Violence and Abuse in Personal Relationships:



Violence and Abuse in Personal Relationships:

Conflict, Terror, and Resistance in Intimate Partnerships

Michael P. Johnson[1]

DRAFT: DO NOT QUOTE WITHOUT PERMISSION OF AUTHOR

Prepared for Anita Vangelisti and Daniel Perlman (Eds.), The Cambridge Handbook of Personal Relationships. New York: Cambridge University Press.

Violence and Abuse in Personal Relationships:

Conflict, Terror, and Resistance in Intimate Partnerships

Michael P. Johnson

Violence occurs in all sorts of personal relationships. Parents hit their children, siblings have fights, girls slap their boyfriends, friends get into it, husbands terrorize their wives, and abused wives murder their husbands. Most of this violence receives very little attention from scholars of personal relationships. For example, a 1997 handbook on personal relationships covers violence in only one section of one chapter (Klein & Johnson, 1997). However, in other disciplines such as family studies, social work, criminology, and sociology there are research literatures (some small, some rather large) on most of these forms of violence. In the wake of the 20th century women’s movement and the related cultural emphasis on gender equality, one of the largest of these literatures is focused on violence between intimate partners, including people who are dating, living together, married, or separated (Jasinski & Williams, 1998; Johnson & Ferraro, 2000). As a feminist sociologist, I work within and know best this literature, which is why this chapter focuses on intimate partner violence (IPV).

However, it may well be the case that the major lessons of the IPV literature are relevant for understanding violence in other kinds of personal relationships. I believe the two core lessons to be learned from work on IPV are simple, profound, and broadly applicable to violence in all types of personal relationships. First, one cannot understand violence in personal relationships without understanding its role in the relationship itself. Unlike most other kinds of violence (such as a mugging), which are essentially situational and do not involve a continuing relationship between the parties involved, personal relationship violence arises out of and shapes the dynamics of an ongoing relationship, the violence in some cases—but not always—being a central feature of the relationship.

Second, and more substantively, there are three quite different types of intimate partner violence, identified by their role in the control context of the relationship in which they are embedded. One type involves a violent attempt to take complete control or at least to generally dominate the relationship (intimate terrorism), another involves violent resistance to such a control attempt (violent resistance), and the third is violence that is a product of particular conflicts or tensions within the relationship (situational couple violence). As you will see, the nature of the control context is a major theme in the IPV literature, and although it has as yet received little attention in research on other types of personal relationships, there are hints of it in the parent-child literature.

Types of Intimate Partner Violence

These two core propositions are central to any theory of personal relationship violence. The three types of violence—intimate terrorism, violent resistance, and situational couple violence—have different origins, different dynamics, and different consequences. They therefore require different theoretical frameworks to explain them and different strategies for prevention and intervention. The failure to acknowledge these differences has led to major errors in the empirical literature on IPV, and perhaps in literatures on violence in other types of relationships. Unfortunately, until we carry out a broad program of research to investigate differences among the causes and consequences of the various types of IPV, we cannot know how widespread the errors are. Two examples, however, will illustrate the basic processes by which these errors are produced.

First, inadvertently aggregating different types of violence under one label produces data that are an “average” of the characteristics or correlates of the types that are aggregated. For example, a recent meta-analysis of the literature on the relationship between growing up in a violent home and subsequently becoming part of a violent marital relationship indicates quite small effects (Stith et al., 2000), calling into question what is often claimed to be one of the best established relationships in the IPV literature, the so-called “intergenerational transmission of violence.” The overall conclusion is that childhood experiences of family violence are not strongly related to adult IPV. However, that literature (and therefore the meta-analysis) does not distinguish among types of violence, thus inadvertently aggregating whatever mix of types is found in the studies reviewed. Of course, this would not be a problem if the relationships of different types of violence to childhood experiences were the same. However, a recent study differentiating among the types finds that while situational couple violence is not strongly related to childhood experiences of violence, intimate terrorism is (Johnson & Cares, 2004). The “average” relationship thus does not represent the effect that is of most interest to most audiences: the effect on the likelihood of becoming an intimate terrorist—a wife-beater.

