Citation
Determinants of battered women's destination following a shelter experience

Material Information

Title:
Determinants of battered women's destination following a shelter experience
Creator:
DeMark, Joanne F., 1949-
Publication Date:
Language:
English
Physical Description:
x, 145 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Authoritarianism ( jstor )
Battered women ( jstor )
Child psychology ( jstor )
Dogmatism ( jstor )
Locus of control ( jstor )
Personality psychology ( jstor )
Shelters ( jstor )
Social psychology ( jstor )
Violence ( jstor )
Women ( jstor )
Abused wives -- Southern States ( lcsh )
Abused women -- Southern States ( lcsh )
Counselor Education thesis Ph. D
Dissertations, Academic -- Counselor Education -- UF
Domestic relations -- Southern States ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references (leaves 128-143).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Joanne F. DeMark.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
023247939 ( ALEPH )
19045574 ( OCLC )

Downloads

This item has the following downloads:


Full Text












DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE
















By

JOANNE F. DeMARK


















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFIL114ENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1988





























This dissertation is dedicated in loving memory to my parents, Assunta (Susie) Cini DeMark and
Charles Domenick DeMark, who inspired and nurtured
my educational dreams.














ACKNOWLEDGEMENTS

I extend my appreciation to everyone whose assistance enabled me to complete this manuscript. The encouragement, love, support, and assistance from my committee, intimate friends, and family have been mainstays during my pursuits.

My committee members' attention, direction, and ideas

have been extremely valuable. Dr. Jaquelyn Liss Resnick, my committee chair, ably guided my process, especially assisting me with the feminist and academic scholarship parameters to which I aspired for my work. She was supportive while expecting quality work. Dr. Phyllis Meek provided consistent encouragement and confidence in my abilities throughout my university and community work with her. Dr. Sandra Damico and Dr. Harry Grater offered excellent review of my research and offered important ideas and suggestions for the work. I have enjoyed and benefitted from shared enthusiasms with my committee: feminist theory and scholarship with Drs. Resnick and Meek; psychodynamic theory and practice, specifically Alice Miller's ideas with Dr. Grater; and the joys and intricacies of quantitative and qualitative research with Dr. Damico. Their time, expertise and scholarly review of my work have been greatly appreciated.




iii








In addition to my doctoral committee, Dr. Joe Wittmer and John Dixon provided notable academic and consultative assistance for which I am grateful.

I am thankful to the battered women at the shelters who volunteered their participation in my study and to the staffs at the shelters who agreed to assist me, adding to their already lengthy and stressful responsibilities.

I appreciate my many friends and family, Deb Vingle, Dayna Buskirk, the late Sally Briegal, Margaret Levings, Charlene Shackle, Susan DeMark, Rosemary DeMark, Ann McKain, my group, Lisa Schultz, Ron Remillard, Susan Myers, Kris Billhart, Sarah Boykin, John Bryant, Carol Samuels, Norma Calway-Fagan, Gerri Green, the Dunns, Fleury Means, and others whose kind hearts and nurturing souls have been sustenance upon which I could draw.

Above all, I wish to thank my closest allies and intimates, Phyllis DeMark and Kay Leigh Hagan. The encouragement, support, processing, patience, understanding, scholarly support, spiritual and emotional scrutiny, honesty, nurturing, and love from these two women have been inspiriting for me.














iv















TABLE OF CONTENTS

PAGE

ACKNOWLEDGEMENTS . . . . . . . . . iii

LIST OF TABLES . . . . . . . . . . viii

ABSTRACT . . . . . . . . . . . ix

CHAPTERS

I INTRODUCTION . . . . . . . . 1

Statement of the Problem . . . . . 4
Purpose of the Study . . . . . . 4
Rationale for the Study . . . . . 5
Research Questions . . . . . . . 10
Significance of the Study . . . . . 11 Definition of Terms . . . . . . 13

Battered Woman . . . . . . 13
Battering . . . . . . . . 14
Shelter : I I I * I * I * * 14 Destination . . . . . . . 14
Returning to the Abuser . . . . 14 Not Returning to the Abuser . . . 15 Locus of Control . . . . . . 15
Resources . . . . . . . 16
Number of Previous Separations . . 16 Authoritarianism . . . . . . 16
Monitoring . . . . . . . 17

Organization of Study . . . . . . 17

II REVIEW OF THE LITERATURE . . . . . 18

Emergence of Conjugal Violence . . . . 18

The British Refuge Movement . . . 19 The Women's Rights Movement in the U.S. 21 Police Involvement . . . . . 22
Study of Societal Violence . . . 24






v









Theoretical Models . . . . . . 26

General Systems Theory . . . . 27 Biological Theory . . . . . . 29
Exchange/Social Control Theory . . 30 Feminist Theories . . . . . . 31
Social Learning Theory . . . . 35 Intrapsychic Theories . . . . . 38

Variables Relating to the Battered Women's
Destination . . . . . . . . 43

Frequency and Severity of the Abuse . 44 Abuse in Family of origin . . . . 44
Length of Marriage, Previous Separations,
Religious Affiliation . . . . 45
Love, Affection, Hope . . . . . 48 Self Esteem . . . . . . . 49
Resources/Occupational History . . 51 Children . . . . i i . . 53
Locus of Control, Learned Helplessness 55

Authoritarianism and Monitoring . . . 57

Authoritarianism . . . . . . 57
Monitoring . . . . . . . 59

III METHODOLOGY . . . . . . . . 62

Population and Sample . . . . . . 62
Sampling Procedures . . . . . . 65
Instruments . . . . . . . . 66

Measure of Resources . . . . . 68 Measure of Locus of Control . . . 68 Measure of Authoritarianism . . . 71 Measure of Monitoring . . . . . 72

Data Collection Procedures . . . . . 73 Data Analysis . . . . . . . . 78

IV RESULTS . . . . . . . . . 80

Demographic Information . . . . . 80
Analysis of Hypotheses One Through Five . 85 Analysis of Hypothesis Six . . . . . 88 Summary . . . . . . . . . 88

V DISCUSSION . . . . . . . . . go

Summary . . . . . . . . . go
Limitations of the Study . . . . . 91
Discussion of Results * ' * * 93 Recommendations for Future Resear h . . . 99

vi









APPENDICES

A DEMOGRAPHIC INFORMATION QUESTIONNAIRE . . . 103 B LETTER OF INTRODUCTION . . . . . . . 105

C MEASURE OF RESOURCES . . . . . . . 107

D LEVENSON LOCUS OF CONTROL SCALES . . . . 109 E DOGMATISM SCALE (FORM E) . . . . . . 112

F SELF-MONITORING SCALE . . . . . . . 117

G LETTER OF INSTRUCTION/INFORMED CONSENT LETTER . 120 H COMPARISON DEMOGRAPHICS . . . . . . 123

I COMPARISON NORMS FOR LOCUS OF CONTROL,
AUTHORITARIANISM, AND MONITORING . . . . 126 REFERENCES . . . . . . . . . . 128

BIOGRAPHICAL SKETCH . . . . . . . . . 144


































vii















LIST OF TABLES

Table PAGE

4-1 Age of Sample . . . . . . . . . 81

4-2 Race of Sample . . . . . . . . 82

4-3 Number of Children . . . . . . . . 83

4-4 Previous Separations . . . . . . . 83

4-5 Population of City in Which Subject Resided
Before Shelter . . . . . . . . . 84

4-6 Means and Standard Deviations of Study Measures 86 4-7 Logistics Regression Model of the Relationship
Between Destination and the Research Variables . 89 H-1 Comparison Demographics from Other Battered
Women Studies . . . . . . . . . 123

I-1 Locus of Control Comparison Norms . . . . 126 1-2 Authoritarianism and Monitoring Comparison Norms 127























viii














Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE

By

JOANNE F. DeMARK

April 1988

Chair: Dr. Jaquelyn Liss Resnick Major Department: Counselor Education

The purpose of this study was to examine seven potential determinants of battered women's destination following shelter experience. The variables were drawn from previous research and the psychodynamic theory of Alice Miller. It was hypothesized that battered women who did not return to their abusers would be characterized as less authoritarian, less monitoring of others, have less chance and powerful others locus of control, greater internal locus of control, greater resources, and have left their mates more often previous to this shelter visit than battered women who returned to their abusers.

