Home > Encyclopedia of Drugs, Alcohol, and Addictive Behavior > Family Violence And Substance Abuse

SUBSTANCE ABUSE AND FAMILY LIFE

It has been estimated that abuse is associated with psychological disorders in about 20 percent of cases (Stark & Flitcraft, 1988). The family plays an important role in factors relating to the development, maintenance, and treatment of substance abuse. The fundamental significance of families as dynamic systems has been recognized and studied (Wolin et al., 1980). Today, treatment plans for substance abusers typically involve family members or significant others. The disorganizing impact of alcoholism on families is perhaps the addiction that has been best delineated, but information about the impact of other drug use is increasing (Kosten, Rounsaville, & Kleber, 1985; Bernardi, Jones, and Tennant, 1989).

Disrupted family dynamics can occur irrespective of socioeconomic status and ethnic group membership. Research involving a large cross-sectional sample found that offspring of substance abusers were more likely to experience marital instability and psychiatric symptoms, especially if they had experienced physical and sexual abuse (Greenfield et al., 1993), and it has also been found that alcohol abuse often co-occurs with domestic violence (Fagan, Barnett, & Patton, 1988; Dinwiddie, 1992). Construction of "family trees," or genograms, are now in common use as clinical tools to depict the degree to which abuse of various substances has had effects on several generations in a family, the extent that support is available from family members, and the emotional "valence" of kinship relationships (Lex, 1990). Background factors significant for women include childhood violence experiences, violence from a cohabiting partner, and presence of concurrent antisocial and/or borderline personality disorders (Haver, 1987).

Substance abuse and child abuse may co-occur under similar family conditions and dynamics, or substance abuse can lead to child abuse (Kelley, 1992). Mediating factors, such as social support and education, income, alternative sources of nurturing, and parents' own histories of familial substance abuse and histories of neglect and abuse are important. It is likely, however, that when mothers who use drugs or alcohol are primary caregivers, they will be unable to fulfill some aspects of their children's emotional or physical needs (Tracy & Farkas, 1994).

One typical factor in family lives of substance abusers is the absent father, who usually is affected in some way by substance abuse and whose familial role has had to be reallocated among other relatives (Bekir et al., 1993; Hayes & Emshoff 1993). Often this pattern is transmitted from the grandparental generation to the parental generation. Involuntarily or out of necessity, the missing role is frequently assigned to a child, who has to assume responsibilities inappropriate to his or her age and generation (that is, to act as a spouse or parent). Some children recall having had to raise themselves, since their parents neglected to nurture them or abused or scapegoated them or controlled their activities excessively. Children's responses can include acting out through anger, antisocial behavior, and estrangement, or compliance and assumption of housekeeping, care for siblings, and other domestic tasks. In adulthood, resentment because of the burdens of these childhood role reversals can promote depression in individuals and affect their adjustment to adult roles, and it can, in turn, damage their relationships with their own offspring. In some cases, the onset of substance abuse in children occurs at the age or life-cycle stage when a parent began substance abuse. Substance abusers often appear to expect parental unconditional love from their spouses that includes unquestioned acceptance of their substance abuse and irresponsible behavior (Bekir et al., 1993). Unstated expectations and other communication difficulties occur when the moods and behaviors of substance abusers are closely tied to those of family members (McKay et al., 1993). "Low autonomy" (emotion-ally dependent) substance abusers, however, appear to respond well to treatment if family members provide more nurturing and support. Conversely, male substance abusers whose attitudes and actions are independent and detached from family concerns seem to exhibit a pernicious individualism that is associated with a poor outcome in treatment.

CONSEQUENCES OF ADDICTION IN CHILDREN

Infants exposed to drugs in utero can present problems for caretakers, such as the consequences of prematurity, low birth weight, retarded intrauterine growth, and developmental delays (Blau et al., 1994; Scherling, 1994). Cocaine-exposed infants can be irritable and easily overstimulated, exhibit increased muscle tone, and resist attempts at soothing (Kelley, 1990). There is also a large literature on alcohol effects in utero, which may affect at least 2.6 million infants annually (for review of this literature, see Finnegan & Kandall, 1992). For drug-dependent mothers, these babies sometimes present overwhelming challenges that are often interpreted as "personal" rejection. Mothers' emotions can include guilt about exposure of their child to drugs as well as anger that their efforts at parenting hyperactive babies with feeding difficulties and abnormal sleep patterns seem unsuccessful and only generate more stress. The attachment between mother and child may be disrupted because mothers experience these infants as being highly demanding and ignore and withdraw from them or continue to use drugs. All too often, the consequences of disrupted attachment lead to child neglect and abuse.

