Male Violence against an Intimate Female Partner
Physical aggression against an intimate partner is a serious problem. While both men and women engage in physical aggression in their intimate relationships, male aggression is generally found to have more negative consequences than female aggression, in terms of physical injury, partner fear, and other psychological sequelae (e.g., posttraumatic stress disorder [PTSD], depression). Thus, the focus in this research paper is on male violence toward an intimate partner, or husband violence. The potential causes of husband violence can be considered from a variety of levels, including those of societal and cultural variables, family and dyadic interactions and relationships, and individual characteristics of the violence perpetrator. To date, many studies, examining a wide variety of individual characteristics (e.g., psychopathology, behaviors and social skills, attitudes), have demonstrated differences between maritally violent and nonviolent men, suggesting that it is important to consider individual characteristics of violent men. Such individual characteristics are emphasized in existing batterer typologies.
Initial research in this area generally consisted of studies comparing men who had engaged in violence against an intimate partner with men who had not (i.e., violent vs. nonviolent sample study designs). However, as researchers gained experience with maritally violent men, they began to agree that differing levels and types of husband aggression exist and that maritally violent men differ from one another in a variety of ways. For example, it is clear that some men, often labeled batterers, engage in severe physical violence and usually also engage in other forms of male control and aggression (e.g., psychological and/or sexual abuse); their violence is likely to result in wife fear and injury. In contrast, some men engage in lower levels of violence, to which a variety of terms have been applied (e.g., minor violence, common couple violence, situational violence).
As research has made it clear that samples of maritally violent men are heterogeneous, varying along theoretically important dimensions, it has been assumed that our understanding of husband violence can be advanced by drawing attention to these differences. Comparing subtypes of violent men with each other, and understanding how each type differs from nonviolent men, should help us to identify different underlying processes resulting in violence. By thus better understanding the correlates and causes of varying types of male aggression, batterer typologies might also lead to improved outcome in batterer intervention and treatment, helping us to identify the men most likely to benefit from various interventions and to develop interventions matched to the needs of differing subtypes of violent husbands.
Given such goals, Holtzworth-Munroe and Stuart (1994) conducted a comprehensive review of then available batterer typologies. Across these studies, we observed that batterer subtypes could be classified along three descriptive dimensions: (1) the severity/ frequency of the husband’s marital violence, (2) the generality of the man’s violence (i.e., marital only or extrafamilial), and (3) the batterer’s psychopathology or personality disorder characteristics. Using these dimensions, we proposed that three subtypes of batterers would be identified. First, family-only (FO) batterers were predicted to engage in the least marital violence, the lowest levels of psychological and sexual abuse, and the least violence outside the home. Men in this group were predicted to evidence little or no psychopathology. Second, dysphoric/borderline (DB) batterers were predicted to engage in moderate to severe wife abuse. Their violence would be primarily confined to the wife, although some extrafamilial violence might be evident. This group would be the most psychologically distressed and the most likely to evidence characteristics of borderline personality disorder (e.g., extreme emotional liability; intense, unstable interpersonal relationships; fear of rejection). Finally, generally violent/antisocial (GVA) batterers were predicted to be the most violent subtype, engaging in high levels of marital violence and the highest levels of extrafamilial violence. They would be the most likely to evidence characteristics of antisocial personality disorder (e.g., criminal behavior, arrests, substance abuse).
Holtzworth-Munroe and Stuart (1994) then integrated several intrapersonal theories of aggression into a developmental model of these differing types of husband violence. This model proposed that correlates of male violence that were both distal/historical (e.g., violence in the family of origin, association with delinquent peers) and proximal (e.g., attachment/dependency, impulsivity, social skills in marital and nonmarital relationships, and attitudes [both hostile attitudes toward women and attitudes supportive of violence]) might serve as varying risk factors for differing batterer subtypes.
Based on this model, we predicted that among maritally violent men, FO batterers would evidence the lowest levels of risk factors. The violence of FO batterers was proposed to result from a combination of stress (personal and/or relationship) and low-level risk factors (e.g., modeling of marital violence in the family of origin, lack of relationship skills), so that on some occasions during escalating marital conflicts, these men would perpetrate physical aggression. Following such incidents, however, their low levels of psychopathology and related problems (e.g., low impulsivity, low attachment dysfunction), combined with their lack of hostility toward women and lack of positive attitudes toward violence, would result in remorse and prevent their aggression from escalating. Our newer conceptualization of this subtype suggests that it might also be appropriate to consider the role of dyadic factors in understanding the violence of FO batterers.
