A CRITICAL LOOK AT JANET JOHNSTON'S

TYPOLOGY OF BATTERERS

by R. Lundy Bancroft

c 1998

Given that histories of domestic abuse are present in as many as 75% of custody and visitation disputes, family courts face an overwhelming job in protecting the long-term interests of children. The temptation arises to choose a simple way to identify those batterers who are most dangerous to their children. Janet Johnston's work attempts to make this sort of clear demarcation, and her theories are having a wide influence on court practice; for example, her writings form the basis for the visitation risk assessment distributed nationally by the Association of Family and Conciliation Courts (AFCC).

Johnston's work contains fundamental flaws. The outstanding one is that she identifies only one risk to children from batterers after the couple separates: exposure to new acts of violence towards the mother. She thus fails to address multiple additional concerns. Batterers have rates that are at least 400% higher than non-batterers of abusing their children physically, sexually, and psychologically. These behaviors typically increase, not decrease, post-separation; when batterers feel control over their victims slipping away, their children become an important vehicle for continued intimidation and retaliation. Batterers also tend to be rigid, authoritarian parents who do not adequately nurture their children and who make them a low priority (except during custody or visitation disputes). Batterers also tend to batter serially, a fact frequently noted in clinical experience, so children also are at risk for exposure to violence by the father towards a new partner.

Johnston also overlooks the crucial risks to the mother-child relationship. Abuse is, by nature, undermining to the mother's authority and ability to nurture. In addition, batterers use deliberately divisive tactics, such as sowing divisions within the family and convincing the children that the mother is to blame for the violence. Children are traumatized by domestic violence and therefore need closeness and security with their mothers post-separation; to send them to a batterer for unsupervised visitation, knowing that he is likely to continue to sabotage the maternal relationship, is a decision that has to be made with caution. The quality of children's relationship with their primary caretaker is one of the strongest predictors of long-term well-being after divorce, even without the additional trauma of domestic violence.

Johnston treats a high level of physical violence as defining a batterer. However, it is common for batterers to be highly physically frightening and psychologically cruel with few incidents of actual physical assault; in fact, clinical experience teaches us that they are the majority of abusers. There are other well-documented and clinically-demonstrated indicators of a batterer's risk to his children besides the beatings; the incest-perpetrating batterer, for example, often shows relatively low levels of physical violence, combined with high levels of entitlement and manipulativeness.

Johnston's research is flawed methodologically. Her five types were, as she describes, created before the study; clinicians were instructed to assign each case to one of the types. She does not describe any scales that define the types; in fact, Johnston says that each case was assigned to types based on clinicians' judgement, with no description of any specialized training on domestic violence for these clinicians. The five types suffer both from lack of clarity in the definition (construct validity) and from lack of external validity, as we will see below. Finally, the published table undermines her claims that there are significant differences in outcome for children post-separation based on her types, indicating that it is especially untrue for girls; Johnston is reaching broad conclusions based on small sample sizes and sometimes contradicted by her own data.

Johnston seems to particularly underestimate the well-established risk of incest by batterers (see note 4). One striking section about daughters of batterers warrants extensive quotation:

Johnston shows no sign of recognizing this as possible incest, although it reads like a description from a training course on sexual abuse. It is not at all clear why she feels prepared to label girls who are this young as mutually seductive and provocative of his aggression, a stance that would give considerable comfort to both physical and sexual abusers of children. It is also not clear why she labels the girls "phobic," since she is simultaneously describing their fears of their fathers as realistic. This author was in the audience at a presentation of Johnston's in which she emphasized that there are many reasons why the stress of divorce can sexualize children and that we therefore should not consider this to be an indicator of sexual abuse, an idea she has also promoted in her writing. But how exactly has she determined that incest is not taking place in these other divorce cases to which she is referring? Her failure to identify incest in the above scenario raises questions about whether she is overlooking it in other cases as well.

Johnston also never explains how she distinguishes between "provocation" on a woman's part and the woman's legitimate efforts to defend herself or protect her rights or her children. This is an important gap because judgements about the provocativeness of victims is key to Johnston's formulation. Given that Johnston labels young girls as mutually seductive of their fathers and provoking of their aggression, she may not be the best person to entrust with the determination of which adult women are provoking their partners' violence. She in fact reveals noticeable errors in judgement in this area, as we will see.

Examining Johnston's Five Types

Johnston divides batterers into the following types, each of which calls for careful examination:

Type A: "Ongoing or Episodic Male Battering

Type B: "Female-Initiated Violence"

Type C: "Male Controlled Interactive Violence"

Type D: "Separation and Postdivorce Violence"

Type E: "Psychotic and Paranoid Reactions"

(These categories are called by slightly different names in the AFCC visitation risk assessment, but follow Johnston's formulation precisely in other respects.)