Questionnaires were completed by 72 women during their stay at one of five southern U.S. shelters, three in Georgia and two in Florida. Of the sample, 30 returned to their abusers and 42 did not.




ix









The results indicated no significant differences (p > .05) between those women who had not returned to their abusers and those women who had. A logistic regression analysis identified no significant relationship between the determinants and battered women's destination.

Selected demographics also were collected and, where potential effects were suggested, chi-square analyses were conducted. Nonwhite battered women were less likely to return to batterers than were their white counterparts. Also battered women from small towns were more likely to return to cohabitation with the batterers, and as population of residence previous to sheltering increased, battered women were less likely to return to their mates. The researcher recommended development of specialized instrumentation to operationalize Miller's theory and further investigation of (a) race and city size of battered woman's residence for relationship to destination, (b) battered women's perceptions about authority, and (c) battered women as a heterogenous grouping.


















x















CHAPTER I

INTRODUCTION

In the past decade, domestic violence, particularly

spouse abuse, has been brought increasingly to the public's attention. What was once considered an unspoken family issue, condoned and encouraged by society, is now deemed a social problem (Dobash & Dobash, 1971; Roy, 1977; Stacey & Shupe, 1983; Straus, 1977-78). Considered as one representation of violence against women, spouse abuse was initially addressed through the impetus of the feminist movement. In an effort to better understand, treat, and ameliorate the problem of conjugal assault, a host of disciplines and groups, including helping professionals and feminists, have undertaken the charge to research the problem (Fleming, 1979; Hansen & Barnhill, 1982; Hilberman, 1980; NiCarthy, Merriam, & Coffman, 1984; Walker, 1979, 1984). This involvement has led to a focus on providing safety and treatment for spouse abuse victims: the battered mates and the children who reside in settings where violence occurs.

An indication of the extent of the problem is the

number of individuals affected by domestic violence. From one national survey of violence in American homes, researchers reported that in one household out of six, a


I






2

spouse has committed an act of violence against his or her partner in the past year (Straus, Gelles, & Steinmetz, 1980). The researchers in that survey concluded that an American's greatest risk of being assaulted, injured, or murdered occurs in one's own home by a family member. Forty percent of all female murder victims are killed by their husbands (Dobash & Dobash, 1977-78, 1979). Beatings constitute the most prevalent method of wife murder (Fields, 1977-78).

Conjugal abuse affects not only the marital dyad, but also the children and family unit. When violence occurs in the family, children are seeing models for how to handle relationships and disagreement. A cultural tradition exists to utilize hitting as punishment to curb unwanted behavior. Theoretically, children not only learn to curb behaviors through this model; realistically, they also learn that violence and love are linked, that violence is morally right, and that violence is justifiable when something is really important (Gelles, 1977). Thus, an attitude that violence can be exercised for "the good" of the recipient is promulgated (A. Miller, 1983). These lessons in childhood are transferred to the context of other social relationships, and violence in families and relationships becomes a way of life (Gelles, 1977; A. Miller, 1983, 1986). The intergenerational cycle theory of violence, i.e., that children who are recipients of violence will grow up to be perpetrators of violence, has been validated repeatedly (cf.







3

Bakan, 1971; Gil, 1970; Gillen, 1946; Maurer, 1976; Palmer, 1962; Steele & Pollack, 1974; Welsh, 1976).

Also, domestic violence has significant costs for

society. As examples, public and private sector funds are used to pay for safe shelters, counseling, police intervention, legal avenues, and other resources. Police face greater injury and death at the scenes of domestic violence than at any other crime scene at which they intervene (Bard & Zacker, 1974).

Ratios of wife battering compared to husband battering range from 11:1 through 13:1 in the United States (Gaquin, 1977-78; Levinger, 1966; Steinmetz, 1977-78). Straus, Gelles, and Steinmetz (1980) supported social policy, attention, and treatment specifically on wife abuse for the following reasons: (a) Wife battering involves more dangerous and injurious forms of violence, and it results in greater physical damage and injury; (b) male abusers are much more often repeaters; (c) wives often act in selfdefense; (d) much abuse by the husband occurs when the wife is pregnant, posing additional danger to the fetus; and (e) the mores, laws, and traditions of our society lock women into marriages in a more substantial way than men.

Wife abuse is a chronic crime which escalates in

severity and poses a serious threat to the safety of the women involved (Pagelow, 1977a, 1977b, 1981; Walker, 1979). Battered women commonly report receiving murder threats from their abusers, as well as perceiving the batteries,






4

capabilities to kill them (Walker, 1979). The violence in wife abuse is often excessive and relentless, with beatings continuing after the victims are either unconscious or dead (Okun, 1986; Walker, 1979; Wolfgang, 1958). The severity of the threat to abused women's lives is substantiated by the statistics on women who are murdered by their partners (Dobash & Dobash, 1977-78, 1979). Many more women seek safe shelter than abusers seek treatment for their problem (Fleming, 1979). Batterers find it extremely difficult to acknowledge their behavior as a problem or to take responsibility for the outcome of their brutality (Walker, 1979). The chances are quite slim that battered women who return to their households will experience an improved conjugal relationship with less threat of violence (Pagelow, 1981). Concern for the safety of victims is of utmost importance.

Statement of the Problem

Greater theoretical understanding of wife abuse is

needed. one aspect of this, whether the battered woman will return to the abuser or not following a shelter stay, is important, especially with the strong likelihood that returning to the battered means continuation of the woman's abuse. Theoretical explanations have yet to uncover fully the factors related to battered women's destinations after shelter stays.

Purpose of the Study

The purpose of this study was to reexamine key factors from previous empirical and theoretical research, in







5

combination with factors from a new theoretical perspective, as potential indicators of battered women's destination, i.e., not returning to the abuser or returning to the abuser, following a shelter stay.

Rationale for the Study

In several studies researchers have examined various factors related to women remaining in or leaving a relationship where they are battered, though the researchers did not formulate global theories which guided their choice of factors. For example, in early research frequency of abuse and severity of the abuse were found to be related to the battered woman leaving the relationship (Gelles, 1976); however, all investigators since then have failed to empirically resubstantiate that finding (cf. Pagelow, 1977c, 1980, 1981; D.K. Snyder & Scheer, 1981). Abuse in the woman's family of origin (Pagelow, 1977c, 1980, 1981), the age of children, and number of children (Cristall, 1978; D.K. Snyder & Scheer, 1981) are not related to battered women's exit from a battering relationship. D.K. Snyder and Scheer (1981) found that a higher number of previous separations, shorter length of marriage, and a religious affiliation other than Roman Catholic were indicative that a battered woman would not return to the abuser after a shelter stay (D.K. Snyder & Scheer, 1981). The length of the relationship was found not to be a significant variable relating to dissolution of a battering relationship in another study (Okun, 1986). Of these factors, many of which






6

are demographic, previous separations and Roman Catholic denomination have proved to be significant. D.K. Snyder and Scheer (1981) suggested that perhaps religiosity or authoritarianism would be more revealing than denomination. Neither previous separations nor denomination is a foolproof method to determine likelihood of remaining in or leaving the relationship.