In contrast, DB batterers were hypothesized to come from a background involving parental abuse and rejection. From such childhood experiences, these men were expected to have difficulty forming a stable, trusting attachment with an intimate partner. Instead, they would be very jealous and highly dependent upon, yet fearful of losing, their wives. They would tend to be impulsive and lack marital skills and have hostile attitudes toward women but positive attitudes toward violence. Their early traumatic experiences might lead to borderline personality organization, anger, and insecure attachment which, when frustrated, result in violence against the adult attachment figure (i.e., their romantic partner or wife).
Finally, GVA batterers were predicted to resemble other antisocial, aggressive groups. Compared with the other subtypes, they were expected to have experienced high levels of family-of-origin violence and association with deviant peers. They would be impulsive and lack both marital and nonmarital relationship skills, have hostile attitudes toward women, and view violence as acceptable. Their marital violence was conceptualized as a part of their general use of aggression and engagement in antisocial behavior.
To test the proposed typology, we conducted a study of men, recruited from the community, who had been physically aggressive toward their wives in the past year. We included men who had engaged in a wide range of violence. Using measures of the three descriptive dimensions (marital violence, general violence, personality disorder), the three predicted subgroups of violent men (FO, DB, and GVA) emerged, along with one additional subgroup, which we labeled low-level antisocial (LLA). The three predicted subgroups generally differed as hypothesized on both the descriptive dimensions and the model correlates of violence (i.e., childhood home environment, association with deviant peers, impulsivity, attachment, skills, and attitudes). The LLA group had moderate scores on measures of antisociality, marital violence, and general violence. On many measures, this group had higher scores than FO men but lower scores than GVA men.
In addition to our own study, other recent batterer typologies have usually been consistent with the proposed typology. Thus, across studies from multiple laboratories, the existing research suggests that using some or all of the proposed descriptive dimensions (i.e., severity/frequency of marital violence, generality of violence, psychopathology/ personality disorder), the proposed batterer subtypes will be identified and generally will differ in a theoretically consistent manner.
Given that the proposed subtypes can be identified at one point in time, then an important followup question becomes: How stable, over time, are the violence and related characteristics and behaviors of the differing batterer subgroups? Some researchers initially suggested that the subtypes of maritally violent men, identified cross-sectionally, represent different phases of marital violence, with violent husbands progressing from a less violent subtype to a more severely violent subtype. Such predictions were derived from the assumption that husband violence, once begun, inevitably escalates in frequency and severity. Contradicting this notion, however, are several longitudinal studies suggesting that the initial severity level of husband violence is the best predictor of violence continuation and that men engaging in low levels of violence do not necessarily escalate their violence level over time. Thus, rather than hypothesizing that the subtypes represent phases in the escalation of relationship violence, we suggested that the violence levels of subtypes of maritally violent men should continue to differ over time. More severely violent men would continue to engage in more marital violence and related relationship abuse (e.g., sexual and psychological aggression), while less severely violent men would maintain lower levels of abuse or even desist from relationship violence over time.
To examine such issues in our typology study, we reassessed men approximately three years after their initial assessments. As predicted, the aggression of men who initially engaged in lower levels of violence was less stable over time than that of more severely violent men. In fact, among men still having at least monthly contact with their partners (i.e., the opportunity for continued violence), 40 percent of FO batterers and 23 percent of LLA batterers desisted from violence over a three-year period, but only 7 percent of GVA batterers and 14 percent of DB batterers desisted from violence.
