder, 2010). However, there has been a call for emotionally focused rather than cognitive models (Babcock et al., 2004). Gottman (1999) highlighted the ineffectiveness of talking about conflict once the pulse rate goes beyond 100 beats per second and uses a pulse rate–monitoring device to provide feedback to couples for the times when further discussion is pointless or may be damaging physically or psychologically. Couples are then advised to decrease the heart rate before continuing the conversation. Thus, managing emotion can be related to physiological arousal, and people may be taught to manage this arousal better when speaking about emotional topics. Goldner (1998) also worked with perpetrators, encouraging them to learn to talk about their emotional experiences without losing control as an initial evaluation in assessing readiness for couple treatment. Thus, focusing on emotion as well as cognition may be an important initiative in perpetrator programs.
Generally, thinking about perpetrators through the lens of clients needing treatment rather than reeducation fits well with the perspective of the rejection–abuse cycle. Effective treatment requires a focus on the client–therapist relationship (Dutton & Corvo, 2006; Wampold et al., 1997). Research suggested that perpetrators tend to feel comfortable talking to someone who is trustworthy, and they are usually willing to deal with their problems with someone who is nonjudgmental when it comes to matters of IPV (Campbell et al., 2010). This issue has frequently been minimized or ignored in treatment models currently being used and should be intentionally addressed.
Given that much of the violence experienced in relationships is situational couple violence, the use of couple therapy needs to be encouraged. Contrary to previous thinking, perpetrators’ behavior should be considered within the context in which it takes place. Systemic perspectives highlight the importance of conjoint treatment, particularly in cases where coercive control is not present (Goldner, 1998; Murray, 2006). In the past, couple therapy was frequently proscribed where there was abuse, when it was thought that men were the only perpetrators. Accepting the possibility of female perpetration and male victims highlights the importance of couple sessions as a necessary part of the diagnostic process. However, where there is clearly coercive control by one partner, the advisability of couple therapy should still be questioned.
Finally, the need for readiness before attempting change has been highlighted by the TTM of change that was discussed above (Eckhardt et al., 2008). Although all these variables may stimulate change in perpetrators, it still needs to be addressed only when the client is ready for change to occur. Of course, readiness may be stimulated by the acknowledgement of the rejection–abuse cycle and the establishment of an effective client–therapist relationship.
The rejection–abuse cycle (Diagram modified from Brown et al., 2010)
In many ways, these 10 implications provide a new level of complexity that may appear to be confusing, making treatment initiatives difficult. A meta-theory to facilitate a better integration as advocated by Brown (2010) may provide some clarity. He suggested that intervention could be understood along a continuum of content (topics discussed in therapy) and process (the way in which the therapy was managed). Applying this meta-theory to the 10 implications demonstrates one way as to how it might be used in focusing on an integrated treatment approach. By focusing on process, 3 of the implications for treatment become salient: applying a treatment lens (Implication 8), considering when to use couple therapy (Implication 9), and working with readiness for change (Implication 10). The other 7 implications are related to the content end of the continuum and the five elements of: self, affect, beliefs, meaning, and behavior. Self is central to the core personality and is formed by the often unconscious experiences in the past and influences how the individual construes the world. This element is addressed in the rejection–abuse cycle (Implication 4) and in the consideration of attachment styles (Implication 5). Affect is related to how emotion is experienced and managed. This element is addressed in the rejection–abuse cycle (Implication 3), management of emotion through attachment styles (Implication 4), dealing with shame (Implication 6), and managing feelings generally (Implication 7). Beliefs are communicated both consciously and unconsciously through current and past experience. This element is addressed in how individuals manage rejection (Implication 3). Meaning is often shaped by beliefs that are held, as well as by feelings in the present moment and impact on what actions are taken. This element is addressed in how rejection is managed (Implication 3) as well as in the rejection–abuse cycle (Implication 4). Finally, behavior is influenced by interactions with others. This element is addressed by a consideration of coercive control (Implication 1), violence in all its forms (Implication 2), and the rejection–abuse cycle (Implication 3). Thus, a new model of perpetrator intervention could comprehensively target self, affect, beliefs, meaning, and behavior, providing a much broader focus than has been used in past perpetrator programs.
A number of targets within a meta-theory have been identified for working with perpetrators that move beyond the current commonly accepted approaches, many of which have been related to the rejection–abuse cycle. This work has been hampered by the dominance of the currently accepted models of treatment of CBT and the Duluth model. As funding for alternative models is not readily available, these new ideas have not yet been evaluated in any systematic way. As the current models of treatment have shown questionable effectiveness, it is time to develop an approach based more on the results of extensive research that has emerged from the family research perspective represented above. The principles underlying these strategies are at odds with many, but not all the ideas that inform current treatment programs. Thus, the radical shift in thinking may require major change strategies, as many of the current programs are strongly advocated. However, this change is important if we are to provide all perpetrators with the best forms of treatment for IPV.
Footnotes
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.