Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
VIOLENCE IN FAMILIES 221 5 Violence in Families Recently, we have come to realize realize that our homes may be as dangerous as our streets. Family violence in its various formsâspouse assault, elder abuse, sibling violence, child abuseâis more prevalent than the public or officials ever suspected. The National Commission on the Causes and Prevention of Violence report issued 23 years ago does not mention spouse assault in its discussion on violent crime (Winslow, 1972). And although incest has long been considered a crime (more against the natural order than against the law), it has only recently been labeled a violent crime justifying the intervention of the courts. On the terrain of family violence, elder abuse is just being recognized as a phenomenon and sibling violence remains largely invisible. Child abuse has received the most attention, having been documented since the 1950s. However, in common with other forms of family violence, we know more about the barriers to accurately counting it than we do about its rate and prevalence. Long considered family matters rather than criminal matters, we are just beginning to probe the causes and consequences of family violence and to test the effectiveness of interventions designed to reduce its severity, frequency, and impact. CONCEPTIONS OF FAMILY AND FAMILY VIOLENCE In Chapter 2 we draw attention to the ways that changes in the social construction of violent behaviors pose difficulties for classifying
VIOLENCE IN FAMILIES 222 and counting violent crime. Nowhere in the criminal law and in its administration is the social construction of violent crime changing more rapidly than in what constitutes family violence and society's responses to it. Because family law is legislated and administered under different statutes and regulations in each state as well as by the federal government, there is no national legal definition of a family. Similarly, there is no standardized and generally accepted scholarly definition of what constitutes a family. Information on family violence is usually based on identifications of people by their current marital status (married, separated, divorced, or single), by their spouse status (spouse/ex-spouse; husband/wife), or by relationships among members of a household (e.g., cohabitants; child/parent; brother/sister; father/mother). Statistics about families are not collected according to theoretical constructs; rather researchers sample address locations and identify households; families are then designated within households either by respondents themselves or by interviewers. By this procedure, two or more families of several generations may reside within a household with a single head.1 Substantial recent changes in family structure in the United States, which may affect counts of violent behavior by "family" members, are disclosed by some contemporary statistics on marital status and living arrangements (Bureau of the Census, 1990, 1991):2 â¢ The proportion of all households accounted for by two-parent families declined from 40 percent in 1970 to 26 percent in 1990. â¢ The number of unmarried-couple households almost tripled between 1970 and 1980 and grew by 80 percent between 1980 and 1990, from 1.6 to 2.9 million. â¢ The proportion of children under 18 years of age living with two parents declined from 85 percent in 1970 to 73 percent in 1990, an estimated 15 percent of whom are stepchildren. â¢ In 1990, 19 percent of white, 62 percent of black, and 30 percent of Hispanic children under age 18 lived with only one parent. Trends in family violence must be interpreted against a decline in the fraction of households containing exclusively married couples and their natural children. Violence between growing numbers of same-sex and opposite-sex cohabiting partners is increasingly regarded as family violence regardless of legal marital status. Those who record statistics may or may not classify violence between increasing numbers of divorced or separated ex-couples as family violence.
VIOLENCE IN FAMILIES 223 Proportionally more of today's children live in households with a cohabiting couple or at least one stepparent. Legally, a cohabitant unrelated by blood or marriage who physically or sexually abuses children in the household is an unrelated person, although programmatically the abuser may be treated as a family member. Children today are far more likely to grow up in a female-headed household than heretofore, especially in black households. Parents today are more likely to have never married or divorced. In a purely statistical sense, a child growing up with only one parent has only half the risk of violent victimization by an adult family member; however, the tensions surrounding family breakup may increase the chance that an adult family member or cohabitant will act violently toward a child. Moreover, patterns of divorce, separation, remarriage, and cohabitation expose a fairly sizable proportion of children to several parents. Although there is no common agreement on when violent behaviors that occur within a household are to be regarded as family violence, the panel considered all violent behaviors within households made up of married or cohabiting persons and their relatives and minor children as domestic or family violence. Specifically, domestic (family) violence includes spouse assault (often referred to as marital violence), physical and sexual assaults of children, sibling assaults, and physical and sexual assaults of other relatives who reside in the household. These violent behaviors do not include all violent behavior that occurs within families. Among those excluded are nonconsenting sexual behavior among adult family membersâespecially incest and marital rapeâand consensual sexual behavior among spouses or family members that at law is regarded as violent behavior. Attempts have also been made to construct parental physical punishment of children as child abuse. Because so little is as yet known about the nature and extent of these behaviors in families, they are not discussed further in this chapter. DYNAMIC NATURE OF FAMILY VIOLENCE Family violence has a number of distinguishing features. Unlike victims and offenders who are strangers, victims and offenders in family violence cases have some form of continuing relationship such as spouse or partner, parent and child, or siblings.3 Interacting almost daily and sharing a common domicile increase the opportunities for violent encounters.4 Because they are bound together in a continuing relationship, it is quite likely that the victim will be violated repeatedly by the offender. Usually one
VIOLENCE IN FAMILIES 224 person in the continuing relationship is relatively powerless and more vulnerable to the aggression and violence of the more powerful one. The more powerful victimizer often threatens the victim with additional violence if the incidents of violence are disclosed to others, or the victim may refrain from disclosure anticipating stigmatization and denigration. Moreover, because many of these episodes occur in private places where they are invisible to others, they are less likely to be detected or reported to the police. For these reasons, family victimization continues for want of intervention by others. INSTITUTIONAL BARRIERS TO OBSERVING FAMILY VIOLENCE Major institutional barriers to understanding family violence are limited access to observe violent behavior in the private places where it occurs and the need to obtain consent to gather information on family matters, including violence. The problems of securing access and obtaining consent are especially acute for children and dependent family members who lack the competence to give consent or to report physical and sexual violence by members of their household. Data collectors are limited by the necessity to obtain consent from adults before collecting information about their own, another adult's, or their dependents' victimization. Law enforcement officers likewise must obtain consent for intrusion into private matters or have legal grounds for intervention or investigation in criminal mattersâgrounds that derive primarily from complaints by persons who have private knowledge of intrafamily violence. Institutions of privacy are not the only barriers to obtaining information about physical and sexual assaults of minors in the household. There are substantial social and psychological processes that lead family members to conceal intrafamily violence, especially sexual violence. Victims of spouse assault and of sexual violence are often threatened with additional violence if they disclose occurrences to the police. Threats based on financial dependency, uncertainty about their proper role, and family loyalty may all make it difficult to verify the victimization of minors by members of their own householdâ especially for sexual assault. The cognitive difficulties people have in separating occurrences of the same event create problems in counting repeated victimization, especially of the same kind of violence. In addition, self-blame and fear of the stigma associated with battering among
VIOLENCE IN FAMILIES 225 intimates create a climate for even greater underreporting. Frieze and Browne (1989:182), summarizing what is known about self-reports of those involved in spouse assault, conclude that women who have been victimized over a long period of time tend to underestimate the frequency and severity of the violence they experience when their reports are compared with hospital and other records or the reports of witnesses. Male batterers also underreport their violent actions and minimize their own responsibility for battering, projecting blame onto the battered spouse. Research is needed on whether and how self-reports of victims and offenders could be used to measure the prevalence of family violence (i.e., the proportion of the population that experienced at least one event during some period of time) and its incidence (i.e., estimates of the number or rate of events that occur within a given period, usually a year). Research is also needed on the extent to which valid information on physical and sexual violence can be obtained from family members, especially from self-reports of minors and from their adult victimizers.5 In episodes of intrafamily violence, it is sometimes difficult to determine the respective roles of victim and offender, since participants often appear to engage in mutual combat. Though the reports were not verified, Straus et al. (1980) found that half of their male and female respondents who reported violence said that both partners had been assaultive. The determination of victim and offender roles is therefore generally judged in terms of consequences of the combat, such as the severity of injuries or the dangerousness of the violent means used by each participant. Although comparable data are lacking for the severity of injury for men and women in domestic assault, the consequences of injury appear more severe for women. Stark and Flitcraft (1991) conclude that a battering may be the most common cause of injury among women seeking emergency department treatment. And men use more dangerous means in domestic assaults. Straus and Gelles (1986) found that men more often used the most injurious and dangerous forms of violence, such as using a gun or knife or physically assaulting the partner. PATTERNS AND TRENDS IN FAMILY VIOLENCE Because of the institutional barriers described above, as well as problems stemming from the design of our major national systems that collect and record information on violent behavior, it is not possible to estimate reliably the prevalence and incidence of