Others (Gelles, 1972, 1976; Martin, 1976; Pagelow, 1981) have proposed that women who are able to leave battering relationships are those who are economically independent from the abuser. Several researchers (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981) have empirically validated this proposition; financial resources available to a woman have served as predictive indicators of her remaining in or leaving a battering relationship. Although the availability of resources is a key factor, it does not always explain who extricates herself from an abusive marriage versus who remains in such a marriage. Some battered women in shelter programs or therapy have faced tremendous economic and social hardship, at times with minimal career skills, yet they left their abusers and established themselves and their children in new locations on their own (Fleming, 1979; Martin, 1976; Stacey & Shupe, 1983).

Several theories have been proposed to explain whether a woman will remain in or leave an abusive relationship. One theory for the battered woman's entrapment in an abusive







7

relationship is that a patriarchal society perpetuates abuse and coercive control of women (Brownmiller, 1975; Dobash & Dobash, 1979; Schechter 1982). Rights of chastisement accorded to men as legal property owners of women and children resulted in the legal and social sanctioning of male violence against women and children. Patriarchal values, such as traditional gender identity development which supports the goals of individuation and separation for males and attachment for females, reinforce women staying in relationships unsatisfying to them (Gilligan, 1982). Women view success of the spousal relationship as their primary responsibility, and the definition for success is dictated by the male heads of the household in a traditional family hierarchy (Dobash & Dobash, 1981). Consequently, battered women are conditioned within the patriarchal system to maintain relationships as long as their husbands and society deem them important.

Although the characteristics of both batteries and

battered women very often include traditional values about families and stereotypic ideas regarding gender roles, not all abusers and abused women fit in this category (Walker, 1979). In fact, research to distinguish between women who remained in abusive relationships via the variable of traditional ideology has not supported that more traditional women remain in abusive relationships (Pagelow, 1981). Although the traditions of a patriarchal society contribute to an environment "ripe" for wife abuse, the patriarchal






8

legacy theory does not completely define the determinants of women's destination.

The most relied upon theoretical explanation regarding battered women is that some women experience a greater degree of learned helplessness than others, perceiving themselves as powerless to alter their situations for the better, regardless of what they do (Walker, 1979). The battering cycle of tension building, acute incident, and loving contrition powerfully reinforces the women's learned helplessness. Walker (1984) expected that battered women who remained with abusers would perceive themselves as having less control over their situations than those who did not remain with abusers, because they perceived others as exercising greater control over them. Instead, the findings indicated that battered women in or out of battering relationships saw themselves as having a great deal of control over what happens in their lives, more so than a general population norm. They did not view powerful others, e.g. the abuser, as having a lot of control over their lives (Walker, 1984).

Additional investigation of potential determinants of

the destination of domestic violence victims following their shelter stays is needed. A potential source for additional variables is A. Miller's (1983) theoretical framework for violence in society. She has proposed that violence in society occurs because parents and others, exercising the prevalent child rearing practices of the past several






9

centuries, have shown neither validation nor respect for the feelings, needs, and desires of children. This "spare the rod, spoil the child" tradition supports the authority of the parent, whether right or wrong. A. Miller (1981, 1983, 1986) labeled the training of children to meet consistently the parents' needs and respond without question to the parents' and society's authority while the children's own needs go unacknowledged as "poisonous pedagogy." She suggested that where this tradition is particularly extreme, children grow up with a greater vulnerability to violence, either as the perpetrators or the victims. In adulthood, they are most likely to deny or repress traumatic childhood experiences, maintain needs to idealize someone else, and often skillfully monitor and attend to the needs of others to gain love and acceptance (A. Miller, 1981, 1983, 1986).

This theoretical perspective applied to battered women suggests that some battered women are more vulnerable to return to battering relationships than others because of the following: (a) They monitored and responded more to their parents' needs than their own, and (b) they strongly idealized parents and authority in general.

Various factors previously identified and studied had not satisfactorily accounted for determining whether battered women's destination following shelter experience would back with their abusers or not. Additional determinants were considered by selecting the variables of authoritarianism and monitoring from the violence theory







10

proposed by A. Miller (1983, 1986). The previous two variables already noted in the literature as determinants, i.e. resources and number of previous separations (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981), were investigated. Finally, the locus of control determinant which Walker (1979, 1984) proposed from the learned helplessness theory was reexamined. Reinvestigation of this was considered important in light of the strong reliance in this field on her theory and her surprising findings disputing locus of control as a factor (Walker, 1984).

Research Ouestions

The specific research questions were

1. Do battered women who do not return to their

abusers following shelter stays and battered women who do return to their abusers differ regarding their authoritarianism?

2. Do battered women who do not return to their

abusers following shelter stays and battered women who do return to their abusers differ regarding their monitoring?

3. Do battered women who do not return to their

abusers following shelter stays and battered women who do return to their abusers differ regarding their resources?

4. Do battered women who do not return to their

abusers following shelter stays and battered women who do return to their abusers differ regarding their number of previous separations?






11

5. Do battered women who do not return to their

abusers following shelter stays and battered women who do return to their abusers differ regarding their locus of control (internal, powerful others, chance)?

6. Which combination of the above mentioned variables provides the best determination of battered women's destination following a shelter stay?

Significance of the Study

This study has implications for those researching domestic violence, shelter personnel, community agency personnel, and battered women. A workable theory which accurately describes the social problem of wife abuse and which can be used to shape a better family unit and society has yet to be defined. Whether refinement of current theories or evolution of a new theory occurs, research is required on the battered, the family unit, society, and the battered woman. Despite the heritage of this problem as a socially taboo subject, and the paucity of funding provided to it, the findings from this research effort will add to what is known about battered women and be a part of the base for future research efforts on conjugal violence against women.

If there are identifiable social and psychological factors that determine the destination of battered women following a shelter experience, then shelter program planners could design and administer programs most suited to the two different populations. Rather than assuming that






12

all programs are appropriate for all shelter residents, better attempts at identifying the appropriate recipients for certain interventions could be made. The woman who is setting up independent living arrangements for herself and her children needs housing assistance, career guidance, oftentimes school transfers for children, continuing welfare aid and food stamps, and social networks. The woman returning to the abuser might benefit from self-defense training, conflict resolution skills, and strategies to encourage marital and family counseling. Depending on her long-range desires and the degree to which her return is based on economic dependence, the abused woman returning to the battered might benefit from strategies for acquiring career skills or employment opportunities.

Community agency personnel who come in contact with battered women include those from food and financial assistance agencies, churches, legal aid, state prosecution and public defender offices, vocational rehabilitation programs, health care agencies, school programs, and sometimes community mental health or crisis counseling agencies. With more insight into who is likely to return to the abuser and who is likely to separate longer-term from the battered, personnel from each of these agencies can more appropriately intervene. Oftentimes, staff working with battered women can get invested in the hope that a client will decide her destination in a particular direction. A more definitive profile of who will stay away and who will






13

return will aid concerned helpers in setting their expectations while doing what they know to be helpful in each case.

Finally, battered women can gain greater selfunderstanding and mutual appreciation by more clearly knowing the challenges they face to establish violencefree lives. These findings will not provide them with the causal explanations for why some return to abusers and some do not. Yet, the findings may provide them with clues to explore what in their situations are defensive responses to their abuse and which have roots in both external and internalized oppression. Greater understanding can help minimize self-blame and guilt.

