theas This is to certify that the dissertation entitled A Study of Internalized Shame and Adult Models of Attachment in Two Samples of Male Batterers presented by Timothy R. Strang has been accepted towards fulfillment of the requirements for Ph.D. mgmem Counseling, Educational Psychology and Special Education v u‘ ’ Major professor Date 0154 8/; ma: MSU is an Affirmatiw Action/Equal Opportunity Institution 0-12771 . Illilllllll illliliillllliflil Ill'ilifllll 1 3 1293 01025 9053 LIBRARY Michigan State University PLACE ll RETURN BOX to romovo this chookouttrom your rooord. TO AVOID FINES roturn on or botoro dot. duo. DATE DUE DATE DUE DATE DUE MSU lo An Afflrrnottvo Action/Emu Opportunity Institution ‘ A STUDY OF INTERNALIZED SHAME AND ADULT MODELS OF ATTACHMENT IN TWO SAMPLES OF MALE BATTERERS BY Timothy R. Strang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Educational Psychology and Special Education 1994 ABSTRACT A STUDY OF INTERNALIZED SHAME AND ADULT MODELS OF ATTACHMENT IN TWO SAMPLES OF MALE BATTERERS BY Timothy R. Strang Men who batter their intimate female partners have begun to receive wide attention in the empirical literature over the past twenty years. Until then, male batterers had been understood primarily through the eyes of their victims. Beginning with the research of Ganley and Harris (1978), the personality characteristics of these men have come under closer scrutiny. A number of researchers have proposed various typologies of male batterers, with the most promising being the categorizing of batterers into dependent versus dominant types by Saunders (1992). Previous domestic violence research has not examined the preposed dependent versus dominant typology along the dimensions of internalized shame affects and models of adult attachment. In an effort to explore these dimensions, a study was conducted to examine the relationships between internalized shame, abusive behavior, and adult models of attachment in samples of dependent and dominant male batterers referred for treatment. Subjects were 100 men enrolled in a domestic violence treatment program in Fort Wayne, Indiana. A package of questionnaires containing measures of internalized shame, frequency and severity of physical abuse, and relationship styles corresponding to adult models of attachment were administered to subjects. Internalized shame was measured by the Personal Feelings Questionnaire (Harder & Zalma, 1990), and the Internalized Shame Scale (Cook, 1988). Abuse was measured by the Woman Abuse Scale (Saunders, 1992). The adult attachment style was measured by the Relationship Questionnaire (Bartholomew & Horowitz, 1991). Significant differences were found between dependent and dominant batterers on measures of internalized shame, with dependent batterers showing higher levels of shame. Higher levels of shame were associated with lower frequencies of abuse. Dependent batterers primarily occupied the preoccupied category of adult attachment. Subjects occupying the secure and dismissive categories of attachment scored lower on the shame scales than did subjects occupying the fearful and preoccupied categories. Subjects in the preoccupied attachment category showed lower frequencies of abusive behavior than subjects in the other three categories. Results were discussed with implications for further research and clinical practice. ACKNOWLEDGEMENTS The completion of this dissertation is the culmination of a lengthy and arduous process that began somewhere in the "Reagan years". I have learned that an undertaking of this magnitude is never the result of one person's hard work, but rather, the efforts of many. Teachers, friends, and colleagues have been kind enough to provide information, advice, support, and encouragement throughout this project, and I can truthfully write that without them, I could not have produced a finished dissertation. I would like to thank the former Battle Creek Adventist Hospital for giving me the autonomy to develop a treatment program for male batterers, allowing me to experience what it is like to work with this difficult population. Dr. John Berecz provided valuable insights into the personality dynamics and behaviors of wife assaulters, and gently guided me into possible theories of rage and shame. I would also like to thank the Center for Nonviolence in Fort Wayne, Indiana, for all their help in collecting the data for this research project. I would like to thank Dr. John Powell, my chairperson, for his encouragement, optimism, and patient guidance. He showed an astonishing confidence in my ability to complete this dissertation, often when my own confidence was waning. iv My appreciation also goes to Dr. Richard Johnson, for his support and insightful comments. I would like to thank Dr. Lee June for substituting for another committee member on short notice and adding his thoughtful comments. Dr. Ellen Strommen provided valuable insight as well as emotional support. I also wish to thank my wife Pat, who helped in numerous ways. Without her support, completion of this dissertation would have been much more difficult. T. R. S. 1994 TABLE OF CONTENTS LIST OF TABLES . . . . . . . . Chapter I. II. III. INTRODUCTION . . . . . Statement of the Problem Purpose of the Study Importance of the Study Definition of Terms 0 Overview of the Study . REVIEW OF THE LITERATURE Research into the Characteristics of Male Batterers Psychometric Research: Profile . . . . Toward a Typology of Male Batterers Childhood Exposure to Domestic Violence Shame and Anger . . Shame, Attachment, and Domestic Violence Research Questions . METHODOLOGY . . . . . Description of the Setting . Description of the Sample . Instruments . . . . Demographic Questionnaire vi Search for a Unitary 10 10 13 14 23 24 28 32 34 34 35 36 36 Harder Personal Feelings . . . . The Woman Abuse Scale . . . . . . Internalized Shame Scale . . . . . Personal Reaction Inventory . . . Relationship Questionnaire . . Procedures for Collecting Data . . . Research Design and Analysis of the Data Hypotheses . . Summary . . . . IV. ANALYSIS OF THE DATA Overview of the Research Questions . Preliminary Analysis of the Data . . . Description of Description of Main Analysis . Hypothesis Hypothesis Hypothesis Hypothesis Hypothesis Hypothesis Hypothesis Summary . . . . 1 7 the Research Sample Nonparticipators . V. SUMMARY AND CONCLUSIONS . . . . . . . . . Summary of Findings Demographic Variables . . . . . . Shame and Male Batterers . . vii 37 38 38 39 39 .40 41 .42 45 47 47 48 48 50 50 50 56 57 63 64 65 66 68 7O 73 73 75 Wife Assault and Adult Categories of Attachment . . . . . . . Shame and Attachment . . . . . . Abuse and Attachment . . . . . . Abuse and Shame . . . . . . . . . Discussion of Findings . . . . . . . . Demographic Variables . . . . . . Shame and Male Batterers . . . . Wife Assault and Adult Attachment Shame and Attachment . . . . . . Abuse and Attachment . . . . . . Abuse and Shame . . . . . . . . . Limitations of the Study . . . . . . . External Validity . . . . . . . . Design and Methodology . . . . . Implications of the Study . . . . . . Theoretical Implications . . . . Practical Implications . . . . . Recommendations for Future Research . APPENDICES O O O O O O O O O O O O O O O O O O O O A. B. Demographic Questionnaire . . . . . . . . . . Personal Feelings Questionnaire . . . . . . . Woman Abuse Scale . . . . . . . . . . . . . . Internalized Feelings Scale . . . . . . . . . Personal Reaction Inventory . . . . . . . . . Informed Consent Form . . . . . . . . . . . . Instructions to Participants . . . . . . . . viii 75 76 76 76 77 77 83 85 88 89 .90 92 92 93 93 93 95 97 99 99 101 103 104 107 108 109 H. Instructions for Administration . . . . . . . . 110 I. Relationship Questionnaire . . . . . . . . . . . 111 BIBLIOGRAPHY.....................112 ix Table Table Table Table Table Table Table Table Table 4.7 4.9 LIST OF TABLES Summary of Subject Demographics Item by Item Means for Woman Abuse Scale Group Means for Two Measures of Shame Item by Item Means for HPFQ . . Item by Item Means for 188 . . Group Means for Four Descriptions of Relationship Style . . . . Attachment Category and Shame--Spearman Coefficients . . . . . . . Level of Abuse for Attachment Categories Level of Abuse and Level of Shame 53 55 58 59 61 64 66 67 67 CHAPTER I ;, at: 1, INTRODUCTION The American family is not immune to violence. People are more likely to be beaten, hit, slapped, punched, or killed in their own homes than anywhere else in their environment. Violence among family members may be at least as common as feelings of love (Straus, 1977). Domestic violence affects nearly two million marriages in the United States (Gelles & Cornell, 1990). More than one in four couples reports violence at some point in their relationships. Severe assaults (slapping, kicking, biting, hitting with a fist or object, beating with or threatening use of a weapon) occur in approximately 13% of all marriages and are repeated two-thirds of the time (Dutton, 1988). Data indicate that 21% of all women who use hospital emergency room services are injured by their male partners. Battering is the single major cause of injury to women, more frequent than auto accidents, muggings, and rapes combined (Charron, 1991). Domestic violence is clearly a problem that requires the attention of the therapeutic community. Scientific investigation into the personality char— acteristics of male batterers first appeared in the litera- ture in the 1970s with the work of Ann Ganley (1978). Until then, information about men who batter was obtained through reports of their female victims (Douglas et. al., 1984). Similarly, treatment for the male batterer is a fairly 1 2 recent occurrence. The treatment programs that exist have used a variety of techniques that have included cognitive- behavioral modification, assertiveness training, anger management, communications skill-building, relaxation, and education into nonsexist attitudes and behavior (Saunders & Hanusa, 1984; Pirog-Good & Stets-Kealey, 1985; Barrera et. al., 1987; Farley & Magill, 1988; Edleson & Syers, 1991). Statement of the Problem Although it is common knowledge that anger and rage are often involved in male violence, the importance of shame is not as well understood. It is known that, for some individuals, anger serves the defensive function of short— circuiting anxious feelings of vulnerability (Wessler, 1981; Dutton, 1988). Built into the expression of anger are feelings of energy and potentiation that serve as reinforcement for the further display of angry affects. Given that male batterers in treatment programs often report feelings of power following battering incidents, for a significant subset of the male population that batter their female intimates the use of anger and violence can be a powerful, even addictive, antidote to feelings of emotional vulnera- bility and shame (Walker, 1979; Dutton, 1988). Although shame appears to be essential for protecting the boundaries of the self, and thus normative (Schneider, 1987), when magnified through repetitive infancy and childhood 3 experiences it can lead to the development of a self- structure dominated by or based upon shame (Kaufman, 1989). Thereafter, whenever intense negative emotional states become triggered in such individuals, it becomes necessary for defenses to be erected to negate or contain painful, shame-based emotions. For example, anger and rage may be looked at as protective measures employed to protect the self from the emotional vulnerability that typifies shame. The physical aggression employed by some male batterers against their partners could be considered as an extreme way of defending against feelings of shame. There are multiple etiological pathways leading to an understanding of physical violence used by males against their intimate female partners. The most cogent explanation is that battering is caused by powerful cultural and societal forces (such as patriarchy and male privilege) interacting with individual forces such as learning history, exposure to violent role models, and personality characteristics. This dissertation concerns itself with one such variable, internalized shame, and the quality of attachment experiences that may make the formation of a shame-based personality more likely. If shame becomes magnified in childhood it can lead to the formation of a shame-based self—structure (Kaufman, 1989). Whenever intense emotional states are triggered, some means must be devised by the self to defend against painful emotions associated with shame. These defenses 4 include but are not limited to addiction to chemicals, foods, and violence (Cook, 1991). Behind the development of most addictive disorders we would expect to find magnified shame emotions and the type of early experiences that are known to trigger shame. These conditions may be true for certain types of male batterers. \: Purpose of the Study This study will examine internalized shame emotions in a population of two different "types" of male batterers; those that confine their violence to their immediate families and those that are violent outside of their family as well. Shame will be viewed as a correlate but not a cause of male spousal violence. To gain a better understanding of the types of family of origin experiences that would give rise to both shame and physical aggression, research participants will also be assessed according to the quality of attach- ment behavior shown as an adult (Shaver et. al., 1988; Bartholomew, 1990). This kind of attachment research is fairly recent and has not yet been applied to a pOpulation of male batterers. It involves the theory that children internalize their attachment experiences, and that these experiences-—good or bad--serve as models for future adult relationships. It must be cautioned that this theory assumes a continuity from infancy to adulthood that has not yet been proven to exist. 5 Importance of the Study This study is important in the following ways: 1. This study could help depict a clearer picture of the personality characteristics of the male batterer. Since research has not been fruitful in establishing a homogenous personality profile for male batterers, it has moved toward an examination of discrete personality variables such as self—esteem, assertiveness, and emotional expression as a means of differentiating male batterers from the more normative population. Further differentiation along the lines of internalized shame and models of attachment could prove useful. 2. Results from this study could be helpful to pro- fessionals desiring to understand the male batterer and improve the effectiveness of their treatment efforts. This author's direct experience with group treatment of the male batterer as well as conversations with other treat- ment providers suggests that clinically significant change is difficult to produce and recidivism of violence is high. Professionals must attempt to produce change in their clients against the backdrop of a society and culture often resis- tant or at best lukewarm to the idea of equality of the sexes. Results from this study could lead to the evolution of treatment approaches that are more effective in producing or stimulating desired changes in the male batterer. 3. This study could help shed light on several issues of theoretical interest. Different theories have been advanced 6 to explain wife assault. Straus (1976) viewed the marriage license as a "hitting license", meaning that abuse toward women has its roots in the structure of society and the family, in our cultural norms, and in its sexist organi- zation. Other perspectives explaining wife assault include the idea of generational transmission of violence, abandon- ment anxiety, addiction to violence, deficient anger control skills, and provocation by the victim (Deschner, 1984). It is generally agreed upon by writers in the field that wife assault is rooted in and maintained by sexist, patriarchal, social, cultural, and institutional norms and practices (Hamberger & Hastings, 1989). One position often stated is that male batterers are basically no different than non-assaultive males in terms of their personality characteristics. Batterers may then be seen as oversocialized males or as representatives of one end of the continuum of male privilege and ownership. Yet any systematic theory of why some males batter must also explain why the majority of males do not. It must also provide explanations as to why there are differences in patterns of aggression between what is proving to be emerging types of batterers. This study could be helpful in this regard. <1: Definition of Terms Following is a list of terms designed to add clarity to this study: 1. Wife assault: This will be defined as any act of physical 7 violence by a male toward his intimate female partner. The term "wife" will include the male's female partner, married or unmarried. 2. Male batterer: This will be defined as any male who employs physical violence toward his female intimates. This violence usually occurs along a continuum ranging from mild acts such as grabbing or pushing to severe acts such as hitting with a closed fist or using a weapon. It must be acknowledged that these males may "psychologically" batter their victims as well. This could include behaviors such as verbal coaxing, pleading and coercion, withholding favors or sex, and threatening suicide. As a term used in this study, however, a male batterer must have physically aggressed against his female partner. 3. Shame: Following Morrison's (1989) definition as a feeling of inferiority, inadequacy, or incompetence. Shame is the feeling that one is a failure as a human being and has been exposed to others as a failure. 4. Internalized shame: Defined as feelings of inferiority, inadequacy, or incompetence that have been internalized to become basic to the person's sense of identity. When internalized, shame is experienced as a deep sense of defectiveness. Shame becomes internalized when the child is disparaged, ridiculed, and humiliated without later reparation being made by the one who shames. The self becomes, in effect, shame-based, and whenever a triggering event occurs (such as further ridicule, dis— paragement, or even questioning of one's abilities and 8 competence), the self becomes engulfed by shame. 