VIOLENCE IN FAMILIES 226 domestic violence in the United States. Below we present some crude estimates from currently available information, but there is good reason to conclude that intrafamily violence is substantially underreported. Beginning in 1988, the National Crime Survey (NCS) (Bureau of Justice Statistics, 1990) reports annual estimates of the extent of family violence for persons age 12 and older. Of the violent victimizations reported in 1989 in which the offender was a relative of the victim, 59 percent of the assaults were by a spouse (41%) or ex-spouse (18%). Other relatives accounted for another 29 percent and parents (7%) and children (5%) for the remaining offenders. The NCS lacks the necessary information, however, to reliably determine how much of the reported violence is classifiable as family violence. For example, in considering violence only among married spouses, it excludes violence among cohabitants. Likewise, the NCS does not collect information on children under age 12. The Uniform Crime Reports (UCR) system does not currently report family violence separately, except for information on the relationship of victim and offender in homicides and for some arrests. The new incident-based reporting system of UCR will have the capability of tabulating information on felony and misdemeanor domestic violence and on other assaults involving family members.6 The panel urges UCR to utilize the new system to provide information on the status of victims and offenders in all violent offenses so that we may better understand violence in terms of social relationships among victims and their victimizers.7 Information on other forms of intrafamily violence is even more problematic. National estimates of violence toward children are based largely on reports of cases of maltreatment known to investigatory bodies such as juvenile probation, coroners, and law enforcement agencies; to major health, education, and welfare organizations whose professional and semiprofessional workers have contacts with childrenâphysicians, nurses, teachers, mental health, public health, and social service workersâand who observe visible injuries;8 and to child protective service agencies that process referrals from these agencies and from citizen complaints.9 Most of the prevalence and incidence estimates for specific types of intrafamily violence come, however, either from special surveys or from samples of victims or offenders known to some agency. In this discussion we rely not only on these special studies and official reports but also on a commissioned review of research on
VIOLENCE IN FAMILIES 227 spouse abuse (Fagan and Browne, Volume 3) and previously published reviews by Frieze and Browne (1989) on violence in marriage, by Garbarino (1989) on child maltreatment, and by Pagelow (1989) on violence toward and among other family members. Marital Violence The main measure of domestic violence used in surveys, the Conflict Tactics Scale developed by Straus (1979), includes verbal and physically aggressive acts, including violent acts, ranging in severity from hurling objects to the use of a deadly weapon such as a gun or a knife. Using it in an initial national telephone survey of couples in 1975, and in a repeat of the survey in 1985 (Straus and Gelles, 1986), Straus et al. (1980) reported that 16 of every 100 couples reported at least one incident of physical aggression during the year before the survey. The prevalence of severe violence in both surveys was about a third of that; about 4 in 100 females and 5 in 100 males reported experiencing severe violence during the year. There is reason to conclude that these estimates may be somewhat low since the sample excludes unmarried couples and misses segments of the population that do not have phones. The national rate reported by Straus et al. is well below that found in local surveys of selected segments of the population. Our commissioned review indicates that there is considerable variation in the prevalence rate from local sample surveys (Fagan and Browne, Volume 3). However, almost all of the high estimates were obtained from select populations such as divorce filings, students, or batterers or from small samples. The larger samplesâover 1,000 women or couplesâhave rates around 10 in 100 reporting some form of family violence as measured by the Straus and Gelles scale. Assaults on Children National estimates of assaults on children, including physical and sexual abuse and physical neglect, are derived from a series of studies commissioned by governmental sponsors. The U.S. Advisory Board on Child Abuse and Neglect, a governmental advisory board created by the 1988 amendments to the Child Abuse and Prevention and Treatment Act, estimates that in 1989, at least 1,200 and perhaps as many as 5,000 children died as a result of
VIOLENCE IN FAMILIES 228 maltreatment, and over 160,000 children were seriously harmed (U.S. Department of Health and Human Services, 1990: 15).10 Another indicator of the severity of the problem of child maltreatment is the rapid escalation of reports of child maltreatment that are received by municipal, county, or state social service or child protection agencies. The advisory board has stated that in 1974 there were about 60,000 cases of child maltreatment reported. This figure rose to 1.1 million in 1980 and more than doubled to 2.4 million in 1988 (U.S. Department of Health and Human Services, 1988: x). Part of this increase can be attributed to more inclusive definitions of abuse and neglect resulting from broader public awareness of different forms of child maltreatment, and an increase in the likelihood that professionals will recognize maltreatment, rather than an increase in incidence per se (U.S. Department of Health and Human Services, 1988:xxv). However, it is also likely that many cases of child maltreatment are reported to public health or educational agencies that are not known to social service personnel who often supply the âcountable" case figures. There are also many cases of interfamilial or third-party assaults on children that are never reported to any professionals concerned with the health or welfare of children. Under the expanded definitions of child abuse and neglect, the National Incidence Survey published in 1988 (NIS-2) and commissioned by the National Center for Child Abuse and Neglect estimated that the majority of countable cases involved neglect (63%, involving 15.9 per thousand, or 1,003,600 children). Less than half (43%) of these cases involved abuse (10.7 per 1,000, or 675,000 children) (U.S. Department of Health and Human Services, 1988:3-4). Under the revised definitions, reports can be made on behalf of children who may be subject to demonstrable harm but who have not (yet) experienced any demonstrated injury or impairment. Using the revised definitions, NIS-2 estimated that physical neglect represented the "largest subcategory of countable neglect, representing 57% of all neglected children within these standards (or 571,600 children)" (U.S. Department of Health and Human Services, 1988:3-9). Educational neglect was the second most frequent subcategory (29% of the total neglect cases, or 292,100 children) and emotional neglect was the least frequent subcategory (22% of the total neglect cases, or 223,100 children) (U.S. Department of Health and Human Services, 1988: 3-8). Within the category of abuse, physical abuse is the most frequent subcategory, accounting for 358,300 children reported in the NIS-2 study (5.7 children per 1,000). Emotional abuse was the
VIOLENCE IN FAMILIES 229 next most frequently occurring subcategory of abuse (211,100 children, or 3.4 children per 1,000) and sexual abuse was the least frequent of all three major subcategories of abuse in the NIS-2 study (155,900 cases, involving 2.5 children per 1,000) (U.S. Department of Health and Human Services, 1988: 3-8). In comparing changes between the NIS-1 study (conducted in 1980) and the NIS-2 study conducted in 1986, there are significant increases in the incidence of both physical and sexual abuse. In particular, the incidence of countable sexual abuse more than tripled since 1980, an increase of high statistical significance (U.S. Department of Health and Human Services, 1988: 3-8). In contrast, however, one other national incidence survey conducted by Richard J. Gelles and Murray A. Straus, funded by the National Institute of Mental Health, reported a decrease in the self-reported incidence of physical abuse by parents between 1975 and 1985 (U.S. Department of Health and Human Services, 1988: xi). Their study involves information about patterns of violence reported by persons contacted through a telephone survey of two-parent families with children over 3 years of age. To what extent the reports reflect an actual decline as opposed to less candid responses by the parents is simply not known. Assaults Against the Elderly Surprisingly little is known about violence toward the elderly in families. A main reason for knowing so little is that previous research has not distinguished between elder abuse and elder neglect, terms that include different types of mistreatment of the elderly (Pedrick-Cornell and Gelles, 1982). Another is the small and selective data base. National victim surveys in the United States do not report the prevalence and incidence of domestic violence toward the elderly. Conclusions consequently are based either on surveys of local communities or assessments by professional caregivers. The more reliable of the surveys conducted in recent years conclude that domestic violence toward the elderly may be less than presumed from surveys of professionals. A stratified random sample of all persons 65 and older resident in households in the Boston metropolitan area (Pillemer and Finkelhor, 1988) estimated that between 2.5 and 3.9 percent of Boston elderly persons had experienced physical violence, verbal aggression, or neglect. Similar results, with some differences in definitions of abuse, were found for a national survey of elder abuse in Canada (Podnieks and Pillemer, 1989).