The second type of error arises because different sampling strategies have different biases in terms of the types of violence they include. This is the error that produced the decades-long, and continuing, debate over the gender symmetry of domestic violence. Researchers using agency samples find domestic violence to be almost entirely male-perpetrated, while those using general survey samples find domestic violence to be gender-symmetric. As it turns out, general survey samples are dominated by situational couple violence, which is roughly gender-symmetric, while agency samples are dominated by intimate terrorism, which in heterosexual relationships is almost entirely male-perpetrated (Johnson, 1995, 2001). It is important to remember that to most audiences intimate terrorism is what the term “domestic violence” is all about. Thus, when we present conclusions about situational couple violence under the general rubric of domestic violence, we mislead the public in important ways, such as giving them the false impression that there are as many battered husbands as there are battered wives.

The typology of IPV presented below has its roots in this debate about gender symmetry. For decades, feminist theorists have argued that domestic violence is largely male-perpetrated and rooted in the patriarchal traditions of the Western family (Dobash & Dobash, 1979). Family violence theorists, although acknowledging some role of gender in family violence, have argued that domestic violence is rooted in the everyday tensions and conflicts of family life and that women are as violent as men in intimate relationships (Straus, 1999). We now know that they were both right, but were not studying the same phenomenon. Two of the major types of IPV (intimate terrorism and violent resistance) are rooted in the dynamics of control and resistance that have been the focus of feminist theorists. Intimate terrorism and violent resistance comprise the bulk of the violence in the agency samples with which feminist theorists work. The third major type, situational couple violence, is rooted in the dynamics of family conflict that have been the focus of family violence theorists. Situational couple violence comprises the bulk of the violence in the general survey samples with which family violence theorists work. The three types constitute a typology of individual violence that is rooted in information about the couple and their relationship, and defined by the control context within which the violence is embedded.

The Nature of Intimate Terrorism (IT)

In IT, the perpetrator uses violence in the service of general control over his or her partner. The partner does not. The “control” that forms the basis of this typology of IPV, and that is the defining feature of IT, is more than the specific, short-term control that is often the goal of violence in other contexts. The mugger wants to control you only briefly in order to take your valuables and move on, hopefully never to see you again. In contrast, the control sought in IT is general and long-term. Although each particular act of intimate violence may appear to have any number of short-term, specific goals, it is embedded in a larger pattern of power and control that permeates the relationship. This is the kind of violence that comes to mind when most people hear the term “domestic violence.” Figure 1 is a widely used graphical representation of partner violence deployed in the service of general control.

INSERT FIGURE 1 ABOUT HERE

This diagram and the understanding of domestic violence that lies behind it were developed over a period of years from the testimony of battered women in the Duluth, Minnesota area, testimony that convinced the staff of the Duluth Domestic Abuse Intervention Project that the most important characteristic of the violence they encountered was that it was embedded in a general pattern of power and control (Pence & Paymar, 1993). A pattern of power and control cannot, of course, be identified by looking at violence in isolation. It can only be recognized from information about multiple control tactics, allowing one to find out whether a perpetrator uses more than one of these tactics to control his or her partner, thus indicating an attempt to exercise general control.

A brief tour of the wheel, starting with economic abuse and moving through the other forms of control, might help to capture what Catherine Kirkwood calls a “web” of abuse (Kirkwood, 1993). It is not unusual for an intimate terrorist to deprive his[2] partner of control over economic resources. He controls all the money. She is allowed no bank account and no credit cards. If she works for wages, she has to turn over her paychecks to him. He keeps all the cash and she has to ask him for money when she needs to buy groceries or clothes for herself or their children. He may require a precise accounting of every penny, demanding to see the grocery bill and making sure she returns every bit of the change.