5. Models of attachment: Modern attachment research has focused on the importance of attachment relationships throughout the life span (Bartholomew & Horowitz, 1991). Infants and children construct mental models of themselves and their major interactional partners. These internal models serve to regulate the person's social behavior during each subsequent relationship. Although varying according to the researcher, these models parallel Ainsworth's (1978) secure, anxious-resistant, and avoidant categories. Overview of the Study This study will examine internalized shame in a popu- lation of male batterers who are enrolled in a group treat— ment program. This will be investigated within the context of attachment theory. Two types of batterers will be differentiated according to whether or not violence is confined to members of their family or generalized to persons outside of the intimate relationship. It is proposed that levels of internalized shame will differ with respect to the type of wife assaulter. Adult attachment styles will also differ according to the type of male batterer. In Chapter Two the review of the literature will be developed. Attention will be given to research addressing the typology of wife assault and the application of shame issues to violent behavior. In addition, theoretical considerations will be taken up and research hypotheses preferred. In Chapter Three the methodology for this study will be discussed. This discussion will be organized around the description of the setting, the description and selection of the samples, instrumentation, procedures for collecting the data, the design of the research, and the analysis of the data. In Chapter Four the results of the data analysis are presented. Preliminary results are discussed, after which the results of the hypothesis testing are presented. In Chapter Five the study is summarized, with the research questions discussed in light of the findings. Theoretical and practical implications of the findings are developed and limitations of the study are addressed. Recommendations for future research are provided. CHAPTER II REVIEW OF THE LITERATURE In this chapter, research relevant to the personality attributes of male batterers will be reviewed. The review will begin by summarizing efforts to identify personality characteristics setting male batterers apart from the normative male population. Following that, efforts toward establishing a typology will be reviewed, with a closer look taken at Saunders' behavioral typology (1987). Next, studies establishing the impact of childhood exposure to domestic violence will be discussed. This chapter will conclude with the presentation of a theoretical per- spective involving shame, attachment, and domestic violence, with a series of research hypotheses based on that perspective and the literature review. ’qi Research into the Characteristics of Male Batterers Ganley and Harris (1978) were among the first researchers to make the observation that men who batter women enter into dependent relationships with their victims. Appearing anxious and jealous, these men go to great lengths to control every aspect of their female partners' lives. The authors also observed that these men had many problems identifying and consequently expressing emotions other than anger, were deficient in relationship-specific 10 11 assertiveness, and employed tactics such as minimization, denial, and blame when referring to their own violent behavior. Other studies support Ganley and Harris' (1978) findings concerning deficits in spouse-specific assertiveness in male batterers. In addition, these studies discovered higher power needs and dependency as well (Davidson, 1978; Dutton & Strachen, 1987). Interestingly, some researchers have suggested that male batterers' attitudes toward women are no more patriarchal or traditional than are the attitudes of nonviolent males (Neidig et. al., 1984; LaViolette et. al., 1985). Although these findings would seem to lend support to an explanation of battering behavior based upon personality characteristics as opposed to an explanation involving only societal and cultural forces, much additional research needs to be done before this issue is resolved satisfactorily. Wetzel and Ross (1983) found the male batterer to be jealous and controlling. They stated that batterers tend to project the blame for their violence, use denial when referring to violent episodes, have an explosive temper, and possess a family history positive for episodes of domestic violence. Bograd (1983) found that batterers are likely to deny the seriousness of their violence and view it as justified. Given that jealousy frequently precipitates acts of spousal homicide (Hilberman & Munson, 1978), support has been found for the presence of severe to pathological 12 levels of jealousy in populations of male batterers (Pagelow, 1981; Barnett & Peck, 1987). There appears to be ample evidence suggesting that men who batter lack insight into their own feelings. They also tend to misperceive both their own and their female partner's degree of dominance in marital conflicts (Barnett & Lindsay, 1985), are more likely to interpret provocations as being the result of their female partner's "annoying" characteristics (Biggio & Brownell, 1986), and resemble experts in persuasion, appearing able to convince their partners to stay with them despite the existence of high levels of violence in the relationship (Schutte et. al., 1988). In addition to the above, Geffner et. al. (1984) observed lower levels of self-esteem, an external locus of control, and poor communication skills in their popu- lation of male batterers. A number of studies examined risk factors specific to this clinical population. Correlates of domestic violence have included unemployment or underemployment (Arias, 1988; Hamberger & Hastings, 1989), high stress level (Barnett et. al., 1983), sex role socialization (Straus, 1973; barnett et. al., 1983), alcohol abuse (Barnett et. al., 1983; Hotaling & Sugarmen, 1986; Hamberger & hastings, 1989; Saunders, in press), and learned history of violence (Ganley & Harris, 1978; Barnett et. al., 1983; Caesar, 1986; Hamberger & Hastings, 1989; Saunders, 1992; Saunders, in press). 13 Psychometric Research: Search for a Unitary Profile In pursuit of a unitary male batterer's "profile" a number of studies have used personality assessments that have included the Minnesota Multiphasic Personality Inventory, the Millon Clinical Multiaxial Inventory, and other lesser known instruments. In one study (Hale et. al., 1988), 60 male batterers given the MMPI scored low in ego strength and high on impulsivity. Subjects also displayed a lack of respect for societal standards, had difficulties with the criminal justice system, conflicts with their families, and low self-esteem. A cluster analysis performed on the data did not statistically support a MMPI profile that was typical for male batterers.- In a replication of an earlier study using the Millon Clinical Multiaxial Inventory, Hamberger and Hastings (1985) concluded that the vast majority of the male batterers tested showed evidence of a personality disorder. No unitary male batterer profile emerged, however. In other studies these same authors found mild support for personality types consistent with antisocial, narcissis- tic, and borderline pathology (1988, 1991). Another study using the same test found that scores clustered about three distinct areas: (1) antisocial and narcissistic pathology, (2) passive-aggressive, avoidant, and borderline pathology, and (3) nonpathological types (Lohr et. al., 1986). The authors concluded that battering is best understood as a product of multiple determinants which may or may not 14 include significant personality pathology. Yet another study using the Millon found significant levels of narcissism and borderline personality pathology as well as concern with autonomy issues and higher needs for affiliation (Cogan & Faulkner, 1988). Allen et. al. (1989) assessed 100 male batterers with the FIRO-B. Males who were abusive were more likely to fall into test categories such as loner, rebel, and pessimist. These men had more difficulty forming relation- ships with others and in expressing needs for intimacy. Results did not indicate high needs to control others. The results of further studies using personality tests are inconclusive in so far as yielding support for a unitary batterer's profile (Rosenbaum & O'Leary, 1981; Bernard & Bernard, 1984; Goldstein & Rosenbaum, 1985; Maiuro et. al., 1986; Hastings & Hamberger, 1988). A review by Faulkner and his associates (1988) described the central difficulty encountered by researchers in determining a unitary personality profile. The authors concluded that male batterers were both pathological and nonpathological along several dimensions. The pursuit of a unitary personality profile for men who batter certainly has not yielded promising results. Toward a Typology of Male Batterers A number of researchers have observed enough hetero- geneity in the characteristics of male batterers to begin 15 to establish typologies (Hamberger & Hastings, 1989). Hanneke and his associates (1981) concluded that batterers are not all members of a distinct and homogeneous group. The authors found that their observed population contained one group of males who were only abusive within their family. These men were fairly well educated, law abiding, and con- cerned about their marriages. This finding contrasted with a group of men who showed violence towards spouse and others outside of their immediate family. These men held hostile attitudes, tended to have more instability in their marriages, and engaged in more severe violence then the other group. Using a modified version of the Straus Conflict Tactics Scale, the MMPI, the Michigan Alcohol Screening Test, and a two hour clinical interview, Caesar (1986) found support for three distinct types of male batterers. One group of subjects produced a MMPI profile containing antisocial and paranoid tendencies, and were labelled "ty- rants". These subjects were depressed, irritable, and suppressed their hostility. Their need for support from their partners was pervasive. Alcohol was a problem for these men. Few of these males were arrested by police more than once. They perceived their parents as limited in their ability to nurture, with the father seen as tyrannical and the mother as ineffectual. All members of this group either observed marital violence as children or were physically abused themselves. As adults, they 16 were seen by their family members as controlling, suspicious, and rigid. The "exposed rescuer" (exposed to violence in their family of origin) showed elements of antisocial behavior on the MMPI as well. This group contained the most alcoholic members. Members of this type also required the most attention from the criminal justice system. Again, all members of this type were exposed to violence in their family of origin. Dependent upon their female partners for their sense of environmental mastery and competence, these men felt unappreciated and exploited by their wives. The "non-exposed altruist" (no exposure to violence in their family of origin) produced a MMPI profile similar to the exposed rescuer. Fewer men in this group had drinking problems. Avoiding confrontation with their female partners, they tended to go out of their way to please them. Most of the males in this group had difficulty with intimacy, creating personal space through their battering behaviors. The difficulty with the above typology is that it is constructed from only 26 subjects total. Another is that the categories are somewhat artificial, given that the central characteristic separating groups appears to be the presence or absence of domestic violence in the family of origin. Finally, Caesar's typology is of little use to the treatment provider who must look at matching the most effective treatment to the appropriate group. 17 Gondolf's (1988) work remedies what he believes are inherent weaknesses in prior efforts to establish a typology of male batterers. Citing various methodological shortcomings such as biased clinical samples, inappropriate instrumentation, and inclusion of questionable variables, Gondolf argues for a behavioral typology that addresses all of the above and provides for a more substantial theoretical approach. For Gondolf, behavioral patterns are more established and less varied than are personality attributes, more predictive of future violence, and more practical to research. Gondolf based his behavioral typology on the self- reports of 525 Texas women admitted to shelters during an 18-month period. The study drew variables from an intake interview that included questions about the male partner's background, his kind of abuse, and range of antisocial behavior. A cluster analysis revealed support for a typology that consisted of three types of batterers: (1) sociopathic, (2) antisocial, and (3) typical. The "sociopathic" batterer is abusive to the extreme toward his wife and children. It is likely that he has used a weapon on his spouse and has been sexually abusive as well. His antisocial behavior is pervasive enough to have resulted in a wide range of arrests for property, violent, and drug or alcohol related crime. The "antisocial" batterer is also extremely abusive. Although also generally violent, his arrest record is 18 either minimal or nonexistent. The "typical" batterer has engaged in less severe abuse than the sociopathic and antisocial types. This individual is more likely to be apologetic following abusive incidents, with his victim being more likely to return to him. His level of general violence and number of arrests are also much lower than the other two types. The pattern of abuse with the typical batterer can either be chronic or sporadic. One type of batterer that is defined for researchers by the criminal justice system are those individuals mandated into treatment by the courts. Court referred male batterers seem to possess higher levels of denial than those volunteering for treatment (Douglas et. al., 1988). This "type" of batterer appears more resistant to the treatment methods currently being employed by treatment programs. Research has yet to focus on court- mandated versus voluntary male batterers. The rationale for seeing male batterers as falling into a dominant (generally violent) versus dependent (family only violence) typology is generated from four key studies. In 1980, Hofeller found indications of dependency among family only violent males, whereas generally violent men displayed a number of dominant traits. Hofeller also found the dependent type to be more remorseful following abusive episodes and the dominant type more 19 likely to have been physically abused as children. Hanneke (1981) investigated behavioral patterns of violence as part of a study examining violent husbands. She and her associates interviewed 85 violent husbands assigned to three groups: "family only" violent husbands, "non-family" violent husbands, and "generally" violent husbands. Distinct differences were found between the family only and generally violent husbands. Overall, family only violent husbands were better educated, listed fewer arrests, and appeared more concerned about their marriages than did generally violent husbands. Violence, as used by members of the family only group, was used as a means of dealing with conflict within the relationship. By contrast, generally violent husbands tended to complain of more instability in the marriage, held violent attitudes, and engaged in more serious violence in addition to illegal activities. Members of this group used violence as a general interpersonal strategy. In an attempt to replicate Hofeller's and other studies dealing with this dichotomy, Saunders (1987) examined 182 males being assessed for admission to a treatment program for men who batter. Subjects were measured on a number of variables that included severity of violence, views of the role of women in society, dominance in decision making, jealousy, and depression. A cluster analysis of the data found partial support for dependent and dominant types of male batterers. Factors 20 attributed to the dependent type were family confined violence, an overall decrease in the intensity and severity of violence, and the ability to express affection. Dominant batterers were more often severely abused as children, vio- lent outside of the home, and involved in a criminal lifestyle. They also endorsed a more traditional view of the woman's role in family and society. Saunders (1992) refined his earlier study by using 165 subjects and postulating three types: family only aggressors, generalized aggressors, and emotionally volatile aggressors. The family only aggressor, or Type One, reported the lowest rate of abuse received as a child, the most marital satisfaction, and the least-assertiveness. A general suppression of feelings may characterize this type, as this group reported low levels of anger, depression, and jealousy and high scores on measures of social desirability. The generally violent male, or Type Two, is more likely to have been severely abused as a child. They were the most frequent users of severe violence (which often occurred in conjunction with alcohol use) and had the highest number of arrests. Their low scores on measures of depression and anger suggests the presence of some sort of affective "numbing" that is perhaps a response to childhood trauma. The emotionally volatile male, or Type Three, is described by Saunders as closely resembling Hofeller's 21 dependent type. Members of this group scored high on measures of jealousy, depression, and anger. Less physical abuse was received in childhood. They also reported holding rigid sex-role attitudes. It is this researcher's view that the literature to this point supports a tentative typology of batterers based on a family-only (dependent) versus generally violent (dominant) paradigm. Saunders' three-fold typology demands careful consideration, although it is likely that the Type Three batterer possesses characteristics similar to the dependent batterer. It is for this reason that this study will advance the view that a two-fold typology is approp- riate given research findings. The findings from the studies discussed above may be summarized as follows: Family-only (Dependent) Generally Violent (Dominant) --less severe levels of abuse --severe levels of abuse towards wife, children towards wife, children (Gondolf, 1988; Saunders, (Gondolf, 1988; Saunders, 1987, 1992) 1992) --concerned about relationship, --violence outside the home, apologetic (Hofeller, "criminal" lifestyle, 1980; Hanneke, 1981; Gondolf higher number of arrests 1988) (Hanneke et. al., 1981; Saunders, 1987, 1992; Gondolf, 1988) Familygonly Type --arrest record much lower (Gondolf, 1988) --dependent traits (Hofeller, 1980) --general suppression of feelings, minimal assertiveness (Saunders, 1992) --emotional volatility, rigid sex-role attitudes (Saunders, 1992) --higher level of education (Hanneke et. al., 1981) 22 Generally Violent Type --higher incidence of severe abuse suffered as a child 1980; (Hofeller, Saunders, 1987, 1992) —-dominant traits (Hofeller, 1980) --more traditional views of women (Saunders, 1987, 1992) —-affective numbing as a response to early child- hood trauma (Saunders,1992) --violence used as a general interpersonal strategy (Hanneke et. al., 1981) --violence is used to deal with conflict in the relationship (Hanneke et. al., 1981) —-more likely to do well in therapy (Cadsky & Crawford, 1988) 23 p/‘, Childhood Exposure to Domestic Violence Before turning to a discussion of shame, anger, and attachment issues, the role that early exposure to domestic violence plays will be discussed. A growing number of studies has implicated childhood exposure to domestic violence as a prominent risk factor. Ganley and Harris (1978) found that 63% of their sample of males were either battered as children or had witnessed marital violence. Other studies have yielded similar or greater percentages (Kalmuss, 1984; Rouse, 1984; Sugarman & Hotaling, 1986; Barnett et. al., 1987; Briere, 1987; DeMaris & Jackson, 1987; Hastings & Hamberger, 1987; Stahly, 1987; Waldo, 1987). Several studies found that males exposed to severe childhood abuse became generally assaultive in adulthood (Fagan et. al., 1983; Saunders, 1992). A study by Rouse (1984) looked at childhood exposure to several types of violence. The most powerful predictor of subsequent domestic violence was observing the perpetrator of family aggression. Other studies have yielded similar results (Kalmuss, 1984; Barnett et. al., 1987). Regardless of whether experienced firsthand or only observed, early exposure to domestic violence appears to have a profound impact on those males who later batter their female intimates. The nature of the relation- ship with the mother may also influence later battering behavior. In a study comparing male batterers with groups 24 containing generally assaultive, maritally conflicted, and happily married males, Dutton (1988) found a correlation between early humiliation and verbal abuse by the mother and later battering behavior in a group of males. Dutton explained that a son is dependent upon maternal nurturance but vulnerable to and humiliated by maternal anger and rejection. Mixed messages from the mother that contain the above elements might be the result of a patriarchal system that has frustrated women and produced ambivalent reactions to their sons. Thus, society and its influence on contemporary family organization has perhaps produced a mother—son relationship that contributes to disparagement and fear of women (Chodorow, 1978). It seems obvious that conditions such as humiliation, verbal and physical abusiveness, and the observation of domestic violence would contribute to the formation of a shame-based personality as well as battering behaviors. That issue will be examined next. Shame and Anger Unacknowledged shame has been proposed to play a role not only in interpersonal conflict but in depression (Lewis, 1981), narcissistic disturbances (Morrison, 1989), anorexia (Scheff, 1987), addictive and codependent relationships (Kaufman, 1989), crime (Lansky, 1987; Katz, 1988), and dysfunctional families (Fossum & Mason, 1986). Indeed, hardly a session of psychotherapy goes by 25 without the appearance of some expression of shame, humiliation, or embarrassment (Wurmser, 1981). Shame has been defined as a feeling of inferiority, inadequacy, or incompetence. It is a sense of personal defect, of failure, of being scorned by others (Morrison,. 