CHAPTER TWO (part 2)

Summary.

This review of research investigating the relationship between socio-demographic variables and the experience of violence within intimate relationships confirms that individuals caught in abusive relationships can be defined in terms of their place in society and the environmental agents impacting upon their lives. Researchers committed to studying sociologically based variables have contributed to the understanding of the family violence issues.  However, notwithstanding Gelles and Straus's (1988) claim that more than 90 percent of family violence could be accounted for by socio-demographic variables, there is a body of literature which also suggests that factors such as alcohol consumption and dependence and personality play a large role in its occurrence.  The focus of the next two sections will be a review of relevant research in those areas.

Alcohol Consumption

It has been suggested that the consumption of alcohol is not limited to incidents of family violence but is associated with other forms of violent conduct ranging from common assault (Gerson, 1978) to homicide (Chimbos, 1978). Although Hauser's (1982) review of the literature estimated the presence of alcohol in domestic assaults as ranging between 40 and 90 percent, Roberts (1988) pointed out that most of the research indicated a 60 to 70 percent rate of alcohol abuse among batterers.  In spite of the strong association between acts of domestic violence and the consumption of alcohol just presented, considerable variation in reported rates is still evident.   In an effort to account for this variability, the following explanations have been put forth:

1. The amount a person drinks, the degree of his or her alcohol dependence and the extent of the violence experienced in a relationship can vary greatly according to the sample in question (i.e., a clinical sample of batterers and/or alcoholics v. a community sample).

2. The method of data collection (i.e., reports of victims and assailants v. objective physiological measures such as obtaining blood alcohol levels following an abuse incident) will have an impact on the degree of association between these variables.

3. The strength of the relationship between alcohol consumption and partner abuse will be influenced by whether or not researchers make distinctions among the independent and combined effects of drug and alcohol abuse.   Quite understandably, it is also these factors that have also contributed to the confusion surrounding the link between alcohol and domestic violence.

Theoretical basis of alcohol-related violence.

While researchers generally agree on the magnitude of the problem of alcohol related violence, there is less consensus concerning a theoretical basis for this phenomenon.  Two schools of thought predominate the literature on alcohol abuse and dependence; one adheres to a medical model, while the other provides a psychosocial paradigm of alcohol and violence.

Proponents of the former believe that the pharmacological effects of alcohol produce physiological and cognitive responses in some individuals that are manifested in aggressive behaviour (Frieze & Schafer, 1984; Shapiro, 1982). Examples of such responses include: increased heart rate, skin flushing, decreased ability to experience pain, impaired memory and reduced reaction time. Furthermore, the application of the medical model to alcohol related violence views alcoholism as a disease and because of this, absolves the alcoholic of the responsibilities for his or her conduct while under its influence.

The psychosocial interpretation of alcohol induced violence also releases the alcoholic perpetrator of violence from the responsibility for his or her actions. However, whereas the former interpretation relies upon a disease process as the basis for excusing violent conduct, this interpretation posits that the ingestion of alcohol provides individuals with a socially acceptable excuse for engaging in socially unacceptable behaviours.

Kantor and Straus's (1990) description of an integrated model based on this perspective strongly supports the influence of factors such as the symbolic meaning attached to alcohol use and associated expectancies about its effects. It is thought that situational factors present in the interactional process of individuals and perceptual and cognitive changes stemming from the effects of alcohol combine to produce alcohol related violent outcomes.  Shapiro (1982) emphasized the role of the family as a force that maintains the drinking behaviour in order to preserve what it perceives as family stability.

Variables derived from the psychosocial model combine to create an environment that provides the opportunity for alcohol related violence to occur and be sustained.  Because individuals are not held accountable for their conduct while under the influence of alcohol, it is thought that many get drunk so that they can behave violently. In this way, the consumption of alcohol is viewed as providing "time out" for violence (Frieze & Schafer, 1984; Livingston, 1986; Richardson & Campbell, 1980: Russell, 1982).