"Ongoing or Episodic Male Battering":

The first type appears to represent Johnston's notion of a real batterer; it involves violence at "life-threatening levels," and "most closely resembles the battering spouse/battered wife syndrome, which has been well described in the literature." The problem is internal to the man, and the women who are victims of this chronic battering "did not generally provoke, initiate, or escalate the physical abuse", although "a subgroup of them did not tolerate the abuse. Instead, they left the marital relationship early, soon after the abuse was first manifest. These were assertive women with high self-esteem and good reality testing." No other factors are mentioned as playing any role in determining which women leave and when they leave, such as economic obstacles, degree of manipulativeness of the batterer, terror, or law enforcement response in her community. Johnston thus appears to ignore almost the totality of domestic violence research.

Although Johnston is agreeing that batterers in this category are dangerous for unsupervised visitation, only a small percentage of domestic violence cases, even of the serious ones, will fit her restrictive criteria. Those batterers who use an escalating pattern of lethal threats but do not often attack, for example, are apparently not placed here.

In fact, the majority of batterers will not fit into any of Johnston's types, because they have low to moderate levels of physical violence on Johnston's scale, but are highly controlling, intimidating, and psychologically abusive.

Johnston claims as fact many questionable statements about batterers. She states that their behavior is a product of insecurity, low self-esteem, high vulnerability to humiliation, poor impulse control, and other emotional factors. Johnston mentions, but minimizes, the role that the entitled and domineering belief-systems of batterers play in their actions, thereby overlooking the great majority of clinical experience of specialists in men who batter.

"Female-Initiated Violence":

The second type involves women who always initiated the physical attack, and whose outbursts were repetitive during the marriage:

However, Johnston also states, "Few of these women did much physical damage with their violence: broken cups, torn clothing, and scratched faces were common." She does not state what percentage of the cases that were studied fit this category.

Johnston asserts that these women did not describe themselves as fearful. Such a case would be less complex for the court to deal with. In practice, courts are sometimes faced with cases where the woman has been the first to hit on many occasions, or even most, but continues to assert that she is physically afraid of her partner and that he has torn her down psychologically. A careful assessment needs to be made of which partner has had power over which aspects of the relationship, what the patterns of psychological abuse have looked like, and who has received the more serious physical injuries. Johnston's description could lead us to judge these cases too quickly.

"Male Controlled Interactive Violence":

Johnston's third category is the most problematic. This type involves violence said to be resulting from "mutual verbal provocation." Both sides hurl insults and yell, building towards a violent episode:

It is difficult to explain how Johnston manages to construe this scenario as being one of mutual abuse. Notice that a woman who attempts to leave or refuses to communicate is deemed to be engaged in mutual provocation of violence; thus almost any batterer could be assigned to this type. Johnston does not explain why she has accepted at face value the abuser's account that he slapped his partner in order to calm her hysteria, rather than to control or intimidate her; this is an excuse familiar to any batterer counselor, and one that breaks down quickly under scrutiny. The fact that the batterer sees the woman as a child to be disciplined and controlled (with violence) is not enough, for some reason, to qualify him as a batterer who is focused on power and control (which would put him in Type A); neither is the fact that "the exercise of physical control was seen as legitimate by the male." These are difficult contradictions to account for.

Johnston also has overlooked an apparent risk to his children: If the reason he hits her is because he sees her as a child to be disciplined and controlled, isn't he quite likely to hit his children under circumstances when they need to be disciplined or controlled,or when he feels that they do?

In this category, the woman is deemed to be equally responsible for the violence; she only becomes the victim because she loses the fight. This conclusion does not follow from the description, however; in Johnston's own words, the men were being violent in order to gain "their goal of compliance," while the wives, to varying degrees "resisted their efforts at control." The fact that Johnston defines this as a condition of equal responsibility leaves no way for an abusive woman to resist control or protect her rights (or her children's). This failing in turn raises questions about Johnston's judgement in characterizing the cases she studied.

The description continues:

It is not clear how the determination was made that the woman in this case was unafraid; battered women, like other people, may fight fiercely in self-defense, or in defense of their children, while terrified; in fact, they may fight precisely because of how frightened they are of what may happen if they fail to protect themselves. Johnston appears to be uninformed regarding the typical responses to trauma described in the literature in that field.

