One view of Intimate Partner Violence (IPV) that has been influential in terms of public policy is the “gender perspective,” which is associated with feminist analyses (e.g., Dobash and Dobash 1979). This view holds that IPV is asymmetrical, with men as the primary perpetrators, who use violence in a bid to control and dominate their female partner. Men’s violence arises from patriarchal values and should be studied in isolation away from general models of aggression. In contrast, there are other researchers who support studying IPV within the context of both family violence and other forms of aggression outside the home. This “gender inclusive” approach (Hamel 2007) has been supported by a wealth of studies that have found that IPV perpetration is more symmetrical, with men and women reporting physical aggression perpetration towards their partner at similar rates, or in the female direction (e.g. Bates et al. 2014).

A key aspect of the gender symmetry and gender asymmetry debate revolves around the extent to which IPV perpetration can be considered to be unilateral or bidirectional. Early theorists of IPV (e.g. Dobash and Dobash 1979), focused on examining unilateral violence of men against their female partners; this was often labelled as wife abuse or termed violence against women, to highlight the specific focus. Early IPV research as a consequence focused on men’s violence and neglected the victimisation of men and boys, as well as women’s violence towards their male partners (Langhinrichsen-Rohling et al. 2012). Since large scales studies revealed the extent of the symmetry between men and women’s perpetration, it has been important to consider the dynamics that exist within violent relationships. Studies around bidirectional or mutual IPV can further aid our understanding of the context of men and women’s aggression, which gives insight into motives and risk factors.

Prevalence of Bidirectional IPV

Studies have suggested that bidirectional violence is the most common type experienced in relationships. For example, Stets and Straus (1989) found that in couples where violence occurred, both partners were violent in around half the cases, then female-only and male-only in about a quarter of the time each. Females were more frequently the perpetrator in unilateral aggression in this and other studies (e.g., Gray and Foshee 1997). This was also found cross-culturally; in a sample of over 13,000 students across 32 nations, Straus (2008) found that the most frequent pattern of abuse is bidirectional, followed by female only. Male only was least frequently reported and this was from both men and women’s reports. This study also showed that there is an overlap of risk factors for men and women, with dominance by either partner being found to increase the probability of violence; this is in contrast to feminist theories that assert male dominance is the cause of IPV.

Traditional, gendered approaches see IPV perpetrators and victims as being relatively homogenous groups. This has not been found to be the case, and indeed there is also heterogeneity found within bidirectional abuse. Consequently, a number of typologies have been proposed. For example, Langhinrichsen-Rohling (2010) presented three subtypes of bidirectional violence between couples. The first involved the motive of control and coercion with both partners displaying these behaviors. The second involved violence because of issues regulating their emotions and controlling their behavior, referred to as dyadic-dysregulation or mutually dysphoric; here the conflict and aggression is as a function of the level of interdependency that exists between partners. The third subtype is believed to be the least severe IPV perpetration with violence restricted to partners and with little evidence of personality disorders or psychopathology; this was also discussed in line with retaliatory violence.

The importance of exploring the dyadic nature of behavior in a relationship was also conceptualised by Michael Johnson. Johnson (1995) created his original typology to address the conflicting findings presented within the feminist and family violence approaches to IPV. He labelled the first situational couple violence (formerly common couple violence), which encompasses low-level violence with little use of control. Intimate terrorism (formerly “patriarchal terrorism”) involves the use of severe and coercive violence as part of a range of behavior that men use to dominate and control their female partners. His later work expanded the typology from an individual to a dyadic one to encompass all combinations of controlling aggression, non-controlling aggression and no aggression (Johnson 2006). He added “violent resistance” to represent violence of a non-controlling kind in response to controlling aggression from the partner; this often encompasses women’s violence in self-defence. The other, labelled “mutual violent control”, represents a destructive relationship where both partners use controlling aggression. Heterogeneity in bidirectional violence is supported by his typology that indicated two bidirectional categories, with different levels of violence and control. Whilst there has been support found for the categories within his typology, evidence suggests his assertions around gender are not substantiated (e.g. Bates et al. 2014). For example, in their sample of 14,252 student couples, Straus and Gozjolko (2016) discovered that more female only intimate terrorists (33 %) were found than men only (16 %) but that the majority saw this behavior in both members of the couples (51 %). Despite the evidence presented here about the frequency, severity and risk factors for bidirectional violence, it is largely ignored when it comes to treatment and interventions.