This economic abuse may be justified through the next form of control, male privilege: “I am the man of the house, the head of the household, the king in my castle.” Of course, this use of male privilege can cover everything. As the man of the house, his word is law. He doesn’t have to explain. She doesn’t disagree with him. She is to do his bidding without question. And don’t talk back. All of this holds even more rigidly in public, where he is not to be humiliated by back-talk from “his woman.”

How does he use the children to support his control? First of all, they too know he is the boss. He makes it clear that he controls not only them, but their mother as well. He may use them to back him up, to make her humiliation more complete by forcing them into the room to assist him as he confronts her, asking them if he isn’t right, and making them support his control of her. He may even have convinced them that he should be in charge, that he does know what is best (father knows best), and that she is incompetent or lazy or immoral. In addition, he may use her attachment to the children as a means of control, by threatening to take them away from her or hurt them if she isn’t a “good wife and mother.” Of course, being a good wife and mother means doing as he says.

Then there’s isolation. He keeps her away from everyone else. He makes himself her only source of information, of support, of money, of everything. In a rural setting he might be able to literally isolate her, moving to a house trailer in the woods, with one car that he controls, no phone, keeping her there alone. In an urban setting, or if he needs her to go out to work, he can isolate her less literally, by driving away her friends and relatives and intimidating the people at work, so that she has no one to talk to about what’s happening to her.

When she’s completely isolated, and what he tells her about herself is all she ever hears about herself, he can tell her over and over again that she’s worthless—humiliating her, demeaning her, emotionally abusing her. She’s ugly, stupid, a slut, a lousy wife, an incompetent mother. She only manages to survive because he takes care of her. She’d be helpless without him. And who else is there to tell her otherwise? Maybe he can even convince her that she can’t live without him.

If she resists, he can intimidate her. Show her what might happen if she doesn’t behave. Scream at her. Swear at her. Let her see his rage. Smash things. Or maybe a little cold viciousness will make his point. Kick her cat. Hang her dog. That ought to make her think twice before she decides not to do as he says. Or threaten her. Threaten to hit her, or beat her, or pull her hair out, or burn her. Or tell her he’ll kill her, and maybe the kids too.

Pull all these means of control together, or even a few of them, and the abuser entraps and enslaves his partner in a web of control. If she manages to thwart one means of control, there are others at his disposal. Wherever she turns, there is another way he can control her. Sometimes she is ensnared by multiple strands. She can’t seem to escape—she is trapped. But with the addition of violence there is more to power and control than entrapment. There is terror.

For this reason the diagram does not include the violence as just another means of control, another spoke in the wheel. The violence is depicted, rather, as the rim of the wheel, holding all the spokes together. When violence is added to such a pattern of power and control, the abuse becomes much more than the sum of its parts. The ostensibly nonviolent tactics that accompany that violence take on a new, powerful, and frightening meaning—controlling the victim not only through their own specific constraints, but also through their association with the general knowledge that her partner will do anything to maintain control of the relationship, even attack her physically. Most obviously, the threats and intimidation are clearly more than idle threats if he has beaten her before. But even his “request” to see the grocery receipts becomes a “warning” if he has put her into the hospital this year. His calling her a stupid slut may feel like the beginning of a vicious physical attack. As battered women often report, “All he had to do was look at me that way, and I’d jump.” What is for most of us the safest place in our world—home—is for her a place of constant fear.

The Nature of Violent Resistance (VR)

What is a woman to do when she finds herself terrorized in her own home? At some point, most women in such relationships do fight back physically. For some, this is an instinctive reaction to being attacked, and it happens at the first blow—almost without thought. For others, it doesn’t happen until it seems he is going to continue to assault her if she doesn’t do something to stop him. For most women, the size difference between them and their male partner ensures that violent resistance won’t help, and may make things worse, so they turn to other means of coping. For a few, eventually it seems that the only way out is to kill their partner.