As noted, the theoretical perspectives just described, share in common the notion of absolved responsibility for the alcoholic perpetrator of violence. However, these viewpoints differ in the mechanisms that exempt persons from being held accountable for their own conduct.  The medical model explains alcohol induced violence by way of internal factors (i.e., physiological effects) whereas the psychosocial model utilizes both internal and external factors (i.e., attributions of blame, social attitudes toward alcohol and violence and cognitive effects) in its interpretation.

Variables linking alcohol consumption to partner abuse.

In spite of the extensive literature on alcohol consumption as it relates to various forms of family violence, there remains much confusion about the nature of its linkage.  Some of the issues in need of clarification relate to the timing of drinking relative to violence, the relationship between the amount of alcohol consumed and violence, and the relationship between alcohol consumption and violence perpetrated by women.

A neglected aspect of this research area concerns the limited number of investigations exploring the possible relationship between the occurrence of a violent episode and the timing of alcohol ingestion.  Most of the research examining the relationship between alcohol consumption and the occurrence of domestic violence has been correlational in nature, and has failed to consider alcohol as an immediate antecedent of a violent episode.

Kantor and Straus's (1987) study on a nationally representative sample of 5,159 couples addressed this issue. They found that, while a linear relationship was found between alcohol consumption and wife abuse, alcohol was not used immediately prior to a conflict in 76 percent of the cases.  Findings of this kind provide the opportunity to make additional inferences concerning the role of alcohol in the occurrence of domestic violence based on objective measures of alcohol consumption and abuse (i.e.,   quantity-frequency), drinking at the time of violence and wife abuse (CTS).   Kantor and Straus (1990) concluded that alcohol use at the time of violence was neither a necessary nor a sufficient cause of wife abuse.

Leonard, Bromet, Parkinson, Day and Ryan (1985) arrived at a similar conclusion.  In a sample of community based blue collar workers (n=484 males), the authors concluded that the amount of alcohol consumed most recently is less relevant to marital assault than is a diagnosis of alcoholism based on pathological alcohol consumption patterns and indices of long term drinking.  It appears that the lifestyle associated with a well established pattern of drinking may be a more important determinant of violent outcomes than a drinking episode prior to an incident of partner abuse.

An earlier study by Gerson (1978) provided contradictory findings. Incidents of marital assaults drawn from police data all were preceded by the consumption of alcohol by the offender (43.7%), victim (13.6%), or both offender and victim (42.8%).   It is possible that these conflicting findings are largely due to the nature of this sample since it likely that alcohol related violence incidents come to the attention of law enforcement agencies.

The research reviewed suggests that the long term effects of alcohol consumption and dependence are important in understanding the mechanisms of violence between intimates.  According to Hamilton and Collins (1981), chronic drinking has the potential to change its consumer.  The physiological and cognitive effects of alcohol alter the behaviour of the individual and shape personal interactions through its influences.  The cognitive distortions caused by the consumption of alcohol lead to a redefinition of situational norms and increases the likelihood for violence to occur.

The amount of alcohol consumed prior to episodes of partner abuse is another variable in need of clarification.  Kantor and Straus's (1987) demonstration of a linear relationship between the amount of alcohol consumed and reports of wife abuse suggests that the more alcohol consumed, the greater the likelihood for abuse to occur due to the disinhibiting effects of alcohol on peoples' behaviour.   However, Russell (1982) suggested that alcohol produces paradoxical effects depending upon the biochemistry of its consumer, as well as upon the amount of alcohol consumed.

Consistent with the notion that alcohol's effect on violence may be dependent upon individual differences and the amount of alcohol consumed, Frieze and Schafer (1984) indicated that the highest rates of violence have been reported among moderate drinkers.  Hamilton and Collins's (1981) review noted a curvilinear association existed between these variables.  The lowest rates of violence were found among men who drank the least, followed by heavy drinkers and then finally by moderate drinkers.  Support for this relationship was also provided by Coleman and Straus (1983) who found almost no violence for the most alcoholic group.  Hamilton and Collins (1981) suggested that for some men, high alcohol consumption appears to undermine the capacity for violence.  There may be a point at which the disinhibiting effects of alcohol diminish and its depressive effects increase.   However, it has not been determined whether this relationship is a function of dosage level, individual differences (i.e., biochemistry and personality) or the chronicity of a drinking problem.