VIOLENCE IN FAMILIES 230 A recent review of domestic violence against the elderly (Pillemer and Frankel, 1988) concludes that only a few findings consistently emerge from these studies of the abused elderly population. Elderly victims are disproportionately over age 75. They are more vulnerable to victimization because of illness or impairment and usually reside with the perpetrator of abuse. Although most studies report that women are more at risk of elder abuse than men, they do not take into account women's greater risk exposure because of their longer life expectancy. Because of the broad definitions of elder abuse and the absence of information on its prevalence in family settings, little is known about the prevalence of different forms of violence toward the elderly or its consequences. Given this current state of knowledge, priority should be given to the collection of more precise information about the prevalence and incidence of violence toward the elderly and its consequences. Moreover, studies of the perpetrators of violence toward the elderly and a better understanding of the situations that give rise to their behavior are essential to develop and implement programs of violence prevention and intervention. Violence Among Related Family Members Although the available data are severely limited, the view is firmly held that a considerable amount of violence goes on amongst other members in a household (Pagelow, 1989). Of particular concern is violence of children toward parents and of siblings toward one another. Available estimates for these forms of violence vary substantially and are subject to considerable error. It is generally concluded that physical violence of children toward parents is less severe than is violence of parents toward children (Pagelow, 1989), but there are no precise estimates of its magnitude. Violence between children is regarded as the most common form of family violence. In the national survey by Straus et al. (1980), 82 percent of the parents of more than one child aged 3 to 17 reported sibling violence, the highest rate for all types of intrafamily violence. There are few studies of either physical or sexual violence among siblings, however. Risks of Intrafamily Violence Current reporting systems are not designed to estimate the magnitude of the risks of different types of family violence or to isolate
VIOLENCE IN FAMILIES 231 the principal risk groups. The highly aggregated nature of current classifications of violent offenses masks much of the variation important for understanding the risks involved. For example, the national reporting systems rarely use the available classifications of sexual abuse and sex crimes. When the sole classification is "sexual abuse," it is impossible to understand the nature of the sexual abuse of childrenânor, with such gross categorization, can one understand the differences in risk for male and female children. The classification of intrafamily violence is similarly crude. Classifications of homicide and of assault cloak considerable variation in types of family violence. Moreover, the failure to distinguish the relationships among persons involved in family violenceâsuch as cohabitants of the same or opposite sex, married, separatedâlimits understanding. Recognizing these limitations, the panel recommends that efforts should be made to develop classifications of intrafamily violence that lend themselves to understanding risks of victimization and offending and provide a basis for intervention and control. Assault According to the National Crime Survey, 9 of every 10 violent victimizations by family members or relatives of the victim are assaults (Bureau of Justice Statistics, 1990). Women, however, experience a higher rate of both simple and aggravated assault by relatives than do men. It is generally assumed that this greater vulnerability of women to assault by relatives is due to their greater vulnerability to assault by a spouseâmen are victims in only about 1 in 10 misdemeanor spouse assaults. Yet it may also be a function of women's greater vulnerability to violence by male partners or cohabitants. Unfortunately, the data available to the panel did not permit exploration of the general vulnerability of women to all forms of male family violence, although they are consistent with an explanation of general vulnerability. Although women are more vulnerable to simple and aggravated assault by relatives than are men, women and men are both more vulnerable to assaults by nonrelatives well known to them and by strangers (Table 5-1). Men are not only more vulnerable to assault by strangers than are womenâthey are also more vulnerable to assault by strangers than by nonrelatives well known to them. These comparisons do not take into account presumed differences in reporting of assaults by relationship to the victim.
TABLE 5-1 Family Violence, 1989: Victimization Rates by Victim-Offender Relationship and Type of Assault Rate per 1,000 Persons Age 12 and Over Aggravated Assault Simple Assault Characteristic Population Relatives Well Casual Strangers Relatives Well Casual Strangers VIOLENCE IN FAMILIES Known Acquaintances Known Acquaintances Sex Male 96,875,920 0.3 2.2 1.2 7.4 0.5 3.1 3.0 10.3 Female 104,499,700 0.7 1.2 0.6 2.0 2.6 3.6 1.7 4.1 Race White 172,071,010 0.5 1.6 0.9 4.5 1.7 3.4 2.3 7.3 Black 23,378,200 0.4a 2.5 1.3 4.9 1.0 3.4 2.5 5.1 Other 5,926,410 0.4a 1.6a 0a 5.0 2.0a 2.9 0.7a 9.3 Age 12-15 13,256,460 0.5a 4.4 1.6 6.5 1.7 11.2 9.7 14.4 16-19 14,235,270 0.6a 4.9 2.6 13.3 1.8 10.7 7.0 17.2 20-24 18,084,190 1.1 3.1 1.8 9.9 3.9 5.6 4.1 15.7 25-34 43,335,460 0.7 1.7 1.2 5.6 2.6 3.4 2.1 9.1 35-49 50,293,180 0.6 1.2 0.5 3.5 1.5 1.9 1.1 4.6 50-64 32,774,300 0.1a 0.4a 0.2a 1.2 0.4a 0.7 0.3a 2.1 65 and over 29,396,730 0a 0.2a 0.2a 0.6 0.2a 0.2a 0.1a 0.4a 232
TABLE 5-1 Family Violence, 1989: Victimization Rates by Victim-Offender Relationship and Type of Assault Rate per 1,000 Persons Age 12 and Over Aggravated Assault Simple Assault Characteristic Population Relatives Well Casual Strangers Relatives Well Casual Strangers VIOLENCE IN FAMILIES Known Acquaintances Known Acquaintances Marital statusb Married 110,124,950 0.3 0.8 0.5 2.7 0.8 1.1 1.0 3.5 Widowed 13,407,180 0a 0.1a 0.6a 0.4a 0.1a 0.4a 0.3a 0.4a Divorced or 18,786,270 2.3 2.5 2.0 5.7 8.6 5.0 1.4 10.4 separated Never married 58,618,550 0.3 3.5 1.4 8.8 1.2 7.7 5.5 14.2 Family incomec Less than $7,500 20,425,690 0.7a 3.8 1.6 5.5 3.8 6.7 3.1 9.0 $7,500-$9,999 8,374,160 1.0a 2.7 0.8a 4.0 1.7a 6.2 2.5 5.3 $10,000-14,999 19,790,200 0.6a 1.8 1.0 4.8 3.2 3.7 2.5 8.0 $15,000-$24,999 35,690,810 0.5 2.1 1.0 5.1 1.6 4.2 2.7 6.8 $25,000-$29,999 15,302,260 0.4a 1.4 1.1 4.5 1.3 3.8 2.2 7.3 $30,000-$49,999 45,673,340 0.4 0.7 0.6 4.2 1.0 2.2 2.0 7.4 $50,000 or more 28,905,330 0.3a 1.1 0.6 3.9 0.6 1.3 2.1 5.9 a Estimate is based on about 10 or fewer sample cases. b Excludes data on persons whose marital status was not ascertained. c Excludes data on persons whose family income was not ascertained. 233
VIOLENCE IN FAMILIES 234 A number of studies conclude that women who are separated from a spouse or divorced are most at risk of violence and that women who cohabit are more at risk than are those who are married. Available NCS data provide indirect support for this conclusion (Table 5-1). The rate of simple and aggravated assault by relatives is more than 10 times greater for divorced and separated people than it is for the married and widowed ones.11 Unfortunately, marital status risks are not reported separately for women and men, nor are they related to other statuses in the table, such as their age and income. Information is likewise lacking on assaults among cohabitants of both the same and the opposite sex. Such reporting would make more refined risk assessment possible. Homicide Assessments of the risks of intrafamily homicide are more accurate than for other forms of assault, and several patterns are noteworthy: â¢ Newborns, infants, and children ages 1 to 4 are more vulnerable to homicide than are children ages 5 to 9 years (Federal Bureau of Investigation, 1990:11).12 â¢ Infants and small children are more likely to be killed by their mothers than their fathers, perhaps in part as a result of differential risk exposure. â¢ The risk of homicide for children under age 5 is greater for male than female children, according to a recent case control study (Winpisinger et al., 1991:1053-1054). â¢ Although men's overall homicide risk is three times that for women, women face a greater risk of homicide by their spouse than do men (Federal Bureau of Investigation, 1990:8). â¢ At a minimum, intrafamily violence accounted for 15 percent of all homicides in 1989. Violence between husbands and wives accounted for less than half (44%) of the family homicides; family members in male rolesâfathers, sons, and brothersâwere more likely victims of family homicides than were female membersâmothers, daughters, and sisters. Brothers were five times more likely to be killed than sisters (Federal Bureau of Investigation, 1990:12). Because the data are either unavailable or subject to considerable error, calculation of lethality for types of intrafamily assaults seems unwarranted. The higher homicide death rate for the female