1989). When feeling shamed, the self experiences sudden and unexpected exposure; one has failed as a human being (Kaufman, 1985). According to Kaufman (1985), shame originates interpersonally, with one's earliest relationships. The preverbal child experiences his or her shame as a sense of abandonment. The parent has suddenly become un— available to the child in an emotional way. -Later in childhood, shame is manifested through such things as disparagement, ridicule, and instances of humiliation. When internalized, shame is experienced as a deep sense of defect. This belief in one's defectiveness lies at the core of the self and becomes basic to the child's develOping sense of identity. The self becomes what Kaufman describes as "shame-based" when shame is triggered without later reparations being made by the "shamer". After that, whenever a triggering event occurs the sense of shame deepens, causing the self to become engulfed in shame. Morrison's (1989) conceptualization of the formation of a shame-based identity is similar to Kaufman's. Failures in early object relations, such as those caused by active 26 humiliation or disruptions of empathic attunement, lead to shame sensitivity in the later construction of the self. If unrepaired, i.e., if not addressed within the inter- personal relationship by the participants, shame is interna- lized and becomes a part of one's sense of identity. Shame may be followed by any affect, but the most typical are fear, distress, and rage (Kaufman, 1989). Rage seems to be a naturally occurring reaction to shame, serving to insulate the self against exposure and to keep others away. Anger and rage foster an illusion of power and activity, reversing the sense of passivity and helplessness generated by shame (Morrison, 1989). The self attempts to "purge" shame by attacking other objects (persons). A frequent source of shame is need, about which the self experiences a sense of intense humiliation. Shame is generated by either the recognition that others may be aware of one's neediness or the perceived failure of others to meet one's needs. Shame may be a response to this neediness, with anger and rage aimed at the needed person for making the self feel insignificant, dependent, and unworthy. If shame is invoked but not acknowledged, anger or withdrawal is likely to follow (Lewis, 1971). Anger is aroused by indications that the shaming other does not value the self or else imputes damaging thoughts toward the self. The other who shames is almost always seen as 27 a source of hostility by the one who has been shamed. Humiliation, which has been associated with shame, has been found to underlie many homicides (Katz, 1988). A study of a population of assaultive men found them more likely to respond with violence to humiliation than to physical pain (Toch, 1969). Lansky (1987) noted that perpetrators of domestic violence usually feel shamed by the victim, with the resulting violence being an extreme form of self-defense. In the midst of shame, there seems to exist an ambivalent longing that is either expressed as a wish for reunion or for revenge towards whoever shamed us (Kaufman, 1989). Suzanne Retzinger has published a number of studies exploring shame and rage within the context of the marital relationship (1985, 1991). In one study involving the videotaping of marital partners during psychotherapy sessions, she observed that whenever a participant described a situation that caused anger, shame was present also. Retzinger concluded that shame signalled a threat to the relationship, with the threat involving either too much or too little intimacy. The one who is the "shamer" is perceived by the one who is "shamed" as caring less about the relationship. Protracted conflict is marked by social- emotional separation and unacknowledged shame, leading to anger, which in turn is expressed with disrespect, which then leads to further separation. Both research and theory seem to associate at least 28 some degree of anger with shame. The next section will attempt to present the idea that male batterers can be differentiated along both shame and attachment dimensions. (f ; Shame, Attachmentyyand Domestic Violence Bowlby (1988) describes attachment behavior as part of a system evolved to maintain proximity between infants and their caretakers in times of danger or threat. The theory has been advanced by Ainsworth and her associates so that attachment is seen as functioning continuously to provide children with a sense of security (Ainsworth et. al., 1987). On the basis of a structured laboratory procedure, Ainsworth identified three patterns of infant attachment: secure, anxious-resistant, and avoidant. A basic principle of attachment theory is that attach— ment relationships continue to be important throughout the life span (Bowlby, 1980; Bartholomew & Horowitz, 1991). According to Bowlby, children internalize experiences with parental figures over time in such a way that early attach- ment relationships form a model for other relationships outside the family. Infants and children construct mental models of themselves and their major social-interactional partners; these internal models then regulate the person's social behavior throughout her or his life. It has been only recently that the relationship between what Bowlby calls these "working models of attachment" and subsequent adaptation in adults has been investigated (Shaver et. al., 1988; Bartholomew & Horowitz, 1991; Brennan et. al., 1991). 29 The model of adult attachment to be employed in this study is that advanced by Bartholomew and Horowitz (1991). The authors combined the model of the self and the model of the other as conceptualized by Bowlby to describe forms of adult attachment. A person's image of self and image of other can be dichotomized into positive and negative, yielding four combinations of adult attachment. 1. Secure. The secure person possesses a sense of self- worth plus the expectation that others are for the most part accepting and responsive (positive view of self and other). 2. Preoccupied. This person possesses a sense of unworthi- ness combined with a positive evaluation of others. This combination will lead this person to strive for self— acceptance by arranging for others to accept her or him (negative view of self and positive view of others). 3. Fearful-avoidant. The fearful—avoidant person combines a sense of unworthiness with the expectation that others will be rejecting and untrustworthy. This leads to an avoidant style of interacting with others (negative view of self and others). 4. Dismissive-avoidant. This person possesses a sense of self—worth coupled with negative feelings about others. This combination would lead to a detached style of relating to others (positive view of self and negative view of others). In two studies using 146 introductory psychology students as subjects, the authors found support for their model of 30 adult attachment. Subjects completed self—report measures describing interpersonal scenarios based on the proposed types of attachment, as well as structured interviews. Results of their research showed that self—models and models of others are separate, important dimensions of an adult's orientation to close relationships. The two dimensions were found to vary independently. The dimension that will be assessed in this study is the self-model of adult attach— ment. This model of adult attachment differentiates what has been called the avoidant type into fearful-avoidant and dismissive-avoidant. Although both the fearful- avoidant and dismissive-avoidant showed difficulties with intimacy, they differed on measures reflecting self—worth. Only the fearful-avoidant group displayed lack of assertiveness and social insecurity. This categorization seems flexible and better suited to an exploration of attach- ment patterns in battering males, given the differing characteristics of family—only and generally assaultive males. To summarize, it is useful to study shame and attach- ment issues in a population of male batterers. Cook (1991) argues that the psychological roots of many problems, including chemical and other addictive experiences, can be discovered in the internalization of shame and how that is related to attachment issues. If, as Tomkins (1963) has stated, affects are the primary motivators of human 31 behavior, then shame must play a key role in the develop- ment of attachment relationships. Shame certainly has a place in the formation of the person's internalized working model of the self, possibly leading to negative perceptions of the self. As Cook notes, shame is associated with social relationships (within which context attachment issues take place) and is frequently triggered when what Kaufamn (1989) terms "the interpersonal bridge" is broken. Breaking this interpersonal bridge, or rupturing the social bond, as it is called by Retzinger (1991), is the critical event that activates shame. One cannot examine shame without some understanding of the internalized model of attachment that creates the context. Berecz (1992) theorizes that shame is one of the driving forces contributing to assaultive behavior, and that an increase in felt shame may exacerbate violent behavior. If indeed the ability to tolerate shaming is one of the' major determinants of the quantity of violence in a person (Coleman, 1985), then perhaps some measure of internalized shame would be useful in guiding treatment for certain types of male batterers. Since shame makes its appearance early in human relationships, some measure of adult adaptation relating to attachment issues would be equally valuable to assess as well. 32 / Research Questions The following research questions will address issues pertaining to family—only (dependent) versus generally assaultive (dominant) batterers that were raised in the review of the literature. Research hypotheses based on these questions are presented in Chapter Three. 1. Are there significant differences in demographic variables between the above two types of batterers? Based on the review of the literature, significant differences are expected on variables such as level of education, degree of involvement in the criminal justice system, and level of violence manifested toward female intimates and others. 2. As represented by scores on two shame scales, will subjects who have membership in the family-only group evidence higher levels of shame than subjects having membership in the generally assaultive group? 3. As represented by scores on a relationship questionnaire, will subjects having membership in the family-only group of batterers primarily describe themselves as falling within the secure type of adult attachment? Will family-only batterers describe themselves as occupying the preoccupied group, or instead view themselves as fitting into the fearful— avoidant or dismissive-avoidant group? 4. Will persons who have membership in the generally assaul- tive group see themselves as fitting into the secure classi- fication of adult attachment? Will they view themselves as preoccupied, fearful—avoidant, or dismissive-avoidant? 33 5. What is the relationship between levels of shame and categories of adult attachment? Is there a category (or categories) associated with a higher or lower level of shame? 6. What is the relationship between level of abuse and categories of adult attachment? 7. What is the relationship between level of abuse and intensity of shame? Is more abuse associated with higher shame? Is less abuse associated with higher shame? CHAPTER III , METHODOLOGY This chapter presents the operating plan for this study. The study is correlational, although it is hoped that the results will lead to a beginning understanding of the causes of battering behavior in men. The methodology for this study will be discussed under the following sections: description of the setting, description and selection of the samples, instrumentation, data collection procedures, research design, and statistical analysis. The chapter will conclude with the presentation of the research hypotheses. Description of the Setting The study was conducted with the cooperation of a program treating male batterers operating in Fort Wayne, Indiana. This program provides assessment and treatment services to men who have been violent with their female partners. Services are provided to men who voluntarily seek treatment and to those who are ordered by the judicial system into therapy. Assessment interviews are performed by those staff members qualified to execute that duty. The majority of those staff members hold graduate degrees in counseling, social work, and psychology. The program provides group treatment that lasts about six months. Group 34 35 participants who re—offend are usually terminated from the program. Altogether, this researcher contacted four treatment programs during the Fall of 1992 to enlist their partici- pation in the study. Two of the program directors declined to participate in the study, one of them expressing concerns that the study was attempting to show that shame caused battering behavior. The other declined due to lack of staff interest in the subject matter of the study. A program in Michigan agreed to participate, but was only able to produce two subjects. av Description of the Sample The samples in this study were composed of males in treatment for domestic violence issues. Beginning in September, 1993, all clients who completed the assessment process and entered treatment at the research site were asked to participate in the study. Participants were limited to those in treatment for two weeks or less in order to minimize any impact treatment might have had on internalized shame affects. Participation was completely voluntary, with the confidentiality of all the participants protected. Data collected from each participant were identi- fied only by number. Data were collected until there were 50 participants in each of the two sample groups (family—only and generally assaultive). A number of precautions were taken 36 in an effort to minimize nonparticipation and its attendant sample bias. First, the cooperation of all participants was solicited by the program staff. In addition, a cover letter explained to participants the value of the research. Second, research participants were not asked to identify themselves by name and received written confirmation to this effect. Third, completion of the instruments used in this investigation did not take place until after each participant was enrolled in treatment. Fourth, the instru- ments used in this study took no more than 30 minutes to complete. Fifth, some demographic data will be available for those clients who refused to participate in the study. The program routinely collects demographic data on each person presenting for services. These data were used to analyze demographic differences between participants and nonparticipants. Instruments In this section, the instruments that were employed in the study are described. A copy of each instrument appears in the appendices. Demographic Questionnaire This is a two page form created by the researcher for the purposes of collecting demographic information. Variables such as age, treatment history, cohabitation status, occupation, race, and arrest record were collected. Also 37 collected was information pertaining to physical abuse received while a child and the witnessing of domestic violence within the family of origin (see Appendix A). Harder Personal Feelings Questionnaire The Harder Personal Feelings Questionnaire (Harder & Zalma, 1990) is a 22 item research scale measuring both shame and guilt proneness. The instrument possesses a satisfactory test-retest stability (Cronbach's alpha, .78 for shame and .72 for guilt; test-retest, .91 for shame, .85 for guilt). The authors administered this scale to 63 college students enrolled in an introductory psychology course. The age of the sample ranged from 17 to 22 years old. In addition, the authors administered a number of other instruments such as the Beck Depression Inventory, the Rotter Locus of Control Scale, the Social Anxiety Scale, the Public Self-Consciousness Scale, and the Zimbardo Stanford Shyness Inventory. The Personal Feelings Questionnaire showed the ability to correlate successfully with social desirability, self—consciousness, self-derogation, and depression. Subjects must base their responses to the items along four levels of endorsement (from 0: you never experience the feeling, to 4: you experience the feeling continuously or almost continuously). The questionnaire includes items such as "embarrassment", "feeling humiliated", and "feelingchildishfi. Only the items having to do with shame were analyzed (item numbers 1,3,6,7,10,12, 16,17,18,21). (see Appendix B). 