PRECIPITATING FACTORS

Alcohol, Drugs, and Aggression.

It is popularly believed that alcohol use facilitates the commission of violent acts. Although there is an association between alcohol (and drug) use and aggression, it is not appropriate to attribute all family violence to substance abuse, and substance abuse does not inevitably result in violence (Hayes & Emshoff, 1993; Taylor & Chermack, 1993). Individual, familial and environmental factors are all implicated in family violence. Controlled studies in research laboratories constitute one means of disentangling the important interrelationships of these factors. One series of laboratory experiments that used electric shocks between competitors as a proxy for aggressive behavior (see Taylor & Chermack, 1993) showed that both the quantity of alcohol that has been consumed and the social environment encouraging aggression are two major contributing factors. Results should be interpreted cautiously, since the extent to which controlled laboratory conditions, and the stimulus of a shock, can be generalized to the events in daily domestic life in households with a person who meets the diagnostic criteria for substance dependence or abuse remains to be demonstrated (Leonard & Jacob, 1988).

Experiments were designed to identify factors that could instigate aggression in persons intoxicated with alcohol. In an interactive setting, research subjects were tested while sober and while intoxicated (i.e., about 0.10 blood alcohol level, or the limit for intoxication while driving in many jurisdictions). Since actual violence could not be condoned ethically, the experiment could only give the illusion that a subject would compete with an "opponent" who could signal intention to send a shock of intense magnitude.

Unless their opponents indicated willingness to administer a strong shock, 80 percent of the sober subjects and 40 percent of the intoxicated subjects were reluctant to retaliate by increasing the magnitude of the shock presumably to be received by the opponent. An additional important factor was pressure from bystanders. In another experiment, two accomplices of the experimenter encouraged both sober and intoxicated subjects to use high-magnitude shocks against their opponents. Under this condition, escalation of shock strength occurred for 10 percent of sober subjects and 50 percent of intoxicated subjects. Once escalation had occurred, however, intervention by a third party was generally ineffective. Instead, the strategies best suited to averting aggression in intoxicated persons were to show the opponent to be nonthreatening, to announce a conventional limit on aggressive behavior (in this instance, magnitude of shocks), or to divert attention from aggression to more socially acceptable behaviors. Although intoxicated subjects expected opponents to be more aggressive than did sober subjects, using a video camera to project an image of the sober opponent's behavior diminished the aggressive responses.

Effects of other drugs on aggression also were evaluated by using this type of laboratory experiment. These studies are important because some tranquilizers are prescribed for anxiety and irritable behavior (Ratey & Gordon, 1993). Low doses of marijuana could result in aggressive behavior, but high doses suppressed it. The use of low doses of benzodiazepines increased aggression, but amphetamines did not augment aggression, and these results were contrary to prevailing expectations. Other studies showed that pretreatment with nicotine, dextroamphetamine, or propranolol (which lowers blood pressure) inhibited aggressive behavior. Furthermore, when individuals were evaluated on an aggression rating scale, the nonaggressive group did not respond to provocation while intoxicated with alcohol, but persons in the moderate- and high-aggression groups responded with aggression.

Thus, pharmacological action of drugs, dosage, characteristics of the consumer, and the social factors surrounding drug taking are all important factors contributing to aggressive behavior. Disturbance of higher-order information processing, or reasoning, appears to be the factor that best explains escalation in aggression while intoxicated. Intoxicated subjects were likely to continue aggressive behavior once it had begun, unless they were strongly prompted to engage in self-reflection. Weak suggestions to limit aggressive behavior apparently are not perceived. Having crossed a behavioral boundary may make it easier to continue to do so.