VIOLENCE IN FAMILIES 235 partner in spouse assault is consistent, however, with the higher rates of aggravated assault and serious injury for female victims of assault. Child Abuse Although there is a paucity of information on the prevalence and incidence of child abuse, a national profile of risk factors can be derived from the National Incidence Surveys. Female children are three times as likely as males to be sexually abused: 2.9 females per 1,000 compared with 0.9 for males (Sedlak, 1991a,b). Physical abuse is more prevalent than sexual abuse for both sexes. Among males, their less frequent experience of sexual abuse is offset by the greater prevalence of serious injuries among them (Sedlak, 1991a,b). Although female children are at greater risk of sexual abuse, one-fourth of all sexually abused children reported in NIS-2 were male (Sedlak, 1991a,b). Others report a somewhat higher proportion of sexual abuse of male childrenâ about 29 percentâand an even higher rate of abuse for cases in day care centers â38 percent (Finkelhor and Associates, 1986; Finkelhor et al., 1988). They suggest that the higher rate for male children in day care may result from fewer pressures on younger boys against reporting assault by day care providers. Black children were one and one-half times more likely to be physically abused than white children and five times more likely to die of physical abuse or neglect (Sedlak, 1991a,b). There is disagreement about how the severity of the physical assault of children varies with their age. The national survey of Straus et al. (1980) found a curvilinear relationship with age but that parents use more severe and potentially lethal violence against their older than younger children. NIS-2 found however that both fatal and more serious injuries were more numerous among young children (Sedlak, 1991a,b) perhaps because of their greater fragility (Sedlak, 1991a,b). Social status (measured by family income) is substantially related to children's risk of injury from abuse and neglect. For children from families with incomes less than $15,000, the rate of physical abuse was three and one-half times greater and the rate of sexual abuse six times greater than for other children. The general pattern that physical abuse was more frequent than sexual abuse held for both high-income and low-income families (Sedlak, 1991a,b). Moreover, the seriousness of injury or impairment was
VIOLENCE IN FAMILIES 236 substantially related to family income: the rate of serious injury was almost seven times greater, of moderate harm nearly five times greater, and of probable injury seven times greater for children from lower-income than from higher-income families (Sedlak, 1991a,b). There is relatively little information on the consequences of child sexual abuse apart from that provided by clinicians for small numbers of cases (Wolfe et al., 1988:163-166; Bolton et al., 1989: Tables 3.3 and 3.4). In general, that literature concludes that a major consequence is sexual dysfunction when abused children reach adulthood, but that result is not unexpected, given the limitation of the samples to treatment cases. Also, clinicians report that victims of child sexual abuse display fear, immaturity, and neurotic behavior as well as high levels of aggression and antisocial behavior (Finkelhor, 1991:87). Given the general lack of information on child sexual abuse and its consequences, the panel recommends that research be funded to secure information for cross-sectional samples of families and of children in different settings. Although such studies preferably should be part of prospective longitudinal designs, the cost may be prohibitive. Efforts should be made both to improve the reporting of information on victims and offenders and to determine how to improve the quality of retrospective and surrogate reporting of child sexual abuse. Particular attention should be given to ways of securing self-reports from children who are sexually abused in families. Repeat Family Violence A distinguishing feature of family violence is that it occurs among intimates who are bound in recurring, often dependent, relationships that are not easily broken and who share a private space where much behavior is treated as private. Using aggregated data from NCS surveys for the years 1978 to 1982, Langan and Innes (1986) concluded that 32 percent of those assaulted at least once by a domestic partner in the previous six months were victimized again within six months of the incident.13 By contrast, only 13 percent of the victims of violence by strangers were victimized again within six months. To the extent that intrafamily violent acts are more repetitive than other violent acts and tend to be omitted from arrest records and self-reports of violent offending, the extent of repeated violent behavior may be understated. Recent research on intrafamily
VIOLENCE IN FAMILIES 237 violence provides evidence that there is substantial repeat offending among perpetrators of spouse assault (Dunford et al., 1990; Hirschel et al., 1991). The national survey of child abuse (Straus and Gelles, 1990) found that in 94 percent of the child abuse cases more than a single assault was reported. Evidence also suggests that assaults on children in day care centers and nursery schools frequently involve repeated victimization by the same offender (Finkelhor et al., 1988:91-92). Perpetrators of Family Violence Surprisingly little is known about perpetrators of family violence. The panel recommends that more research be undertaken on spouse assaulters' characteristics. Both clinical and case studies and police reports draw attention to the seemingly high rate of drug and alcohol use by both victims and offenders in incidents of family violence. In a recent Milwaukee, Wisconsin, replication of an experiment to test the effect of arrest in preventing recurrences of spouse assault, one-fifth of the female victims and almost two-fifths of the accused males were judged by the officers to be intoxicated at the time of the arrest (Sherman, 1992). Unfortunately, there is no baseline for drug and alcohol use in the general population or in the geographic areas in which the experimental cases occurred. Consequently, how this level of use compares with that among nonabusive husbands cannot be ascertained. Both victims and their batterers may substantially underreport the use of drugs. Urinalyses of randomly selected accused batterers indicate alcohol use by three-fourths and some kind of drug use by at least half of those accused (Sherman, 1992); almost 3 in 10 showed some evidence of cocaine use. The drug use was not necessarily immediately prior to the incident; marijuana in particular leaves substantial traces that can be detected over a period of time. Also, it is not clear that cocaine use precipitated any violent events, since its effects depend on the time since ingestion and are confounded by the concurrent use of alcohol. Consequences of Family Violence A common immediate consequence of family violence is physical harm of the victim requiring emergency department treatment or outpatient medical attention. In a recent review of spouse abuse, Stark and Flitcraft (1991:139) conclude that spouse assault
VIOLENCE IN FAMILIES 238 may be the single most common cause of injury for which women seek emergency medical attention. From their investigation of emergency treatment of women in a metropolitan hospital, they report that battered women were 13 times more likely than other women receiving emergency treatment to be injured in the breast, chest, and abdomen and three times as likely to be injured while pregnant (Stark and Flitcraft, 1991:140; Stark and Flitcraft, 1982). No estimates of the health consequences of other types of family violence are available, nor are there estimates of health and treatment costs for any type of family violence, although the need for such information to guide policy and interventions seems apparent. Suicide and attempted suicide have been linked to domestic violence. Although there appear to be no differences between battered and nonbattered women in the risk of attempted suicide, the risk increases substantially with repeated victimization (Stark and Flitcraft, 1982). The risk of suicide and attempted suicide has not been estimated following other types of family violence. Given the high rates of suicide in the United States, their relationship to types of family violence would appear worthy of investigation. The different forms of family violence have also been linked to mental illness and personality disorders, although these links have been established for clinical populations rather than by using case control methods or general population samples. Studies of psychiatric (Stark, 1984) and women's shelter (Frieze and Browne, 1989:197) populations report that depression is quite common among women who are repeat victims of domestic violence. Battered women are also reported to require significantly higher levels of psychiatric care than nonbattered women (Stark, 1984). It is difficult to determine what role depression has in abusive behavior since there are reported biological correlates of affective disorders such as depression. Because there appear to be generational links in depression, the question of social versus genetic transmission of depression and concomitant violence is also at issue. Some evidence suggests that depressed persons may be more prone to violence, especially toward family members. A number of studies report that abusive mothers often suffer from depression (Zuravin, 1989). Studies report that males who physically abuse their partners show signs of depression. Yet the causal direction is not clear. While some sources of depression (e.g., repressed anger toward others) may cause the abuse, the depression may result from becoming labeled as abusive, or other consequences of the violent act.