38 The Woman Abuse Scale This is a 26 item adaptation of the Straus Conflicts Tactics Scale, a scale that has been widely used in research about incidence and types of violence perpetrated against women (Saunders, 1992). This scale is a measure of the frequency and severity of verbal and physical aggression in a relationship. One of its advantages is the pro- gression it makes from relatively positive behaviors to increasingly aggressive behaviors. A disadvantage of the scale as it is used in this study is the fact that only the male in the relationship completes it, thus making it somewhat prone to social desirability and minimization effects (see Appendix C). The Internalized Shame Scale This is a 35 item experimental scale assessing shame (Cook, 1988). The author administered the scale to three different sample groups. The first group consisted of 603 undergraduates (331 males, 272 females). The next group was an older adult sample of 198 (75 males, 123 females). The final sample was a group of 64 subjects drawn from an outpatient chemical dependency treatment program and a shelter for battered women (equal number of males and females). The internal consistency reliability coefficient for the undergraduate sample was .95. Reliability coefficients for the other samples were .95 and .93, respectively. To provide construct validity for his measure, the author had his subjects complete a problem history and family of origin 39 questionnaire. His measure was found to correlate favorably with a sense of feeling fragile, out of control, empty, and feelings of abandonment. For this scale subjects respond with a choice of five answers ranging from "never" to "almost always". Items include statements such as "I feel like I am never quite good enough", "I feel intensely in- adequate and full of self-doubt", and "I think others are able to see my defects" (see Appendix D). The Personal Reaction Inventory This is a 10 item adaptation of the Marlowe—Crowne Scale of Social Desirability (Saunders, 1991). Since dis- tortion, mostly in the form of under-reporting, is common in research with populations of male batterers, this instru— ment is included in order to make statistical adjustments for social desirability bias. Subjects scoring high on this scale can be removed from the data pool; a more desirable alternative would be to make a statistical adjustment of the sort recommended by Saunders (1988). This procedure would make an adjustment to raw scores based on responses to the Personal Reaction Inventory. The inventory has a reliability coefficient of .91. It has an average correlation with the MMPI clinical scales of .28, and correlates with the Lie Scale of the MMPI, the coefficient being .54 (see Appendix E). Relationship Questionnaire This is a self-report measure developed by the authors for use in their study of adult attachment models 40 ’(Bartholomew & Horowitz, 1991). The subject is asked to read and respond to short descriptions of the four general relationship styles consistent with the attachment categories of secure, preoccupied, dismissing, and fearful. Responses to each style range from "not at all like me" to "very much like me". The order of presentation was varied to avoid order effects. The authors computed alpha coefficients to assess the reliability of this measure in their study. Reliabilities ranged from .87 to .95 (see Appendix I). Procedures for Collecting Data The data collection procedures used in this study were developed with the assistance of the treatment program director in order to maximize subject interest and minimize staff time. The director appointed staff members to solicit participation from the client population. All individuals who agreed to participate in the study were asked to complete the questionnaires. All individuals approached had been in treatment for less than two weeks. Staff members were given standardized instructions for the administration of the questionnaires (see Appendix H). This consisted of the staff member reading aloud the informed consent form and insuring that the willing partici- pant understood and signed the form (see Appendix F). Clients were advised of the confidential nature of their responses and were encouraged to call the researcher if they had any concerns about their participation in the 41 study. Subjects were provided with a stamped mailing envelope addressed to the researcher. Each questionnaire packet contained a cover letter that served as further in- structions to the participants (see Appendix G). Participants were instructed not to place identifying information on the questionnaires. Confidentiality was further insured by having subjects mail questionnaires directly to the researcher, eliminating program staff members from handling the questionnaires after they were completed. It was estimated that 20 to 30 minutes would be sufficient for the completion of the questionnaires. -§ Research Design and Analysis of the Data This investigation was an exploratory, correlational study of internalized shame and adult models of attachment in two samples of male batterers presenting themselves for treatment at a specialized treatment program. The objective of the study was to provide support for differences in shame emotions and adult attachment categories between two proposed types of male batterers: family-only (dependent) and generally assaultive (dominant). The dependent variable or criterion was membership in one of the groups. The dependent variables were subject demographic characteristics, scores on the two shame scales, and scores on the Relationship Questionnaire. The analysis of the data involved the calculation of descriptive statistics and formal hypotheses testing. Univariate analysis of variance (one-way ANOVA), t-tests 42 for independent samples, and chi-square tests were used to test observed group differences relative to the variables of interest. The specific analyses done in connection with the testing of each hypothesis are described in the following section. Hypotheses Hypothesis 1. There will be no difference in the following demographic characteristics between the family—only (dependent) and generally assaultive (dominant) samples of male batterers: 1. age 2. race 3. educational level 4. previous mental health treatment 5. marital status 6. history of substance abuse treatment 7. occurrence of violence in prior relationships 8. frequency and intensity of violent behavior In this analysis, eight demographic variables were analyzed with regard to group membership (family-only or generally assaultive). Group membership was the independent variable; age, race, educational level, previous mental health treatment, marital status, history of substance abuse treatment, occurrence of violence in prior relationships, and frequency and intensity of violent behavior were the dependent variables. The chi—square test of association 43 was done on the categorical data associated with race, education, previous mental health treatment, marital status, substance abuse, and occurrence of violence in prior relation- ships. Age data were analyzed using measures of central tendency. The t—test for independent samples was performed on data associated with the frequency and intensity of violent behavior. Hypothesis 2. There will be differences in the following demographic characteristics between the family-only (dependent) and generally assaultive (dominant) groups of male batterers, with the generally assaultive group showing significantly higher criminal justice system involvement, aggressiveness against others, and receipt of physical abuse and/or observation of domestic violence in childhood. Again, group membership was the independent variable; criminal justice involvement, aggressiveness against others, and receipt of physical abuse or observation of domestic violence as a child were the independent variables. The chi-square test of association was performed on data associated with criminal justice involve- ment, aggression against others, and physical abuse/obser- vation of domestic violence. Hypothesis 3. As measured by scores on the Internalized Shame Scale and the Personal Feelings Questionnaire, there will be no difference in levels of shame between subjects belonging to the family-only group and subjects belonging to the generally assaultive group. Group means for this hypothesis 44 were tested with t-tests for independent samples, with the tests for significance being two-tailed. Hypothesis 4. As represented by scores on the Relationship Question- naire, the family-only (dependent) group of male batterers will show no clear preference for any one adult attachment category (secure, preoccupied, fearful, dismissive). In other words, there will be no association between family- only batterers and adult attachment categories. In this analysis, the rating of agreement with each attachment category was the dependent variable and group membership was the independent variable. Group means were analyzed by t-tests for independent samples. The chi-square test of association was performed to further investigate the relationship between group membership and adult attachment category. Hypothesis 5. As represented by scores on the Relationship Question- naire, the dominant group of male batterers will show no preference for any one adult attachment category. In other words, there will be no association between dominant batterers and adult attachment categories. The t-test for independent samples was used to analyze group mean scores, followed by further analysis with the chi-square test of association. The rating of agreement with each attachment category was the dependent variable, with group membership being the independent variable. 45 Hypothesis 6. There will be no association between levels of shame as measured by scores on the Personal Feelings Questionnaire and the Internalized Shame Scale and the category of adult attachment as measured by scores on the Relationship Questionnaire. Data were analyzed for correlation by Spear- man rank-order correlation coefficients. Hypothesis 7. There will be no association between level of abuse as measured by the Woman Abuse Scale and the category of attachment as measured by the Relationship Questionnaire. Data were analyzed for correlation by Spearman rank-order correlation coefficients. Hypothesis 8. There will be no association between level of abuse as measured by the Woman Abuse Scale and level of shame as measured by the Personal Feelings Questionnaire and the Internalized Shame Scale. Data were analyzed for corre- lation by Pearson Product-Moment correlation coefficients. gja Summary Volunteer subjects for this study were recruited from participants in a treatment program for male spouse abusers located in northern Indiana. Staff members asked subjects to complete a questionnaire packet that contained measures assessing their level of shame, frequency and intensity of violent acts toward their female intimates, and the 46 perception of their own relationship style (attachment category). They were also asked to provide basic demographic information. The results were used to test a number of hypotheses developed to answer several research questions posed about male spouse abusers, shame affect, and category of adult attachment. Correlational tests, ANOVA, and chi- square tests of association were performed to analyze the data. CHAPTER IV ANALYSIS OF THE DATA 1 In this chapter the results of the data analysis are reported. The research questions, around which the investi- gation was developed, are presented first. Next, the research samples are described and the correlations among the main variables are presented. The results of the main analysis of the data consume the remainder of the chapter. Overview of the Researchpguestions This study investigate the relationship between a number of variables and male spouse abusers. A number of research questions were formulated to explore the possibility of the above relationships. First, are there differences between family-only and generally assaultive batterers on demographic variables such as race, educational level, previous mental health treatment, marital status, history of substance abuse treat- ment, occurrence of violence in prior relationships, and frequency and intensity of violent behavior? Second, are there differences between family-only and generally assaultive batterers in demographic variables such as criminal justice system involvement, aggressiveness against others, and observation of domestic violence or being the recipient of physical abuse as children? 47 48 Third, are there differences between the sample groups on levels of shame as measured by the two shame scales? Fourth, are there clear preferences shown by the family- only group for a certain category of adult attachment? Fifth, are there clear preferences shown by the generally assaultive group members for a certain category of adult attachment? Sixth, is there an association between levels of shame as measured by scores on the two shame scales and the cate- gory of adult attachment as measured by responses to the Relationship Questionnaire? Seventh, is there an association between levels of abuse as measured by the Woman Abuse Scale and the category of adult attachment as measured by the Relationship Question— naire? Eighth, is there an association between levels of abuse and levels of shame? Preliminary Analysis of the Data Description of the Research Sample One hundred fifty-four clients of the treatment program agreed to complete and return the instruments used in this study. Of this group, 111 returned their questionnaires, which resulted in a return rate of 72.1%. The high rate of return is thought to reflect the commitment of program staff to this research project. Questionnaires received from 11 respondents could not be used, mostly due to 49 omission of large amounts of information. Only three respon- dents scored high on the measure of social desirability. Their protocols were discarded rather than statistically adjusted. The remaining 100 questionnaires supplied the data analyzed in this investigation. The 100 subjects whose questionnaires were retained for use in the study had the following demographic profile. They ranged in age from 18 to 65, with a mean age of 34.4 years. Most were Caucasian (96%), with African—American and Hispanic subjects comprising the remaining four per- cent of the sample. Married subjects made up 52% of the sample, 14% were unmarried, and 34% were living alone at the time of their participation in the research project. Forty-six percent of the sample graduated from high school, with 36% having at lease some higher educational experience. Subjects were employed in a wide range of occupations. Some were students and some unemployed or laid-off from their place of employment. Fifty-four percent of the subjects were experiencing their first psychotherapy. Approximately two-thirds of the sample has never received professional help for problems related to the level of violence in their relationships. Seventy-nine percent had been arrested, with 50% being convicted of a crime. Most of the subjects (82%) had problems related to violence in their previous relationships. Thirty-six percent of the subjects claimed to have been abused as children, and 60% of the sample claimed to have observed their own parents become violent with one another. Eighteen percent 50 identified themselves as having received treatment for substance abuse. Thirty-seven percent of the men were told by others that they had a problem with alcohol or drug use. Overall, this was a group of men who displayed frequent violent acts toward their female intimates. Most (88%) had pushed, grabbed, or shoved their partner at least once during the past year. Forty-one percent had slapped, kicked, or hit their female partners and 11% admitted to having beaten their spouses. No subjects in the study admitted to having used a weapon on their partners. Description of Nonparticipators Limited demographic data were obtained from the 43 clients who initially agreed to take part in the study but never returned their questionnaires. When separated into dominant and dependent batterers, no significant differences were discovered when they were compared to the participant groups on variables such as age, race, marital status, level of education, arrest history, treatment history, violent acts, and learning history of aggressive behavior. Main Analysis In this section, each hypothesis is restated, followed by the statistical procedures performed and the results obtained. Hypothesis 1: There will be no difference in the following demo- graphic characteristics between the family-only (dependent) 51 and generally assaultive (dominant) samples of male batterers: 1. age 2. race 3. educational level 4. previous mental health treatment 5. marital status 6. history of substance abuse treatment 7. occurrence of violence in prior relationships 8. frequency and intensity of violent behavior Hypothesis 1 was analyzed to determine whether certain demographic variables were related to membership in the family-only or generally assaultive groups. The chi—square test of association was used to test the data associated with each demographic variable. The t-test for independent samples was used to analyze data regarding the frequency and intensity of violent behavior. ’ Hypothesis 1 was rejected at the .05 level of signifi- cance for educational level, occurrence of violent behavior in previous relationships, and frequency and intensity of violent behavior. As displayed in Table 4.1, subjects in the generally assaultive (dominant) group were less educated than were subjects in the family-only (dependent) group. Of the subjects not completing high school, the generally assaultive group contained 77.8% of them. The generally assaultive group also contained less college graduates (4.0% to 12.0%). The generally assaultive group also claimed more problems related to violence in previous relationships 52 with female intimates. Of the subjects responding positively to the existence of violent behavior in prior relationships, the generally assaultive group contained 58.5% of them. Ninety-six percent of the subjects in the assaultive group had been violent in previous relationships, compared to only 68% of the family-only group. The generally assaultive group evidenced a higher frequency and intensity of abuse directed against female intimates. Group mean responses to items on the Woman Abuse Scale are shown in Table 4.2. Hypothesis 1 was retained relative to respondent age, race, previous mental health treatment, marital status, and history of substance abuse treatment. None of the observed differences in these variables was statistically significant. 53 Table 4.1 --Summary of Subject Demographics-- Variable Dominant Group Dependent Group a a Race Caucasian 47 49 Black 2 0 Hispanic 1 1 Education Not Complete High School 14 4 * High School Graduate 24 22 Some College 10 14 College Graduate 2 6 Graduate School 0 4 Mental Health Treatment None 28 26 Once 10 9 Twice 8 9 Three or More 4 6 Marital Status 'Married 25 27 Unmarried 9 5 Living Alone 16 18 Alcohol/Drug Treatment Yes 9 9 No 41 41 Violence in Prior Relationship Yes 48 34 *** No 2 16 54 -—Table 4.1 (Cont'd)-— Variable Dominant Group Dependent Group E H Arrested Yes 44 35 * No 5 15 Convicted Yes 34 16 *** No 8 33 Aggressive To Others Yes 50 0 No 0 50 *** Abused as Child Yes 22 14 No 28 36 Observed Violence as a Child Yes 33 27 No 17 23 Level of Abuse Mean 10.7 6.7 *** SD 3.2 2.8 Note. N = 100 * = p (1.05 ** = 2< .01 *** = p < .001 Table 4.2 --Item by Item Means for Woman Abuse Scale Items-- 55 Item Dominant Dependent 2 Group Group 1. Angered partner 1.00 1.00 .00 2. Refused to talk .68 .80 -1.37 3. Stomped Out .50 .32 1.84 4. Insulted or swore at partner .62 .32 3.12 ** 5. Threatened to leave partner .42 .54 -1.20 6. Screamed or yelled .88 .58 3.55 ** 7. Hit something .48 .34 1.42 8. Threatened to hit partner .38 .36 .21 9. Pushed or grabbed .94 .82 1.86 10. Slapped partner .66 .16 5.84 *** 11. Threw object at partner .28 .20 .93 12. Kicked or hit .60 .22 4.15 *** 13. Hit partner with object .18 .06 1.86 14. Beat up .22 .00 3.72 *** 15. Choked .26 .12 1.80 16. Threatened with gun .20 .04 2.51 * 17. Used gun or knife .00 .OO .00 Note. * = p<.05 ** = p<.01 *** = 2<.001 56 Hypothesis 2: There will be differences in the following demographic characteristics between the family—only (dependent) and generally assaultive (dominant) groups of male batterers, with the generally assaultive group showing higher frequencies of criminal justice system involvement, aggressiveness against other persons outside the intimate relationship, and receipt of physical abuse and/or observation of domestic violence in childhood. Hypothesis 2 was analyzed to determine the relationship between the above demographic variables and membership in the two groups of male batterers. The chi-square test of association, using the Pearson statistic, was used to analyze the data associated with each variable. Hypothesis 2 was rejected at the .05 level of signifi- cance for the variable associated with physical abuse received and domestic violence observed during childhood. As displayed in Table 4.1, slightly more members of the generally assaultive group were abused as children than were members of the family-only group (22 to 14). There was no statistical difference as well between groups on the variable associated with the observation of domestic violence as a child. Hypothesis 2 was retained at the .05 level of signifi- cance for the variables associated with higher levels of criminal justice system involvement and occurrence of aggressive behavior against others outside of the intimate relationship. Subjects having been arrested comprised 89.8% 57 of the generally violent group and 70% of the family-only group. The groups differed even more significantly in terms of convictions. In the generally assaultive group, 81% of the subjects had been convicted of some crime. This compares with only 32.7% of the family-only group members. The two groups differed in terms of physical aggression against non-family members, with the family-only group showing no instances of aggression outside of the intimate relationship. Hypothesis 3: As measured by scores on the Internalized Shame Scale and the Personal Feelings Scale, there will be no difference in level of shame between subjects having membership in the family-only group and subjects having membership in the generally assaultive group. Hypothesis 3 was analyzed to determine the relationship between shame and group membership. Data were analyzed for this hypothesis by testing the group means with t-tests for independent samples, with the tests for significance being two-tailed. Hypothesis 3 was rejected at the .05 level of signifi- cance. The family—only group of male batterers was found to have significantly higher levels of shame than did the generally assaultive group of male batterers. This result was consistent for both measures of shame (Internalized Shame Scale and Harder Personal Feelings Questionnaire). The results are displayed in Table 4.3. 