The critical defining pattern of violent resistance is that the resistor is violent but not controlling, and is faced with a partner who is both violent and controlling, i.e., he is an intimate terrorist. Violence in the face of IT may arise from any of a variety of motives. She may (at least at first) believe that she can defend herself, that her violent resistance will keep him from attacking her further. That may mean that she thinks she can stop him right now, in the midst of an attack, or it may mean that she thinks that if she fights back often enough he will eventually decide to stop attacking her physically. Even if she doesn’t think she can stop him, she may feel that he shouldn’t be allowed to attack her without getting hurt some himself. This desire to hurt him in return even if it won’t stop him can be a form of communication (“What you’re doing isn’t right and I’m going to fight back as hard as I can”) or it may be a form of retaliation or payback, along the lines of “He’s not going to do that without paying some price for it.” In a few cases, she may be after serious retaliation, attacking him when he is least expecting it and doing her best to do serious damage, even killing him. But there is another, more frequent motive for such premeditated attacks—escape. Sometimes, after years of abuse and entrapment, a victim of IT may feel that the only way she can escape from this horror is to kill her tormenter.

The Nature of Situational Couple Violence (SCV)

Probably the most common type of partner violence does not involve any attempt on the part of either partner to gain general control over the relationship. The violence is situationally-provoked, as the tensions or emotions of a particular encounter lead someone to react with violence. Intimate relationships inevitably involve conflicts, and in some relationships one or more of those conflicts may escalate to violence. The violence may be minor and singular, with one argument at some point in the relationship escalating to the level that someone pushes or slaps the other, is immediately remorseful, apologizes and never does it again. Or it could be a chronic problem, with one or both partners frequently resorting to violence, minor or severe. The motives for such violence vary. A physical reaction might feel like the only way one’s extreme anger or frustration can be expressed. It may well be intended to do serious injury as an expression of anger. It may primarily be an attempt to get the attention of a partner who doesn’t seem to be listening. Or there could be a control motive involved, albeit not one that is part of a general pattern of coercive control. One partner may simply find that the argument is not going well for him or her, and decide that one way to win this is to get physical.

The critical distinctions among types of violence have to do with general patterns of power and control, not with the ostensible motives for specific incidents of violence. Thus, many of the separate violent incidents of SCV may look exactly like those involved in IT or VR. The difference is in the general power and control dynamic of the relationship, not in the nature of any one assault. If it appears that neither partner is generally trying to control the other, i.e., it is not the case that the relationship involves the use of a range of control tactics by one or both of the partners, then we are dealing with SCV. It is simply that one or more disagreements have resulted in violence. The violence may even be frequent, if the situation that provokes the violence is recurring, as when one partner frequently feels that the other is flirting, and the confrontations over that issue regularly lead one or the other of them to lash out. And the violence may be quite severe, including even homicide. What makes it SCV is that it is rooted in the events of a particular situation rather than in a relationship-wide attempt to control.

How Do We Know About These Types?

The descriptions of the three types of partner violence that you have just read are derived from 30 years of social science research on violence between intimate partners, most of which did not make the distinctions that I describe. How, then, can we manage to come to conclusions about these different types of partner violence from a research literature that doesn’t distinguish among them? There are two answers to that question. First, some of the more recent research does operationalize the distinctions. Second, there are some “tricks” that can be used to tease information regarding the types out of the “old style” research that didn’t make distinctions. One involves the sampling biases noted above. The violence in general survey research is almost entirely men’s and women’s SCV. Thus, any survey research that compares violent with nonviolent men or women—or victims with non-victims—can be reliably assumed to tell us mostly about SCV. In contrast, the violence in agency samples is almost all men’s IT and women’s VR. Thus, agency-based studies can be used to inform us regarding those two types of violence. The second trick we can use with the old literature is to look for patterns that are associated with violence that shows the characteristics of each of the types. Here is an example. Violence that is frequent and severe is most likely to be IT (see below). Violence that is infrequent and mild is more likely to be SCV. Thus, if studies show that anger management therapy is only effective in the treatment of men whose violence is infrequent and mild, we have indirect support for the conclusion that it is effective for SCV but not effective for IT.