As the above points demonstrate, "mutual verbal provocation" is a perilous construct, lending itself easily to blaming the victim for her own trauma. Johnston never defines the term, which increases its dangerousness. If a woman says angrily to her partner that he never helps with the children, is she provoking violence? Suppose she yells at him that she is sick of being called names in front of the children? Does a woman invite violence when she swears at her husband for giving her a sexually transmitted disease? These are all common scenarios in the lives of abused women. How "assertive" or "feisty" is a battered woman permitted to be in attempting to stand up for herself and for her children (as even the most terrorized victim will periodically do)? Johnston does not explain how provocation to violence is distinguished from legitimate efforts to stand up for one's rights or to defend one's children; thus the choice between this type and the true "battering" type seems to come down to the clinicians' personal feelings about the parties. This risky way of proceeding is likely to play out in the court process if judges and custody evaluators attempt to use this typology.

Continuing:

But which person became sexually excited, and how was this measured? And how common is she suggesting that this is? (By choosing to mention it Johnston is implying that a significant number of cases are involved.) Some batterers find violence sexually exciting (and they may be among the most dangerous), and many batterers insist on having sex after violence to reassert their control. The woman is not well able to decline sex shortly after a beating, because of the danger and because of her trauma; thus sex following violence can never be considered consensual. This section would have to be termed irresponsible on Johnston's part.

Johnston assumes that children are relatively safe in unsupervised visitation with batterers from this category, since they are seen as not being focused on power and control. But in reality, an abuser's level of physical violence is just one indicator among many of his level of abusiveness and the pervasiveness of his control. His risk to abuse the children physically, sexually, or psychologically, or to expose them to years of sustained animosity towards their mother, is just as dependent on certain other factors as on his level of violence.

This category encourages us to determine the dangerousness of an abuser by assessing the degree of victimization of the battered woman. This leads courts to underestimate the danger to the children of those women who have fought back the most, who have experienced the most support and recovery, or who have taken the greatest steps to protect their children. No method of calculating risk could be more mistaken.

"Mutual verbal provocation" is also a dangerous construct. Is complaining to your partner that he never helps with the children provocation to violence? How about telling him that you are sick of being called names in front of the children? Is it provocative to tell him that he is blaming you for things that are his own fault? Does a woman invite violence when she swears at her husband for giving her a sexually transmitted disease? An abuser doesn't distinguish well between aggression and self-defense, so when his partner stands up for herself he defines those acts as attacks on him. (The same is true of child abusers, by the way.) Johnston's own descriptions portray women who are making various kinds of efforts to escape their partners' domination, yet Johnston holds them equally responsible for the violence that occurs. Analyzing a particular confrontation without examination of the underlying dynamics of the relationship can lead to serious errors in assessment.

If this category had any validity, studies of battered women would find that they are more likely to have aggressive or hostile personality types than non-battered women. But Hotaling and Sugarman's comprehensive review of research on domestic violence found that the woman's personality type had no statistical impact on her likelihood to be abused. In other words, the increased hostility and anger that are often observed in battered women are the result, not the cause, of her abuse.

This category (and each of Johnston's other categories) assumes that the only danger to children from unsupervised visitation is the risk of exposure to continued violence between the parents. This is a tremendous oversight, ignoring other risks that are actually far more pervasive. Anywhere from one-half to two-thirds of batterers also physically abuse their children. Batterers are frequently sexually abusive as well; one study found daughters of batterers six times as likely to be incest victims as girls growing up with non-batterers, and another study indicates that a batterer is three to four times more likely than a non-batterer to perpetrate child sexual abuse. The profile of the batterer is, in fact, very similar to the profile of the incest perpetrator, making the overlap a predictable one. Finally, batterers engage in various forms of psychological abuse of their children during unsupervised visitation that cannot be predicted by their level of violence; their level of overall abusiveness in the relationship has to be measured in order to make a good prediction of what their future behavior will be.

Johnston has, through the creation of this category, re-created the most obsolete notions of victims of domestic violence being to blame for their abuse because of how "provocative" they are, an idea taken right from the mouths of the abusers themselves. And even if such mutually abusive relationships did exist, she gives no supporting evidence for her claims that the children are in less danger from unsupervised visitation in these cases.

The woman who has taken the greatest steps to protect herself and her children is the one who, ironically, is most likely to be held mutually responsible by this formulation. As common sense might dictate, the risk from a batterer can only be assessed by looking at him, not by looking at his victim.

"Separation and Postdivorce Violence":

This group is said to become violent only as a result of the stress of divorce:

As in the mutual provocation category, the children are assumed not to be in great danger once tensions cool off.

Several problems are present. First, the separation may have resulted precisely from an escalation in the violence. Second, separation makes abusers irate, because they do not believe that their partners have a right to leave them; much of the literature on lethality assessment with batterers indicates that the abuser who refuses to accept the end of the relationship is particularly dangerous. Third, abusers believe that all of their needs should come first, due to their overarching sense of entitlement; since separation is a time when important issues are in dispute regarding finances, visitation schedules, and other matters, an abuser may escalate to physical violence at precisely this time. Johnston is implying that there is no practical difference between a person who becomes violent under the stress of divorce and one who does not; this is an especially risky stance to take, given that she includes "serious" violence and "several incidents" within this type.