VIOLENCE IN FAMILIES 239 In a study of 50 mothers and fathers who had abused and neglected their children and lost custody of them by court order, 30 percent of the mothers and 25 percent of the fathers were diagnosed as suffering from major depression. An additional 17 percent of the mothers and 15 percent of the fathers showed clinical signs of a minor depression. Despite many social service contacts with almost all of the abusing families, only one parent had received any pharmacological treatment and most of the parents had no previous diagnosis of a specific affective disorder. By contrast, the comparison group appeared to have no signs of depression. Depression may affect the severity of maltreatment of children as well as lead to their neglect. If a significant subgroup of abusing parents or partners suffers from affective disorders, especially major depression, then chemical and other forms of treating depression may be a means of controlling some family violence. This assumes that reasonably effective means are available for controlling affective disorders, particularly any volatile mood swings associated with them. A significant consequence of family violence is its transmission from one generation to the next. A substantial body of studies concludes that the experience of violence passes from generation to generation. For example, one study concludes that, among adults who were abused as children, more than one-fifth later abuse their own children (Straus et al., 1980). Unfortunately, methodological limitations in these studies, especially their retrospective designs, restrict the validity of conclusions about the long-term consequences of abuse in childhood (Widom, 1989:161). Widom concluded from a prospective longitudinal design controlling for age, sex, and race, that people who were physically abused or neglected as children had a significantly greater likelihood of arrest for a violent offense than did persons in a control group: 15.8 percent for the physically abused compared with 7.9 percent for the controls (Widom, 1989:163). She also found that the rate was significantly higher for neglected children at 12.5 percent, raising questions about whether the correlation reflected intergenerational transmission of violence or some more complex causal chain that involves both abuse and neglect. EXPLANATIONS OF FAMILY VIOLENCE No explanation, to our knowledge, attempts to explain all the forms of violence within families that we have discussed. Most
VIOLENCE IN FAMILIES 240 explanations are partial in one of two senses: either they attempt to explain a single type or a few types of family violence, such as partner assault, or they seek to identify a particular factor or set of variables that account for some of the observed variation in behavior between violent and nonviolent persons or acts. Gelles (1983) is one of the few to have attempted an integrated theory of violence. We review below some of the explanations that might lead to the development of a more integrated theory or explanation or that seem especially useful in considering forms of intervention to prevent or control family violence. Cultural and Structural Determinants and Social Learning A major assumption of feminist theory is the unequal distribution of power between men and women in most societies. In this view, women are subject to male dominance, which keeps them in subordinate positions in all spheres of family, work, and community life. The many ways in which coercion is exercised in relationships between men and women depend primarily on men's use of both their physical power and their social power to maintain a dominant position (Finkelhor, 1983). Power extends to the sexual relationship between men and women as well as to their ordinary work and social relationships. Feminists argue that marital rape occurs when male coercion rather than consent is the basis of the sexual union (Russell, 1982). The unequal distribution of power is also the basis for explaining parental physical and sexual maltreatment of children. The exercise of parental power over the child victim leads to disempowerment of the child, rendering the child helpless (Finkelhor and Browne, 1986:183). Children as dependents are particularly vulnerable to the exercise of power by women as well as men. Similarly, violence toward the elderly family members is seen as an exercise of youthful power over the weak. The theory does not explain why some men exercise this power and others do not, or why some women do so toward children and the incapacitated. In this framework, growing up in a patriarchal society that emphasizes male dominance and aggression and female victimization, children are socialized into their respective sex roles (Dobash and Dobash, 1979). Moreover, they are said to learn through the intimate experiences of the family. These are reinforced in the larger community when male and female roles similarly rest on elements of macho culture. In this sense, male dominance and aggression and female acquiescence are said to be learned most critically in family and peer relationships.
VIOLENCE IN FAMILIES 241 Major recent changes in family organization and structure may account for some family violence as well. Among those believed to be of significance are changes that affect the social and moral bonding among family members. One such change is deinstitutionalization, in the past 15 to 20 years, of children in foster care, the mentally ill, and the disabled. Distinct subpopulations of the deinstitutionalized are at risk of violent victimizationâthe homeless and those cared for in homes. The temporary placement of children in foster homes, adoption, and the informal placement of children with relatives exposes some of them to the risk of violence from caretakers for whom the constraints on parenting are less controlling. Social service agencies responsible for such placement attempt to screen out potentially abusive and neglectful surrogate parents, but the means for doing so are quite crude and unreliable. A second major change is the increase in the number of children who are not living with their natural parents. Their numbers are substantial owing to serial cohabitation, divorce, and desertion. Lacking the bonds of natural parenthood increases the risk of violence to minors in the family, especially of female adolescents to sexual violence. There is some conjecture also that younger children in these households are more often left without adult supervision and hence subject to more physical violence from older siblings. Social Isolation Although there is some evidence that social isolation is an indicator of family violence, it is unclear whether it is largely an antecedent of family violence or a consequence of it. There are no models of how social isolation develops prior to family violence or of how it contributes to particular violent incidents. Moreover, there are no empirical studies of the extent to which social isolation is generally present for all forms of family violence. Social isolation is a characteristic of some families that are at high risk of physical and sexual abuse of a spouse or children (Garbarino and Crouter, 1978; Pike, 1990). Some batterers enforce social isolation on their partners, and shame may lead the visibly battered spouse to further withdrawal and further social isolation. Members of such families lack the basic social support systems for family life, integration or embeddedness in a network of support systems. They are isolated from friends, acquaintances, or anyone acquainted with what goes on within the family. Such families isolate themselves from others in many different
VIOLENCE IN FAMILIES 242 and often subtle ways. They have an unlisted telephone number or no phone, they lack any means of transportation, and their homes or apartments may be physically shuttered from the gaze of outsiders. They are socially isolated from normal social interactions in their neighborhood and lack organizational ties of any kind (Garbarino and Sherman, 1980). Resick and Reese (1986) suggest that violent families rarely invite others to their home, do not engage in social and recreational activities of any kind, and place less emphasis on personal growth and development. INTERVENTIONS In recent years, a large number of interventions has been developed to prevent intrafamily violence, to reduce the likelihood of its recurrence, and to repair its consequences. We discuss the prospects for detecting and preventing child abuse. We then discuss two responses to spouse assault and conclude with a proposed comprehensive strategy to reduce it. Child Abuse For the same reasons that accurate measurement is so difficult, the problem of detection for intervention purposes is especially acute for physical and sexual violence toward children. The NIS, for example, limited its inquiry into child abuse and neglect to three major sources: (1) cases reported to state and/or local child protective service agencies; (2) cases known to other investigatory agencies such as police, courts, and public health departments; and (3) cases known to professionals in other major community agencies such as schools, hospitals, day care centers, and social services and mental health agencies (Sedlak, 1991a,b). The NIS survey developers recognized that it would be possible to develop a fourth source of cases: survey interviews of the general population to secure reports from persons who know others who are abusing their children, from other family members, or from self-reports of the abusers and the abused. However, acknowledging the difficulties in securing reliable information on uniquely identified persons who are abusing and being abused, the NIS opted not to attempt such surveys (Sedlak, 1991a,b). How to identify the victims of intrafamily violence and their assailants remains an important issue for the design, implementation, and testing of programs to prevent the occurrence or recurrence of intrafamily violence. Attention should be given to ways to detect victims of intrafamily violence and their assailants.