Implications of Overlooking Bidirectional Violence

By focusing on unilateral violence, in particular men’s violence towards women, there are significant implications for research, risk assessment and interventions. There are implications of erroneously focusing on unilateral violence in terms of understanding the consequences of bidirectional IPV. The impact of bidirectional violence is considered to be more serious, and most likely to result in injury and mental health problems (e.g. Whitaker et al. 2007). Rhodes et al. (2009) found men disclosing both perpetration and victimisation had a greater prevalence of adverse health conditions, including PTSD symptoms, depression and suicidality. However, even in acknowledging bidirectional abuse, some scholars still choose to only focus on the impact for men or women (e.g. Hellmuth et al. 2014), rather than both within the same sample. To be able to provide effective intervention services to both reduce the violence, and manage the consequences, there needs to be research and practice that focuses on the holistic view of the relationship. This includes exploring the prevalence, severity and impact on men and women within the same studies. Only by studying IPV in this context and asking about both perpetrator and victim behavior can the nature of bidirectional abuse emerge.

With research suggesting that bidirectional violence is the most common pattern of aggression found (e.g. Langhinrichsen-Rohling et al. 2012; Straus 2008), it challenges the more traditional and gendered approaches to studying IPV. Whilst patriarchy could be an explanation of some men’s violence towards women (though it seems a small proportion), it is unlikely to be the main etiological factor influencing women’s IPV, and especially so when IPV is mutual within the relationship. When both members of the couple are being aggressive it suggests causes could be in dyadic areas, for example around conflict management (Langhinrichsen-Rohling et al. 2012). It further highlights the importance of considering women’s aggression in uni- or bidirectional relationships.

Risk Assessment and Interventions

Existing risk assessment and IPV intervention programs treat perpetrators and victims as distinctly separate; largely they do not consider the context in which the violence exists or the dynamic of the couple, instead often choosing to focus on men’s violence towards their female partners. Considering women’s own violence is essential to understanding IPV, as women’s perpetration has been found to be the strongest predictor of their victimisation (Stith et al. 2004). Indeed, retaliation may be a factor that increases the violence, and therefore the likelihood of being injured (Whitaker et al. 2007). Proponents of the gender paradigm often choose not to examine women’s perpetration and behaviour, instead focusing solely on their victimisation. However, those that have explored the behavior of women in shelters find them to be heterogeneous as a group; for example Langhinrichsen-Rohling (2006) found a quarter of her shelter sample had engaged in stalking their ex-partner and that this group had higher levels of depression and self-blame, and also were more likely to be the victim of stalking behavior.

Traditional, gendered approaches to interventions have their roots in the Duluth model (Pence and Paymar 1993), which treats IPV as unilateral and focuses on men’s patriarchal use of violence towards their female partners. Critics of this approach have long argued that this approach neglects women’s violence, violence within same-sex relationships and bidirectional abuse (e.g. Bates et al. in press). Dutton and Corvo (2007) state that the Duluth influenced programs still purport the “gender-political assumptions that male violence is always unilateral and any mention of female violence is ‘victim blaming’” (p.661). Consequently these gender politics inhibit asking about female violence and in some settings it is prohibited. Indeed, for the Respect accreditationFootnote 1 procedures within the UK, risk management makes assumptions that the violence is unidirectional and that a man is to be held fully accountable for his violence. It does not include reference to asking about women’s behavior and forbids “denial and minimisation of abusive behaviour or any justifications for using abusive behaviour including the use of drugs or alcohol” (Respect 2012; p 29). Within this model, men’s own experience of victimisation is not seen as a risk or causal factor; yet in contrast, women’s perpetration is seen as wholly a factor of their victimisation. Dutton and Corvo (2007) describe the “two totally different gestalts for male and female violence” (p.660); any and all risk factors of men’s IPV are dismissed as excuses (Dutton and Corvo 2006).