Though these data suggest that the FO group may be a stable one, they should not be interpreted to mean that all men who engage in low levels of physical aggression will not escalate to more severe violence. Instead, the data suggest that among men engaging in low levels of violence, those who resemble FO batterers, in terms of evidencing low levels of other risk factors (e.g., little concurrent psychopathology or generally violent behavior, low risk from other risk factors such as impulsivity and negative attitudes), may continue to have a low risk of marital violence over time. In contrast, we would predict that a man who has, to date, engaged in only low levels of marital violence but who resembles a more violent subtype, such as the GVA subgroup, on risk factors (e.g., criminal behavior, delinquent friends, substance abuse, impulsivity, problematic attitudes) is at high risk to perpetrate escalating levels of marital aggression over time. A related question regards the stability of batterer characteristics that are theoretically linked to a man’s use of violence. We predicted that the initial subtype differences, observed on measures of theoretically relevant individual characteristics, would continue to differentiate the subtypes over time.
This prediction was based on our implicit hypothesis that the relationship violence of severely violent men is related to stable individual characteristics of these men. Indirectly, our model suggests that the individual characteristics of some men put them at high risk for perpetrating severe relationship violence; in these relationships, the man is viewed as the cause of the relationship violence and is likely to carry his violence forward, across time and across relationships. Thus, the individual characteristics of these men that are theoretically linked to their use of violence (e.g., antisociality, insecure attachment, impulsivity) should remain relatively constant across time. In contrast, the low levels of violence perpetrated by our least violent subgroup, while reflecting some individual characteristics of the man, may also be related to dyadic factors (e.g., marital conflict), life stressors (e.g., job loss), and the cultural acceptability of low levels of relationship aggression. If correct, then men in the least violent subgroup are not necessarily men in the early phases of developing a lifelong pattern of escalating relationship aggression, but rather, over time, should use violence inconsistently and should continue to evidence low levels of risk factors for relationship violence (e.g., reporting less psychopathology, less positive attitudes toward violence, less impulsivity than other subtypes).
Many of our longitudinal study findings supported our hypotheses that the subtypes would continue to differ in men’s individual characteristics assumed to be related to their use of violence. While not all group differences reached statistical significance, across time and as predicted, men in one or both of the more severely violent subtypes (GVA and DB) reported more continuing problems than the less violent men (FO and/or LLA) on a variety of measures, including:
- highest level of wife injury (for wives of both GVA and DB batterers),
- highest levels of impulsivity (both GVA and DB batterers),
- most positive attitudes toward violence and most hostile attitudes toward women (both GVA and DB batterers),
- the highest scores on measures of borderline personality disorder characteristics, fearful/ preoccupied attachment, jealousy and dependency on the spouse, and being most likely to seek psychological help (DB batterers), and
- highest levels of problems resulting from substance use and the highest number of new arrests (GVA batterers).
One related issue emerging from recent typology research is the question of whether the GVA and DB subgroups are distinct. Outside of our lab, one group of researchers identified a pathological subtype that scored high on measures of both antisocial and borderline personality disorders. Similarly, in our study, the GVA and DB groups did not always differ significantly. Part of the problem in differentiating these two subgroups is that both are predicted to be similar in their levels of marital aggression and on several model correlates of violence, including impulsivity, positive attitudes toward violence, and negative attitudes toward women. Yet, over time, the GVA and DB groups in our study did differ on some variables. For example, GVA batterers were the most likely to be arrested, and DB batterers scored highest on a measure of jealousy and spouse-specific dependency and were the most likely to have been treated for depressive symptoms.
While our study findings do not conclusively resolve this debate, they do suggest the potential importance of two types of personality-related characteristics (i.e., antisociality and borderline) when studying husband violence. Indeed, considering antisocial behavior, it is possible to conceptualize three of our violent subtypes (FO, LLA, and GVA) as falling along a continuum of antisociality (e.g., FO batterers have the lowest levels of violence, antisocial behavior, and risk factors; GVA batterers have the highest; and the LLA group has intermediate levels). The DB group, however, could not be easily placed along this continuum, as it had the highest scores on variables that clustered in a theoretically coherent manner (i.e., fear of abandonment, preoccupied/fearful attachment, dependency). This has led us to argue that both dimensions (antisociality and borderline personality characteristics) are needed to describe all of the subgroups.