VIOLENCE IN FAMILIES 243 Foster Care Placement A major intervention in child abuse cases is to remove the child from the home and arrange for foster placement. An estimated 15 percent of victims of child maltreatment are placed in an unrelated foster home (American Humane Association, 1979; Runyan et al., 1981). One of the recurring findings is that the more changes in placement a child experiences, the greater the likelihood of adult criminality and violent criminal behavior (Hensey et al., 1983; Lynch and Roberts, 1982; Widom, 1990). Large numbers of placement changes, however, may be both an effect and a cause of children's emotional and behavioral problems that precede adult violent crime. Long-term longitudinal studies are necessary to determine how placement experiences mediate the consequences of early childhood victimization, how other behavior problems affect the frequency of changes in foster placement, and how these distinct processes affect the probability of violent adult criminality. Home Nurse Visitation A home visiting program for infants and preschoolers is one proactive means of detection that also provides a basis for intervention. National programs of this kind exist in the United Kingdom and some European countries, although most are unevaluated. Only one randomized experiment involving a visiting nurse program has been reported in the United States. Olds (1988) and his collaborators randomly assigned 400 first-time mothers at high risk for child maltreatment to one of four treatment groups that offered different services. Among the highest- risk groupâpoor, unmarried, teenage mothers having their first childâthe most comprehensive treatment was found to decrease the incidence of child abuse in comparison with the groups that did not receive nurse home visits. There were other positive effects as well. At 12 and 24 months, infants of poor, unmarried teenagers assigned to the nurse-visited treatment showed improved intellectual functioning on development tests and there was some evidence of improved family function with less evidence of conflict and scolding and less punishment of infants. Olds and Henderson (1989) suggest caution in generalizing from these findings. Although the home nurse condition resulted in a lower rate of child abuse than did comparison treatment groups, the rate of child abuse or neglect was nonetheless around 5 percent, suggesting the necessity for additional preventive or ameliorative
VIOLENCE IN FAMILIES 244 interventions. They also concluded that although the nurses can link families to community and social servicesâto meliorate the effects of poverty, violence, and drug useâthe lack of employment opportunities in the neighborhoods where these families live poses severe constraints on their improvement efforts. Spouse Assaults Common responses to spouse assaults include arrest by police, individual or joint counseling of the perpetrator and the victim, education and training programs for batterers, and relocating the victim to a shelter in a concealed location. Except for police arrests in misdemeanor cases, these interventions have not been systematically evaluated. We review available findings about two of the strategiesâpolice interventions and services in battered women's shelters. We then explain some of the difficulties in evaluating and comparing the effects of interventions and conclude by recommending a comprehensive initiative for improving the understanding and control of spouse assault. Police Responses Most research has focused on police interventions in misdemeanor spouse assault, particularly on the effectiveness of arrest in preventing recurrence. The National Institute of Justice (NIJ) sponsored a program of randomized field experiments and replications to test the relative effectiveness of different forms of police intervention in preventing recurrences of spouse assaultâthe Minneapolis Domestic Violence Experiment. The randomized treatments included arrest of the offending spouse, police counseling of both parties, and temporary separation of the assailant. Sherman and Berk (1984) concluded that arrest was the most effective of these three methods. To further test the validity of that finding, replications were carried out in six areas: Atlanta, Ga.; Charlotte, N.C.; Colorado Springs, Colo.; Metro Dade County, Fla.; Omaha, Neb.; and Milwaukee, Wis. The experiments replicated the arrest treatment but tested other police interventions in each jurisdiction. On the basis of results available to the panel at this writing, it appears that, as with violence outside the family, arrest is not usually a deterrent to repeat domestic violence. Indeed, it may well increase the incidence of domestic violence of unemployed males with low socioeconomic status (Sherman, 1992). There is some evidence that a particular treatment may be effective for
VIOLENCE IN FAMILIES 245 some subgroup of the population. For example, in Omaha, Nebraska, the issuance of a prosecutorial warrant for the appearance of an assailant who left the household before the police arrived appears to have had a modest deterrent effect (Dunford et al., 1989). Yet overall no treatment had a major impact on reducing domestic violence. The panel was impressed by the relatively small impact that these police responses have on repeat domestic violence. Moreover, it is clear that most other interventions to reduce domestic violence have not been effectively evaluated. Given the substantial amount of violence inflicted in domestic situations, it seems imperative to the panel that agencies allocate resources to comparative tests of existing and new interventions that show promise of reducing domestic violence. Because police are in a strategic position to detect some domestic violence situations, the search for and testing of police-administered treatments should continue. But incidents detected by other authorities should also be studied, and interventions that involve police referrals to social service and substance abuse treatment agencies and to battered women's shelters should also be tested. Shelters and Other Services In recent decades, programmatic efforts have focused primarily on providing shelters for battered women, residences where abused women and their children can reside safely and receive emotional support. Starting in the early 1970s, shelters were established in the United States, following a model developed in England. There are now approximately 1,200 shelters offering temporary, emergency housing (typically families stay from two days to three months) to more than 300,000 women and their children each year. And, in response to housing shortages and the desire of many women to establish independent households, transitional housing (where families can stay for up to 18 months) has been developed. The New York City Victim Services Agency opened the first permanent housing for battered women in 1990. A safe harbor is the first goal of shelters. Other services, provided to help women become self-sufficient, include relocation assistance, day care for children, and welfare advocacy. Services directed at increasing self-esteem include support groups and courses in parenting, job readiness, and budgeting. Increasingly, therapeutic services for children who have witnessed violence are incorporated into shelter programs. Many battered women have reported that the most effective
VIOLENCE IN FAMILIES 246 strategy preparing them for self-sufficiency and counteracting the isolation and powerlessness that resulted from being abused was the support they got from other women. Formal and informal groups reinforce for an abused woman that she is not to blame, that she can be a nurturing mother, and that she can develop healthy relationships. Although services for battered women are not limited to shelter, nonresidential programs offering counseling and practical help have been slower to develop and have had even more difficulty than shelters in securing stable funding. Law enforcement offers one possibility for early intervention; health institutions offer another. Studies have shown that abused women are routinely underidentified in hospital emergency departments (Stark and Flitcraft, 1991). Whether early more complete detection would be effective in reducing the risk of further violence or in encouraging women to seek services is unknown. A Comprehensive Initiative to Reduce Partner Assault A wide array of interventions is available to prevent family violence or reduce its escalation. Strategies include public education and awareness campaigns and programs for batterers. Education targets youngsters to develop nonviolent ways of coping with anger and frustration; public awareness speaks to men and women with the message that family violence is a crime and that help is available. Programs for abusive men seek to teach batterers, often coerced to attend by a court mandate, alternative ways to behave. Psychotherapy may have preventive value. Alcohol or drug abuse treatment may reduce violent attacks by some batterers. Where other interventions fail, isolationâby arresting and incarcerating the perpetrator or by relocating the victim to a shelterâat least reduces the opportunity for recurrences. With the exceptions of arrest in misdemeanor cases and programs for batterers (both of which have generated mixed favorable effects), few of these interventions have been systematically evaluated. Perhaps more important, there has been no systematic investigation of their comparative effectiveness or of how they might be used togetherâin part because these interventions are offered by different public authorities and/or community-based organizations. This leads us to recommend a comprehensive initiative against spouse assault, with the following components: 1. risk assessment: better estimates in local surveys and the NCS of the incidence and prevalence of all types of family violence, including attention to cohabiting families; special attention
VIOLENCE IN FAMILIES 247 to the incidence of spouse abuse, especially to repeat victimization; tabulations of more information on types of family violence within legal/statistical categories (e.g., disaggregating specific family relationships, capturing sexual assaults on minors); 2. case control or randomized experimental studies that follow two groups of families over several years: one group would be exposed to a public awareness and education campaign; another offered an infusion of services including shelter, parenting, vocational assistance, assertiveness training, and day care; these treatments could be provided in communities with different law enforcement policies; 3. a jurisdictional comparative analysis of shelter availability with assault and homicide rates; 4. evaluation of batterers' programs through experiments using randomized assignment; and 5. evaluation of law enforcement and criminal justice interventions including arrest as well as less expensive interventions: (a) restraining orders, (b) police warnings in cases without probable cause, (c) police training coupled with intense public education campaigns, and (d) sanctions varying from mandated courses to work programs. Extensive collaboration will be required, involving all the agencies concerned with spouse assault. And considerable care will be needed to overcome the methodological barriers discussed below. Barriers to Evaluation A number of difficulties arise in evaluating strategies to prevent family violence, to deter its recurrence, or to meliorate its consequences. We illustrate these difficulties with reference to the NIJ-sponsored domestic violence experiments (Sherman and Berk, 1984). First, local eligibility criteria for inclusion in the experiment eliminated a substantial proportion of cases and introduced selection biases. Among the exclusions from eligibility for a randomized treatment were the following: â¢ Arrest was mandated for all felony assaults in all experimental sites, so eligibility was limited to misdemeanor domestic assault. â¢ State laws vary considerably in what qualifies as misdemeanor domestic violence; some jurisdictions, for example, excluded same-sex couples and others excluded cohabiting couples.