It also should be noted that alcohol and other drugs have a pharmacological effect on sexual arousal and sexual behavior. Among men, alcohol can cause secondary impotence and heroin use can delay ejaculation. There also is evidence to support the notion that cocaine use can increase sexual interest for men and women, and marijuana use has become associated with uninhibited sexual activity. Some women find that heroin use by their partner prolongs intercourse, and once heroin is used as an adjunct to sexual activity, couples are prone to relapse to drug use (Lex, 1990).

Pharmacological effects of alcohol and drugs can also distort communication. For example, large doses of alcohol consumed in short periods of time can result in blackouts, or disrupted short-term memory. A person in a blackout is unlikely to remember what was said and done during the episode. Excessive cocaine consumption can result in suspicion, hostility, and paranoia. A person in a state of withdrawal from alcohol or drugs can be irritable, and oscillation between withdrawal and intoxication distorts communications, thereby leading to inconsistency, unpredictability, and mistrust (Hayes & Emshoff, 1993).

Social Context of Domestic Violence.

Many sociologists have assumed that domestic violence is a relatively rare event, and until the 1980s anthropologists had only a limited perspective on the occurrence of family violence in other societies. In a major analysis of data from ninety societies (Levinson, 1987), it was found that wife beating was nearly ubiquitous and predictably associated with social and cultural factors. The frequency of wife beating was analyzed, and societies were classified according to whether wife beating was absent or rare, occurred in less than half of households occurred in more than half of households, or was present in almost all households. Using these criteria, it was found that wife beating occurred in 84 percent of the societies in the sample. Occurrence of this behavior was best explained by both social acceptance of violence and economic dominance of men. In a restudy by Erchak and Rosenfeld (1994), additional societies were selected for analysis and when wife beating was coded as simply being either present or absent; it was found that that it occurred in 80 percent of the sample. However, social isolation occurred in 47 percent of societies without wife beating, in contrast to occurrence in 94 percent of nonisolated societies. Socially isolated societies were typically smaller, and their members need to be mutually interdependent for the purposes of survival. In comparison, societies where raiding or warfare against outsiders was common—that is, where disputes with outsiders were resolved by physical force—had a wife-beating rate of 85 percent, versus 29 percent for societies without warfare. In societies that strongly emphasized men's role as warriors, rates of wife beating were 94 percent, in contrast to rates of 56 percent in societies lacking these attitudes and behaviors. Neglect or abuse of children co-occurred with wife beating. Other associated values were beliefs about women's inferiority, the lack of value of women's lives, and a widow's ability to choose a new spouse. Additional associated behaviors included tolerance for homosexuality, control of female sexuality, and competition for economic resources. Thus, the current prevailing desire of women for equality between men and women in the United States may be counter productive and result in more violence, because of increased economic competition between the sexes and increased confusion about appropriate gender-related social behaviors (Erchak & Rosenfeld, 1994).

For impoverished members of minority groups, attributes of the community and neighborhood can adversely affect family life (Wallace, Fullilove, & Wallace, 1992). In a number of urban areas, deterioration of housing, decreases in levels of services such as housing inspections and response by fire-fighting and arson units, and diminished police presence have permitted the dynamics of urban decay to operate. As buildings deteriorate, are further damaged by vandalism, and are destroyed by fire, the impact is much like the spread of a contagious disease. Adjacent buildings may be affected as landlords abandon housing stock and businesses leave or fail. Whole blocks may be damaged, and, finally, entire districts of a city may deteriorate completely.

The quality of life diminishes accordingly. Abandoned buildings are taken over by substance users and sellers or used for other illicit activities such as prostitution. Adolescents can gain ready access to drugs and alcohol, and their behavior may go unchallenged. As people move away, there remain fewer persons available to notice children's behavior, and more unsupervised locations become available where children can engage in disapproved acts. When an area lacks former types of social control, such as sanctions from neighbors, acts such as smoking tobacco cigarettes may escalate to greater deviance, such as using marijuana or crack cocaine. As a consequence, antisocial behaviors may go unchecked, and feelings of anger and hostility can grow. It should be noted, however, that urban settings are not the only locations in which deviance can increase. Contexts that permit anonymity, including ready accessibility of transportation, can also separate perpetrators from persons who know them or would report deviance to authorities.

Perpetrators of Domestic Violence.