The alcohol literature has tended to focus on females as abused coalcoholics and males as alcoholic perpetrators of violence. Although studies have implicated women's drinking in the occurrence of partner abuse, it has nevertheless, been limited to the examination of their own victimization.  To date, there have been no empirical works investigating males as coalcoholics in households where females are perpetrators of alcohol-related violence.

The research on female alcoholic perpetrators of partner abuse is limited. Bland and Orn (1986) found that for those with a DSM-III diagnosis for alcoholism, only 3.8 percent of women were violent against their partners compared to 22.3 percent of men. This study appears to suggest that the mechanism underlying the association between alcohol consumption and partner abuse is quite different for men and women.  Coleman and Straus (1990) reported that although men get drunk more often than women, "women who are drunk rarely or occasionally are much more likely to be violent than males at the same frequency of alcohol use (2.6 times more likely for severe violence and 1.2 times for all violence)" p. 116.

Frieze and Schafer (1984) proposed a cognitive model of male and female reactions to alcohol.  According to this model, "a drinker's reaction will depend on the social context in which drinking occurs and the prior expectations of the person about how alcohol will affect him or her" (p. 277).  Because women may interpret the sensation of warmth associated with vasodilation as emotional warmth and men may interpret the same as power, given one's prior socialization together with the right circumstances, men are more likely to react aggressively following the consumption of alcohol.

Summary.

The research reviewed has demonstrated that in spite of the strong evidence linking alcohol consumption and partner abuse, there is still much confusion when it comes to fully understanding the nature of this association. This review has delineated the sources of such confusion.

The notion of causation with respect to the relationship between alcohol and partner abuse is one issue that deserves comment. Writers have cautioned that the presence of an association between alcohol and domestic violence must not be interpreted as implying causation (Shapiro, 1982; Van Hasselt, Morrison & Bellack, 1985; Kantor & Straus, 1990).  The use of experimental and longitudinal designs will be important in establishing a model of causality.

Finally,  there is a need to consider alcohol's interaction effects with other variables in determining its link to violent outcomes.  For example, research by Gustafson (1985) found that the interaction between frustration and alcohol was associated with an increase in aggression such that in a frustrative condition, alcohol increased both the probability and strength of an aggressive response. While studies have investigated the interaction effects of alcohol and personality in the occurrence of partner abuse (Barnes et al., 1991; Van Hasselt et al., 1985), there is the likelihood that other variables will interact with alcohol in its prediction (i.e., expectancy of effects, attitudes toward partner, attitudes toward violence).  The challenge remaining is to determine whether alcohol is symptomatic of deeper problems such as violence (Everett, 1985), or linked to dysfunctional interactions involving personality, genetics and environmental influences.

Personality

The abandonment of an earlier conceptualization that the perpetration of incidents of family violence was the result of psychopathology has resulted in a limited number of epidemiological studies examining psychological risk factors. For the most part, sociologically based research has discounted the role of psychological variables in the occurrence of violence in the family (Gelles & Straus, 1988).  Elliot (1982), on the other hand, defended biological determinants of intrafamilial violence by stating "... for many years most sociological studies of the subject have paid little more than lip service to biological factors such as genetics, hormonal disorders, developmental defects and acquired damage..." (p. 37).