Definition of Terms

Operational definitions for terms relevant to the research are provided to enhance understanding. Battered Woman

A battered woman is any married or unmarried woman over the age of 16 who has been physically abused in ways which caused pain or injury on at least one occasion at the hands of an intimate male partner.1 Battered wives and wife abuse


1A definition for battered women has not been universally agreed upon by researchers. The most notable distinctions about this definition are as follows: A battered woman would not need to present evidence of injury; selfreport of physical battering is sufficient. The battering need not be a repeated occurrence as preferred by some authors (Michigan Women's Commission, 1977; Parker & Schumacher, 1977). Although abuses other than physical may be just as devastating, and support exists to include psychological abuse as a component of battering (Moore, 1979; Walker, 1978b, 1979, 1980a,), the definition herein has been limited to physical harm alone. The purpose is two-fold: (a) Physical abuse can be documented more






14

victims will also mean battered women as defined here. Abused, harmed, or beaten may be substituted for battered. Battering

Battering refers to intentional physical abuse which causes pain or injury. Spouse abuse, spousal violence, conjugal assault, wife beating, or domestic violence will be used interchangeably with battering. Shelter

A shelter is an emergency refuge for safety available for more than one battered woman at a time, operated generally by women's groups, volunteer organizations, a governmental or non-profit agency, and other than a safe place to stay which the woman provides for herself or is provided for her by family members, in-laws, or friends. Destination

Destination, in this investigation, refers to either returning to the abuser or not returning to the abuser following a shelter stay.

Returning to the Abuser

For the purposes of this study returning to the abuser means being in residence with the assailant during the


readily; operationalizing a definition for psychological
abuse is, at best, extremely difficult; and (b) due to space
considerations and various other exigencies, shelter
programs for battered women most often provide safe refuge
services utilizing this same delimited definition as an
admissions requirement. Although battered women could also
include those who have been abused by family members
(Pagelow, 1977A), those groups have been excluded from this study. This sample may include women who are cohabitation with men although not legally married to them, or separated or divorced partners who are living with each other.






15

follow-up contact which occurs from 2 to 6 weeks following departure from the shelter, regardless of whether the battered woman left the shelter to return to him, or in the interim weeks decided to move back with him. Not Returning to the Abuser

Women who are in the category of not returning to the abuser are those who decide to leave the shelter and do not live with the abuser and are not residing with him as of follow-up within the time frame of 2 through 6 weeks after departure from the shelter.

Locus of Control

Locus of control refers to how an individual perceives contingency reinforcement upon his or her actions(s) and generally includes an internal and/or an external dimension. Attributional style may be interchanged with locus of control. The three types of locus of control in this investigation are internal, powerful others, and chance. Internal locus of control refers to when an individual perceives that a reinforcement has been contingent upon his or her action(s) (Rotter, 1966). Powerful others locus of control is defined as when as individual believes that a reinforcement has not been completely contingent on his or her action(s), but externally due, to some extent, to the control of powerful others (Levenson & Miller, 1976). Chance locus of control means that an individual perceives that a reinforcement has not been fully contingent on his or her action(s), but due, in part, to the result of chance,






16

fate, or luck (Levenson & Miller, 1976). The Levenson Locus of Control scales (Levenson & Miller, 1976) was be utilized in this investigation.

Resources

Resources refers to an indication of the battered woman's objective dependence or independence from her abuser, especially as it relates to the woman's employment status, the presence of children age 5 or younger at home, and 75% or more of the household income being earned by the mate. In this study it was measured by the Resources Index (Kalmuss & Straus, 1982).

Number of Previous Separations

Number of previous separations is defined as any time the woman has left the mate previously. In this study that was indicated in response to two questions: (a) Have you left your mate before, and (b) if yes, number of times. Authoritarianism

Authoritarianism, also known as dogmatism, is defined by Rokeach (1960) as an organization of belief-disbelief systems constructed to both satisfy a need for a cognitive framework for knowing and understanding the world and to distance threatening aspects of reality. In this study it was measured by the Dogmatism Scale (Form D), a generalized measure of authoritarianism, free of political bias

(Rokeach, 1960) .






17

Monitoring

Monitoring is defined as the observation of others and mediating of one's presentation as guided by social cues (M. Snyder, 1974), as measured by the Snyder Self-Monitoring Scale (M. Snyder, 1974).

Organization of Study

This study will be presented in five chapters. The

current chapter is a brief introduction of the subject, the purpose and rationale for the study, and a description of relevant terms. Chapter II is a review of the related literature. The research methodology is detailed in the third chapter. Data analyses and results are presented in the fourth chapter. Finally, in Chapter V the researcher offers a discussion and interpretation of the results, a discussion of the limitations of the study, and further implications.














CHAPTER II

REVIEW OF THE RELATED LITERATURE

Compared to other psychological topics, the literature focusing on wife abuse is quite recent, appearing only in the last 15 years (Hilberman, 1980). Described initially in this chapter is how conjugal violence emerged as a social issue and how the beginnings of the public's acknowledgement of the problem shape the theories which developed. The socio-psychological theories of violence, particularly those which have relevance to this investigation, are presented. Next, the pertinent research in regard to factors related to the battered woman's destination is detailed. Finally, previous research findings are given on authoritarianism and monitoring, the new variables in the current investigation.

Emergence of Conjugal violence

Woman battering was slow to be recognized as a problematic social issue for several reasons: the sanctity and privacy of the home, the theory of sadomasochism in humans, and the subordinate role of women to men in both home and society (Gelles, 1977; Straus, Gelles, & Steinmetz, 1980; Walker, 1979). In the past two decades, societal issues have changed substantially resulting in an environment favorable to examining both the dynamics of conjugal violence and a society which has supported, excused, and


18






19

tolerated such abuse. From examination of the early literature several main forces in the early 1970s reflect a Zeitgeist which has brought wife abuse to public attention:

(a) the British refuge movement, (b) the women's rights movement in the U.S., (c) examination of police injuries and fatalities, and (d) the study of societal violence. The British Refuge Movement

Although the British were the first to open battered women's shelters, their initial discovery of the social problem was quite unintentional. In 1971 Erin Pizzey went to the Chiswick City Council to seek support for establishing a home as an advice center for women. When she opened the doors to Chiswick Women's Aid (WA), she discovered that the majority of women coming for help were seeking a safe place away from husbands who beat them. Within a year and a half she collected demographic and incidence data from the WA program, including 3,000 requests from battered women seeking refuge. She gave the information to England's social service administration thereby documenting the widespread problem of wife abuse and the lack of support from legal and governmental communities. Thus, the shelter movement began (Dobash & Dobash, 1971; Pizzey, 1974).

Dobash and Dobash (1971) recorded the founding and

growth of WA groups throughout Britain. They suggested that the establishment of shelters was a natural evolution from the British women's movement. In addition, they suggested






20

that the patriarchal attitudes which fostered husbands beating wives also frustrated women's groups in society at large as they sought help for the battering problem. A contradiction existed in society between protecting the privacy of home life and preventing the battering of spouses.

Becoming aware of WA's growth, the psychiatric and

legal communities were next to get involved. In Memorandum on Battered Wives (1974), the Royal College of Psychiatrists scholars used case studies to depict the complexity of factors surrounding spouse abuse. They described wife battering as an adaptation failure of inadequate acquisition of appropriate social learning skills by the abusers. Citing the fact that battered wives use help when available, the Royal College scholars recommended more services, research, and education. They proposed that child abuse often occurs in homes where wife battering exists.

Scott (1974) described persons involved in spouse abuse as likely to come from many clinical classifications, rather than just one or two. Common diagnoses he cited were immature personality, dependence, aggression, jealous reactions, and drug or alcohol addictions. Scott asserted that although sadomasochism is often assumed as the reason for battered wives returning, more often dependency, ignorance of choice, and fear of loneliness exist as the actual reasons. By 1975 British pamphlets outlining legal options and explaining the difficulties and inadequacies of






21

the law began to proliferate (Gill & Coote, 1975; M. Kemp, Knightly & Norton, 1975).

The Women's Rights Movement in the U.S.