In summary, current batterer typology research suggests that identifying subtypes of maritally violent men is a useful method to account for the heterogeneity among maritally violent men on theoretically important variables that may remain relatively consistent across time. Yet one can ask whether these subgroups are true diagnostic categories that identify underlying taxonomical differences across subtypes of batterers, or whether, while some men are prototypes of the different subtypes, the majority of batterers fall along dimensions of theoretical importance rather than forming distinctly identifiable groups. The answers to such questions await further research, but we currently believe that both dimensional and categorical approaches to conceptualizing the data should be considered.
Assuming that one wishes to use a categorical approach, and thus to identify subgroups of men in clinical or research work, we have been asked to provide subgroup cut-off scores on the measures we used to identify subgroups of maritally violent men. Based on pilot work with various samples, however, we are concerned that cut-off points may not be generalizable from one setting to another. Thus, at this time, we believe that absolute cut-off scores cannot be provided to identify subtypes until further research is conducted. Instead, cut-off scores for each sample need to be established.
Clinical Implications of the Typology
Given concerns about the potentially low overall effectiveness of batterers’ treatment, one hope is that a batterer typology might allow us to distinguish subtypes likely to benefit from treatment from those unlikely to improve following clinical intervention. Indeed, recent findings suggest that standard batterer treatment programs may be ineffective for certain subtypes of maritally violent men. Specifically, researchers have found that GVA batterers are the most likely to drop out of treatment and the most likely to recidivate following treatment. In contrast, FO batterers are the most likely to complete treatment and to remain violence free afterward.
Another idea is that treatment outcome might be improved by matching interventions to batterer subtypes. For example, in a study comparing cognitive-behavioral-feminist treatment to a process-psychodynamic treatment, researchers found that batterers scoring high on an antisocial measure did better in the structured cognitive-behavioral intervention, while batterers scoring high on a measure of dependency did better in the other intervention. While it is premature to recommend particular interventions for various subtypes of batterers, we can generate hypotheses for future testing.
When considering the DB subtype, the one study discussed suggests that process-psychodynamic interventions might be useful, as might interventions developed for borderline personality disorder or that focus on past traumas and affect regulation. In addition, given their high levels of psychological distress, it may prove beneficial to provide DB batterers entering batterer programs with adjunct medication, such as antidepressants, or individual therapy for their immediate psychological distress. When considering GVA batterers, the available research on treatments for violent offenders and individuals with antisocial personality disorder suggest that many currently available interventions (e.g., insight oriented) are not effective, while other interventions (i.e., cognitive-behavioral approaches) deserve more study. In addition, it may be important to consider new interventions developed in other fields, such as criminal justice efforts (e.g., close monitoring of offenders in the community), and to directly address the other potentially relevant problems prevalent among this subgroup (e.g., substance abuse). Given their lower levels of behavioral, personality, and criminal problems, we have hypothesized that FO, and perhaps LLA, batterers would be the most likely to benefit from existing batterer treatment programs (e.g., cognitive-behavioral approaches including anger management and feminist approaches encouraging these men to examine gender roles). Finally, couples with an FO-batterer husband may be the only physically aggressive couples for whom conjoint couples therapy would be appropriate, as this group of men often resembles maritally distressed men who do not perpetrate violence, and the risk factors that characterize FO batterers (e.g., skills deficits) are often directly addressed in available conjoint therapy programs.
Issues for Further Consideration
Many important issues in this area remain unresolved and require further examination. One is the question of the meaning and impact of low-level husband violence. Low levels of physical aggression (e.g., FO batterers) are so prevalent as to be almost statistically normative in our culture, and we do not understand how less violent men differ from men who are experiencing marital conflict but who do not engage in physical aggression. It is thus tempting to assume that low levels of aggression do not lead to any more negative outcomes than does marital distress alone. However, recent longitudinal studies of newlyweds suggest that even relatively low levels of physical aggression may be highly detrimental to relationship functioning, even perhaps predicting marital dissolution better than does marital distress or negative marital communication. Thus, we believe that lower levels of male physical aggression continue to deserve attention.
Also, it is now time to examine the potentially differing impacts on women (and children) of being in a relationship with differing subtypes of batterers. For example, in our study, we found increasing group differences in wives’ marital satisfaction over time, with the wives of more severely violent men reporting more relationship distress. Study of other negative consequences for women (e.g., depression) should be conducted, as should studies of possible negative consequences for the children (e.g., child abuse) in these homes.