58 Table 4.3 -—Group Means for Two Measures of Shame—- Shame Measures Dominant Dependent p Group Group HPFQ (Shame 1) .68 1.91 -13.32 *** ISS (Shame 2) .96 1.95 -9.35 *** Note. HPFQ ISS Harder Personal Feelings Questionnaire Internalized Shame Scale *** p .001 An item by item analysis on group differences for each shame scale is summarized in Table 4.4 and Table 4.5. 59 Table 4.4 --Item by Item Means for Harder Personal Feelings Questionnaire-— Item Dominant Dependent t Group Group 1. Embarrassment 1.3 2.6 -8.05 *** 2. Mild guilt 1.5 3.1 -9.51 *** 3. Feels ridiculous .2 1.7 -11.93 *** 4. Worry about hurting someone 2.4 1.5 5.61 *** 5. Sadness 2.1 2.6 12.71 ** 6. Self—conscious .9 2.3 -8.76 *** 7. Humiliated 1.2 2.7 -9.31 *** 8. Intense guilt 1.2 2.4 —6.71 *** 9. Euphoria .4 .8 -3.69 *** 10. Feeling stupid .8 1.8 -6.08 *** 11. Regret 1.5 3.3 —11.56 *** 12. Childish .4 1.4 -5.96 *** 13. Mild happiness 3.5 3.1 3.44 ** 14. Helpless 1.1 1.6 -3.61 *** 15. Depression 1.6 1.9 -1.90 16. Blushing .2 1.2 -5.53 *** 17. Deserves criticism 1.3 2.9 —9.40 *** 18. Laughable .2 1.3 -7.09 *** l9. Rage 2.5 1,7 6.16 *** 60 —-Table 4.4 (Cont'd)—— Item Dominant Dependent p Group Group 20. Enjoyment 3.1 3.0 .95 21. Disgusting to SGlf .2 1.3 -8.46 *** 22. Remorseful .7 3.3 -16.85 *** Note. ** = p<.01 ***: p_<.oo1 61 Table 4.5 --Item by Item Means for Internalized Shame Scale-- Item Dominant Dependent 5 Group Group 1. Never good enough 1.4 2.4 -5.16 *** 2. Left out .8 1.7 —5.48 *** 3. People look down 1.5 1.9 -2.72 ** 4. Never measure up 1.4 2.1 -3.59 ** 5. Scold self 1.7 2.7 —5.30 *** 6. Feel insecure 1.1 2.2 -5.10 *** 7. Insignificant .9 1.8 -5.79 *** 8. Inadequate 1.2 2.2 —5.51 *** 9. Defective 1.2 2.4 -5.85 *** 10. Faults revealed .9 2.1 -6.02 *** 11. Painful gap .4 1.9 -8.35 *** 12. Secret parts .4 2.5 -11.60 *** l3. Emptiness .8 1.8 -6.55 *** 14. Not important 1.3 2.1 -5.09 *** 15. Emptiness .8 2.6 —8.15 *** 16. Something missing 1.2 2.0 -4.74 *** 17. Do not know self .8 1.4 —2.91 ** 18. Replay events 1.3 2.2 —4.53 *** 19. Break into pieces .6 1.7 —5.66 *** 20. Lost control/body .3 2.0 -10.52 *** 62 --Table 4.5 (Cont'd)—- Item Dominant Dependent p Group Group 21. No bigger than pea .5 1.2 -4.22 *** 22. Exposed .5 1.3 —4.71 *** 23. Confused by guilt .6 1.6 -5.70 *** 24. Watches self .4 1.3 -4.82 *** 25. Strives for perfection 1.9 2.6 -3.64 *** 26. See defects 1.1 2.4 —6.21 *** 27. Deserve bad things 1.1 2.6 -7.80 *** 28. Watching others .3 .8 —3.71 *** 29. Look directly .8 1.1 -1.58 30. Accept compliment 1.6 2.1 —2.19 * 31. Beat up self 1.5 3.0 -7.21 *** 32. Avoid event 1.8 3.1 -5.39 *** 33. Distress excites .1 .3 -1.69 34. Feel messy/dirty .2 1,0 -5,45 *** 35. Shrink from mistakes .9 2.3 -8.05 *** Note. * = p<.05 ** = p<.01 *** = p<.001 63 Hypothesis 4: As represented by scores on the Relationship Question- naire, the family-only (dependent) group of male batterers will show no preference for any one adult attachment category (secure, preoccupied, fearful, dismissive). In other words, there will be no association between family—only batterers and adult attachment categories. Hypothesis 4 was analyzed for the strength of the relationship between the above variables with the t-test for independent samples and the chi—square test of association. Hypothesis 4 was rejected at the .05 level of signifi- cance. Thirty—seven members of the family-only group responded to the Relationship Questionnaire item requesting them to choose the style that best described their way of relating to others. The preoccupied category of adult attachment was chosen by 54.1% of the family-only group. The fearful category was chosen by 21.6% of the group, while 16.2% chose the secure group. The remaining three subjects picked the dismissive category. The Pearson coefficient for the chi-square test of association was used for this analysis. The family-only group also differed from the generally assaultive group in the extent to which each description corresponded to their general relationship style. Family— only group subjects were significantly less likely to des— cribe themselves as belonging to the secure category than were members of the generally assaultive group.- Family- 64 only members were significantly more likely to see themselves as falling into the preoccupied category than were generally assaultive group members. These differences are summarized in Table 4.6. Table 4.6 —-Group Means for Four Descriptions of Relationship Style—— Category Dominant Dependent 5 Group Group SECURE 3.9 2.8 2.61 * FEARFUL 3.4 3.4 . .00 PREOCCUPIED 3.0 4.5 -3.54 ** DISMISSIVE 2.9 2.2 1.75 Note. * = p<.05 ** = p<.01 *** = p<.001 Hypothesis 5: As represented by scores on the Relationship Question— naire, the generally assaultive group of male batterers will show no clear preference for any one adult attachment category (secure, fearful, preoccupied, dismissive). In other words, there will be no association between generally assaultive batterers and adult attachment categories. Group mean scores were analyzed using the t-test for independent samples. The chi-sqaure test of association was also used to further investigate the relationship between the variables. 65 Only six participants in the generally assaultive group responded to the item in the Relationship Questionnaire asking one's primary approach to close relationships. Two subjects chose the secure category, one chose the fearful category, and the remaining three picked the dismissive category. No subjects described themselves as preoccupied. The generally assaultive group did differ with the family— only group in how they rated each adult attachment category. Those results are summarized in Table 4.6. Hypothesis 6: There will be no association between levels of shame as measured by scores on the Personal Feelings Questionnaire and the Internalized Shame Scale and the category of adult attachment as measured by scores on the Relationship Questionnaire. Data were analyzed for correlation by Spear- man rank—order correlation coefficients. Hypothesis 6 was rejected at the .05 level of signifi- cance. Shame was found to be correlated with category of adult attachment for both shame scales (Harder Personal Feelings Questionnaire, Internalized Shame Scale).. Lower levels of shame were found to correlate with the secure and dismissive categories of adult attachment, while higher levels of shame were found to correlate with the fearful and preoccupied categories of adult attachment. These relationships are summarized in Table 4.7. 66 Table 4.7 —-Attachment Category and Shame--Spearman Coefficients—— Attachment HPFQ 183 Category SECURE -.3461 *** -.4485 *** FEARFUL .1993* .3158 ** PREOCCUPIED .3649 *** .4617 *** DISMISSIVE —.2703 ** -.3546 *** Note. * = p<.05 ** = p<.01 *** = p<.001 Hypothesis 7: There will be no association between level of abuse as measured by the Woman Abuse Scale and the category of adult attachment as measured by the Relationship Question- naire. Data for this hypothesis were analyzed for corre- lation by Spearman rank-order correlation coefficients. Hypothesis 7 was rejected at the .05 level for one of the adult attachment categories. Identification with the preoccupied category of adult attachment was associated with significantly lower levels of abuse. Identification with the other categories (secure, fearful, and dismissive) .did not result in a significant association with levels of abuse. The results are summarized in Table 4.8. 67 Table 4.8 -—Level of Abuse for Attachment Categories—- Categories Spearman Coefficient SECURE .1195 F FEARFUL .1340 $ 4 PREOCCUPIED —.3027 ** i Lt. .. DISMISSIVE .0517 Note. ** = p<.01 Hypothesis 8: There will be no association between level of abuse and level of shame. A correlational analysis was performed on the data obtained from the subjects, with the Spearman rank-order correlation coefficients being reported in Table 4.9 Table 4.9 --Level of Abuse and Level of Shame—- Scale Coefficient HPFQ (Harder Personal Feelings) -.4765 *** ISS (Internalized Shame Scale) -.4260 *** Note. *** = p<.001 68 Summary The research sample of this study consisted of 100 subjects. Most were white (96%) with an average age of 34.4 years. The sample consisted of dependent and dominant male batterers. Each sample group had 50 members. The groups differed on level of education and frequency/intensity of violence. The dominant group showed more criminal justice involvement and occurrences of violent behavior in prior relationships. Although not statistically significant, the dominant group held slightly more members who had been physically abused as children. The family-only (dependent) group of male batterers was found to have significantly higher levels of shame than did the generally assaultive (dominant) group. This result was consistent for both measures of shame employed in this investigation. Family-only group members described themselves as resembling the preoccupied and fearful categories of adult attachment. About 54% of group members chose the preoccu- pied category, while the fearful category was chosen by about 22% of the group. Only six of the dominant group responded to the questionnaire item asking them to choose the category of adult attachment closest to their own relationship style. Three of these subjects chose the dismissive category. Shame was found to correlate with the category of adult attachment. Lower levels of shame were associated with the secure and dismissive categories of adult 69 attachment. Higher levels of shame were associated with the preoccupied and fearful categories of adult attachment. There was an association between level of abuse and category of adult attachment for one of the classifications. Lower levels of abuse were associated with the preoccupied category of adult attachment. There was no association between abuse and the other three attachment categories. There was a significant association between shame and abuse. Higher levels of shame were associated with lower levels of abuse, with lower levels of shame being associated ‘ with higher levels of abuse. CHAPTER V SUMMARY AND CONCLUSIONS Severe assaults occur in 8.7% to 12.6% of marriages and are repeated about two—thirds of the time (Dutton, 1988). The demand created by wife assault on both police and hospital emergency services is considerable and warrants more effective intervention to reduce the substantial social costs involved. Explanations for wife assault began by attributing it to a single factor, such as pathology within the male batterer, biological predisposition toward violence, and the existence of patriarchal norms in the society in general. Empirical evidence now suggests multiple factor or inter- actional causal models. The search for a unitary psychologi- cal profile of the wife assaulter has given way to an emerging typology resting in many respects upon behavioral and per- sonality characteristics. Several matters of theoretical interest were examined by this study. First, demographic information was gathered respective to the typology of male batterers proposed by Saunders (1992). Men who confined their battering behavior to their female intimates, family-only (dependent), were compared and contrasted to those men who exhibited a pervasive pattern of aggression toward others besides their intimate partners. This demographic information was then considered in light of differences observed in other studies. Second, 70 F. be 1‘6 tr in he 71 it was possible to study the relationship between dependent and dominant batterers and shame affects. Shame has emerged as an important consideration in the treatment of many mental health clients, and this study suggests that shame could be a powerful variable of interest in the treatment of wife assaulters. Third, it was possible to investigate the compli— cated relationships between shame and abusive behavior. Fourth, the concept of adult attachment was applied to this two-fold typology of male batterers. Some of the battering behavior observed in these men can be viewed as an aspect of the relationship between working models of attachment and resulting social and emotional adaptation. Fifth, the interaction between those categories of adult attachment, abuse, and shame affects was investigated. The setting for this investigation was a facility for treating male batterers located in Fort Wayne, Indiana. The facility is well established in that community and is recognized for its competent and humane approach to the treatment of the male batterer. Staff members are experienced in both the assessment and treatment process. Staff members hold graduate degrees in counseling, social work, and psychology. Assessment of the male batterer occupies a significant portion of staff time. The sample for this investigation consisted of 100 subjects recruited from participants in the program. White males made up 96% of the sample, with African-American and Hispanic men making up the remaining four percent. Married men comprised 52% of the sample; 34% of the participants 72 were living alone at the time of their participation in the study. This was the initial counseling experience for 54% of the subjects. Most of the men (82%) admitted to problems related to violence in their previous relationships. When assessed as a group for their violent behaviors, 88% of the participants had pushed, grabbed, or shoved their partners. About 12 % of the sample had threatened their partners with a weapon, although nobody had actually used a weapon while assaulting their female partner. The mean age for the sample was 34.4 years. The main variables included in this investigation were frequency and intensity of abusive behavior, shame affect, subject identification of adult attachment categories, and demographic characteristics, which included age, race, marital status, education, criminal justice system involvement, and presence of abuse or domestic violence in the family of origin. The demographic variables in the study were collected on a two page form created by the researcher for this purpose. Shame was assessed by two instruments: the Harder Personal Feelings Questionnaire (Harder & Zalma, 1990) and the Internalized Shame Scale (Cook, 1988). Each instrument was developed for research involving shame affects, and in the case of the Harder Personal Feelings Questionnaire, guilt. The frequency and intensity of abusive behavior was measured by the Woman Abuse Scale, an adaptation of the Straus Conflicts tactics Scale (Saunders, 1992). Social 73 desirability bias was assessed by the Personal Reaction Inventory, an adaptation of the Marlowe-Crowne Scale of Social Desirability (Saunders, 1991). The categories of adult attachment were measured by the Relationship Question— naire (Bartholomew & Horowitz, 1991), a self—report measure developed by the authors to assess perceptions of relation- ship style. The basic design of the investigation was correlational. This design allowed relationships between all variables of interest to be examined to determine statistical signifi- cance and strength of relationship. The design did not permit conclusions regarding cause-and-effect relationships between the variables. The chi-sqaure test of association, ANOVA, and t-tests for independent samples were used to analyze the data obtained from each subject. I); Summary of Findings Dempgraphic Variables It was hypothesized that there would be no differences between family—only and generally assaultive male batterers in relation to the following variables: age, race, edu- cational level, history of previous mental health treatment, marital status, history of substance abuse treatment, occurrence of violence in prior relationships, and frequency and intensity of violent behavior. Although several studies found evidence for the use of severe violence in batterers possessing antisocial tendencies, it is not clear what level of abuse to expect in generally assaultive batterers. 74 No significant differences were found between the groups with respect to age, racial composition, marital status, previous mental health treatment, and history of substance abuse treatment. Differences in educational level, occur- rence of violence in previous relationships, and in the frequency and intensity of abusive behavior were significant at the .05 level. About 28% of the generally assaultive group did not complete high school, compared to only 8% of the family—only group. The family—only group showed a higher percentage of college graduates than did the generally assaultive group, 12% to 4%. The generally assaul- tive group of males showed a significantly higher incidence of relational violence present in earlier relationships. In that group, 96% of the subjects admitted to the presence of violence in previous relationships with female intimates. This compares to 68% of the family-only group claiming vio- lence in prior relationships. The family-only group scored significantly lower on the Woman Abuse Scale than did the generally assaultive group. Approximately 60% of the assaultive men had kicked or hit their female intimates, compared to 22% of the family-only group. Only four percent of the family-only group threatened their partners with a weapon, compared to 20% of the generally assaultive group. It was also hypothesized that there would be significant differences between groups with respect to criminal justice system involvement, aggressiveness (violence) outside of the relationship, and being the recipient or observer of no [\J 75 physical abuse within the family of origin. A significant difference was found between groups on criminal justice system involvement. The generally assaultive group contained more subjects arrested (89% to 20%) and convicted (81% to 33%) than did the family-only group. A significant difference was also found between groups on violence outside the relationship, which is not surprising since this distinction was used to differentiate the two groups. No difference was discovered between groups with respect to being the recipient of physical abuse as a child or in observing domestic violence within the family of origin. Although no differences between groups were found, this dynamic was prevalent within the sample as a whole. Twenty- two members of the generally assaultive and 14 members of the family-only groups were abused as children, with an even greater number observing domestic violence as a child (33 and 27 respectively). Shame and Male Batterers Given the lack of previous research examining shame and wife assaulters, itywas hypothesized that there would C be no difference in level of shame between the family-only and generally assaultive groups of male batterers. There was a significant difference at the .05 level, with the family-only group scoring higher on both measures of shame than the generally assaultive group. Wife Assault and Adult Categories of Attachment It was hypothesized that the two groups would show no difference in their choice of which adult attachment 11., 76 category best fit their relationship style. Only six members of the generally assaultive group responded to the Relation- ship Questionnaire item asking for this choice. On the basis of such a low number of responses, no conclusions could be drawn for the generally assaultive group. The family—only group had 37 members who responded to the question— naire item. Of that number, a significant amount (20) chose the preoccupied category as best fitting their relationship style. The fearful category was chosen by eight, and six subjects chose the secure category of adult attachment. Shame and Attachment It was predicted that there would be no association between shame and the category of adult attachment in this sample of male batterers. There was, however, a relationship discovered, with lower levels of shame being associated with the secure and dismissive categories and higher levels of shame being associated with the fearful and preoccupied attachment categories. Abuse and Attachment It was hypothesized that there would be no association between level of abuse and categories of adult attachment. This hypothesis was retained with the exception of one attachment category. Lower levels of abuse were associated with the preoccupied category of adult attachment, signifi- cant at the .05 level. 