Correlates of the Types of Intimate Partner Violence

Here we face something of a dilemma. There is no shortage of comprehensive causal models, inventories, reviews, or meta-analyses of correlates of intimate partner violence (K. L. Anderson, 1997; Archer, 2000; Hotaling & Sugarman, 1986, 1990; Jasinski & Williams, 1998; O'Leary & Slep, 2003; Stith et al., 2000; Sugarman, Aldarondo, & Boney-McCoy, 1996; Sugarman & Frankel, 1996; Sugarman & Hotaling, 1989, 1997), but none of these papers or books makes the distinctions among types of intimate partner violence discussed above. Thus, we cannot be certain about which generalizations apply to which types of violence. For example, although there are perhaps hundreds of studies that purport to show that women are as violent as men in intimate relationships (Archer, 2000; Straus, 1999), a breakdown of IPV into types (Johnson, 2001) indicates that in heterosexual relationships IT is almost entirely male-perpetrated, violent resistors are almost all women, and SCV is not heavily gendered (at least not in terms of incidence; see below).

Antecedents

Gender. Let me begin with a reminder that the discussion above indicates that perhaps the most important antecedent of IPV is gender, with IT being almost entirely male, VR female, and SCV more gender symmetric. Within each of these types, men’s violence is more frequent and more severe than women’s. It would not be wise, however, to think of gender only as a characteristic of perpetrators and victims if IPV. For over two decades, now, feminist sociologists have admonished us to treat gender as an institution, not an individual characteristic. It has manifestations at all levels of social organization from the most macro of organizational contexts through the “meso” level of social interaction on down to the individual level of identities and attitudes (Ferree, Lorber, & Hess, 2000; Risman, 2004).

Thus, the role of gender in IPV is almost impossibly complex. First, sex/gender is important in heterosexual relationships simply because of average sex differences in size and strength. Second, the gendering of individual attitudes, values, knowledge, and skills affect partners’ goals in their relationships and the means they use to attain them. Third, individual partners’ attitudes regarding differences between men and women and the role of gender in relationships each play a part in the development of any particular relationship. Fourth, intimate partnership norms are heavily gendered, certainly in the midst of considerable historical change but rooted in a patriarchal heterosexual model that validates men’s power (Dobash & Dobash, 1979, 1992; Yllö & Bograd, 1988; Yllö & Straus, 1990). These norms affect the internal functioning of all relationships, regardless of the individual attitudes of the partners. For example, when a couple sets out to plan a wedding, they find their individual interests embedded in a larger social context that cannot be ignored. Finally, the gendering of the social context within which the relationship is embedded affects the resources the partners can draw upon to shape the relationship and to cope with or escape from the violence. As an example, consider the major changes in the way the criminal justice system has reacted to domestic violence over the last 30 years (Dobash & Dobash, 1992).

There certainly is not enough space here to explicate the complex interaction of these gender-related factors in the shaping of different types of heterosexual IPV, but I would like to at least provide a few examples. First, in intimate terrorism, the use of violence as one tactic in an attempt to exercise general control over one’s partner requires more than violence. It requires a credible threat of a damaging violent response to noncompliance. Such a threat would be more credible coming from a man than a woman because of both the size difference and the cultures of masculinity and femininity. Second, with regard to SCV, the damage will be greater when perpetrated by a larger against a smaller partner (Felson, 1996), and the cultures of masculinity and femininity ensure that whatever the level of violence, its meaning will differ greatly depending upon the gender of the perpetrator (Straus, 1999). And third, in cases of either IT or SCV, the reactions of criminal justice personnel are likely to differ as a function of the gender structure of the organization to which the officers belong, as well as the gender of the perpetrator (Buzawa & Buzawa, 1996).