An important criticism of existing batterer typologies is that they focus so heavily on individual characteristics of violent men (e.g., psychopathology) that are relatively distal predictors of violence. While typologies tend to emphasize characteristics of the individual, one must remember that husband violence occurs in the context of interpersonal relationships, communities and subcultures, and society. For example, it is important to note that batterer typologies have been developed only in Western countries, primarily the United States. Whether such typologies will be replicated in other cultures is a question requiring empirical investigation. It is possible, given the role of societal factors (e.g., patriarchy, availability of guns) in shaping the expression of husband violence, that existing typologies will require modification in other cultures. For example, in societies in which husband violence is widely viewed as acceptable, there may be a weaker link between husband violence and psychopathology. Similarly, minimal attention has been paid to the possible impact of ethnic and socioeconomic group variability on typologies. As another example, the generalizability of existing batterer typologies to same-sex relationships has not been examined.
At the opposite end of the spectrum, one could examine more immediate variables related to how stable characteristics (e.g., personality variables, motivations) are expressed in a given situation. Consider the DB subgroup: How is an insecure attachment style activated in a particular situation? And once activated, how do fears of abandonment translate into violence? Does the man act in rage, experiencing emotional dysregulation, or in a calculated manner, to prevent his partner from leaving him? It will be important to understand why, for each subtype of batterer, violence emerges within an ongoing dyadic relationship and within particular situations.
The field needs prospective studies to identify the developmental pathways resulting in different subtypes of violent husbands. In such studies, researchers could examine constructs assumed to predict the use of violence among samples of adolescents or children (e.g., characteristics of antisocial and borderline personality disorders) and then observe the relationship between these variables and the emergence of relationship violence as study participants enter intimate relationships. Similarly, future researchers could study the process of desisting from violence. Several longitudinal studies suggest that some men, particularly those who engage in ‘‘minor’’ violence, desist. This issue might have clinical implications if it can suggest techniques that men have found useful in refraining from violence. Indeed, longitudinal studies of the predictors of violence and violence desistance could help guide future violence prevention efforts. Ultimately, rather than waiting to intervene with cases of serious levels of partner violence, it would be better to work to prevent the onset and escalation of physical aggression within intimate relationships. To do so, we must continue to better understand the predictors of varying types of husband violence.
- Holtzworth-Munroe, A., and J. C. Meehan. ‘‘Typologies of Men Who Are Maritally Violent: Scientific and Clinical Implications.’’ Journal of Interpersonal Violence 19 (2004): 1369–1389.
- Holtzworth-Munroe, A., J. C. Meehan, K. Herron, U. Rehman, and G. L. Stuart. ‘‘Testing the Holtzworth-Munroe and Stuart (1994) Batterer Typology.’’ Journal of Consulting and Clinical Psychology 68 (2000): 1000–1019.
- ———. ‘‘Do Subtypes of Maritally Violent Men Continue to Differ Over Time?’’ Journal of Consulting and Clinical Psychology 71 (2003): 728–740.
- Holtzworth-Munroe, A., and G. L. Stuart. ‘‘Typologies of Male Batterers: Three Subtypes and the Differences among Them.’’ Psychological Bulletin 116 (1994): 476– 497.
- Lynam, D. R. ‘‘Looking Earlier in the Lifecourse for the GVA: Comment on Holtzworth-Munroe and Meehan.’’ Journal of Interpersonal Violence 19 (2004): 1410–1440.
- Prentky, R. A. ‘‘Can Sex Offender Classification Inform Typologies of Male Batterers?: A Response to Holtzworth- Munroe and Meehan.’’ Journal of Interpersonal Violence 19 (2004): 1405–1411.
- Saunders, D. G. ‘‘The Place of a Typology of Men Who Are Maritally Violent Within a Nested Ecological Model.’’ Journal of Interpersonal Violence 19 (2004): 1390–1395.
- Widiger, T. C., and S. N. Mullins-Sweatt. ‘‘Typology of Men Who Are Maritally Violent: A Discussion of Holtzworth-Munroe and Meehan.’’ Journal of Interpersonal Violence 19 (2004): 1396–1400.