7 Abuse and Shame Given the lack of research evidence to the contrary, it was hypothesized that there would be no association 77 between level of abuse and level of shame in a sample of male wife assaulters. This hypothesis was rejected at the .05 level. Results from this study suggest the existence of a moderate relationship between abusive behavior and shame affects. Discussion of Findings In this section, the results of the data analysis are discussed in reference to the research questions introduced in Chapter Two. The discussion proceeds in the order with which results were presented in the preceeding section. Demographic Variables As respected from the review of the literature, no differences were observed with respect to age, race, and marital status between the two samples of male batterers. Both groups were remarkably similar with respect to these variables. The overall percentage of minority subjects was only four percent. There was a significant difference between groups on level of education, with the family-only group being better educated than the generally assaultive group. Approximately 28% of the generally assaultive group did not complete high school, compared to 8% of the family-only group. The family- only group also contained more college graduates as well as subjects receiving some college education. Several researchers have found that, as a whole, male batterers are less educated than are males who do not batter (Hotaling & Sugarman, 1986; Saunders, 1992). This fact is mentioned 78 as a risk factor by Saunders (1992). There was no difference between groups in terms of previous counseling and psychiatric hospitalizations. A little over one-half of each group had no previous mental health treatment. In the generally assaultive group, 18 subjects had received previous help for abusive behavior toward their female intimates; 16 subjects in the family- only group had received prior services for battering behavior. For the entire sample, only 13 subjects had ever been hospitalized for psychiatric reasons. Thus, this appeared to be a reasonable sample with respect to psycho- pathology, supporting several researcher's claims that male batterers are more normal than not (Arias, 1988; Saunders, 1992). It is also possible that the more disturbed clients were referred elsewhere for treatment, this being due to a fairly stringent admission criteria that the treatment program employed. A person who is psychotic, addicted, or clearly dangerous to himself or others is not admitted to the treatment program. Although clinical syndromes such as depression or anxiety are not ignored, the primary focus is on battering behavior and other related domestic violence issues. The two groups showed a significant difference on the variable involving violence in previous relationships. Ninety-six percent of the generally assaultive group admitted to the existence of battering behavior in their previous relationships compared to 68% of the family-only group. 79 Approximately 36% of the subjects were first time offenders, nearly all of them family-only batterers. This result is not surprising considering that the generally assaultive group contains subjects who have a distinctly antisocial component to their personalities. As a group, these men had a higher degree of criminal justice system involvement and a history of violence toward others outside of their immediate families. It is also possible that the dominant group's higher criminal involvement might reflect middle versus lower class inequities. It is reasonable to assume that an individual belonging to the lower class would have less resources with which to avoid arrest or imprisonment. As expected then, assaultive subjects showed a higher incidence of criminal justice system involvement than did subjects in the other group. This finding supports the observations of researchers that dominant batterers show more evidence of a "criminal lifestyle" and a higher number of arrests (Hanneke et. al., 1981; Saunders, 1987; Gondolf, 1988). In the family-only group, 30% of the subjects had never been arrested, compared to only 10% of the assaultive group. Arrests in the assaultive group were for crimes such as assault, assault and battery, theft, breaking and entering, and armed robbery. The dependent group was much more likely to be arrested for the actual assault of their female partner. 80 Arrest and conviction for wife assault seems to have an impact on the rate of recidivism (Dutton, 1988). Of the men who assault their female partners, about two-thirds repeat their abuse if not arrested. If arrested, the rate of recidivism drops to about 20%. It is not known, at present, whether men arrested for wife assault do not repeat because of fear of re—arrest or whether the original arrest serves a didactic function of demonstrating to them that wife assault is unacceptable behavior. Identical numbers of each group of male batterers admitted to problems with substance abuse. Only nine sub- jects from each group stated that they had received treat- ment, although more from each group indicated that their intimate partners believed they had a problem with alcohol (assaultive=16, dependent=21). Alcohol abuse had been associated with severe domestic abuse in a number of studies (Gondolf, 1988; Saunders, 1988, 1992). The number of subjects who admitted to problems with alcohol was lower than one might expect based on the above studies. There are several possible reasons as to why this study had few subjects with substance abuse problems. First, problems with alcohol may have been subject to minimization by the respondents. Respondents may have been uncomfortable disclosing information about their use of alcohol, given that they were participating in a program treating spouse abuse. Second, men with significant substance abuse problems would have been detected during the intake process and referred to an appropriate treatment facility. Although 80 ba IE a1 51; 81 some programs treat substance abuse simultaneously with battering behavior, the facility used in this study will refer the substance abuser to another provider for treatment, allowing the individual to return to the program following successful alcohol or drug treatment. Alcohol can be a convenient scapegoat for men attempting to excuse their battering behavior. Batterers tend to externalize the cause of their assaults by attributing them to their wife's behavior or to alcohol. It is this researcher's experience that many batterers will "not remember" assaults, laying the responsibility on alcohol. Others will state that they never engage in abusive behavior unless they "have a few too many". There were no significant differences between the two groups on either being the recipient of abuse as a child or observing domestic violence within the family. For the entire sample of 100 men, 36 had been abused as children, with approximately 61% of them being members of the assaultive group. About 60% of the entire sample had witnessed their parents engage in violent behavior. A majority of studies found that if the male was abused by his parents, he was more likely to be severely violent in his marriage (Fagan et. al., 1983; Hofeller, 1980; Dutton, 1988; Saunders, in press). A questionnaire assessment by Dutton (1988) discovered that wife assaulters had higher scores than all other men (happily married, maritally conflicted) for both observation of parent-parent violence and for being the recipient of violence from both their parents. Par chi Chi Chi to to EXp of wit sex in not are Won wit bec and no 31; ab. 1e 82 There are several cogent reasons for the above findings. Parents who behave violently toward or in the presence of children are providing role models of behavior that the children readily learn. Children who witness violence not only learn specific, aggressive behaviors but are also likely to acquire the belief that violence is a legitimate way to solve personal problems. They are therefore likely to expect that they will be involved with violence as part of their adult relationships. Furthermore, children who witness violence between adults may develop attitudes and sex-role orientations that predispose them to become involved in violent relationships as adults. Men who saw their mothers being beaten may develop an attitude that women are second-class citizens and deserve to be ill-treated. Women who observed domestic violence may come to identify with their mothers as the victims of aggression. They may begin to see themselves as powerless and deserving of scorn, and may come to see the world as a place where they have no control over what happens to them. The lack of significant differences between the groups on this variable may be due to a number of reasons. First, the question of abuse in the family of origin , as asked in the demographic questionnaire, was vaguely worded. People tend to have their own definitions as to what abusive behavior is and how it is manifested. Many men might regard slapping, spanking , or pushing as disciplinary rather than abusive. The question might have needed to be specifically defined, with examples provided. Second, the similarity 75 gm lrn >3 (I) it Or 83 between generally assaultive and family-only groups of male batterers on this variable may be greater than previous research suggests. A growing body of literature cites the observation of family violence as an important influence on the growing child's perception of human relationships. Perhaps this is a dimension that underlies battering behavior in general and cuts across all types of batterers. Shame and Male Batterers 1 Due to the lack of research on this question, it was J hypothesized that the two groups of batterers would not i show differences in shame as measured by two shame scales. The instruments chosen were research scales that correlated quite well (Pearson = .82). The scales were administered as part of the questionnaire packet given to the research subjects. Shame concerns the whole self. It involves sudden exposure of deficiency in one's own eyes as well as the eyes of the other. In shame the self feels helpless, not in control; the reaction is to hide. The word itself refers to a family of emotions with certain characteristics. It includes many variations, from social discomfort and mild embarrassment to intense forms such as humiliation and mortification. Variations belonging to the shame family include embarrassment, dishonor, disgrace, humiliation, and chagrin. The nature of shame is such that it always involves the self in its relationship to another, even if only in the imagination. Retzinger (1991) suggests that shame plays a major 8! Ti 84 role in the escalation of marital conflict. She believes that it is common knowledge that conflict involves some amount of anger, but the presence of anger itself does not explain why anger sometimes escalates and sometimes does not. Shame is often experienced as an attack coming from the other, and when not acknowledged, perceived as hostility. This creates a type of entrapment that can easily lead to escalation. If intense shame is evoked but unacknowledged, rage is quick to follow. This rage then threatens the social bond that exists between two persons. The inability to acknowledge shame and bondlessness go hand in hand. Shame- bound conflict characterizes rigid relationships; instead of acknowledging shame, one holds fast to a relationship or abruptly severs it. Anger is then seen as a protective measure used to protect the self from the vulnerability that is shame. Viewed within the context of the above, the group of batterers that is less abusive should show higher levels of shame. That is to say, they should acknowledge their shame more readily than males who are more violent. The family—only group of subjects did show higher levels of acknowledged shame. This group differed from the assaultive group on every item of the two shame scales. This difference was in the direction of feeling "more" shame, or feeling it more intensely. The dependent type of male batterer appears more involved in the relationship than does the dominant type of batterer. It might be fair to suggest that the dependent batt rel: socf in : on T the and fac fac abt fr an ti 0f 85 batterer is more sensitive to shifts and changes in the relationship that might signal a potential rupture of the social bond. The dominant batterer may be less invested in a relationship, or only invested in the relationship on his terms. The antisocial tendencies and traits of these males may signal an absence of guilt for their actions and a denial of shame. This unacknowledged shame may in fact act as a catalyst for their anger and rage. Other factors may be at work to translate that rage into physically abusive behavior. It is clear from scores obtained on the shame scales that male batterers, at least in this sample, do express some recognition of shame affect. This recognition is significantly more apparent with the dependent batterers. Wife Assault and Adult Attachment It has been proposed that individuals carry forward from infancy and childhood "representations" of self, others, and relationships that tend to persist across developmental time and mediate behavior and emotions. In the context of attachment theory, Bowlby (1988) refers to these repre- sentations as "working models" of the self and of caregivers and significant others. Bartholomew and Horowitz (1991) have proposed a four group model of attachment styles in adulthood that was used for this research study. Due to the lack of previous research into the attachment styles of male batterers, it was hypothesized that no clear association would exist between male batterers and their choice of adult attachment style as determined by a 86 relationship questionnaire. The generally assaultive sample had only six members who responded to the questionnaire item requesting them to choose the one relationship style that best fit their own way of relating to others. It is not clear why so many subjects from this group failed to respond to the directions of the questionnaire. The directions were clearly written directly underneath bold-faced letters. It is possible that the level of education displayed by this group had an impact on the way they answered the questionnaire in general, and that item in particular. At any rate, it is not possible to determine with confidence this group's preferential relationship style. The family-only group of male batterers had 37 members who responded to the questionnaire item. Approximately 54% of them, or 20 subjects, indicated that their relation- ship style was most like the preoccupied category of adult attachment. Persons in this category are preoccupied with their personal relationships. They tend to reach out to others for the fulfillment of their own dependency needs. Bartholomew and Horowitz (1991) describe this attachment style as a sense of unworthiness coupled with a positive evaluation of others. Persons in this category strive for self-acceptance by gaining the acceptance of valued others in their lives. Eight subjects, comprising about 22% of the family- only group, described themselves as falling into the fearful category of adult attachment. These are people who are af: otl by wi‘ 0t] su‘ ca (ii [8 PI a: b< f i 87 afraid of intimacy and social involvement. They possess a sense of unworthiness coupled with the expectation that others will be untrustworthy or rejecting. The secure category of adult attachment was chosen by six subjects in the family—only group. A basic comfort with relationships along with a positive view of self and others is prevalent with this category. The remaining three subjects chose the dismissive category. PeOple in this category have a basic sense of worthiness and a negative disposition toward people. They tend to avoid close relationships and act independent and invulnerable. The rejection of this hypothesis of no difference seems to suggest that the dependent batterer fits best into the preoccupied category of adult attachment. This takes into account the dependent batterer's closer attention to the relationship and his interest in maintaining the social bond. The family-only male is often more "apologetic" following a battering incident and more persuasive in talking his female partner into returning to the relationship. This repairing of the social bond is essential to the family- only batterer, as it is his partner who fulfills his own dependency needs. The selection of the fearful category by a percentage of the subjects is not surprising, given the fear of intimacy that some batterers express (Barnett et. al., 1987). This, too, is an attachment category that would presuppose relational difficulties among its members. A member of this attachment category would feel a basic sense of lll'iVOl b811t issu 5129 the adul thee disr of s DGa 88 unworthiness, or shame, and would avoid intimacy in the belief that others would be rejecting and indifferent. Only three subjects in the family-only group chose a relationship style corresponding to the dismissive category of adult attachment. This is a relationship style that seems best suited for the assaultive batterer. This is the person who fails to acknowledge his shame, who Retzinger (1991) might describe as "bond—less". r Shame and Attachment l__..__.._.... 