From 1973-1976, news of what was going on in Great

Britain surfaced in popular magazines in the United States, many of which had female audiences. On July 9, 1973, Newsweek ("Britain: Battered," 1973) included a report on the opening of Chiswick Women's Aid. Ms. Magazine (Search, 1974) and McCall's ("Wife Beaters," 1975) ran similar stories. Ladies Home Journal (Durbin, 1974) reported on wife-beating as discussed at the National Organization of Women's (NOW) national conference. NOW's perspective suggested that battered wives stayed in destructive relationships due to finances and the positive societal identity of married females rather than due to masochism.

As coordinator of the NOW Task Force on Battered Women, Martin wrote Battered Wives (1976), one of the earliest major works on spouse abuse in the U.S. Martin faulted society for maintaining gender and marital inequities, the legal system for being unresponsive, and social service agencies for poor coordination of services to battered women.

The organized women's rights movement had already

addressed the plight of rape victims in America, and the focus on wife abuse was a direct result of the attention of the women's movement (Hilberman, 1980). The obvious parallel between the two issues showed that both are actions






22

involving male violence perpetrated against female victims. In investigating further, Pagelow (1977b) listed the parallels: Instances of sexual assault and wife abuse were both underreported to police; they rarely went to court and had abysmally low rates of conviction when they did. Both rapists and wife abusers were likely to be repeat offenders; few offenses evidenced higher rates of recidivism than these two. The victims of these crimes were pictured stereotypically as masochistic, provoking the crime, reporting it for hidden reasons, and not dedicated to seeing the perpetrator prosecuted. Thus, in a "blame the victim" cycle, women in wife battering cases, as well as rape cases, had to prove that they were "worthy victims" even though the need to establish worth was not present in other crimes. Pagelow reported that these crimes were not predominantly lower class occurrences as often thought, identifying high incidence in middle and upper classes also. Police Involvement

With the addition of computers which stored and

categorized vast amounts of crime scene data, law enforcement agencies gained greater awareness of police fatalities and injuries at "domestic disturbance" response calls (National Institute of Law Enforcement and Criminal Justice, 1978). Twenty-eight percent of the assaults perpetrated against police officers have been while the officers were responding to domestic disturbance complaints (Stephens, 1977). A substantial proportion of all homicides on police






23

are related to domestic disturbance calls (Rochester Police Bureau, 1974).

Addressing the American Bar Association, Detroit Police Chief Bannon (1975) commented that in addition to fearing injury, police avoid domestic violence for two other reasons: (a) They lack the skills needed to mediate interpersonal conflicts, and (b) they view domestic conflicts as private problems, not a public matter. He commented:

of all the nonathletic occupations, none is as absorbed
with the use of physical coercive force as that of the
police officer. Nor are any more thoroughly socialized in their masculine role images. This . suggests to
me that traditionally trained and socialized policemen are the worst possible choices to attempt to intervene
in domestic violence. (p. 8)

The lack of police skills to intervene in conjugal

violence was addressed in 1967 by Bard and the New York City Police Department. They jointly developed and implemented the first family crisis intervention program within a police unit (Bard & Zacker, 1971). Other literature detailed other early efforts at police training in crisis intervention, referral, and conflict management (Olsen, 1972; Rochester Police Bureau, 1974; Spitzner & McGee, 1975).

Examination of police work brought additional domestic violence factors to light. The incidence and degree of interpersonal aggression of domestic disturbance calls cannot be isolated to lower socioeconomic, minority neighborhoods as a comparison between Harlem, New York, and Norwalk, Connecticut, indicated (Bard & Zacker, 1971). Evidence contradicted the common beliefs that family







24

disputes are usually associated with alcohol and that assaultive behaviors by family members usually ceased before the arrival of police (Bard & Zacker, 1974). Greater caution was advised for the common practice of diverting wife beaters from the criminal justice process (Brakel, 1971).

The Study of Societal Violence

By the late 1960s society in general felt more concern about the rising crime rate, especially of violent crimes. In 1969 the Eisenhower Commission (Goode, 1969) looked into the causes and prevention of violence, and in 1971 the U.S. Surgeon General conducted a similar study (Roy, 1977). The Surgeon General's examination linked media violence and violence in real life, suggesting a causal relationship between the two (Roy, 1977). Goode (1969), a member of the National Commission on the Causes and Prevention of Violence, theorized that young children learn violence and the appropriate use of force through observations of parents during early childhood. Goode suggested that social class differences exist for spouse abuse due to lower class acceptance of violence and conflict in their environment in general. With more conflict situations arising in lower class neighborhoods, limited places to escape conflict, and fewer resources (e.g., money, power) available to get what counts, violence is a more likely option.

Gelles (1972) coined the family as the "training ground for violence,, (p. 169) with family members modeling violent






25

acts that have much more impact than either television violence or school discipline in producing future generations of assaultive people. In his research, family members from both violent and nonviolent families approved of certain violent acts as "normal" in that they served the purpose of dissipating a husband's tension and/or appeasing an otherwise hysterical mate. Gelles further supported the family structure and the family's societal position as more important factors than individual pathology in the occurrence of family violence. Gelles estimated that intrafamily violence occurs in 37% of any general population which compares similarly to other estimates of this nature (Levinger, 1966; O'Brien, 1971). Gelles (1975) was the first researcher to note the higher likelihood of violence against a wife when she is pregnant.

Straus established the Family Violence Research Program at the University of New Hampshire in 1970. He found that the cultural and social signals from a violent society combine with familial, psychological factors to incite violence in the home, contradicting the myth of the family as a warm and loving system (Straus, 1974).

Straus (1973) looked at the balance of power in

families in relationship to the occurrence of violence. Where equal power existed between the husband and wife, the husband to wife violence was lowest. High power in either wife or husband was related to high levels of husband-to-






26

wife violence. Where the wife's power was recognized as high, the wife-to-husband violence was also high.

In studies where college students recalled parental conflicts, Straus (1974) disproved the popular notion of "cathartic violence," i.e., if pressure is built up and not released through some means, it will eventually erupt as violence. Instead, he found that increased verbal battling escalated physical aggression, rather than dissipating such stresses. College students' parents who utilized more intellectualizing during marital conflicts experienced lower levels of aggression towards spouses.

Straus, working with others (Allen & Straus, 1975;

Owens & Straus, 1975), found a moderate correlation between observing, committing, or being victim of violent acts as a child and adult approval of interpersonal or political violence. He also found that the greater the husband's resources, the less likelihood that the husband would perpetrate violence in the home.

Theoretical Models

These four preceding forces focused on the problem of domestic violence, building a body of knowledge which brought multiple facets of wife abuse to light. Theoretical models of family violence resulted. In some cases, theories applied to human behavior and aggression in general have been overlaid on the topic of family violence, e.g., the theory of learned helplessness (Seligman, 1975) or the frustration-aggression hypothesis (Dollard, Dobb, Miller,






27

Mower, & Sears, 1939). In other cases theories have evolved explicitly from the study of domestic violence as in Walker's (1979, 1984) cycle theory of battering. The following major theoretical areas are examined: (a) General systems, (b) biological, (c) exchange/social control, (d) feminist, (e) social learning, and (f) intrapsychic. General Systems Theory

Straus (1973) developed a theory to explain family

violence known as the "general systems theory." The family unit is a social system with a purpose, seeking to meet goals and adapting to the environmental context in which it exists (Straus, 1973). A violent family system is described as follows: A probability for conflict occurs in the family due to involuntary membership, intensity of relationships, generation and gender differences, and parental rights of influence. The probability of violence can be positively reinforced due to a highly violent society and socialization of family members in violence vis a vis parental models, physical punishment, tolerance of sibling assaults, and macho values, especially for boys. As a result the culture legitimizes a norm of violence and family members integrate assaultiveness into their personality characteristics (Straus, 1973). The sexism of society, through its limited roles and career opportunities for women, unequal pay, assumptions regarding wives and mothers, assumptions regarding males as head of households, and socialization of women into subordinate roles, reinforces that a woman should






28

"stay in her place" regardless of what abusive treatment she receives. All of these socialization imprints provide for a high level of violence in the family (Straus, 1973).