The role of individual differences as predictors of partner abuse has been researched.  Studies examining factors such as personality (Barnes et al., 1991; Gondolf et al., 1990; Hamberger & Hastings, 1991), motivation (Rouse, 1990) and attribution (Shields and Hanneke, 1983; Williams, 1990) are now evident in the literature. The research conducted has primarily employed male clinical samples and as a result limits the generalizability of results. Nevertheless, according to Sommer (1990), it is these clinical studies that form the only knowledge base upon which profiles of the spouse abusers have been built.  In spite of this serious limitation, some insight has been provided into the psychological factors differentiating abusers from nonabusers. Moreover, their findings have also provided direction for general population based research and the development of psychologically based theories.

Hamberger and Hasting's (1988) founded their definition of the batterer on the following DSM-III description of personality disorder:

"personality traits that are inflexible and maladaptive and cause either significant impairment in social or occupational functioning or subjective distress p. 764."

Expanding upon this characterization, the authors noted that many personality disordered individuals experience mood disturbances such as anxiety and depression, and do not view these behavioral manifestations as either "desirable or ego-syntonic".

The aim of researchers interested in describing the partner abuser has been to separate perpetrators of partner abuse from the general population based on the evidence of psychopathology.  Gondolf et al.'s (1989; 1990) investigations involving a sample of violent psychiatric admissions (n=389) found that forty-one percent of the sample was involved in incidents of violence with a family member. Nineteen percent of them represented incidents of spousal abuse of which 67 percent were perpetrated by males.

In a replication study involving a sample of 99 men attending a domestic violence abatement program, Hamberger and Hastings (1986) found that 88 percent of them demonstrated some indication of psychopathology.  Although multiple personality measures (eg., Millon Clinical Multiraxial Inventory, Novaco Anger Scale, Beck Depression Inventory) were employed, no single profile of the abuser emerged.

A factor analysis conducted by Hamberger and Hastings (1986) confirmed the existence of three major personality categories observed in their first study. According to the authors, these categories are consistent with the diagnoses of schizoidal/borderline, narcissistic/antisocial and passive dependent/compulsive personality disorders.  Using a different set of assessments (16 Personality Factor, Psychological Screening Inventory, Michigan Alcoholism Screening Test), Schuerger and Reigle (1988) found that higher levels of violence were associated with greater pathology in the form of anxiety, depression, schizoid tendencies, social nonconformity, as well as signs of alcoholism among 250 men in treatment for wife abuse.  These findings suggest that the spouse abuser is someone who presents with psychopathology along a number of dimensions rather than by a distinct set of personality features.

Hale, Duckworth, Zimostrad and Nicholas (1988) investigated the MMPI scores of 67 men participating in a batterers' treatment programs.  Their findings confirmed previous clinical ones suggesting the existence of an antisocial personality characterized by the following characteristics: impulsiveness, lack of respect for social standards, frequent difficulties with the law and with their families, situational depression, feelings of inadequacy or low self-esteem or both, and a tendency toward substance abuse.

Gondolf (1988) performed a cluster analysis of battered women's self reports (n=525).  While the results raise questions about the validity of women's assessments of their partners personalities, the following three types of batterer personalities emerged from the analyses: antisocial, sociopathic, and typical. The latter type is characteristic of the individual who regrets his behaviour and asks for forgiveness (Walker, 1979).  The other profiles are indicative of more serious abuse in which there is no caring or remorse.  Moreover, sociopathic abuse is also characterized by a wide range of arrests related to crimes against property, violence of a non-family nature, and alcohol-related crimes.

Kalichman (1988) examined male and female perpetrated domestic homicides. As in Hale et al.'s (1988) study, MMPI scores were used to assess the personality profiles of these offenders.  Analysis of scale scores provided two separate (although somewhat similar) profiles.  For males convicted of domestic homicide (n=19), the most distinguishing characteristic was a single elevation on the Pd scale. High scores on this dimension are indicative of sociopathy depicted by pathological lying and chronic difficulty with the law.  Whereas both male and female murderers (n=16) shared this personality characteristic in common, females scored significantly higher than males on the Pa scale, a measure of paranoia. Moreover, female murderers were also found to score low on the Mf scale signifying higher masculinity.  Based on this MMPI profile, Kalichman (1988) described these women as "presenting difficulties in social relations while experiencing high levels of dependency in their relationships" (p. 852).

