Current Perceptions

Findings from hundreds of reviewed studies (Fiebert 2010) and meta-analyses (Archer 2000) indicate that men and women self-report perpetrating physical violence at similar rates, and that approximately 50 % of relationship violence is bidirectional, meaning both partners report using violence (Langhinrichsen-Rohling et al. 2012). While these data have led some researchers to conclude intimate partner violence (IPV) is gender symmetrical (Straus 2010), others have disputed this conclusion (Loseke and Kurz 2005). Thus controversies about gender symmetry related to the causes, individual consequences, and societal implications of partner violence remain unsettled (Coker et al. 2002; Stark 2007; Tjaden and Thoennes 2000). Further, methodological differences with regard to sampling (e.g., use of university and community samples over agency samples; Johnson 2008), measurement (e.g., reliance on CTS which fails to capture the context of violence; Dobash et al. 1992), and classification (e.g., labeling all violent behaviors as “domestic violence” without consideration of power and control) make it virtually impossible to compare across studies.

Johnson (2008) developed a classification system with multiple IPV categories that might ultimately lead to empirical resolutions of the gender symmetry debate. Johnson argued that different IPV types vary in their use of power and control and reflect different patterns of gender symmetry. Johnson (2008) attributed inconsistent gender symmetry findings to the likelihood that community and clinical samples tend to experience and report different types of IPV. Between 1995 and 2008, Johnson refined IPV categories to reflect dyadic patterns within relationships which views one partner’s use of violent and controlling behaviors in conjunction with the other partner’s behaviors on the same constructs. Johnson’s most recent typology includes four unique IPV categories: Intimate Terrorism, Situational Couple Violence, Violent Resistance, and Mutual Violent Control. Intimate Terrorism (formerly called Coercive Controlling Violence or Patriarchal Terrorism) represents relationships where one partner is physically violent and controlling, and the other partner is not violent, which is conceptually similar to feminist definitions of domestic violence/abuse (Dobash and Dobash 1979). Situational Couple Violence (formerly called Common Couple Violence) captures relationships where one or both partners is physically violent, but neither uses controlling behaviors (sometimes referred to as aggression). Violent Resistance includes relationships where one partner is physically violent and controlling (abuse) and the other partner is physically violent but not controlling (aggression), capturing self-defense strategies, among other reactions to abuse. Mutual Violent Control represents relationships where both partners are physically violent and controlling (mutual abuse). Subsequent research supports gender symmetry patterns in some IPV categories (e.g. Situational Couple Violence and Mutual Violence Control) and gender asymmetry in others (e.g. Intimate Terrorism perpetrated at higher rates by men, Violent Resistance used more commonly by women) (Johnson 2006).

While Johnson’s work helps to differentiate IPV relationship patterns, some problematic relationships fall outside of the typology. Specifically, the typology classifies relationships where one or more partners are controlling but not physically violent as IPV-free. Many IPV theorists view control as the defining constant feature of abusive relationships with physical violence use more likely to occur episodically (Smith et al. 1995), thus underscoring the need to understand relationships that are controlling but not physically violent.

Implications

Feminist IPV conceptualizations view physical violence as a tactic used to achieve power and control within abusive relationships (Dobash and Dobash 1979). However because control can be achieved through non-violent strategies (e.g. threatening to hurt or kill a loved one), relationships that are highly controlling but not physically violent can still be considered abusive. Johnson (2008) analyzed the level and direction of control within relationships where past year violence was reported; however his analysis excluded relationships marked by unidirectional or bidirectional control exerted in the absence of violence. Thus Johnson’s typology in its’ current form constrains our ability to assess IPV comprehensively, particularly as we seek to compare gender symmetry rates across IPV categories.

Physically violent acts are often preceded by credible threats of violence--a type of control. Fearful partners will comply with demands in the face of threats quickly in order to avoid physical violence (Dutton 1992). Paradoxically, relationships with the highest levels of fear and control may be among the least physically violent. Alternately, non-violent control strategies may escalate to physical violence when non-violent strategies are no longer effective in achieving power (Pence & Paymar, 1993). Categorically, controlling non-violent relationship patterns include both relationships with a physical violence history or the likelihood of violence in the future. Both these scenarios represent abusive relationship dynamics that are potentially missed by IPV categories, which only consider control behaviors when physical violence is present within the past year.