Much recent attention has been focused on the psychopathology of both perpetrators and victims. One review (Dinwiddie, 1992) suggested that perpetrators had poor communication skills, higher levels of hostility, and, predictably, less control over their anger. Perpetrators studied for personality problems were more likely to be antisocial, passive-aggressive, or narcissistic. The picture is less clear regarding substance abuse, although men meeting criteria for alcohol abuse or dependence (American Psychiatric Association, 1980) were more likely to hit or throw objects at their wives. Studies of community samples have generally found that perpetrators also meet the criteria for diagnoses of depression and antisocial personality disorder.

In one study, rates of spousal abuse and other problem behaviors were studied in 380 married male relatives of alcoholics (Dinwiddie, 1992). Only 16 percent of the men were self-reported spouse abusers, and 30 percent of these were separated or divorced at the time of the interview, in contrast with 14 percent of the nonabusers. When effects of single diagnoses were examined, alcoholism was the most commonly diagnosed psychological disorder (87%) and was associated with an almost fourfold increase in likelihood of abuse. Diagnoses of antisocial personality disorder (46%) or major depression (33%) were associated with an almost double increased likelihood of spousal abuse. Only four abusers (7%) had no psychological disorder. Most abusers, however, had more than one diagnosis of psychological disorder. Antisocial personality disorder or depression usually co-occurred with alcoholism. Among nonabusers, 65 percent were alcoholic, 23 percent were drug dependent, 20 percent had major depression, and 31 percent had an antisocial personality disorder. Aggressive childhood behaviors were poor predictors of abuse in adulthood, but as adults 95 percent of all abusers reported having physical fights, about half reported marital infidelity, 23 percent had been divorced one or more times, and 17 percent had made attempts at suicide.

Alcohol problems and marital distress appear to be highly interrelated (Halford & Osgarby, 1993). Drinking outside of the home increases marital dissatisfaction, and marital disputes can provoke a relapse in abstinent alcoholics. Divorce rates for alcoholics are thought to be highest among persons with psychological disorders, and divorce or marital problems diminishes the likelihood that alcohol treatment will succeed for individuals. Treatment efforts directed at increasing marital stability, however, can successfully promote abstinence (McCrady et al., 1979). Accordingly, many therapists who treat people for alcoholism suggest conjoint treatment for alcoholism and marital problems. In contrast, few marital therapists address issues of alcohol abuse (Halford & Osgarby, 1993).

A sample of eighty-four women and fifty-six men seeking marriage counseling were identified in a marriage guidance clinic (Halford & Osgarby, 1993). All subjects were still married and cohabiting. The subjects were mainly in their thirties, had about two children, and had been married about nine years. One-third were involved in second or later marriages. The subjects completed questionnaires that probed for information about amounts of alcohol consumption, occurrence of physical violence, and frequency of disputes about alcohol use. About half of the men, but less than 20 percent of the women, met the criteria for a diagnosis of alcoholism. More than 80 percent of the entire sample reported having repeated arguments about alcohol intake, and almost 70 percent reported the occurrence of physical violence. Men and women taking steps leading to divorce were more likely to report disagreements about alcohol use. Women mentioned male violence as a factor in marital dissatisfaction, but men who had been abusive were more likely to seek divorce. In this sample, alcohol abuse was significantly associated with couples taking steps toward divorce, but few other common sources of marital dissatisfaction, such as allocation of household tasks, communication, finances, use of leisure time, and parenting issues, were reported to any significant extent. At the very least, these data suggest that marital therapists should routinely screen their clients for alcohol intake and alcohol-related problems, and that they should assess the extent to which these factors interact with domestic violence. It also is possible that abuse by a husband signals a desire to terminate the relationship rather than to exert greater control over the wife's behavior within the context of marriage.