VIOLENCE IN FAMILIES 248 â¢ Some jurisdictions consider minors ineligible, even though they may be in a cohabiting or marital relationship that would otherwise qualify. â¢ Most jurisdictions require that an arrest be made if the victim requests it or if the officer regards it as necessary to ensure the victim's safety, thereby excluding those cases from any randomized treatment evaluation. Second, treatments were limited for a variety of reasons, among them the fact that no treatment not provided at law can be mandatory unless it is ordered by a court of law. Consequently the police had to rely on voluntary compliance with some counseling treatments; voluntary compliance with repeat counseling sessions is generally so low as to preclude effective evaluation. Moreover, the police cannot effectively administer most treatments (other than arrest) when the subject is intoxicated with alcohol or other drugs. Some treatments are limited by the presence of children if the parent has responsibility for the care of those children. What constitutes a realistic treatment then depends very much on whether it is voluntary or mandatory and in what circumstances it can be administered. Third, intrafamily violence cases may be detected or reported by a large number of different agencies that appear to have different case populations. There can be considerable selection bias, then, when cases are drawn for intervention from just a single source. Therefore, findings from the experiment and its replications may not apply to different populations, such as events that come to the attention of emergency departments, women's outreach groups, or other programs rather than police. Victims in new immigrant families may be especially difficult to account for. Fourth, determining the outcome of treatment is difficult. Changes in family structure, for example when the cohabiting assailant moves out of the household, affect the extent to which one can evaluate the effect of assailant treatments. Moreover, while it is difficult enough to locate victims and secure their self- reports, it is far more difficult to do so for their assailants. We know relatively little, therefore, of the effect of the treatment on the assailant. NOTES 1 Current procedures are based primarily on U.S. Bureau of the Census rules (1988). Its definition is âa family or household requires the
VIOLENCE IN FAMILIES 249 presence of at least two persons: the householder and one or more additional family members related to the householder by birth, adoption, or marriage." Any householder living alone or exclusively with persons who are not related are defined as living in a "non-family household." By this definition, all cohabiting couples are excluded from family statistics. Under these rules, to avoid a sex bias, any adult member of the household who responds may be designated as the householder. 2The Census Bureau currently identifies families, family households, family groups, married couples, and unmarried couples, among other living arrangements (U.S. Bureau of the Census, 1991: Appendix A). 3These characteristics are shared by a number of other continuing relationships, such as teacher and pupil; child and caretaker; custodian and inmate in institutions such as jails, prisons, and mental hospitals; and employer and employee (Finkelhor and Associates, 1986; Finkelhor et al., 1988). 4Although most intrafamily violent incidents occur within the home, each type of intrafamily violence occurs elsewhere as wellâfor example, in bars and on streets, in the workplace, and in school and playground settings. 5The panel made no effort to deal with the many measurement and consent issues that arise in self-reporting of family violence, especially for physical and sexual offenses against children. The fact that many households have a substantial turnover in membership, especially of the male members in cohabiting relationships, poses substantial sampling and response problems, in addition to the measurement problems associated with self-reporting. 6 Because the information is based on incident reports for offenses known to local police, assaults, especially misdemeanor assaults, will be substantially underreported in comparison with the assaults reported in self-reports for the NCS. 7 As of 1991, UCR provided no information on intrafamily violence for any violent crime either for offenses known to the police or for arrests. The combined arrest classification of "offenses against family and children" includes "abuse of family and children" but that category also includes nonviolent misdemeanors such as arrest for nonsupport, neglect, and desertion. (See the definition of âoffenses against the family and children" in Federal Bureau of Investigation, 1991: Appendix II.) The number of these arrests grew by 18 percent between 1980 and 1990 to a rate of 29.2 arrests for every 100,000 inhabitants in 1989 (Federal Bureau of Investigation, 1991: Tables 26 & 27). 8Parenthetically we note that countries such as the United Kingdom that provide for a health care worker to routinely make home visits to preschoolers detect considerably more physical child abuse because they regularly conduct a physical examination of the child. 9 The reporting rates of child protective services increased between 1980 and 1986. Despite the increase, the majority of children reported still are screened out of consideration. Sedlak (1991a,b) concludes that
VIOLENCE IN FAMILIES 250 this failure is due to the use of more stringent screening standards when accepting cases for investigationâstandards that have become more restrictive i n the face of an increase in cases reported and a decrease in resources to deal with them. Research is needed to determine what effects this change in eligibility for allocation resources for child protective services may have on the recent gains in the detection and reporting of physical and sexual abuse of children. 10The sources cited by the U.S. Advisory Board on Child Abuse and Neglect include the following: Daro et al. (1990), Christoffel and Samler (1981), and U.S. Department of Health and Human Services (1988:3-11). 11 Unfortunately, cohabitants and cross-sex friendships are included in the "well-known" relationship in Table 5-1, as are persons of the same sex who are well known to one another. Consequently, we cannot determine the rate for cohabiting and former cohabiting spouses, although presumably it is also very high. 12There is some error in attributing all of these deaths to family violence since UCR does not tabulate the relationship between victim and offender by age of victim. 13 The design of the NCS does not control for prior victimization by any type of crime, so that repeat victimization can be measured only in the context of the "first" report of a sample respondent during the period he or she is in the sample; an unknown proportion of these first reports are not an initial victimization by the offender. Moreover, because the NCS does not take the unique identity of the offender into account, it is not always possible to determine if it is the same offender for a given victim. As for all forms of violence, different offenders can victimize the same family member. References American Humane Association 1979 National Analysis of Official Neglect and Abuse Reporting. Denver: American Humane Association. Bolton, Frank, L. Morris, and A. MacEachron 1989 Males at Risk: The Other Side of Child Sexual Abuse. Newbury Park, Calif.: Sage Publications. Bureau of Justice Statistics, U.S. Department of Justice 1990 Criminal Victimization in the United States, 1988. A national crime survey report, December, NCJ-122024. Washington, D.C.: U.S. Government Printing Office. Christoffel, Liv, and Samler 1981 Epidemiology of fatal child abuse: International mortality data. Journal of Chronic Diseases 34:57-64. Daro, Deborah, Kathleen Casey, and Nadine Abrahams 1990 Reducing Child Abuse 20% by 1990: Preliminary Assessment. National Center for Prevention of Child Abuse. Working Paper No. 843.
