Double Binds Facing Mothers in Abusive Families: Social Support Systems, Custody Outcomes, and Liability for the Acts of Others

. . . Although I agree that a very high standard of maternal responsibility is appropriate [in child abuse cases], there is little point in merely holding mothers responsible under high standards. Mothers in families in which husbands or boyfriends are abusive face many obstacles to protecting their children, a number of which are within the legal system itself and many of which could be eliminated or softened if taxpayers and legislatures were willing to spend resources to make exit possible and safe for mothers and children. I focus on obstacles facing mothers in abusive families, with an eye toward identifying places in which the situation of both women and children could be improved by giving more power to women so that they are more likely to protect their children effectively. I argue that, in addition to holding mothers responsible, any sincere commitment to the well-being of children must include making it easier for mothers to escape abusive households.

I therefore discuss the double binds facing mothers in abusive families. Although a mother may be found criminally or civilly liable for injuries caused by an abusive man in the household, she is offered little in the way of social support necessary to escape with their children. Part of the problem is the legal system itself. In divorce cases, judges often dismiss evidence of a husband’s violence against or abuse of their children as fabricated by the mother to give her an edge in the divorce negotiations. Sometimes, to be sure, there are conflicts between the interests of these mothers and their children. But often their needs overlap, and in this Article I try to identify some such areas.

In Section I, I discuss the relevant empirical evidence on why women stay and how children are hurt by living in violent or abusive households. In Section II, I discuss a mother’s legal liability for abusive acts of others. In Section III, I identify double binds facing women in abusive households in light of the combination of legal liability and the difficulty of protecting their children effectively through separation. Finally, I suggest legal and social reforms that would improve the chances of children escaping abusive households.

Corporal Punishment by Parents: Implications for Primary Prevention of Assaults on Spouses and Children

The National Family Violence Survey shows that each year more than 3.4 million couples experience physical violence involving severe assaults such as punching, kicking, or biting and that an additional 5.1 million couples are involved in less serious violence such as slapping or throwing things at a partners. They also show that each year a minimum of 1.7 million children are severely assaulted by their parents and that an additional 5.4 million children are hit with objects. Given numbers of this magnitude, it is unlikely that sufficient resources can be allocated to deal with family violence through intervention. Family violence alone could require almost the entire current U.S. expenditure on the criminal justice and social service systems. Thus, it is important to focus at least part of the effort to end family violence on primary prevention.

The concept of primary prevention is borrowed from the fields of public health and mental health. To paraphrase a definition from Caplan, primary prevention lowers the incidence of family violence by counteracting harmful circumstances before they have a chance to produce violence. Primary prevention does not seek to prevent a specific person from committing a violent act; instead, it seeks to reduce the risk for a whole population. The outcome envisioned as a result of primary prevention is that although some individuals may continue to be violent, their numbers will be reduced.

In this Article, we discuss the results of two studies on corporal punishment. The first study tests the hypothesis that corporal punishment of an adolescent increases the probability that the adolescent will physically assault a spouse later in life. The second study tests the hypothesis that corporal punishment of an adolescent increases the probability that the adolescent will physically abuse a child later in life. Both studies are relevant to a discussion of primary prevention because they investigate a circumstance that, according to the hypothesis, causes intra-family violence and because they are based on community samples rather than clinical samples. A clinical population is obviously essential for research intended to evaluate the effects of a treatment method. However, for intervention to prevent family violence in the first place, the most appropriate sample is a representative cross-section of the community in which the prevention steps are to take place.

Who Speaks for the Child?

Child advocacy is a recent phenomenon. Fifty years ago there was little concern about who would speak for children in legal proceedings that affected them. Not until the 1967 decision in In Re Gault were children guaranteed the right to an attorney in delinquency proceedings. A few years later, Congress passed the Child Abuse Prevention and Treatment Act (CAPTA), which, in part, provided federal financial assistance to state child protective services agencies so long as the state enacted legislation ensuring that every child involved in a child welfare proceeding had a court-appointed guardian ad litem. While neither Gault nor CAPTA has resulted in representation for all children in juvenile delinquency or dependency cases, they at least established the goal of full representation for all jurisdictions to follow. In other legal settings, the law has been even slower to provide for the representation of children.

The dramatic rise in child abuse and neglect reports over the past twenty years has led many persons within the legal system to examine how best to represent children. A growing consensus has rejected the traditional assumption that persons involved in legal proceedings will look out for the interests of the child. Even when a child’s interests are at stake, the other participants in the proceeding cannot be counted on to speak for the child. Parents and those who represent them have their own perspectives to present, while the court, with all of its other legal responsibilities, cannot be expected to focus upon the needs of the child. Moreover, child advocacy is needed both within and outside of the classroom. Many now recognize that the legal system was not designed for children and that a child may be traumatized by the very system designed to provide protection.