Disentangling cause-and-effect sequences between alcohol or drug abuse and family violence is an important and necessary step in understanding factors that promote or maintain any interrelationships. There are several ways of approaching these questions, and researchers with competing theories have attempted to explain the relevant issues (Fagan et al., 1988; Strauss & Gelles, 1990). One theory termed "deviance disavowal" has argued that drinkers are not responsible for their actions while they are intoxicated (McAndrew & Edgerton, 1969). Drunkenness is used as an excuse, and it is possible that some persons seek an intoxicated state so as to be able to engage in violent behaviors (Gelles, 1974). According to another theory, alcohol acts on the central nervous system to create a "disinhibition" that releases aggression. Although this reflects a popular belief about the effects of alcohol, it is the social environment promoting or discouraging aggression that is an important contributing factor (Strauss & Gelles, 1990; Taylor & Chermack, 1993). Social learning theory has been applied to a wide variety of behaviors, and the proponents of this theory argue that social meaning becomes attached to behaviors, such as alcohol use, with the result that people come to expect certain behaviors in association with alcohol. Researchers who support a more focused approach have suggested that drinking and violence become associated within the family context, and that discussion of drinking behavior acts as a cue or trigger that escalates verbal hostility and culminates in physical aggression (Fagan, Barnett, & Patton, 1988).

Characteristics of Perpetrators and Victims.

One study used a Relationship Abuse Questionnaire to assess levels of marital violence among abusive and control subjects, including happily married men, maritally dissatisfied men, and men convicted of a violent offense who had not committed acts of domestic violence (Fagan, Barnett, & Patton, 1988). Men in the marital-violence group were young males from minority groups, with limited education and a high rate of unemployment. All members of these groups had been married for an average of four years, had about two children, and were between one to two years older than their wives. Maritally violent men were more likely to consume whiskey and beer, drink daily, drink at lunch on workdays, and drink at home—after work and in the company of their children or by themselves. In addition, maritally violent men indicated that their female partners also drank, but to a lesser degree than they did. These men in the maritally violent group reported that they drank to "deaden the pain in life," to "cheer up a bad mood," to "relax," to "celebrate special occasions," to "forget worries," "to forget everything," and to allay feeling "tense and nervous." They said their female partners drank to "celebrate special occasions" and to "be sociable." Maritally violent men reported that drinking accompanied abuse about one-third of the time but occurred without drinking occasionally, about one-fourth of the time. Female partners were said to drink on about one-fourth of occasions when abuse occurred. Maritally violent men were most likely to report that in the aftermath of violence they felt "sexy" or "wanted to make love," "tried to stop abuse through reasoning," or "took drugs/had a drink." In sum, these men drank more, drank in many social contexts, perhaps continuously but in low amounts, drank to "escape" unpleasant emotions and events, and had female partners who also drank. Drinking or drug taking could be an outcome, however, rather than the cause of a violent episode. It also should be noted that a violent episode could precipitate sexual activity.

A classic study (Kantor & Strauss, 1989) investigated whether drug or alcohol use by victims increased the likelihood of assault by their partners. Information about violence was obtained from 2,033 married or cohabiting women who responded to the 1985 National Family Violence Survey. Research was stimulated by empirical observations that cultural acceptance of violence was the strongest factor in violence directed at wives. This study was designed to test the hypothesis that victims of violence might in some way precipitate violent episodes. Several studies had indicated that people were more likely to attribute blame for violent episodes to women who had violated the cultural attitude that fosters disapproval of women who are intoxicated and another culturally shaped attitude that excuses intoxicated men from the consequences of their alcohol use, including violence. Specific questions included in the interview asked whether women's alcohol or drug use increased the risk of violence from male partners, whether drinking or drug use by male partners increased the risk of violence, whether intervening variables, such as socioeconomic status, explained the occurrence of violence, and whether minor violence and severe violence had different antecedents.

Events were classified as nonviolent, minor violence (throwing objects, pushing, slapping, or grabbing), and severe violence (kicking, hitting, beating, choking, threatening with knives or guns, or using knives or guns). Subjects also were asked whether they used drugs to the extent of being "high" and alcohol to the extent of being "drunk." Predictably, high rates were obtained for alcohol use. Among nonviolent couples, 16 percent of wives and 31 percent of husbands were reported to use alcohol to the extent of being drunk. In contrast, 36 percent of women and 50 percent of men involved in minor-violence episodes used alcohol, and 46 percent of women and 70 percent of men involved in severe-violence episodes had used alcohol. Correlation of violence with drug use (marijuana) was less than half that of alcohol, but the illegal status of marijuana might have encouraged underreporting. Among nonviolent couples, only 4 percent of wives and 5 percent of husbands were reported to use marijuana. In contrast, 14 percent of women and 18 percent of men involved in minor-violence episodes had used marijuana, and 24 percent of women and 31 percent of men involved in severe-violence episodes had used marijuana. Minor-violence episodes were related to the husband's use of marijuana and to violence in the family of origin of the victim. Drunkenness by the wives and by their husbands, low income, and the wives' acceptance of male violence were significant factors, but wives' marijuana use was unimportant. Severe-violence episodes showed a more restricted pattern. Violence in the women's families of origin and husbands' drunkenness were somewhat stronger factors than husbands' marijuana use. Income level, wives' acceptance of abuse, and wives' drunkenness or being high did not affect the severity of violence. In this study, pregnancy or employment status were not relevant factors.