Assuming, as Johnson (2008) and others (Graham-Kevan and Archer 2003) do, that an individual perpetrates physical violence with a high or low amount of control, four discrete patterns of behavior emerge for an individual when physical violence and control are intersected (see Fig. 1). Individuals can be classified as using: low levels of control/high levels of physical violence (point A); high levels of control/high levels of physical violence (point B); high levels of control/low levels of physical violence (point C), or low levels of control and low levels of physical violence (point D).

Fig. 1
figure 1

Conceptual model for expanded typology

Full size image

The four proposed behavior patterns can be used to operationalize Johnson’s IPV typology, which classifies relationships into categories based on the both partners’ behaviors. As per Johnson’s most recent book (Johnson 2008), Situational Couple Violence includes relationships where at least one partner is physically violent, but not controlling (point A + points A/C/D). Violent Resistance includes relationships where one partner is physically violent and controlling and the other partner is physically violent but not controlling (point A + point B). Mutual Violent Control represents relationships where both partners are physically violent and controlling (point B + point B). Intimate Terrorism represents relationships where one partner is physically violent and controlling, and the other partner is not violent (point B+ point C/D). Relationships where no physical violence was reported by either partner (points C/D + points C/D) were not classified.

Within Johnson’s framework, unidirectionally controlling (point C + point D) and bidirectionally controlling (point C + point C) relationships that do not include physical violence are considered free of IPV. In other words, couples experiencing high levels of control in their relationships would not be classified as experiencing IPV, unless physical violence was reported in the past year. Controlling behaviors, such as threats, stalking, and work interference, are all independently associated with negative health and mental health outcomes (Kulkarni et al. 2016; Basile et al. 2004; Smith, Smith, & Earp, 1999). These controlling behaviors may be experienced as at least as distressing (if not more) than some physically violent behavior (Evans et al. 2016). It is important to capture these experiences within a typology of IPV to fully understand the dynamics of power, control, and violence within relationships.

Resolutions

Given that patterns of control in the absence of past year violence are not accounted for by Johnson’s current IPV typology, a new expanded typology is needed (Langhinrichsen-Rohling 2010; Ross and Babcock 2009). We propose that revised categories can be identified by simultaneously examining each partner’s unique pattern of physical violence and control usage (see Fig. 1). When one person’s behaviors are considered concurrently with their partner’s, ten discrete categories emerge: Relationships with No-IPV, Unidirectional Violence, Unidirectional Control, Bidirectional Violence, Bidirectional Control, Intimate Terrorism, Violent-Control, Violent Resistance, Control Resistance, and Mutual Violent Control (see Fig. 1).

This analytic framework was applied to an existing mixed gender dataset (n = 714) (Dutton et al. 2006) to determine if discrete IPV categories could be identified and if so whether revised categories reflected gender symmetrical or asymmetrical patterns. Data were previously collected from men and women from agencies involved with IPV victims and perpetrators, agencies which provided non-IPV services to demographically similar clients, a community college, and public settings (e.g., fast food restaurants). Participants were asked to report their own violence perpetration and victimization using the Revised Conflict Tactics Scale (Straus et al. 1996). They also reported control perpetration and victimization using the Coercive Control Measure (Dutton et al. 2006). We utilized the same procedure as Johnson (2008) by creating a paired entry in order to capture the experiences of each partner uniquely. This procedure has two important advantages. First, it allows for natural weighting by gender within a heterosexual sample. Secondly, because we are interested in individual IPV perpetration rates this procedure allows us to analyze data at the individual as well as dyadic level.

Using cluster analyses, individuals were categorized as either high or low across four victimization/perpetration categories---physical violence perpetration (individual used physical violence), physical violence victimization (partner used physical violence), control perpetration (individual used control), and control victimization (partner used control)—based on past year physical violence and control behaviors. These data were then used to classify individuals into the 10-category typology described above.

This classification procedure allowed us to identify individuals in all 10 IPV categories (See Fig. 2). More than half (52 %) the sample reported neither partner used high levels of physical violence or controlling behaviors (e.g. the No IPV category). Nearly half (48 %) were in IPV relationships where at least one partner used high levels of physical violence and/or control. Almost two in five individuals in IPV relationships were in either Intimate Terrorism (20 %) or Mutual Violent Control (20 %) categories. Unidirectional Violence (16 %) and Unidirectional Control (17 %) were the next most frequent categories followed by Control Resistance (10 %), Bidirectional Control (7 %), and Violent Resistance (7 %). Violent-Control (2 %), and Bidirectional Violence (2 %) were the smallest categories collectively accounting for about 4 % of all IPV relationships.