Recommendations for the Future

Considering the evidence presented around bidirectional abuse, there is a need to change the way IPV is viewed in terms of risk assessment, risk management and interventions. Langhinrichsen-Rohling (2010), in discussing the heterogeneity of bidirectional IPV, posits a model that encompasses the individual and contextual factors of both partners (e.g. attachment issues, experience of conflict) and notes that this is important in understanding the violent dynamic. She further adds that gender-specific interventions will be unlikely to be successful with men and women in these relationships due to their unilateral focus, which may ignore some of the underlying issues. It is critical to note that interventions will only be successful if they recognise and encompass the fact that a significant number of relationships involve violence by both partners. Langhinrichsen-Rohling et al. (2012) make recommendations about risk management and intervention of IPV. They specifically call on practitioners and clinicians to recognise the heterogeneity of perpetrators and the need to identify subtypes with a “sensitivity and specificity” (p.222); this is in line with other research that states the importance of understanding and considering the relationship violence within the context of the relationship (e.g. Whitaker et al. 2007). This is of paramount importance considering many abusive partnerships remain intact after service interventions (e.g. Koepsell et al. 2006). Other researchers also argue that risk assessment should encompass both perpetrator and victim characteristics (e.g., Kropp 2009) in order to more fully understand an individual’s or couple’s risk and intervention need factors. Furthermore, to predict recidivism and effectively assess risk, there is a need to consider whether the context of the home is violent; as Dutton and Corvo (2007) question “…would it not matter if a group-client was returning to a relationship with a violent woman?” (p.662).

This is important contextual information to understanding the circumstances in which violence is instigated; this should be integrated into intervention strategies. Tailoring the intervention to the specific context of the violence is critical. Whilst interventions created for unidirectional violence will not be suitable for those in a mutually violent relationship, the opposite is also true; treating both members of a couple when the violence is only from one could be harmful (Langhinrichsen-Rohling et al. 2012). Straus (2008) suggests that prevention and treatment of IPV would be more effective if the programs reflected the true nature of IPV. This includes the heterogeneity of both unilateral and bidirectional violence. This consideration of IPV is in alignment with seeing it as part of an interactional model of family violence (Winstok 2007).

Within much of the literature on IPV, especially around treatment and interventions there is a consistent use of the term “perpetrator” and “victim”. Whilst this is clearly the appropriate terminology with unilateral violence, it complicates discussions of bidirectional abuse when both members of the couple often fit into both categories. The language used is reflective of how IPV is treated in practice with there being a focus on separate perpetrator and victim services. Cautions should further be exercised to those working with victims, it is important to recognise that some men and women seeking help may also be perpetrators and it should inform the methods of support put in place. This also involves acknowledging that bidirectional aggression is often perceived as less severe but this is not the case; relationships are in fact often the most aggressive and result in more injuries.

The recommendations described are in line with a plethora of the research that exists and are in accordance with the demand for more evidence-based practice in the area. The impact of changing how we intervene with IPV could improve the success of programs and reduce the risk for men and women in abusive relationships. Dutton and Corvo (2006) questioned assessments in IPV interventions, specifically around the interactive nature of couple’s violence, the power dynamic, lethality potential and treatment/client profile. Their paper is a decade old and yet we still do not consider the dyadic nature of IPV within assessment and intervention. Straus (2010) details the ways in which some members of the academic community have denied the wealth of research that has demonstrated gender symmetry in IPV. These include across the years: misrepresenting data, selective citation and in some cases blocking publication and preventing funding. The politics around this area may stop progression of evidence-based practice in the development of interventions. The recommendations here may not be well received by the proponents of the gendered approach and the Duluth model.