Another general systems model has been advanced by

Giles-Sims (1983). She builds on Straus (1973) and Buckley (1967), as well as Broderick and Smith's (1979) concepts regarding hierarchies of feedback and control in systems. The six stage model includes (a) the establishment of the family system, (b) the first incident of violence,

(c) stabilization of the family system, (d) the choice point, (e) leaving the system, and (f) resolution or more of the same. Both the man and woman creating the system come to it with individual histories and personalities that influence the way they interact with each other. Precipitating events or stressful situations occur in the marriage. Those events, in combination with prior history relating to violence, enter into the execution of violence. If the assault serves to maintain the system or satisfy goals of the battered, the next move is by the woman. If her response to the assault is to ignore, deny, or provide forgiveness, she colludes in a feedback loop which encourages the continuance of such behavior. On the other hand, if the abused woman gets angry and considers the violence as a possible pattern, then she may utilize social supports and examine her alternatives. At this point, the battered woman may leave. If the assaultive behavior did not satisfy some of the system achievement or maintenance goals, then the






29

battered may choose an alternative behavior that could be more useful and acceptable. If the woman leaves and will come back only under the condition that he change, he may choose some alternative behavior as well. Her other two options include not coming back at all, or coming back with no requisite that he change. A weakness of this theory is that Giles-Sims did not incorporate some of the less directly observable influences such as the patriarchal influences or societal gender inequities. Biological Theory

Elliott (1982) suggested that biological explanations have been totally ignored in postulating theories about intrafamilial violence. He detailed how organic and metabolic disorders contribute to domestic disturbances; he specifically focused on episodic dyscontrol syndrome or "explosive rage." His sample of 286 patients with histories of violence was a skewed sample because of the biased selection on the part of the referring physicians, the use of computerized tomography (the CT scan), and the painstaking methods utilized to uncover minimal brain dysfunction. Elliott recommended further research with an unselected group of batteries to uncover neurological causation factors in domestic violence, and by no means suggested biology or genetics as an all encompassing factor. Although neurological dysfunction may be pertinent in a small number of cases, it is highly unlikely that it is a major explanatory feature of wife abuse in our society. The







30

question would arise as to why so much brain dysfunction occurs among men only.

Exchange/Social Control Theory

Exchange and social control apply to all types of intrafamily violence, including conjugal abuse (Gelles, 1983). Exchange theory suggests that the ways humans interact with each other is dictated by the rewards they receive or the negative consequences they avoid (Gelles & Straus, 1979). The social control theory is the question of how much control is exerted through society and social institutions to encourage creation, maintenance, or diminishment of some belief or act such as violent behavior. According to this theory, family members batter other family members simply because they can.

First, the costs of being violent are less than the

rewards; therefore battering occurs. Since effective social controls aimed at preventing family violence are not in place, a family member who batters experiences few consequences due to these lack of controls. Also, inequality within the family, the sanctity of the family unit in society, and the encouraged image of males as macho minimize social control in the family thereby reducing the consequences of violence to the battered and increasing the rewards.

When societies have no legal or normative sanctions regarding family violence, the violence is more frequent (Nye, 1979). Additionally, when norms or laws exist







31

sanctioning one type of family violence but not another, e.g., laws against wife abuse but not child abuse, the violence will be more frequent in the less sanctioned category.

The exchange/social control theory for family violence has application to treatment and policy issues (Gelles, 1983). Treatment interventions and policies that increase both social controls exercised over the family and the costs for violent behavior are primary. The resultant policies should diminish societal norms that glorify violence, increase financial and gender equality, and create more responsive criminal justice and social service agencies. The value in such a theory is that it can be applied across the board to explain all types of familial abuse. Feminist Theories

According to feminists, explanations regarding violence against women are inadequate if concerned with general explanations regarding overall societal violence or family violence (Dobash & Dobash, 1979; Schechter, 1982). By treating all violence between parties as equal, some significant uniquenesses about domestic violence against women are masked (Dobash & Dobash, 1981b). Feminist authors have particularly challenged Straus' (1980) conclusion that there is nearly sexual equality in perpetuation of violence in couples and Steinmetz (1977-78) who suggested an equally prevalent "battered husband syndrome" (p. 501). Pleck, Pleck, Grossman, and Bart (1977-78) have pointed to mnac-






32

curate interpretations by Steinmetz of her own data regarding incidence of husband abuse. Fleming (1979) and Schechter (1982) challenged these premises, making a strong case that the more serious violent incidents, often with more severe physical and psychological repercussions, are perpetrated by men against women in conjugal relationships.

Feminist researchers pinpoint the reliance on the

social survey as contributing to faulty findings (Dobash & Dobash, 1983; NiCarthy, Merriam, & Coffman 1984). The data from survey methods are very informative regarding sociodemographic characteristics, but are of little value for information regarding the sensitive issues involved in family violence about which society and individuals have traditionally kept quiet (Dobash & Dobash, 1983). Consequently, many of the findings and conclusions from this approach have lacked consistency, been proven unfounded, or are replete with contradiction (Dobash & Dobash, 1983; Fleming, 1979; Pagelow, 1980; Pleck, Pleck, Grossman & Bart, 1977). To understand this phenomenon in our society, indepth systematic interviewing of battered women by experts on this subject is preferable to the survey technique as a research method (Dobash & Dobash, 1983; Pagelow, 1981; Walker, 1984). And for feminist researchers, calling "woman battering" or "wife abuse" by the term "spouse abuse" is in itself a sexist act masking the true nature of the problem (Dobash, 1981).






33

There are several theories that represent the feminist analysis of wife abuse. Dobash and Dobash (1977-78) traced the history of violence against women by placing it in the wider context of society, thus supporting that abuse against wives has been acceptable behavior in a patriarchal society. Dobash and Dobash (1977-78, 1979) cited marriage laws throughout history that categorize wives as the possessions of husbands and dictate the husband's obligations to chastise and reprimand wives. Many of the laws clearly reflect double standards. For example, English Common Law denied wives civil rights and legal status. England, Europe, and early America had laws supporting the husband's rights to beat his wife. It was not until the 1700s that this tradition came into question. In the early 1800s written laws were adopted in various states that reiterated the sanctity and privacy of the home unless "permanent injury or excessive violence" were present (Eisenberg& Micklow, 1977, p. 149). Finally, not until 1891 does caselaw reject the husband's legal right to beating or abusing his wife (Dobash & Dobash, 1977-78). Dobash and Dobash (1979) believed that it is the tradition of the patriarchal society that has also impacted on the leaders of the women's movement in their attempts to address wife abuse through both traditional institutions and the grass roots shelter movement. Consequently, nonsupportive laws change too slowly, funding is inadequate, and institutions remain






34

nonresponsive and arduous for battered women and those helping them.

The patriarchal tradition today combines with the needs and realities of a capitalist economy to continue institutional and socialization patterns that perpetuate male dominance and aggression (Schechter, 1982). Battering is one of the many ways that men maintain control in our society. Women stay in abusive relationships for two reasons: (a) Their inferior status under capitalism (lower wages or agelessness and unequal division of labor requiring women to be responsible for maintenance of both home and family) leaves them financially dependent on men, and (b) their sex role socialization as wives and mothers results in a moral obligation to stay and please their husbands and rear their children properly (Schechter, 1982).

One problem with patriarchal society explanations is that they are historical and hypothetical in nature, with minimal empirical work with battered women operationalizing the specifics of the theory. An example of a researcher who has captured empirically some of the aspects of patriarchal society is Pagelow (1981). She has addressed particularly the question, "why do women stay." The hypothesis is that a woman of few resources, receiving negative institutional response (family, law enforcement, social service agencies, society), and very traditional in her beliefs, is unlikely to change or believe she can change her battering situation. Pagelow (1981) supports the proposition that limited






35

resources (as measured by the woman's age, her children's ages, differences in earnings, husband's earnings, and home ownership) correlate strongly with the woman's staying in the relationship. She did not find a strong correlation between institutional response and traditional ideology with the women's length of cohabitation while in a battering relationship.