VIOLENCE IN FAMILIES 251 Dobash, R.E., and R.P. Dobash 1979 Violence Against Wives: A Case Against Patriarchy. Glencoe, Ill.: Free Press. Dunford, F.W., D. Huizinga, and D.S. Elliot 1989 The Omaha Domestic Violence Experiment. Final Report to the National Institute of Justice and the City of Omaha. Institute of Behavioral Science, University of Colorado, Boulder (June). 1990 The role of arrest in domestic assault: The Omaha police experiment. Criminology 28:183-206. Federal Bureau of Investigation 1991 Uniform Crime Reports for the United States: 1990. Washington, D.C.: U.S. Government Printing Office. Finkelhor, David 1983 Common features of family abuse. In D. Finkelhor, ed., The Dark Side of Families: Current Family Violence Research. Newbury Park, Calif. : Sage Publications. 1991 Child sexual abuse. Pp. 79-94 in Mark L. Rosenberg and Mary Ann Fenley, eds., Violence in America: A Public Health Approach. Oxford University Press. Finkelhor, David, and Angela Browne 1986 Initial and long-term effects: A conceptual framework. Pp. 180-198 in David Finkelhor et al., eds., A Sourcebook on Child Sexual Abuse. Newbury Park, Calif.: Sage Publications. Finkelhor, D., and Associates 1986 A Sourcebook on Child Sexual Abuse. Newbury Park, Calif.: Sage Publications. Finkelhor, D., L.M. Williams, and N. Burns 1988 Nursery Crimes: Sexual Abuse in Day Care. Newbury Park, Calif.: Sage Publications. Frieze, Irene Hanson, and Angela Browne 1989 Violence in marriage. Pp. 163-218 in Lloyd Ohlin and Michael Tonry, eds., Family Violence. Chicago: University of Chicago Press. Garbarino, James 1989 The incidence and prevalence of child maltreatment. Pp. 219-261 in Lloyd Ohlin and Michael Tonry, eds., Family Violence. Chicago: University of Chicago Press. Garbarino, J., and Crouter, K. 1978 Defining the community context of parent-child relations: The correlates of child maltreatment. Child Development 43:604-616. Garbarino, J., and D. Sherman 1980 High-risk families and high-risk neighborhoods. Child Development 51:188-198. Gelles, Richard J. 1983 An exchange social control theory. In David Finkelhor, Richard
VIOLENCE IN FAMILIES 252 J. Gelles, Gerald T. Hotaling, and Murray A. Straus, eds., The Dark Side of Families. Beverly Hills, Calif.: Sage Publications. Hensey, O.J., J.K. Williams, and L. Rosenbloom 1983 Intervention in child abuse: Experiences in Liverpool. Developmental Medicine and Child Neurology 25:606-611. Hirschel, J.D., I.W. Hutchison III, C.W. Dean, J.J. Kelly, and C.E. Pesackis 1991 Charlotte Spouse Assault Replication Project: Final Report. University of North Carolina at Charlotte. Langan, P.A., and C.A. Innes 1986 Preventing Domestic Violence Against Women. Bureau of Justice Statistics Special Report NCJ-102937. Washington, D.C.: U.S. Department of Justice. Lynch, M.A., and J. Roberts 1982 Consequences of Child Abuse. London: Academic Press. Olds, David L. 1988 The prenatal/early infancy project. Pp. 9-32 in R.H. Price, E.I. Cowen, R.P. Lorion, and Ramos-McKay, eds., Fourteen Ounces of Prevention: A Casebook for Practitioners. Washington, D.C.: American Psychological Association. Olds, David L., and C.R. Henderson 1989 The prevention of maltreatment. Pp. 722-763 in D. Cicchetti and V. Carlson, eds., Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect. New York: Cambridge University Press. Pagelow, Mildred Daley 1989 The incidence and prevalence of criminal abuse of other family members. Pp. 263-313 in Lloyd Ohlin and Michael Tonry, eds., Family Violence. Chicago: University of Chicago Press. Pedrick-Cornell, C., and R.J. Gelles 1982 Elder abuse: The status of current knowledge. Family Relations 31:457-465. Pike, K.M. 1990 Intrafamilial sexual abuse of children. Paper prepared for the National Research Council Panel on the Understanding and Control of Violent Behavior. Pillemer, Karl, and David Finkelhor 1988 Prevalence of elder abuse: A random sample survey. Gerontologist 28:51-57. Pillemer, Karl, and Susan Frankel 1988 Domestic violence against the elderly. Pp. 158-183 in Mark L. Rosenberg and James A. Mercy, eds., Violence in America: A Public Health Approach. New York: Oxford University Press. Podnieks, E., and K. Pillemer 1989 Final Report on Survey of Elder Abuse in Canada. Ottawa: Health and Welfare. Resick, P.A., and D. Reese 1986 Perception of family social climate and physical aggression in the home. Journal of Family Violence 1:71-83.
VIOLENCE IN FAMILIES 253 Runyan, D.K., C.L. Gould, D.C. Trost, and F.A. Loda 1981 Determinants of foster care placement for the maltreated child. Child Abuse and Neglect 6:343-350. Russell, D.H. 1982 Rape in Marriage. London and New York: MacMillan. Sedlak, Andrea J. 1991a National Incidence and Prevalence of Child Abuse and Neglect: 1988. Washington, D.C.: Westat, Inc. (Revised September 5, 1991). 1991b Supplementary Analyses of Data on the National Incidence of Child Abuse and Neglect. Washington, D.C.: Westat, Inc. (Revised August 30). Sherman, Lawrence W. 1992 Policing Domestic Violence: Experiments and Dilemmas. New York: Free Press. Sherman, Lawrence W., and Richard A. Berk 1984 The specific deterrent effects of arrest for domestic assault. American Sociological Review 49:261-272. Stark, Evan 1984 The battering syndrome: Social knowledge, social therapy, and the abuse of women. Ph.D. dissertation, State University of New York-Binghamton. Stark, E., and A. Flitcraft 1982 Medical therapy as repression: The case of the battered woman. Health and Medicine 1:29-32. 1991 Spouse abuse. Pp. 123-154 in Mark Rosenberg and Mary Ann Fenley, eds., Violence in America: A Public Health Approach. New York: Oxford University Press. Straus, M.A. 1979 Measuring family conflict and violence: The conflict tactics scale. Journal of Marriage and Family Living 36:13-19. Straus, M.A., and R.J. Gelles 1986 Societal change and change in family violence from 1975 to 1985 as revealed by two national surveys. Journal of Marriage and the Family 48:465-479. Straus, M.A., R.J. Gelles, and S.K. Steinmetz 1980 Behind Closed Doors: Violence in the American Family. Garden City, N.Y.: Doubleday. U.S. Bureau of the Census 1988 Households, Marital Status, and Living: March, 1988. Advance Report. Washington, D.C.: U.S. Department of Commerce. 1990 Household and family characteristics: March 1990 and 1989. Current Population Reports, Population Characteristics. Series P-20, No. 447. Washington, D.C.: U.S. Government Printing Office.
VIOLENCE IN FAMILIES 254 1991 Marital status and living arrangements: March 1990. Current Population Reports. Series P-20, No., 450. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Health and Human Services 1988 Study Findings: Study of National Incidence and Prevalence of Child Abuse and Neglect. National Center on Child Abuse and Neglect, Washington, D.C. 1990 Child Abuse and Neglect: Critical First Steps in Response to a National Emergency. Washington, D.C.: U.S. Government Printing Office. Widom, Cathy Spatz 1989 The cycle of violence. Science 244:160-166. 1990 Research, clinical, and policy issues: Childhood victimization, parent alcohol problems, and long term consequences. National Forum on the Future of Children, Workshop on Children and Parental Illegal Drug Abuse, National Research Council: Institute of Medicine, March 8. Winpisinger, Kim A., Richard S. Hopkins, Robert W. Indian, and Jeptha R. Hostetler 1991 Risk factors for childhood homicides in Ohio: A birth certificate-based case-control study. American Journal of Public Health 81 (August):1052-1054. Winslow, Robert W. 1972 Crime in a Free Society. Selections From the President's Commission on Law Enforcement and Administration of Justice, the National Advisory Commission on Civil Disorder, the National Commission on the Causes and Prevention of Violence, and the Commission on Obscenity and Pornography. Encino, Calif.: Dickenson Publishing Company. Wolfe, David A., Vicky V. Wolfe, and Connie L. Best 1988 Child victims of sexual abuse. Pp. 157-185 in Vincent B. Van Hesselt et al., eds., Handbook of Family Violence. New York: Plenum Press. Zuravin, S.J. 1989 The ecology of child abuse and neglect: Review of the literature and presentation of data. Violence and Victims 4(2):101-120.