Finally, it is important to note that much of this discussion of gender is relevant only to heterosexual relationships. In same-sex relationships, some aspects of gender will still be important (e.g., gender differences in individual attitudes and skills), others will be largely irrelevant (e.g., heterosexual relationship norms), and some will play themselves out in quite different ways (e.g., sex differences in size and strength will not be as likely to create differences within couples). Although we know considerably less about same-sex relationships than we do about heterosexual relationships, there is a growing literature that is important not only in its own right, but also because it sheds light on some of the inadequacies of theories rooted in research on heterosexual relationships (Renzetti, 1992, 2002; Renzetti & Miley, 1996).

Personality and attitudes. The work of Amy Holtzworth-Munroe and others on male perpetrators of IPV (Gondolf, 1988; Hamberger, Lohr, Bonge, & Tolin, 1996; Holtzworth-Munroe, 2002; Holtzworth-Munroe, Meehan, Herron, Rehman, & Stuart, 2000, 2003; Jacobson & Gottman, 1998; Ornduff, Kelsey, & O'Leary, 1995; Saunders, 1992, 1996; Waltz, Babcock, Jacobson, & Gottman, 2000) converges on the identification of two types of intimate terrorists. Holtzworth-Munroe refers to them as “borderline/dysphoric” and “generally violent/antisocial,” while Jacobson and Gottman use the more colorful terms, “pit bulls” and “cobras.” (I am going to call them “dependent intimate terrorists” and “antisocial intimate terrorists.”) The dependent intimate terrorists score high on measures of borderline personality organization, dependency, and jealousy, and seem to need general control in order to assuage their fear of losing their partner. The antisocial intimate terrorists are more generally violent and involved with delinquent peers, substance abuse, and criminal behavior; they are broadly willing to employ violence to have their way in many contexts. Both of these types score high on impulsivity, acceptance of violence, and hostile attitudes toward women, and low on measures of social skills (Holtzworth-Munroe et al., 2000). The finding that intimate terrorists score high on hostility toward women is consistent with the meta-analysis of Sugarman and Frankel (1996), and of course harkens back to my discussion of gender as an institution. Sugarman and Frankel found that intimate terrorists score higher on traditional gender attitudes than nonviolent men (effect size = .80 for studies that are probably dominated by IT, .54 when all studies are considered). Unfortunately, this important meta-analysis’s support for feminist theory is often missed because Sugarman and Frankel’s conclusions and abstract dismiss the finding.

SCV is included in Holtzworth-Munroe’s work in what she calls the “family-only” cluster of perpetrators, a group of violent men whose violence is less severe than that of the intimate terrorists, and who do not differ in personality or attitudes from nonviolent men. This pattern is supported by Sugarman and Frankel’s finding that there are no differences in gender attitudes between violent and nonviolent men in studies that are probably dominated by SCV. Thus, it is clear that some elements of gender are not implicated in SCV.

As far as I know, there is no research on personality or attitudinal correlates of women’s SCV, but Murray Straus has written a useful discussion of factors that might increase women’s likelihood of being violent towards their partners. These factors include many features of gender as an institution, such as cultural norms regarding femininity, gender norms for conflict, sources of identity, and reactions of the criminal justice system (Straus, 1999, p. 31), some of which influence the personality and attitudes of women.

So-called intergenerational “transmission.” Although childhood experiences of family violence are often touted as “the most widely accepted risk marker for the occurrence of partner violence” (Kantor & Jasinski, 1998, p. 16), in fact most studies find effect sizes on the order of .17 (Stith et al., 2000). It appears that “transmission” is hardly an appropriate metaphor (see also Johnson & Ferraro, 2000). However, a careful look at the Stith et al. meta-analysis and some recent research suggests that perhaps this is another case of mis-specification through aggregation. A case can be made for a moderate effect of childhood experiences on IT, and a minimal effect on SCV, as follows. Because the literature is dominated by general surveys, the overall average effect size, which is quite small, reflects the impact of childhood experiences on SCV. However, Stith et al. find an interaction with sample type, in which the effect is much stronger (.35) for agency samples, i.e., samples dominated by IT. More direct evidence comes from Johnson and Cares’ (2004) finding of odds ratios of 2.40 (ns) and 7.51 (p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download