4 .o. .hm' There is no literature examining shame and attachment issues in a sample of male batterers. Thus, it was hypothe- sized that there would be no significant association between the above two variables. Given the characteristics of the adult attachment categories and with Retzinger's (1991) theory in mind, it might be expected that the secure and dismissive categories would be associated with lower levels of shame, while the preoccupied and fearful categories would be associated with higher levels of shame. This hypothesis was rejected at the .05 level of significance. The secure and dismissive categories of adult attachment did show lower levels of shame, while the fearful and preoccupied categories showed higher levels of shame. This was true for both shame scales used in the study. The two attachment categories associated with high shame, the fearful and preoccupied, have at their root a basic sense of personal unworthiness. The person views himself as defective, unlovable, and insignificant. They differ 0f as ar Ill Ci gt. 0 l 0" , 89 only in how they view others; the fearful person sees others as untrustworthy, the preoccupied views others as positive, and to be depended upon for acceptance. These two categories may be thought of as shame-based (Kaufman, 1989). The two categories associated with low shame, the secure and dis- missive, have at their root a positive evaluation of the self. These may be shame-resistant personalities, with each differing on how they view others. The secure person will have a positive evaluation of others in his life; the dismissing person will see others in a negative way. Kaufman (1985) writes that when shame is magnified through repeated experiences in infancy and childhood it can lead to the development of a self-structure dominated by shame emotions. The shame-bound personality is equivalent in many ways to the preoccupied and fearful attachment cate- gories. Given the results of this investigation, the persons occupying those two categories are capable of recognizing and expressing shame affect. The dismissive category might represent the ultimate shame-bound personality. Vigorously defending against the intrusion of shame affects, their personalities fend off the painful affects associated with shame. However, this defense contributes to the escalation of anger into rage and violence, which may account for the higher levels of abuse associated with the assaultive batterer. .Ahuse and Attachment It was hypothesized that there would be no association between level of abuse and attachment category as measured ‘ , § 90 by responses to the Relationship Questionnaire. This hypothesis was retained with the exception of the preoccupied category of adult attachment. A correlation coefficient of -.30 indicated a moderate strength of relationship be- tween the two variables. Lower levels of abuse were associated with the preoccupied category. The reason for this result may be found in the nature of the preoccupied person. This is an individual who reaches out to others in an attempt to get them to fulfill his needs. The positive evaluation of the other person, the reliance upon that person to meet dependency needs, and the desire to preserve the social bond might combine to make high levels of abuse unlikely. The batterer who falls into the preoccupied category may be willing to risk much by being abusive, but will be unwilling to risk all by becoming extremely abusive. It could be hypothesized that those men viewing them— selves as relating to others within the dismissive context would show higher levels of abuse than any other attachment category. These men would be expected to value others and relationships less and be more willing to risk higher levels of abuse. Perhaps moderating influences such as threat of punishment or prison work to help lower levels of abusive behavior. - Abuse and Shame It was hypothesized that no relationship exists between frequency and intensity of abuse and shame. This hypothesis 4 em‘m YES was stu of wit ac} Abl acl ac] pe: t01 re; thq fr se In Vi in at 0f is ul me] 91 was rejected at the .05 level of significance. This result was the same for each shame scale used in this research study. The results of this study suggested that for this sample of male batterers, lower levels of abuse were associated with higher levels of shame. Family-only batterers, who acknowledged more shame affect, scored lower on the Woman Abuse Scale than did generally assaultive batterers, who acknowledged less shame. When shame is evoked but not acknowledged, anger is aroused by indications that the other person does not value the self or imputes injurious thoughts toward the self. The shame-denying dominant batterer then reacts with anger and ultimately violence in order to remove the threat to himself. Shame may often be experienced as an attack coming from the other (which may or may not be the case). The self in shame feels like the target of the other's hostility. In spousal abuse, the batterer may feel shamed by the victim's manner, however subtle; the violence he engages in can be seen as a form of self-defense against a perceived attack on the self. The family-only batterer reacts with higher levels of shame and lower levels of violence. Although his shame is more accessible and partially acknowledged, it serves ultimately as an inhibitor of rage and violence. These men do batter, but at a less frequent and severe level. lI-I-C In In 92 Limitations of the Study External Validity Problems of external validity might exist in relation to this investigation due to the use of a volunteer sample. Participants may have differed from nonparticipants in ways that could dilute the generalization of this set of findings to the larger population of male batterers. Although demo- graphic information taken from nonparticipants was not significantly different than information taken from partici- pants, this might not hold true for data with respect to shame, abuse, and categories of adult attachment. Also, wife assaulters represent a diverse target popu— lation. Some treatment programs accept clients that have an active substance abuse or chemical dependency problem. This type of male batterer could differ from the non-addicted batterer in important ways. There might also be important differences between batterers who seek treatment voluntarily and those who are ordered into treatment by the criminal justice system. Some research evidence exists showing that voluntary clients are more amenable to treatment than are those ordered into treatment by the courts (Dutton, 1988; Saunders, 1992). Because of the lack of inclusion of the above groups (or variables), the findings of this study must be cautiously generalized to the target population of male batterers. 93 Design and Methodology Several factors relating to internal validity may have affected the results of this investigation. To begin with, a variety of program staff members introduced and administered the questionnaires to the subjects in the study. Although instructions for administration were standardized, it could not be known whether or not all participating staff members followed them precisely. Optimally, one designated staff member or the researcher should have administered the questionnaires. Another threat to internal validity was the lack of control over the subject's completion of the instruments. Although each subject was strongly encouraged to complete the questionnaires on site, staff members reported that a number of subjects took their questionnaires home with them to complete and mail at their leisure. This introduced the threat that some subjects may have had assistance from friends or female partners in completing the questionnaires. This, of course, would weaken the validity of the results obtained. Implications of the Study Theoretical Implications The results of this investigation support the utility of differentiating male batterers into two distinct types: family-only (dependent) and generally assaultive (dominant). These two types were seen to vary in relation to a number of demographic variables. For example, dependent batterers 94 appear to possess a higher level of education than do domi— nant batterers. They are also less abusive, both in terms of frequency of violent acts and the intensity with which they were delivered. Dependent subjects had less criminal justice system involvement and were convicted of less crimes. The defining characteristic of the dependent batterer is the limiting of violence to the female partner. This group was also less likely to have been violent with previous partners. The results of this study appear to establish shame and shame affects as playing a role in the personality of the wife assaulter. Shame is a construct that can be measured in samples of male batterers. Shame is detectable in higher levels with the dependent batterer and in lower levels with the dominant batterer. A theoretical implication is that shame is more readily acknowledged by dependent batterers and more readily denied by dominant batterers. The acknowledgement of shame may play an inhibiting role in relation to escalation, rage, and violence. To deny one's shame may increase the risk for one to become violently angry in defense of the self. This study discovered a relationship between the level of abuse and the amount of shame acknowledged in this sample of male batterers. The more that shame is acknowledged and przsent in the individual, the less abuse is associated with it. This seems to support theoretical assumptions that shame and anger exist in relationships, and that they exist to protect the self from experiencing potential loss 0) pa 95 of the social bond that exists between one another. Another important theoretical implication is that adult categories of attachment may be useful to study in samples of male batterers. The family-only batterer described his relational style as resembling the preoccupied, and to a lesser extent, the fearful categories of adult attachment. This suggests that this type of batterer values his relation- ship with a female to the extent that he seeks her out for his own self-acceptance. He may be less likely to jeopardize the relationship by engaging in severe levels of abusive behavior. Batterers falling into this category of adult attachment may also be thought of as "shame-based" in that they have a negative perception of themselves, feel inferior, and possess low self-esteem. This study discovered lower levels of shame associated with the secure and dismissive categories of attachment, and higher levels of shame associated with the fearful and preoccupied categories of attachment. This confirms the interpersonal nature of shame: shame is exhibited at a higher level in those categories defined by a negative evaluation of the self and either a positive or negative appraisal of others. Shame is exhibited at lower levels in the cate- gories defined by a positive evaluation of the self and either a positive or negative evaluation of others. Practical Implications The results of this study have implications for the treatment of male batterers. Treatment provides a means through which wife assaulters can learn alternative skills 96 for conflict management, improve their ability to detect and express anger, and have the negative consequences of their violence made salient to them. Batterers tend to express emotions such as hurt, anxiety, excitement, sadness, guilt, humiliation, and helplessness as anger. A refocusing on shame issues could allow for a deeper exploration of issues that lead to violence against others. The evolution of a theory of affect that includes shame has produced some effective treatment approaches (Kaufman, 1985). The distinction that this study makes between two types of male batterers could have important ramifications for treatment programs. It is not unreasonable to suppose that treatment programs could begin to tailor their treatment strategies to best fit each type of male batterer. Dependent; batterers may need a more affective emphasis to their pro- gram, while dominant batterers may require more behavioral or psychoeducational interventions. Results from this study concerning categories of adult attachment could help focus attention on this important variable when initiating treatment of the male batterer. If treatment progressed to the point where the batterer's female partner becomes involved, the man's distinctive style of relating to others could be scrutinized and modified. Bartholomew (personal communication) has developed an attach— ment interview that would be extremely useful in assessing each client's primary relational style more completely. Treatment programs have attempted to cover "all of the bases", providing psychoeducation, stress management, basic 97 assertiveness training, conflict resolution, and cognitive restructuring. The sort of research conducted for this study could be instrumental in sharply focusing treatment strategy to address core issues such as shame. For example, once identified, dominant batterers could be led to an acknowledgement of their shame issues, possibly decreasing their potential for violence. Further research could provide information as to whether or not "too much shame" has a paradoxical effect on treatment. Recommendations for Future Research Replication of this study is encouraged to determine whether or not the results will be the same across other samples of male batterers. Larger samples could be employed, with the inclusion of variables like court mandated treat— ment, income level, and alcohol and drug abuse. It is desirable to repeat the study with substance abusers, as alcohol and drug abuse are present to a large degree in populations of wife assaulters. Cook (1991) argues that addiction to chemicals is, in many ways, a defense against painful shame affects. Addiction to the violence that is prevalent in battering relationships could be a similar defense against shame. Since the testimony of the victim is so important in the research and treatment of the male batterer, future research of this kind should attempt to incorporate that testimony into the design of the study. For example, the Woman Abuse Scale could be administered to the female *n rn n: (n 98 partner, with comparisons made between her responses and his answers. The Relationship Questionnaire could be given to the female partner in order to obtain her perceptions of his relational style. There are advantages to the idea that the researcher should take a direct part in the data collection process. This would allow for the development and employment of a qualitative interview that would expand the Relationship ——-.-- Questionnaire. Bartholomew (1991) has developed just such an interview that could be used for the above purpose. An extension of research with shame scales is also recommended. Shame has been associated with chemical dependency, eating disorders, depression, and obsessive- compulsive disorder, to name a few. Research related to shame could open new avenues into both understanding and treating these disorders. Finally, attachment research should be extended to allow comparisons between groups such as male batterers and a more normative population. For example, the per- centages of batterers choosing among the various attachment categories in this study could have been compared to the percentages found in the population at large as described in Bartholomew's research. It is likely that differences would have been discovered. In conclusion, the present findings suggest that shame and attachment may have utility in understanding the male Wife assaulter. Although exploratory in many respects, this study supports further efforts to examine these processes. APPENDICES APPENDIX A QUESTIONNAIRE INSTRUCTIONS Please complete the following questions by checking, circling, or writing in short answers. Try to be as accurate and thorough as possible. Make your best guess for those items which do not exactly apply to you or which you cannot remember clearly. Please answer all items. Thank you. QUESTIONNAIRE 1. Your age: 2. Your race: White (1) Black (2) Asian (3) Hispanic (4) Native American (5) Other __(6) 3. I am currently: 1) Married to the person I am living with ' 2) Unmarried, living with partner 3) Living alone 4. Your occupation: 5. Your education: - . Did not complete high school 1 2. High school graduate 3. Some college 4. College graduate 5. Graduate or professional school 6. Have you ever sought professional counseling for emotional difficulties? 1. No 2. Once 3. Twice 4. Three times or more 7. Have you ever been so upset that you were hospitalized for emotional difficulties? 1. No 2. Once 3. Twice 4. Three times or more 8. Besides now, have you ever sought help from professional sources for problems related to violence in your relationship? Yes No 99 10. 11. 12. 13. 14. Have you ever been arrested? If yes, please specify how many times and what for: Have you ever been convicted of a crime? If yes, please specify: Have you had problems related to violence in your previous relationships? Yes No g— i : When you were a kid growing up, were you physically { abused? Yes No i Did you ever see your parents become violent toward UL each other? Yes No Have you ever received treatment for a drinking or drug problem? Yes No Does your wife or girlfriend believe you have a drinking or drug problem? Yes No Have you ever been physically aggressive toward any person besides your wife or girlfriend? __Yes __No 100 APPENDIX B PERSONAL FEELINGS QUESTIONNAIRE For each of the following listed feelings, please place a number from 0 to 4, reflecting how common the feeling is to you. "4" means that you experience the feeling all the time, or almost all the time "3" means that you experience the feeling frequently but not all the time "2" means that you experience the feeling some of the time "1" means that you experience the feeling rarely "0" means that you never experience the feeling 1. Embarrassment 2. Mild guilt ____3. Feeling ridiculous ____4. Worry about hurting or injuring someone ____5. Sadness 6. Self-consciousness 7.. Feeling humiliated 8. Intense guilt 9. Euphoria 10. Feeling "stupid" 11. Regret 12. Feeling "childish" l3. Mild happiness ___14. Feeling helpless, paralyzed 15. Depression 101 16. Feelings of blushing 17. Feeling you deserve criticism for what you did 18. Feeling laughable ___19. Rage ___20. Enjoyment 21. Feeling disgusting to others 22. Remorse 102 APPENDIX C WOMAN ABUSE SCALE The following is a list of things you may have done to your partner when you had an argument or at any other time. Please answer "Yes" or "No" as to whether you behaved in each of these ways at any time in your relationship. Also, please write down how often these acts occurred within the past year. 1. I did or said something to anger my partner. . . 2. I sulked or refused to talk about an issue . . . 3. I stomped out of the room or house . . . . . . 4. I insulted or swore at my partner . . . . . . . 5. I made threats to leave my partner . . . . . . 6. I screamed or yelled at my partner . . . . . . 7. I smashed, kicked or hit an object . . . . . . 8. I threatened to hit or throw something at my partner . . . . . . . . . . . . . . . . . . . . 9. I pushed, shoved or grabbed my partner . . . . 10. I slapped my partner . . . . .1. . . . . . . . 11. I threw an object at my partner . . . . . . . . 12. I kicked my partner or hit her with a fist . . 13. I hit my partner with something . . . . . . . . 14. I beat up my partner . . . . . . . . . . . . . . 15. I choked or strangled my partner . . . . . . . 16. I threatened my partner with a knife or gun . . 17. I used a knife or gun . . . . . . . . . . . . 103 891R Cu {VIE APPENDIX D INTERNALIZED FEELINGS SCALE Below is a list of statements describing feelings or experiences that you may have from time to time or for a long time. Try to be as honest as you can in responding. Read each statement carefully and mark the number in the space to the left of the item that indicates how often you feel or experience what is described in the statement. Use the scale below. ~ Scale: "0" - Never "1" - Seldom "2" - Sometimes "3" - Frequently y "4" - Almost always 1 l. I feel like I am never quite good enough. 2. I feel somehow left out. 3. I think that people look down on me. 4. Compared to other people I feel like I somehow never measure up. 5. I scold myself and put myself down. ____6. I feel insecure about others' opinions of me. ____7. I see myself as being very small and insignificant. ____8. I feel intensely inadequate and full of self-doubt. ____9. I feel as if I am somehow defective as a person, like there is something wrong with me. 10. I have an overpowering fear that my faults will be revealed in front of others. 11. I have this painful gap within me that I have not been able to fill. 12. There are different parts of me that I try to keep secret from others. 13. I feel empty and unfulfilled. 14. When I compare myself to others I am just not as important. 104 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. My loneliness is more like emptiness. I always feel like there is something missing. I really do not know who I am. I replay painful events over and over in my mind until I feel overwhelmed. At times I feel like I will break.into a thousand pieces. I feel as if I have lost control over my body functions and my feelings. Sometimes I feel no bigger than a pea. “ At times I feel so exposed that I wish the earth would open up and swallow me. Is I become confused when my guilt is overwhelming because I am not sure why I feel guilty. I seem always to be either watching myself or watching others watch me. I see myself striving for perfection only to continually fall short. I think others are able to see my defects. When bad things happen to me I feel like I deserve it. Watching other people feels dangerous to me, like I might be punished for that. I can't stand to have anyone look directly at me. It is difficult for me to accept a compliment. I could beat myself over the head with a club when I make a mistake. When I feel embarrassed, I wish I could go back in time and avoid the event. Suffering degradation and distress seems to excite me. 105 34. I feel dirty and messy, like no one should ever touch me or they'll be dirty, too. 35. I would like to shrink away when I make a mistake. 106 APPENDIX E PERSONAL REACTION INVENTORY Listed below are a number of statements concerning personal attitudes. Read each item and decide how much each state- ment describes you personally. Place a number to the left to show how strongly you agree or disagree with each statement. "1 u "2" n3" n4" "5" "6" n7" — Agree strongly - Agree — Agree mildly - Neither agree or disagree - Disagree mildly - Disagree - Disagree strongly 1. 2. 