I can think of no better way to make my point-that we simply do not know what the connection is between intimate partner violence (IPV) and child abuse-than by quoting from Appel and Holden's review of the empirical literature on this issue. They come to the following conclusions from a review of 31 studies:

• "In clinical samples of either battered women or physically abused children, the percentage of overlap ranged from 20% to 100%."
• "The base rate of co-occurrence found in representative community samples was about 6%."

One hundred percent overlap? Twenty percent? Six percent? What is wrong with this literature? A major part of the problem is that IPV is not a unitary phenomenon and the different types of IPV are not connected in the same way to child abuse. Furthermore, different samples have different mixes of the major types of IPV, and therefore get wildly varying estimates of the overlap with child abuse.

In my work I distinguish among three major types of IPV. In intimate terrorism, the perpetrator uses violence in the service of general control over his or her partner. (I will be using gender neutral terminology throughout this piece because both men and women are involved in all three major types of IPV. However, in heterosexual relationships intimate terrorism is perpetrated primarily by men.) In violent resistance, it is the partner who is violent and controlling. In other words, the partner is an intimate terrorist, and the resistor's violence arises in reaction to that attempt to exert general control. These two major types of IPV show up mostly in samples from agencies such as hospitals, courts, or shelters. In the third major type of IPV, situational couple violence, the perpetrator is violent (and his or her partner may be as well), but neither of them is using violence to attempt to exert general control. The violence is a product of conflicts that turn into disagreements that escalate to arguments, to verbal abuse, and ultimately to violence. Situational couple violence shows up mostly in general survey samples.

The Appel and Holden review is full of so many complexities that it simply cannot tell a coherent story of the connection between IPV and child abuse. So, in some sense we don't know what the connection between the two is. We need more research and we need it to make distinctions among types of IPV. However, there are substantial clues to the nature of the link. To make that point, I am going to simplify matters by offering three numbers that I have derived from the studies covered by the review. My goal is to provide a rough idea of the likely increase in the risk of child abuse associated with different types of IPV. Data from the 1985 National Family Violence Survey indicate an abuse rate of 15% among children in families in which their parents had not assaulted each other in the previous year. Keeping in mind that violence data in surveys are heavily dominated by situational couple violence, the abuse figure in that same survey for children whose parents were involved in situational couple violence was 31%. Another study, using almost identical measures in interviews with 184 shelter residents (a sample highly likely to be dominated by intimate terrorism) found an abuse rate of 67%. So, what we have is a 15% risk in non-IPV families, double that risk (31%) for couples involved in situational couple violence, and over four times that risk (67%) for intimate terrorism.

What Are the Likely Connections?

Now I'm going to move into "informed guesswork" mode (call it theory if you wish) about what produces these different connections between IPV and child abuse. Why is it that the likelihood of such abuse increases with situational couple violence but only to about one out of three? Well, there are different kinds of situational couple violence. For probably half of couples so identified, the violence involves only one incident in the previous twelve months. There is not much reason to expect their children to be at any more risk of child abuse that those whose parents have never been violent toward each other.

Even the chronically situationally violent are something of a mixed bag. For some, the violence comes from chronic disagreements that escalate to violence primarily due to problems of communication between the parents. If the disagreements and communication deficits are really adult-adult problems, their children may be at no increased risk of abuse. There are, however, sources of chronic situational couple violence that may put the children at considerable risk, including anger management problems and substance abuse. If the source of the couple's violence is that one or both of them frequently turns to violence when he or she gets angry, then children (most of whom at one time or another do anger their parents) may be at considerable risk of abuse. For some adults, this problem with anger and violence is related to substance abuse; for others not. In either case, this is probably the type of situational couple violence that puts children most at risk of abuse.

What about the 67% risk of child abuse in families in which there is intimate terrorism? The risk is high because the violence is rooted in the need to control, and children are often a threat to that control. But even here, there is some important variability. A number of scholars have identified two major sources of the intimate terrorist's need to control his partner. One source is an extreme emotional dependence on one's partner, a desperate need to hang on to him or her. I would venture that a need to control that is rooted in this sort of emotional dependence is not likely to generalize to children. The second source of the need to control, however, is more general, flowing from a sociopathic attitude of "I will have things my way and I will do whatever I need to do to get my way." This type of intimate terrorist has been shown to be likely to be violent outside of the family in addition to inside it, and I would venture that we would also find a high incidence of violence toward children among this type of intimate terrorist.

There is one more thing I want to say about intimate terrorism. Sometimes we can get so focused on abuse as violence that we forget the incredible power of emotional abuse. The level of control sought by intimate terrorists is likely to create an atmosphere in the home that will terrorize the children as well as the partner, even if the controlling parent is never violent toward the children. And that terror may be heightened because even if the children are never physically abused themselves, they are highly likely to have witnessed the violence of intimate terrorism, or even to have been used as a tool in the coercive controlling strategies of the intimate terrorist parent. (If you have the stomach for it, you can find a video of such use of a child in the following YouTube excerpt from a television documentary.)