One notable section says:

But certain behaviors around separation can serve to crystallize the victim's view of her abuser; his behavior may demonstrate openly what has been operating in more manipulative and confusing ways all along. This category is thus a dangerous one for the many women who have not understood previously that they were being abused, due to confusion or to self-blame, and for those who have not disclosed the abuse to other people. Johnston's description encourages the court to conclude that the mother's failure to name the abuse earlier shows that she has rewritten history out of vindictiveness, and to treat the difficulty as one of mutual mistrust. Thus the necessary investigation of the possible history of psychological or physical abuse could be overlooked.

"Psychotic and Paranoid Reactions":

The fifth category addresses violence that comes from psychotic or paranoid reactions due to mental illness or "drug-induced dementia." Psychiatric treatment is recommended as the preferred intervention. However, the fact that a person who batters also has a mental health problem does not prove that the latter causes the former; there is no mental health condition that causes the peculiar constellation of attitudes and behaviors common to domestic abusers. Treating the mental health problem will thus not dependably prevent a recurrence of the battering. The risk is that the court will believe that treating the mental health problem will prevent a recurrence of the battering; psychotherapy rarely reduces abusive behavior and often worsens it. A batterer with mental health problems requires an approach similar to the one needed for the substance abusing batterer; the person has two separate but interrelated problems, both of which need to be addressed.

This discussion includes a concerning case example, where Johnston dismisses a woman's report of terrifying violence because the man appeared in the clinical office to be "shy, quiet, [and] nonassertive." Again, Johnston seems to be failing to draw from the well-established information about how batterers present in public, including some of the most dangerous.

Additional Dangers in Applying This Typology

Johnston states that she considers her work preliminary and exploratory, but unfortunately she simultaneously recommends that custody and visitation arrangements be based on these types. Johnston also appears to speak publicly about these categories as if they were established facts; this author was present at a panel on which Johnston stated that the fact that a mother was knocked to the ground by her partner while holding an infant, causing serious injury to the child, did not eliminate the possibility that the relationship was one of mutual abuse. Johnston's work may, in the aggregate, be contributing to the danger of the children of battered women, and of the women themselves.

In addition to the profound flaws in the types Johnston creates, there are other huge problems in this study. The most glaring is that Johnston has reached conclusions about which types of batterers are least likely to commit repeat acts of violence, without providing a single statistic from her study to support that claim. Second, the great majority of batterers do not fit any of the types, because they exert "chronic pervasive control" but it is not accompanied by the most severe or frequent violence. The most common batterer is one who uses violence a few times a year or less, whose partner has never been hospitalized with injuries, and who shows no evidence of sadism. Nevertheless, his partner and children are exhibiting trauma symptoms due to their fear of the abuser, the repeated denial of their basic rights, and the pattern of psychological attack. Assessing the risk to these children from unsupervised visitation will be a complex process, and attempting to base such an evaluation on Johnston's poorly constructed and sloppily interpreted research is highly dangerous to abused women and their children.

While it is true that batterers are not all equally dangerous to their children, a useful visitation risk assessment has to focus on evaluating the perpetrator of the violence, not the victim, and needs to look at a far greater number of factors. (Effective visitation plans also have to hold perpetrators of violence accountable for their actions, something that Johnston does not appear ready yet to do.)

THE ASSOCIATION OF FAMILY AND CONCILIATION COURTS

VISITATION RISK ASSESSMENT

The AFCC risk assessment improves on Johnston's work somewhat by including a list of questions for evaluating the level of psychological and economic coercion that a batterer is exercising. However, in other ways this tool relies almost entirely on these five highly-flawed and victim-blaming categories. No battered women, battered women's specialists, or specialists in counseling batterers were included in the task force that drew up the assessment tool, which may help to explain why it draws a picture so divorced from reality. The risk assessment tool does open up some questions that have not previously been examined by courts in making visitation decisions, which could be a step forward, but only if a much wider set of factors is drawn into consideration.

Assessment of risk to visitation from batterers cannot be carried out by examining the victim, or by categorizing the nature of the relationship. Creating simple formulas for separating batterers who are dangerous to their children from those who are not is similarly impossible. Only by careful evaluation of the abuser's behavioral history, attitudes, and openness to change, can a visitation plan be constructed that will keep the children and the non-abusive parent safe, and that will provide the abuser incentives to develop a responsible and non-abusive parenting style.

NOTES

[For many additional sources, see Bancroft, L., & Silverman, J. (2002). The Batterer as Parent: Addressing the Impact of Domestic Violence on Family Dynamics. THousand Oaks, CA: Sage Publications.]

Back to Articles, HOME