10. No matter who I'm talking to, I'm always a good listener. I have sometimes taken unfair advantage of another person. I am always courteous, even to people who are disagreeable. I can remember "playing sick" to get out of something. I am always willing to admit when I make a mistake. I would never think 0 letting someone else be punished for my wrongdoing. On occasion I have had doubts about my ability to succeed in life. Before voting I thoroughly investigate the qualifi— cations of all candidates. It is sometimes hard for me to go on with my work if I am not encouraged. I like to gossip at times. 107 APPENDIX F INFORMED CONSENT FORM I am conducting a doctoral study in the Department of Counseling and Educational Psychology at Michigan State University in East Lansing, Michigan. The study involves men and violent behavior. I want to see if men who are violent toward their female partners differ in certain personality characteristics. This will involve your partici- pation in completing some questionnaires. The data from this project will be used to study those personality characteristics. As I mentioned above, your participation will consist of completing some short questionnaires. These can be completed quickly and only require about 20 to 30 minutes of your time. Your help could result in improved therapy services for clients similar to you. If you decide not to be a part of this study, you will still have access to all the therapy services offered at the Center. All information provided by you will be kept confidential. When I write the results of this study, your name will never be used. You can stop your participation in this study at any time without any problems or interference with your therapy. If you have additional questions after today, please feel free to call Tim Strang at 1-219-481-4880. I, have read this statement and have agreed to participate. Date: Signature: 108 APPENDIX G INSTRUCTIONS TO PARTICIPANTS (cover letter) I am a graduate student from Michigan State University in East Lansing, Michigan, pursuing research toward completion of a doctoral degree in psychology. This packet contains a number of short questionnaires requesting information about yourself and how you have been feeling. Your partici- pation is entirely voluntary and confidential. DO NOT WRITE YOUR NAME ANYWHERE ON THIS PACKET. If you decide to participate, please answer all of the items carefully and honestly. The information gathered from you will be helpful in learning how to provide the best possible treatment for clients with problems similar to yours. It should take between 20 and 30 minutes for you to complete this packet. Remember, do not write your name anywhere on these papers. 109 APPENDIX H INSTRUCTIONS FOR ADMINISTRATION 1. Please read the "Informed Consent" form to your clients as an introduction to the research project. 2. For those clients willing to participate in the project, have them sign the "Informed Consent" form. 3. Each questionnaire packet will be provided with a stamped, self-addressed envelope. Instruct each participant to drop his packet into the mail when he has completed it. The idea is to reduce the amount of valuable therapy time required by this project. If the participant prefers to complete his questionnaires while on a break, he must seal them in their envelope. 4. I would like to make myself available at all times should you have questions that need to be answered. My work telephone is 1-219-481-2880. My home telephone is 1—616-964-0103. Please do not hesitate to call. 110 ._ ___l APPENDIX I RELATIONSHIP QUESTIONNAIRE PLEASE READ DIRECTIONS!!! 1) Following are descriptions of four general relationship styles that people often report. Please read each description and CIRCLE the letter corresponding to the style that best describes you or is closest to the way you generally approach close relationships. A. It is easy for me to become emotionally close to others. I am comfortable depending on them and having them depend on me. I don't worry about being alone or having others not accept me. B. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others. C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. D. I am comfortable without close emotional relationships. It is very important to me to feel independent and self— sufficient, and I prefer not to depend on others or have other depend on me. 2) Now please rate each of the relationship styles above according to the extent to which you think each description corresponds to your general relationship style. Not at all Somewhat Very much Style A. 1 2 3 4 5 5 7 Style B. 1 2 3 4 5 6 7 Style c. 1 2 3 4 5 6 7 Style D. 1 2 3 4 5 6 7 111 BI BLIOGRAPHY BIBLIOGRAPHY Ainsworth, M.D., Blehar, M.C., Waters, E., & Wall, 8. (1978). Patterns of attachment: Assessed in the strange situation and at home. Hillsdale, NJ: Lawrence Erlbaum. Allen, K., Calsyn, D.A., Fehrenbach, P.A., & Benton, G. (1989). A study of the interpersonal behaviors of male batterers. Journal of Interpersonal Violence, 4, 79-89. Arias, I. (1988). Economigy social and psychological resources as predictors of victimization. Paper presented at the 38th annual meeting of the Society for the Study of Social Problems, Atlanta, GA. Barrera, M., Byles, J., Palmer, 8., Brown, R., & Kalahen, S. (1987). Effectiveness of a peer group treatment program for violent men. Paper presented at the 3rd National Family Violence Research Conference, Durham, NH. Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social & Personal Relationships, 1, 147-178. Bartholomew, K., & Horowitz, L.M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality & Social Psychology,§l (2), 226- 244. Barnett, 0., Shimogori, Y., Hibbs, C., & Nicoll, B. (1983). Moody stressyyand temperament factors of wife—abusers. Paper presented at the American Society of Criminology, Denver, CO. Barnett, 0., & Linday, A.R. (1985). Therapist's evaluations of batterer's stresses, moods, and power relations. Paper presented at the meeting of the Western Psychological Association, San Jose, CA. Barnett, 0., Butter, K.E., & Ryska, T. (1987). Abuse of batterers and their exposure to violence during‘child- hood: A preliminary report. Paper presented at the meeting of the 3rd National Conference of Domestic Violence Researchers, Durham, NH. 112 Barnett, 0., & Peck, R.J. (1987). Jealous love in male spouse abusers: A preliminary report. Paper presented at the meeting of the Western Psychological Association, Long Beach, CA. Berecz, J.M. (1992). Understanding Tourette syndrome, obsessive—compulsive disorder and related problems. New York: Springer Publishing. Bernard, J.L., & Bernard, M.L. (1984). The abusive male seeking treatment: Jekyll & Hyde. Family Relations, 33, 543—547. Biaggio, M.K., & Brownell, A. (1986). Gender differences in anger expression: The paradox of victim blame. Paper presented at the meeting of the Western Psychological Association, Seattle, WA. Bograd, M. (1983). Excuses andyjustifications: How abusive couples understand violence. Paper presented at the meeting of the American Psychological Association, Anaheim, CA. Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New York: Basic Books. Brennan, K.A., Shaver, P.R., & Tobey, A.E. (1991). Attachment styles, gender and parental problem drinking. Journal of Social & Personal Relationships, 8, 451-466. Briere, J. (1987). Predicting self-reported likelihood of battering: Attitudes and childhood experiences. Journal of Research in Personality, 2;, 61-69. Cadsky, 0., & Crawford, M. (1988). Establishing batterer typologies in a clinical sample of men who assault their female partners. Canadian Journal of Community Mental Health, 1(2), 119-127. Caesar, P.L. (1986). Men who batter: A heterogeneous group. Paper presented at the meeting of the American Psychological Association, Washington, DC. Charron, T.J. (1991). Domestic violence in the criminal justice system. Paper presented at the meeting of the lst National Conference on Domestic Violence, Las Vegas, NV. Chodorow, N. (1978). The reproduction of mothering: Psycho- analysis and the sociology of gender. Berkeley: University of California Press. 113 Coleman, M. (1985). Shame: A powerful underlying factor in violence and war. Journal of Psychoanalytic Anthropology, 8 (1), 67—69. Cogan, R., & Faulkner, K.K. (1988). Core conflictual relation- ship themes of clients in abusive relationships. Paper presented at the meeting of the Southwestern Psychologi- cal Association, Houston, TX. Cook, D.R. (1988). Measuring shame: The internalized shame scale. In The treatment of shame and guilt (PP- 197- 215). Hawarth Press. Cook, D.R. (1991). Shame, attachment, and addictions: Implications for family therapists. Contemporary Family Therapy: An International Journal, 13 (5), 405-419. Davidson, T. (1978). Conjugal crime: Understanding and changing the wife-beating pattern. New York: Hawthorne Press. DeMaris, A., & Jackson, J.K. (1987). Batterer's reports of recidivism after counseling. Social Casework, 68 (8), 458-465. Deschner, J.P. (1984). A review of current theories about battering behavior. Paper presented at the meeting of the Texas Council on Family Violence, Austin, TX. Douglas, M.A., Alley, J., Daston, A.P., Svaldi-Farr, J., & Samson, M. (1984). Court-involved batterers and their victims: Characteristics and ethnic differences. Paper presented at the meeting of the American Psychological Association, Toronto, Canada. Douglas, M.A., Perrin, S., & Strauss, R. (1988). Differential characteristics of court—referred and non- court referred male batterers: Implications for treatment process and outcome. Paper presented at the 40th annual meeting of the American Society of Criminology, Chicago, IL. . Dutton, D.G., & Strachen, L.E. (1987). Motivational needs for power and spouse-specific assertiveness in assaultive and nonassaultive men. Violence & Victims, 2, 145—156. Dutton, D.G. (1988). The domestic assault of women: Psychological and criminal justice perspectives. Boston: Allyn & Bacon. 114 Edleson, J.L., & Syers, M. (1991). The effects of group treatment for men who batter: An 18-month follow-up. Research on Social Work Practice, 1 (3), 227-243. Fagan, J.A., Stewart, D.K., & Hansen, K. (1983). Violent men or violent husbands? Background factors and situational correlates. In The dark side of families. Beverly Hills: Sage Publications. Farley, D., & Magill, J. (1988). An evaluation of a group program for men who batter. Social Work with Groups, 1; (3), 53-65. Faulkner, K., Stoltenberg, C., Cogan, R., Nolder, M., Shooter, G., & Garvin, R. (1988). Are individuals in physically abusive relationships pathological! non- pathologicaly or both? Paper presented at the meeting of the American Psychological Association, Atlanta, GA. Ganley, A.L., & Harris,L. (1978). Domestic violence: Issues in designing and implementing programs for male batterers. Paper presented at the meeting of the American Psychological Association, Washington, DC. Geffner, R., Patrick, J., & Mantooth, C. (1984). Psychological characteristics of men and women in violent relationships. Paper presented at the meeting of the American Psychological Association, Toronto, Canada. Gelles, R.J., & Cornell, C.P. (1990). Intimate violence in families. Newbury Park: Sage Publications. Goldstein, P., & Rosenbaum, A. (1985). An evaluation of the self—esteem of maritally violent men. Family Relations, as, 425—428. Gondolf,E.W. (1988). Who are those guys? Toward a behavioral typology of batterers. Violence & Victims, 3 (3), 187— 203. Hale, G., Duckworth, J., Zimostrad, S., & Nicholas, D. (1988). Abusive partners: MMPI profiles of male batterers. Jouppal of Mental Health Counseling, 19 (4), Hamberger, L.K., & Hastings, J.E. (1985). Personality correlates of spouse abusers: A cross-validation study. Paper presented at the meeting of the American Psychological Association, Los Angeles, CA. 115 Hamberger, L.K., & Hastings, J.E. (1988). Characteristics of male spouse abusers consistent with personality disorders. Hospital & Community Psychiatry, 33 (7), 763-770. Hamberger, L.K., & Hastings, J.E. (1989). Psychopathology differences between batterers and non—batterers: Psychosocial modifiers. Paper presented at the meeting of the American Psychological Association, New Orleans, LA. Hamberger, L.K., & Hastings, J.E. (1989). Personalipy structure of batterers and non-batterers: Qualitative and quantitative differences. Paper presented at the meeting of the American Psychological Association, New Orleans, LA. Hamberger, L.K., & Hastings, J.E. (1989). Counseling male spouse abusers: Characteristics of treatment completers and drop-outs. Violence & Victims, g (4), 275-286. Hamberger, L.K., & Hastings, J.E. (1991). Personality correlates of men who batter and non-violent males: Some continuities and discontinuities. Journal of Family Violence, 3 (2), 131-147. Hanneke, C.R., Shields, N.M., & McCall, G.J. (1981). Patterns of family and non-family violence: An approach to the study of violent husbands. Unpublished manuscript. Harder, D.H., & Zalma, A. (1990). Two promising shame and guilt scales: A construct validity comparison. Journal of Personality Assessment, 33 (3&4), 729-745. Hastings, J.E., & Hamberger, L.K. (1988). Personality characteristics of spouse abusers: A controlled comparison. Violence & Victims, 3 (1), 31-48. Hilberman, E., & Munson, K. (1978). Sixty battered females. Victimology, 3, 460-470. Hofeller, K.H. (1980). Social, psychological and situational factors in wife abuse. D.A. Abstracts International, 21 (l-B), 408. Hotaling, G.T., & Sugarman, D.B. (1986). Violent men in intimate relationships: An analysis of risk factors. Paper presented at the meeting of the Eastern Psychological Association, New York, NY. Kalmuss, D. (1984). The intergenerational transmission of marital aggression. Journal of Marriage and the Family, fig, 11-19. 116 Katz, J. (1988). Seductions to crime. New York: Basic Books. Kaufman, G. (1985). Shame: The power of caring. Rochester, Vermont: Schenkman Books. Kaufman, G. (1989). The psychology of shame: Theory and treatment of shame-based syndromes. New York: Springer Publishing. Lansky, M. (1987). Shame and domestic violence. In D. Nathanson (Ed.), The many faces of shame. New York: Guilfors. ' LaViolette, A.P., Barnett, 0., & Nylund, L.H. (1985). Wife abuser's predispositions on the Attitudes Towards Women Scale. Paper presented at the meeting of the Western Psychological Association, San Jose, CA. Lewis, H.B. (1971). Shame and gpilt in neurosis. New York: International University Press. Lohr, J.M., Bonge, D., Hamberger, L.K., & Hastings, J.E. (1986). Cluster analysis of personality disorder in spouse abusers. Paper presented at the meeting of the American Psychological Association, Washington, DC. Maiuro, R.D., Cahn, T.S., & Vitaliano,P.P. (1986). Assertiveness deficits and hostility in domestically violent men. Violence & Victims, la 279-289. Marshall, L.L. (1987). Preliminary development of a model of the psychological abuse process. Paper presented at the meeting of the Midwestern Feminist Scholars Conference, Akron, OH. Morrison, A.P. (1989). Shame: The underside of narcissism. Hillsdale, NJ: The Analytic Press. Nathanson, D.(Ed.). (1987). The many faces of shame. New York: Guilford. Neidig, P.H., Friedman, D.H., & Collins, 3.8. (1984). Spouse abuse: Mutual combat or wife battering? Beaufort, SC: Behavioral Science Associates. Novaco, R. (1976). The functions and regulation of the arousal of danger. American Journal of Psychiatry, 133 (10), 1124-1128. Pagelow, M.D. (1981). Woman-battering. Beverly Hills: Sage Publications. 117 Pirog-Good, M., & Stets-Kealey, J. (1985). Male batterers and battering prevention programs: A national survey. Response to the Victimization of Women and Chi;dren, 3 (3), 8-12. Retzinger, S.M. (1985). The resentment process: Video tape studies. Psychoanalytic Psychology, 3, 129-151. Retzinger, S.M. (1991). Violent emotions: Shame and ragg in marital quarrels. Newbury Park: Sage Publications. Retzinger, S.M. (1991). Shame, anger, and conflict: Case study of emotional violence. Journal of Family Violence, 3(1), 37—59. Rosenbaum, A., & O'Leary, K.D. (1981). Marital violence: Characteristics of abusive couples. Journal of Consulting & Clinical Psychology, 43, 63-76. Rouse, L.P. (1984). Models, self-esteem, and locus of control as factors contributing to spouse abuse. Victimology, 3(1), 130—141. Saunders, D.G., & Hanusa, D.R. (1984). Cognitive-behaviora; treatment of abusive husbands: The short-term effects of group therapy. Paper presented at the meeting of the 2nd National Family Violence Research Conference, University of New Hampshire. Saunders, D.G. (1987). Are there different types of men who batter? An empirical study wiph ppssible implications for treatment. Paper presented at the meeting of the 3rd National Family Violence Research Conference, Durham, NH. Saunders, D.G. (1988). Issues in conducting treatment research with men who batter. In Hotaling, G., Finkelhor, D., Kirkpatrick, J., & Straus, M. (Eds.), Coping with family violence: Research and policy perspectives. Newbury Park: Sage Publications. Saunders, D.G. (1991). Procedures for adjusting self-reports of violence for social desirability bias. Journal of Interpersonal Violence, 3(3), 336-344. Saunders, D.G. (in press). Prediction of wife assault. In J.C. Campbell (Ed.), Assessingythe risk of dangerousness. Newbury Park: Sage Publications. Saunders, D.G. (1992). A typology of men who batter: Three types derived from cluster analysis. American Journal of Orthopsychiatry. 62, 264-275. 118 Saunders, D.G. (in press). Woman battering. In R.T. Ammerman & M. Hersen (Eds.), Assessment of family violence: A clinical and legal sourcebook. Mew York: John Wiley & Sons. Scheff, T.J. (1987). The shame-rage spiral. In H.B. Lewis (Ed.), The role of shame in symptom formation. Hillsdale, NJ: Lawrence Erlbaum. Schneider, C.D. (1987). A mature sense of shame. In D. Nathanson (Ed.), The many faces of shame. New York: Guilford Press. Schulte, N.S., Malouff, J.M., & Doyle, J.S. (1988). The relationship between characteristics of the victim, persuasive techniques of the batterer, and returning to a battering relationship. Journal of Social Psychology, 128(5), 605-610. Shaver, P., Hazan, C., & Bradshaw, D. (1988). Love as attachment: The integration of three behavioral systems. In R. Sternberg & M. Barres (Eds.), 1p; psychology of love. New Haven: Yale University Press. Stahly, G.B. (1987). Roots of misogyny: Long:term effects of family violence on children's attitudes. Paper presented at the meeting of the Association of Women in Psychology, Denver, CO. Straus, M. (1973). A general systems theory approach to a theory of violence between family members. Social Science Information, 43(3), 105-125. Straus, M. (1976). Sexual inequality, cultural norms, and wife-beating. Vicimology, 1(1), 54-70. Straus, M., Gelles, R.J., & Steinmetz, S.K. (1976). Violence in the family: An assessment of knowledge and research needs. Paper presented at the meeting of the American Association for the Advancement of Science, Boston, MA. Straus, M. (1977). Normative and behavioral aspects of violence between spouses: Preliminary data on a nationally representative USA sample. Paper presented at the meeting of the Symposium on Violence in Canadian Society. Sugarman, D., & Hotaling, G.T. (1986). An analysis of risk markers in husband to wife violence: The current state of knowledge. Violence & Victims, 3(2), 101—124. 119 Toch, H. (1969). Violent men: An inquiry into the psychology of violence. Chicago: Aldine Publishing. Tomkins, 8.8. (1963). Affectlimagerylponsciousness. Volume 2: The negative affects. New York: Springer Publishing. Waldo, M. (1987). Also victims: Understanding and treating men arrested for spouse abuse. Journal of Counseling and Development, 33(7), 385-388. Walker, L. (1979). The battered woman. New York: Harper & Row. Wessler, R. (1981). So you are angry--now what's your problem? Rational Living, 33(1), 29—31. Wetzel, L., & Ross, M. (1983). Psychological and social ramifications of battering: Observations leading to a counseling methodology for victims of domestic violence. Personnel & Guidance Journal, 33(7), 423-428. Wurmser, L. (1981). The mask of shame. Baltimore: Johns Hopkins University Press. 120 nICHian STATE UNIV. LIBRARIES illWillllllll“WWII“WI”llllHlWWllNlWM 31293010259053