Summary.

The research presented supports the inclusion of personality measures in inquiries into family violence.  Although the findings discussed in this review have been predominantly based upon clinical samples made up of male batterers, sufficient evidence has been provided to support the contention that abusers demonstrate levels of pathology higher than those found in the general population (Hamberger & Hastings, 1991).  Nevertheless, while researchers have begun to recognize the need for investigations directed at the influence of individual differences in the occurrence of intimate relationships, current research is limited in a number of ways.  A discussion of some of those issues follows.

Prevailing sampling procedures provide the most obvious example of a methodological restriction.  Kalichman (1988) cautioned that differences can even be found across what are thought to be comparable clinical samples.  For example, he noted that the subjects participating in his study may be more representative of better adjusted inmates because of their willingness to co- operate. Likewise, Hamberger and Hastings (1986) acknowledged that their sample of men arrested and ordered to assessment was exemplary of those abusers who had been caught.  Until comparisons can be made across abusers groups, as well as with those found in the general population, it is unlikely that results of studies can be extended beyond the sample in question.

The research reviewed suggests that as a group, "batterers are heterogenous and fail to conform to a unified batterer profile" (p. 143, Hamberger & Hastings, 1991).  This view has received support from Gondolf (1988), Hale et al. (1988), Straus et al. (1980), and Tolman and Bennett (1990).  The following represents a list of personality characteristics derived from research presented: schizoid/borderline, narcissistic/antisocial, passive dependent/compulsive, antisocial, sociopathic, anxiety prone, depressive, socially nonconforming, neurotic as well as those describing individuals prone to substance abuse, low self-esteem and high masculinity.  The extensiveness of this list attests to broad range of pathology associated with relationship violence and supports Hamberger and Hasting's (1986) bid for a broad-based approach that will allow for the "determination that abusers....exhibit not only a preponderance of personality disorders, but several discrete 'types' " (p. 340).

The implications of personality factors in determining violent outcomes are numerous.  Personality disordered abusers present researchers, law enforcement agencies, and service providers with a particularly difficult challenge.   The range of pathology within the abuser population tests the viability of programs aimed at altering maladaptive family interactions.

Theories of Partner Abuse

As the number of empirical articles published in the area of domestic violence have increased, the story of abuse between intimate partners has begun to unfold.  In an attempt to explain how and why abuse between intimates develops, occurs and is sustained, researchers have relied upon theoretical frameworks to guide their research.

Knight and Hatty (1987) noted that theories advanced by various researchers could be dichotomized into those reflecting the orientations of sociologists on one hand, and psychologists on the other hand.  According to these writers, sociological perspectives place the phenomenon of domestic violence within a macro model of society, while psychologically oriented researchers account for violence within a micro level of analysis.

Differences in the philosophical orientations of these disciplines have resulted in an ongoing controversy with regard to which one provides the most comprehensive accounting of abuse.  From a sociological perspective, violence is seen as an outgrowth of social factors as contrasted to psychological perspectives that attribute violence to such intra-individual factors as aggressiveness, impulsiveness, and paranoia (Straus, 1980). When assigning weight to the role of individual risk factors, Straus (1980) suggested that psychological factors on the other hand,

".... account for only a minuscule proportion of the violence which occurs in families ... at the outside ten percent. At least 90% of the violence which takes place in American families grows out of the very nature of the family and of the larger society, rather than out of individual aberrations" (p. 8, Straus, 1980).

The following discussion will review conceptual frameworks reflecting the dichotomy in research orientations described above. In addition, current theories that integrate the philosophies of both will also be presented. However, the reader is cautioned that in light of the existing research bias that focuses on males as perpetrators of intimate violence and females as their victims, the available theories of domestic violence have been primarily advanced with this belief in mind. 

Next: Chapter 2 Part 3