Some have argued that pregnancy is a factor in the precipitation or escalation of abuse episodes. A recent study examined the extent of physical abuse in a multiethnic sample of pregnant women (Berenson et al., 1991). Of 501 women using services at a prenatal clinic, about 20 percent reported physical abuse, and of this group, 29 percent had been abused while pregnant. However, only 19 percent had ever sought medical help, thus indicating that emergency-room statistics might seriously underreport the prevalence of physical abuse. Abuse occurred typically within the context of a primary relationship, with 92 percent of women reporting abuse by only one person, usually (83% of the time) a male partner. Women who had been abused were more likely to report having a partner who abused alcohol or drugs. Abused pregnant women had significantly more pregnancies and more living children than other pregnant women. Across ethnic groups, white non-Hispanic women were 3.5 times more likely than Hispanic women and 1.6 times more likely than black women to experience physical abuse. Substance abuse increased risk of abuse for white non-Hispanic women to two times that of non-abused women, but for black women, almost four times. Other characteristics were important. Traditional values, as exemplified by speaking Spanish, appeared to be a protective factor for Hispanic women. Divorced or unemployed black women, however, were at higher risk for abuse than either Hispanic or white women. Thus, alcohol or drug use are important factors in the abuse of pregnant women, but black women appear to be at highest risk for abuse when these factors were involved.

There is no single cluster of characteristics that typify men who abuse women. Some studies, however, have indicated that witnessing violence in the family of origin may have taught men to use violence as a coping mechanism. Others have argued that alcoholic abusers also may have had a family history of alcoholism, thereby blurring the relationships between causes and effects in families of origin. In a study of men in a treatment program for family violence (Hamberger & Hastings, 1991), comparisons of marital adjustment, coping with conflict, and personality characteristics were made among alcoholic and nonalcoholic men in treatment and control subjects drawn from the community. The average age of the men was about thirty-five, and they had similar education levels. Nonalcoholic men were more likely to be employed and less likely to have witnessed violence in their families of origin. Alcoholic men who had abused their wives were more likely to have been abused as children, but parental alcohol abuse and parental alcoholism appeared to have no direct role in provoking violence by adult abusers who were alcoholic. As might be predicted, the alcoholic abusers had significantly higher personality-disorder scores for avoidant (passive-aggressive) behaviors, aggression, and negativism, and lower scores for conformity. Both alcoholic and nonalcoholic abusers had a large number of symptoms of pathology, thus scoring high on scales measuring anxiety, hysteria, and depression. Alcoholic abusers had the highest scores on psychotic thinking, psychotic depression, and borderline behaviors. As predicted, abusers had higher scores for personality disorders, and alcoholic abusers had the highest scores in this regard. Alcoholic abusers had witnessed more violence in their families of origin and had themselves been victimized by abusers in their families of origin. Overall, alcohol abuse was significantly related to psychopathology as well as to the degree of harm conferred by abuse. Unemployment as a factor operated in some unknown way to bring abusers to the attention of authorities, but the effect of socioeconomic status was not included in the characteristics examined in this study. Clearly, alcoholic abusers identified through agencies had more severe problems, thus suggesting that treatment programs should carefully assess referral sources of clients. A finding of co-morbidity with depression, anxiety, borderline behaviors, and thought disorders suggests that a program focused on abuse alone would be less successful than a more comprehensive approach that offered services for severe psychological disorders.