Fig. 2
figure 2

Gender ratios by IPV category

Full size image

Gender ratios were used to assess gender symmetry or asymmetry patterns within specific IPV categories (see Fig. 2). Three IPV categories were defined by both partners’ use of physical violence and/or control – Mutual Violence Control, Bidirectional Control, and Bidirectional Violence. Because the sample was comprised of heterosexual couples, these categories, as were Johnson’s, were necessarily gender symmetrical. All other IPV categories were gender asymmetrical. Intimate Terrorism, Violent-Control, and Unidirectional Violence were all perpetrated more frequently by males than females. Violent Resistance, Control Resistance, and Unidirectional Control were perpetrated at higher rates by females than males; however those differences were not as pronounced.

Overall gender symmetrical IPV categories (Bidirectional Control, Bidirectional Violence, Mutual Violence Control) comprised 29 % of all IPV relationships. The remaining gender asymmetrical categories comprise 71 % of IPV relationships. In contrast, when IPV was assessed solely in terms of physical violence perpetration, a more gender symmetrical pattern emerged with 52 % male and 48 % female comprising all physical violence perpetration.

Dyadic analyses accounting for both partners’ use of physical violence and control presented a more nuanced picture of gender patterns than utilizing physical violence perpetration alone. These analyses suggest that females were much more likely than males to use controlling behaviors when in relationships with male partners who used both violence and control (abuse). Males used violence alone (aggression) or violence and control (abuse) against their non-violent female partner more than females did against their non-violent male partner.

Consequences for the Field

The revised typology framework helps us better understand the role of control in IPV as it may operate independently from and in conjunction with physical violence. Perpetration of control behaviors (72 %) was actually more prevalent than physical violence perpetration (43 %) within this sample. Therefore, accounting for control allows us to examine a wider range of unhealthy relationships that may not be physically violent but where partners are mutually controlling or one partner controls another. This framework also allows us to view control as a potential resistance strategy against an intimate terrorist (Control Resistance - 10 %), which is more prevalent than Johnson’s self-defense category (Violent Resistance, 7 % of the sample).

By including both partners’ use of physical violence and control, we better understand the relational context surrounding physical violence perpetration. This analysis supports Johnson’s contention that not all physically violent relationships are the same. Findings also support claims for gender asymmetrical patterns in violence perpetration behaviors. Though both genders are represented in each relationship category, women and men tend to perpetrate physical violence within different relational contexts. Men in this sample engaged in violence perpetration against non-violent partners at higher rates (e.g. Intimate Terrorism, Unidirectional Violence) than women. Women more frequently perpetrated violence and control behaviors in relationships with violent, controlling men. In addition, women were more likely to use control-only strategies against non-violent, non-controlling partners.

Our analysis raises additional questions that can be answered by applying this framework to a broader range of research methodologies. Specifically quantitative, qualitative, and mixed-methods research are all required to make sense of IPV in all its’ complexity. Feminist theory purports that IPV relationships tend to escalate over time from controlling to increasingly violent and controlling (Pence and Paymar 1993). However slightly more women than men perpetrated Unidirectional Control, while men perpetrated Intimate Terrorism more frequently. If Unidirectional Control always reflects an earlier stage of Intimate Terrorism, we would expect to see more similarity in gender ratios across those categories. Longitudinal research is needed to answer questions about stability and/or change in perpetration behaviors over time and whether trajectories are influenced by gender. Similarly, additional research is needed to determine whether categories differ meaningfully in terms of consequences and outcomes. For example, Situational Couple Violence might be considered (as it was by Johnson) a less pathological form of IPV when compared with Intimate Terrorism; however in reality these relationships may be equivalently dangerous in terms of lethality.

IPV perpetration is not a singular phenomenon but rather occurs across different relational categories that reflect unhealthy, conflictual, self-defensive, and abusive contexts. Patterns of gender symmetry emerge when IPV is operationalized as perpetration of physical violence, but the perspective proposed herein demonstrates that when control is accounted for within relationships, asymmetrical patterns arise. Properly operationalized research can determine the risk factors and underlying etiologies associated with specific IPV categories. This research should in turn inform the development and testing of appropriately tailored strategies to reduce IPV perpetration. Reducing IPV in its’ many forms is an ambitious goal that should be grounded in thoughtful and rigorous science. Transcending the gender debate requires us to progress beyond current unidimensional IPV conceptualizations and towards more complex and nuanced understandings. Expanded typologies that capture the ways physical violence and control shape IPV experiences is an important step in that direction.