This Article will address the expanding role of child advocates in legal proceedings. We define a child advocate as a person who speaks on behalf of a child in a legal proceeding, whether that person be an attorney, a guardian ad litem, a hired staff member, or a volunteer. We then discuss the different types of advocates and their roles in legal proceedings. We next review the role of child advocates in a number of different legal settings, including juvenile, criminal, and domestic relations proceedings. Finally, we describe some efforts to improve the quality of child advocacy in local communities and across the country.

Dangerous Decisions: An Essay on the Mathematics of Clinical Violence Prevention and Involuntary Hospitalization

That this Article is on the mathematics of violence predictions and involuntary hospitalization decisions is being prepared for a symposium on "Domestic Violence, Child Abuse, and the Law" requires that I begin by providing readers with an explanation of the connection between what may initially appear to be quite different subjects.

Scientific studies and evaluations of mental health professionals’ predictions of violence typically have focused on the incidence of violence, the characteristics of perpetrators, and the ability of clinicians to distinguish those persons who will be violent from those who will not. In the vast majority of these reports, the identity of the victims of violence is not mentioned. But the relatively few studies that focus on the targets of psychiatric patients’ assaults suggest that family members are the most at risk. One study of three hundred patients admitted to a short-term psychiatric ward found forty-six (fifteen percent) patients assaulted another person in the two weeks before admission, that over half of these forty six patients had assaulted family members, and "that even after assaulting a family member, the majority of patients planned to return to live with their families after discharge." A recent review of studies about clinical violence prediction found similar rates of recent violence among newly admitted psychiatric patients. Thus, decisions about admissions to psychiatric hospitals frequently have important and immediate implications for the bodily safety (as well as the feelings) of the family members with whom patients live.

Especially in public-funded psychiatric facilities, a large fraction of patients are admitted involuntarily in accordance with state civil commitment statutes, which anticipate that mental health professionals will make decisions about initiating involuntary psychiatric hospitalization based on judgments about the imminent risk that their patients pose to others. The inadequacy of clinical predictions of violence has been the subject of extensive literature that consistently criticizes physicians’ ability to gauge individuals’ long-term future dangerousness. Evaluations of this criticism have noted that emergency commitment proceedings are not concerned with long-term dangerousness, but with behavior that will occur over a relatively short time following the prediction. As part of the evolution of "a second generation of thought on violence prediction," recent studies report that involuntarily hospitalized patients have a high degree of violence immediately before and after hospitalization and that professional judgments about these patients’ dangerousness have "a relatively high degree of short-term predictive validity."

The accuracy of violence prediction may be less important to practicing clinicians than a closely related but distinct question: when should mental health professionals predict violence? Two decades after the initial Tarasoff decision, psychotherapists in most jurisdictions can anticipate being held accountable for not having warned victims about, or taken measures to prevent, behavior of patients whose violence should have been anticipated. Even where their liability has been limited through legislative initiatives that define adequate responses to potentially dangerous patients, clinicians are still expected to take action when the potential for violence, as indicated by some combination of their patients’ acts and thoughts, reaches a threshold of "foreseeability." As Dr. Paul Appelbaum has noted, decisions following Tarasoff have emphasized that a duty to protect arises "only when a threshold of probability is crossed, . . . [but] the terms used to define that threshold have varied, and never has it been specified with any precision."

Although professional standards, and not prediction accuracy, provide the criteria by which Tarasoff liability is often judged, these standards ideally should reflect actual features of professionals’ predictions. If clinicians are to be faulted for the violent acts of their patients, then evaluations of their prediction-based decisions should reflect our knowledge about the intrinsic characteristics and accuracy of those predictions coupled with our judgments about the moral and societal interests affected by those decisions.

This Article has two major purposes. First, it provides a mathematical description of an ideal procedure for making clinical decisions about patients’ future violence, a description that provides a context for evaluating clinicians’ "dangerousness decisions." For purposes of illustration, the Article uses a specific clinical situation – deciding whether to hospitalize involuntarily a patient based on his risk of harming another. The Article argues that the decision involves balancing potential risks to third parties (often the patient’s family members) with the "massive deprivation of liberty" and other potential harms to the patient that could result from confinement. The mathematical description of the decision procedure consists of a comprehensive method for describing the accuracy of predictions or prediction instruments, a method for assigning values to correct and incorrect predictions, a method for adjusting predictions based on those values, and, most importantly, an explicit means for expressing uncertainty in those values.