Social Learning Theory

The most relied upon psychological theory regarding battered women is that proposed by Walker (cf. 1977-78, 1979, 1980b, 1984). Because of its well established reference in the battered women's movement and its role as a significant precedent for this investigation, Walker's theory will be considered in detail here.

Walker (1977-78, 1979, 1984) utilizes Seligman's (1975) theory of learned helplessness as a basis to understand the battering relationship and battered women's responses. Learned helplessness is combined with a cycle theory of violence to depict the phenomenon of battering (Walker, 1979). The supporting research (Walker, 1978a, 1978b, 1979, 1980a, 1981) is based on interviews with several hundreds of battered women and some caregivers to battered women (e.g., therapists, shelter workers).

Learned helplessness, according to Seligman (1975), occurs when animals experience negative reinforcement noncontingently, i.e., with no apparent link to their own actions. Eventually they believe that their actions have no






36

effect on what may happen to them. This theory's applicability to humans has also been supported by Seligman and others (Hiroto, 1974; Klein & Seligman, 1976; W.P. Miller & Seligman, 1975; W.P. Miller, Seligman, & Kurlander, 1976; Seligman, 1975; Seligman & Hiroto, 1975). With battered women, as Walker's theory proposed (1977-78, 1979, 1984), battering from an abuser occurs at random, often in no directly rational relationship to the woman's immediately preceding behavior, and according to no set schedule or time frame.

As repeated battering occurs, she comes to believe that nothing she can do will have any impact on the batterer's abusive behavior. All of the actions she has previously tried, e.g., being apologetic and loving, fighting back, leaving, calling the police, have not resulted in an end to the unpredictable abuse. Walker proposed that a woman will more quickly "learn" her powerlessness if she experienced more rigid sex-role stereotyping as a child which already reinforces passivity on her part and the importance of relying on men for help and protection (Dweck, Goetz, & Straus, 1980; Radloff & Rae, 1979, 1981; Walker, 1977-78, 1980a, 1984).

From Walker's interviews, she determined that the battering, although neither constant nor occurring in a consistent manner each time, does occur in a cycle of three distinct phases (Walker, 1977-78, 1979, 1984). The first phase is the tension-building stage. During this stage,






37

tension increases in the home with the battered expressing anger and dissatisfaction, but without any extreme violent expression. The woman tries to please the battered to minimize the hostility, calm him, and prevent any more expression of hostility. However, the tension increases, and her efforts to reduce his angry responses become more unsuccessful. Eventually, the woman withdraws, thus eliciting an insecurity and an increase in possessiveness, jealousy, and oppression from the abuser.

At phase two the tension becomes intolerable and acute battering occurs. The battered explodes into physical and verbal assault. This is the most likely time for the woman to receive injuries or for external parties, e.g., police, to become involved. This incident usually dissipates the tension which had built up, and the violence temporarily ceases.

In stage three loving contrition ensues. The battered may apologize profusely, promise never to repeat the actions, buy gifts and/or be more helpful to the woman. The relationship enters into a honeymoon-like appearance. The battered may believe he will not be assaultive again, and the woman wants very much to believe that he will not.

Empirical findings from Walker's (1984) sample support the three-stage theory. In 65% of the relationships studied by Walker (1984), evidence supported the tension-building stage before an acute battering incident. And with 58% of







38

her subjects, a loving contrition stage occurred after the acute battering incident.

Intrapsychic Theories

Earliest theoretical explanations of spouse abuse are based on the concept that abuse occurs because the participants are individually pathological, i.e., have inadequate personality development (Snell, Rosenwald, & Robey, 1964). This intrapsychic viewpoint suggests that batterers are passive-aggressive, obsessive-compulsive, paranoid, sadistic, immature, dependent, excessively jealous, and/or addicted to drugs or alcohol (Scott, 1974; Shainess, 1977). The wives contribute to the abuse because they are aggressive, masculine, domineering, sexually frigid, and masochistic (Snell, Rosenwald, & Robey, 1964; Shainess, 1979). These theoretical explanations have fallen short because neither explains persons who have these personality characteristics and are not in battering relationships, nor takes into account a myriad of interactional and social factors surrounding domestic violence.

The most prevalent Freudian view applied to battered

women is masochism (Shainess, 1979). Yet within the context of contemporary psychoanalytic theory, Gillman (1980) proposed an object-relations approach to understanding and treating battered women, following from the work on borderline personality (cf. Kernberg, 1975). She (Gillman, 1980) suggested that masochism is not a key factor in battered women since the defense mechanism of repression, usually







39

present in masochistic individuals, is not often present in battered women. Battered women endure far too much abuse for their behavior to be considered "neurotic."

Battered women more often fall into the category of borderline personality, being neither neurotic nor psychotic. The "battered woman personality" is a woman who has two internal, yet separate, representations of herself: one good and the other bad. As battering occurs, she is a worthless person in a destructive relationship to a persecuting husband (mother). She has difficulty leaving the relationship in that the good self and other emerges after the incident is over and as the bad submerges. The splitting means quite opposing views of self and spouse are present, yet the battered woman is unaware of the contradictory nature of her beliefs (Gillman, 1980).

A source for variables for this investigation was the

theory of violence proposed by A. Miller (1983). Because of this theory's application to the current study, the theory is presented in detail.

A. Miller (1983) proposed that violence occurs in society because of the heritage of strict child rearing practices which break children's willfulness when they are very young, and causes them to repress their own traumatic childhood experiences, and leads to their aggressiveness or self-destructive behaviors in adolescence and adulthood. Historically, children have been viewed as mean, demanding, tyrannical, possessing destructive drives and sexual longing






40

for their parents, and having other negative desires and behaviors which society and parents have believed are their responsibilities to ameliorate (Rutschky, 1977). The child's willfulness must be emphatically curbed and the child must view the parents as unquestionably right and superior (Rutschky, 1977). Both physically and emotionally abusive and humiliating methods are utilized by parents to accomplish their responsibilities of rearing socially acceptable children. The parents are unaware of the ways they are abreacting their own suppressed hatred and anger from their own childhood by utilizing their offspring to meet their suppressed needs (A. Miller 1981, 1983, 1986).

In this environment the child's needs are not met but frustrated. The child experiences hurt and anger, but cannot have these feelings validated with parents. Without any validation the child denies the feelings (A. Miller, 1981, 1983, 1986). A proneness to violence is most likely if the child from such an upbringing does not have the following: (a) another human being in whom to confide true feelings, (b) an environment allowing experiencing and expression of pain (whether at home or elsewhere), (c) other objects for abreacting hatred, and (d) an education or intellectual avenue to rationalize hatred (A. Miller, 1983). Depending, then, on the degree to which the child has been traumatized and the unavailability of healthy channeling of negative experiences and feelings which are suppressed,






41

rather than addressed, the likelihood for aggression or self-destruction can be determined.

What follows is that the child represses any memory of the traumata and idealizes the parents and the upbringing. The child, because of love for the parents, takes the responsibility and the blame for the parents' hurtful behavior (A. Miller 1983). The child's feelings of anger, hurt, hatred, helplessness, and pain come forth in other life circumstances, but disassociated from their origin. The child is consciously blocked from realizing their source. In addition to idealizing those who traumatize the child, the child can strongly identify with the perpetrator in an effort to distance the self from the identified "bad child" whom the perpetrator/parent has been disciplining so sharply. The child has thus split-off, projected his or her own weakness, feelings of helplessness, and vulnerability by maintaining a certain derision for others who are younger, weaker, smaller (A. Miller, 1981).