In another line of investigation, researchers examined women's histories of victimization and their alcohol use together with characteristics of their partners. The reasoning behind this approach was the consideration that when abusive behavior was modeled, excused, or condoned, children would perpetuate these behaviors as being appropriate to gender roles. Thus boys would devalue women and consider abuse a conventional way to deal with conflict, and girls would expect to be devalued and would tolerate abuse. One study investigated these background factors among forty-nine abused women and eighteen male abusers (Bergman & Brismar, 1992). Abusers were not identified through their female partners, since many of the women were afraid to permit contact with them and many of the abusers refused to participate. Abusers were selected from men who had been sentenced to prison for assault and battery of their female partners. The extent of injuries inflicted by the selected men and experienced by the women were comparable as a result of matching reports from the abused women and those from the convicted abusers. It was intriguing to find that both the men and the women reported having been raised without fathers in their families of origins, that about half of the absent fathers were alcoholic, and that most of the mothers were abstainers. As children, about 80 percent of both men and women had witnessed domestic violence in their families. Moreover, 29 percent of the women and 11 percent of the men had experienced sexual abuse as children. As adults, almost all of the women (94%) had experienced previous abuse, and 49 percent had been abused by former partners. About half of the men and one-fourth of the women had used marijuana, 62 percent of the women and 44 percent of the men had used sedative-hypnotic prescription drugs, and 55 percent of the women and 61 percent of the men acknowledged that both partners had been drunk at the time of the precipitating episode of abuse (only 20% of the women and 11% of the men had been sober). Roughly two-thirds of the men and of the women indicated that the abusive incident probably would not have happened in the absence of alcohol. Transgenerational perpetuation of abuse patterns seemed likely, since 25 percent of episodes were witnessed by the children of the women and the rate of the parents' alcohol and drug abuse was high. Thus, information about histories of alcohol and drug abuse as well as exposure to domestic violence should be evaluated for each partner in a couple involved in domestic violence.

Less information is available about drug use (see Miller, 1990). Abuse is not uniformly associated with drug use, however. Psychopharmacological factors have been implicated in domestic violence in the case of some drugs, such as cocaine (Maher & Curtis, 1992), and for economic reasons, such as when a drug abuser resorts to appropriation of family funds to purchase drugs. Systemic violence, related to the hazards of illicit transactions, may spill over into the domestic area if a drug abuser is concerned or suspicious that a partner may be an informer or may be adulterating drugs. Female drug users may find themselves devalued on the basis of both their gender and their behavior, and because some women are involved in prostitution to obtain drugs for themselves or their partners, their risk of exposure to violent behavior is increased substantially. Intoxicated women also may be more verbally aggressive and thus violate the cultural norm that values the "soft-spoken" woman (Miller, 1990).

Studies of alcohol abuse as it is associated with the abuse of women have not been able to identify a sequence of cause and events. More definitive studies are needed, but one informative study of alcohol and drug abuse by eighty-two male perpetrators and victims sought important linkages. The perpetrators were parolees, and data about psychological disorders, substance abuse, modes of conflict resolution, and frequency of violent events were obtained from them and their female partners. About three-quarters of the perpetrators, and a surprising 56 percent of their female partners had alcohol problems, and 73 percent of perpetrators and 40 percent of their partners acknowledged using illegal drugs. Similarly, 78 percent of parolees and 72 percent of their female partners reported perpetrating a moderately violent episode, and 33 percent of parolees and 39 percent of their female partners reported perpetrating a severely violent episode at least once during the three months before the interview. About one-third of the episodes were considered severe, and about three-fourths were considered moderate. Neither alcohol nor drug use was involved independently, but concurrent use contributed significantly to violent events, and the separation of drugs into different classes by pharmacological action did not change the effect of alcohol and drug interaction. When combined, however, cocaine and alcohol had a strong effect on violence. In addition, couples with more substance abuse-related problems had a higher incidence of violent episodes, but, overall, alcohol problems most strongly increased the likelihood that violence would occur. Additional studies of women with concurrent alcohol and drug abuse problems are needed to clarify the temporal relationships.