Second, the Article evaluates the actual impact of uncertainty on an ideal decision procedure. When we combine our uncertainty about moral valuations of right and wrong decisions, our uncertainty about base rates, and our uncertainty about the relevant time periods over which predictions should apply, what results is an uncertainty about the correctness of prediction-based decisions that makes most criticism of those decisions untenable. This Article shows that our uncertainty about the factors intrinsic to a hypothetical, best-case prediction procedure usually would preclude valid post hoc criticism of wrong decisions about dangerousness; a fortiori, most real-life prediction errors also should be beyond criticism.

Demonstrating these points will require me to make explicit and (I hope) non-controversial assumptions about a clinical decision process that is often governed by implicit assumptions or unconscious heuristics. This Article’s assumptions include an open recognition that prediction mistakes are inevitable. In describing a mechanism for equitably balancing the negative consequences of prediction errors, this Article takes the viewpoint that such balancing should reflect a decision-making strategy governed by public attitudes about suffering violence and suffering involuntary hospitalization. Because they are explicit, this Article’s assumptions can be evaluated for their normative value or their agreement with empirical findings.

Custodial Choices for Children at Risk: Bias, Sequentiality, and the Law

Children are at risk for many reasons. As Professor Straus has reported to the Symposium, most adults in the United States accept the degrading violence of corporal punishment, with its dangerous potential for escalation, as a routine aspect of discipline and socialization. Overwhelming numbers of children live in poverty. Parental resources are sorely strained by unemployment and disability. The parental role is devalued and often assumed without preparation or planning. Alarming numbers of parents are impaired by addiction or mental illness. These problems are exacerbated by the absence of universal health care and the absence of an effective social safety net. Moreover, some adults whom children encounter are sexually irresponsible or simply cruel. Government has a legitimate obligation, if not a constitutional duty, to protect the youngest members of the national community against neglect and abuse.

Nonetheless, most children thrive in parental care and suffer harm if that care is significantly interrupted. Moreover, the liberty, personal autonomy, and diversity that are the human centerpiece of democracy in the United States are compromised whenever government intrudes excessively in the lives of families. Paternalistic efforts of classical republicanism to develop ideal citizens and statist measures like the Chinese government’s one-child policy, or the French government’s announced intention to forbid post-menopausal in-vitro fertilization, strike the American sensibility as excessive governance. The human capacity to make moral and social meaning implies a human right to be socialized to moral autonomy in the intimate and relatively flexible context of family rather than molded to norms imposed by an impersonal, homogenizing, and all-powerful state . . . It is "cardinal" in the United States that the "care and nature of the child reside first in the parents, whose primary function and freedom include preparation for obligations the state can neither supply nor hinder."

When the state assumes a child-protective function, it takes on the exquisitely difficult task of deciding when intervention is reasonably necessary to the physical or emotional well-being of a child and when it is destructive, both of the bonds upon which the child depends for healthy nurturance and of the child’s right to grow in a community that is open, flexible, and self-defining, rather than state-controlled. Every formal charge of child neglect or abuse presents this difficulty in vivid microcosm . . .

Our critique of decisionmaking regarding children at risk is more charitable than that of critics discussed [in omitted introductory materials]. We do not imagine decisionmakers as thoughtless libertarians or as arrogant paternalists. Rather, we imagine that subtle, systemic factors are responsible for at least some of the costly errors that at times leave children to face unacceptable risks and at other times impose upon children the trauma of unnecessary family disruption.

In the hope of identifying error-producing factors in child-welfare proceedings, we create a simplified model of judicial decisionmaking. After describing our simplified decision model, we identify a "sequentiality effect" that is present when, as in most child protective proceedings, the ultimate question before the decisionmaker is anticipated in one or more preliminary or pre-trial proceedings. We also identify four factors that are likely to produce bias in child protective decisionmaking: the perceived status quo at the time of custodial choice; a heightened emphasis upon risks associated with decisions in favor of the party (usually the respondent) with fewer litigation resources; the fact that the litigation is understandably and inevitably focused upon the possibility that the respondent has caused harm to the child; and the judge’s special vulnerability to negative feedback in the event of adverse consequences from a failure to intervene. We then demonstrate the ways in which the sequentiality effect interacts with bias factors to compound errors made in the early phases of decisionmaking. Finally, we suggest measures that might be taken to guard against systemic bias and against the reinforcement of error by the sequentiality effect.