Witnessing Violence

Whenever violence between parents is chronic, whether it be situational couple violence or intimate terrorism, the children are likely to be witnesses. Imagine the emotional impact on a small child of seeing violence enacted against one of his or her parents; the short-term psychological effects are dramatic, including behavioral and emotional problems, social dysfunction, and cognitive delays. There is also evidence of long-term effects, including stress, depression, aggression, and substance abuse. Thus, intervention for the sake of the children is called for even if the violent parents are "only" abusive toward each other.

What to Do About It?

The link between IPV and risk of either direct child abuse or at least the witnessing of violence suggests that intervention in partner abuse itself is likely to benefit children. The nature of that intervention must take into account the type of IPV involved, but it will also be constrained by the circumstances in which the problem is encountered (individual or family counseling, criminal courts, or family courts). In all cases, the victim of the violence should be made aware of local support services, usually "women's shelters" or "women's resource centers." Although some such agencies serve only women, many offer support to male victims as well.

Let me just briefly address some of the issues involved in interventions in individual or family counseling and the criminal courts that focus on the source of the violence, often with the goal of eliminating the violence while the partners maintain their relationship. There are three major types of education/counseling programs for perpetrators of IPV: individual counseling, couples counseling, and batterer intervention programs-and I will add victim support groups as another alternative. For some cases of situational couple violence, those in which the source of the violence is substance abuse and/or the perpetrator's inability to handle anger without violence, individual counseling is an appropriate response. For situational couple violence in which communication issues are paramount, couples counseling may also help. It is extremely important in such cases, however, to be absolutely certain that you are not dealing with intimate terrorism. Couples counseling can greatly increase the risk to the victim of intimate terrorism, as the couples counseling itself threatens and angers the intimate terrorist. In the criminal courts, where most of the cases do involve intimate terrorism, batterers are typically mandated to batterer intervention programs that focus primarily on issues of coercive control, with some attention to communication and anger management issues.

If an individual or couples counselor suspects intimate terrorism, the victim should be approached privately and guided to victim support services. If it is safe for the victim, the intimate terrorist can be made aware of batterer intervention programs in the area, but in most cases of intimate terrorism, batterers do not enter such programs voluntarily. In cases of violent resistance as well, the best option is a victim support group. One recent study found that batterer intervention groups for women offenders often function much like victim support groups, because the facilitators recognize that many of the women arrested for domestic violence were involved in violent resistance.

Family court interventions most often have to do with custody and visitation arrangements after a violent relationship has ended. The courts have to balance parental rights against concerns for the safety of victim of the IPV and the safety of the children. In cases of intimate terrorism, the breakup of the relationship is probably the most dangerous time the victim has ever faced. Most intimate terrorist homicides take place either while the victim is trying to escape from the relationship or after it has ended. In cases of situational couple violence, the risks may be less, but given the great variability in this type of violence, it cannot be assumed that there is no risk. And in family court there will also be a few cases in which, although there had been no abuse or violence in the relationship prior to the breakup, one or more of the partners has turned violent in response to the breakup. Risks to the children are generally low in such so-called separation-instigated violence. In all cases, of course, court decisions must be made on a case-by-case basis.

The courts have developed some tools for assessing the dangerousness of such situations, with indicators including the characteristics of intimate terrorism, but also the extent to which there is a primary perpetrator, the level of violence involved, the presence of firearms, and so on. Many (but not all) jurisdictions have the resources available to allow finely differentiated custody and visitation arrangements that take into account the potential for further violence. For the less dangerous cases they include co-parenting, generally involving joint custody in which both parents are involved in making cooperative decisions about the child's welfare, and parallel parenting with both parents involved, but arrangements designed to minimize contact and conflict between the parents. In more risky cases, the arrangements might involve supervised exchanges of the child from parent to parent in a manner that minimizes the potential for parental conflict or violence, or supervised access, when one or both parents pose a temporary danger to the child, access provided under direct supervision in specialized centers and/or by trained personnel with the hope that the conditions that led to supervised access will be resolved and the parent can proceed to a more normal parent-child relationship. In the most serious cases, in which a parent poses an ongoing risk to the child, all contact with the child would be prohibited.

In Sum

Differentiation is essential. Although all intimate partner violence does damage to the children in the family who either witness it or who suffer direct abuse related to it, the risks to children and parents vary dramatically from one type of intimate partner violence to another, and interventions must be related to that variability.

A version of this essay complete with bibliography is available on the author's website. His research on types of violence is summarized for a general audience in Michael P. Johnson, A Typology of Domestic Violence: Intimate Terrorism, Violent Resistance, and Situational Couple Violence, Northeastern University Press, 2008.