TREATMENT FOR ABUSERS

Shame, guilt, and denial are powerful emotions that impede both the recognition of problems and the admission of the need for help. It is popularly believed that perpetrators enter treatment only under coercion and with considerable reluctance. Given the strong association between substance abuse and marital violence in some individuals, questions arise as to whether treatment of alcohol or drug abuse alone will concomitantly diminish violent acts. Behavioral marital therapy teaches improved communication skills and has been used to improve the marital relationships of patients as their drinking abates (O'Farrell & Murphy, 1995). This treatment modality, however, does not directly address the problem of violence. A comparison was made between eighty-eight couples with a newly abstinent husband and a nonalcoholic control sample of eighty-eight couples undergoing marital therapy. The study covered the year before treatment and the year after it. Acts of domestic violence occurred between four to six times more frequently during the year before treatment. Rates for violent episodes during the year after treatment remained elevated for both men and their wives, and they were higher than the rates among control couples. In instances of relapse, rates were higher than those for couples who had not relapsed. In turn, rates for couples who had not relapsed were comparable to those for controls. Consequently, effective treatment for alcoholism appears to reduce the frequency of domestic violence, although a study that uses a control group of conjugal pairs not receiving behavioral marital therapy is needed for conclusive results. The cause-and-effect relationships between the release of emotions and relapse still need to be disentangled, however, since the former may provoke the latter or have an additive effect.

Another study examined rates of violent acts among seventy-four persons who completed a treatment program for spousal-abuse abatement and thirty-two who relapsed from this program. Men were referred by themselves or the courts, but neither source of referral nor amount of criminal activity had an effect on outcome. Alcohol problems persisted in 32 percent of the men who completed this program successfully, but 56 percent of recidivists had persistent alcohol problems. Recidivists also had higher levels of drug abuse and less empathy as measured on standardized scales. Recidivists also were found to be significantly more narcissistic (self-centered) and gregarious. These findings suggest that alcohol and drug abuse must be addressed when they occur among perpetrators of domestic violence.

COMMENTARY

Numerous studies that use standardized criteria generally support the prediction that substance abuse and domestic violence co-occur in the majority of violent episodes. Roughly one-fourth to one-fifth of episodes, however, occur without substance abuse as a possible co-factor or precipitant. Some additional studies suggest that verbal hostility can escalate domestic conflict to domestic violence (Lindman et al., 1992), but some episodes of verbal hostility may stem from response to life stress and others may be a result of social learning. In other instances, conflict over a child's or a partner's alcohol or drug consumption may prompt the substance abuser to "protect" the behavior through vehement denial, thereby leading to an escalation of hostility that spirals out of control.

Although any suggestion that women's behaviors might contribute to abuse may seem to take the currently unacceptable position of blaming the victim, there is some evidence that women who express aggression verbally may have had abusive families of origin, and that alcohol abuse may have played a role in fostering a climate of tension and hostility within their households (Gomberg, 1993; Hayes & Emshoff, 1993). This pattern may emerge when women who feel devalued have no behavioral alternative through which to express their frustration. Unfortunately, many potentially interesting and informative laboratory experiments that investigate aggressive behaviors are conducted with undergraduate college students and thus may not disclose important information about effects that stem from income level, social class, educational level, or ethnicity.

Data from alcoholic and drug-abusing women in treatment suggest that younger women may be more verbally aggressive, thus reflecting society-wide changes in gender-role behavior. Other data (Miller, Downs, & Testa, 1993) reveal that women who were victimized as children are more likely to develop alcohol and drug problems in adolescence and adult life. In contrast to women with other psychological disorders, women who require substance-abuse treatment recall more abuse during their childhood. Some contribution to this outcome could be diminished self-esteem and increased alienation from typical childhood socialization processes, as well as limited development of social skills for negotiation and compromise.

It is also possible that the contexts of substance-abuse treatment generate a social expectation that a client must have a family history of substance abuse as well as a background that includes emotional, physical, or sexual abuse. It is clear that additional research is needed and that subject samples need to be drawn from different sources, with different prevalence rates of various types of violence. Longitudinal research that would follow a cohort of children through adolescence, young adulthood, and marital life might hold sorely needed answers. Lacking the answers obtained from definitive research, it is reasonable to continue to screen abuse victims and perpetrators for substance-abuse problems, and to screen substance abusers for perpetration of or victimization through family violence. Because both substance abuse and family violence engender denial that anything is wrong, careful assessment is a prerequisite for effective prevention, intervention, and treatment.

BARBARA LEX