Since society maintains a lack of sensitivity to the denied cruelty, children grow up to become parents who manifest the same types of humiliations and physical and emotional abuse on their own children (A. Miller, 1983). This process means that cruelty and violence become passed on from generation to generation. Only a sensitive discovery of one's own traumatization, in a supportive environment, breaks the intergenerational cycle of violence.






42

Children who receive humiliating and cruel treatment

when they need protection, respect, and honesty will grow up to subject others or themselves to destructive acts. If they become parents, their own children will be the likely outlets. Lacking a weaker spouse or children, their violence may be directed at others in society (A. Miller 1983). Self-cruelty is another possibility. Having received poor treatment as children, adults can continue to treat themselves abusively (A. Miller, 1983).

A. Miller (1986) has argued against Freud's theory of drives in children indicating that his willingness to consistently interpret neuroses as rooted in the child's drive conflicts with Freud's own protection and idealization of the parenting he received. Freud's own narcissistic wounds, and a lack of an empathic environment for him to uncover his own repressed traumatization, led him to collude in blaming the victim and sparing parents' feelings (A.

Miller, 1986).

In addition to being self or other destructive, another variation of behavior can develop in response to such a childhood. As the child idealizes the parents and takes responsibility and blame for traumas, the child comes to understand performing for the parents and parents' needs as a means of winning their love. Thus, some children from this environment have the particular skill of discerning the needs of others (e.g., the parents) no matter how subtle or unconscious the signs are (A. Miller, 1981). The develop-






43

ment of this monitoring skill comes at the cost of underdevelopment of the child's own ability to perceive his or her own needs.

To summarize, A. Miller (1981, 1983, 1986) proposed

that violence occurs because of the following: Parents have narcissistic disturbances from their own childhood traumatic experiences which they have repressed. They treat their own children cruelly and unsympathetically by splitting off their "bad selves" on their weaker and more vulnerable children. Society colludes with the process by mandating parents to rear obedient and conforming children and viewing acting out not as children's reactions to their traumatization but as manifestations of the bad nature of children who have not been properly reared into subservience. society protects parents, provides avenues historically for socially acceptable violence, e.g., beating wives and children as appropriate disciplinary measures, striking students for wrongdoing. Children who are disturbed through such upbringing will either dull their self-sensing abilities, harm others weaker than themselves, become selfdestructive, deny their own traumatization and needs, refine their abilities to monitor and please others in a quest for love and narcissistic fulfillment, or some combination of these qualities.

Variables Relating to the-Battered Women's Destination

Researchers espousing different theories have examined a number of variables relating to the question of battered






44

women leaving abusive relationships. Researching a variety of variables is important to build a solid theoretical understanding of the problem. What follows is a summarization of that research and, where applicable, a discussion of how it applies to the investigation. Frecruency and severity of the Abuse

Gelles (1976) believed that a battered woman would

mobilize on her own behalf the more frequent or severe the abuse. His research supported this hypothesis. Other research did not confirm the same finding (Pagelow, 1977c, 1980, 1981). Nor were frequency or severity significant in defining whether a battered woman would return to the abuser after a shelter stay (Snyder & Scheer, 1981). Abuse in Family of Origin

If a battered woman was exposed to more violence in her family of origin, would she be less likely to stay away from her husband, the abuser? The families of origin of battered women have often been violent families. Twenty-three percent of the family histories of the wives studied by Gayford (1975) were violent. In other studies (Hilberman & Munson, 1977-78; Scott, 1974) that percentage was 50% or more. Parker and Schumacher (1977) found that if the mother in the wife's family of origin was physically abused, it was significantly probable that the wife would be battered by her husband. Half the women in the Hilberman and Munson study (1977-78) were abused themselves as children. Gelles (1976) found that the less a woman was exposed to parents






45

battering each other in her family of origin, the more willing she was to seek outside intervention when she was abused by her spouse. Pagelow (1977c, 1980, 1981) was not able to substantiate the same finding in her research. It is possible that a similar percentage would occur for nonbattered women's families of origin. Study with families of college students indicates violence was present in 28% of the cases.

Length of Marriage, Previous Separations, Religious Affiliation

Snyder and Scheer (1981) examined a number of variables for their ability to suggest whether a battered woman would return to her abuser following a stay at a shelter. They selected variables from data available at admission to the shelter so the results would be helpful for shelter workers in selecting their intervention strategies with the sheltered women. The variables were not selected to support or test any particular theory, but to be helpful in theoretical formulation.

In their study the sampe was comprised of 74 women admitted to a shelter in Detroit, Michigan. Fifty-five percent of the women (41 women) were living with the abuser when follow-up was conducted 6 weeks after they left the shelter; 45% (33 women) were not. When comparisons were conducted between the two groups, six variables from admission were significantly different between the two groups: two of the reasons for seeking admission (seeks short-term separation only and seeks conjoint marital







46

counseling), relationship to the assailant, length of marriage (if married), previous separations, and religious affiliation (Roman Catholic or not). No significant group differences were found for any sociodemographic variables or for measures of the nature or severity of the abuse incident.

These six variables, when entered into a forward

stepwise discriminant function analysis with destination at follow-up as the criterion variable, yielded three predictors. Complete data on all six variables were available for 50 of the 72 abused women; the 50 were subjects for this portion of the research. The length of the marriage, occurrence of previous separations, and religious affiliation best predicted, with an overall classification accuracy of 79.6%, where the woman would be residing at follow-up. The authors support cross validation of their study utilizing independent samples from other settings.

There seems to be no or minimal other research in the literature for the three variables identified by the Snyder and Scheer (1981) research as particularly relevant to why the woman stays or why she returns to an abusive relationship. Walker (1979) indicated that abused women who leave their batteries and return to shelters on multiple occasions are more likely to ultimately leave the battering relationship. Several authors (Martin, 1976; Pagelow, 1981; Walker, 1979, 1984) have reiterated from interviews with battered women that longer marriages are the more difficult relation-







47

ships to leave, despite the abuse. In research which focused on couples who thought about divorce compared to those who did not (Booth & White, 1980), results were that religious affiliation does not have a strong effect on thinking about divorce. Of the couples who had come from lengthier marriages with stronger religious belief there were fewer women who thought about divorce; those couples with shorter marriages or less religious zeal had more wives who thought about divorce. They found that the more intense one's religious belief, the less likely one would think about divorce. Booth and White (1980), nevertheless, did find that in religious couples, even where marital satisfaction was high, persons experiencing abuse were more likely to think about divorce.

Stronger religious belief or activity is often tied to a more traditional way of looking at life (Pagelow, 1981). Walker (1979, 1984) found in her sample of both battered women and batteries, as described by the women, a strong traditional view about the home, family, and sex roles. Pagelow (1981) felt that the more intense the traditional ideology of women who have been battered, the more likely they are to remain in battering relationships and the less likely they will be to take action to significantly better the situation. To test this hypothesis, she looked at whether the women came from a religious family of origin (i.e., family religious activity), as well as five other variables she felt reflected traditional ideology. Her






48

variables in combination did not support her hypothesis. Her variable about a religious family, however, did have a statistically significant correlation to the dependent variable. Wetzel and Ross (1983) supported the idea that battered women are kept captive by the family and religious values that they hold among other things. Love, Af section, and Hope

Research on the impact of love, affection, and hope to battered women's destination is scant. In a sample of battered women who were not sheltered women, but battered women located via social service agencies, legal or medical sources, advertising, and by word of mouth, the most commonly cited reason for staying in abusive relationships was that the women loved their husbands (Cristall, 1978). Forty percent of 542 shelter residents sampled (Stacey & Shupe, 1983) when asked why had they remained with the abusers if they had not left immediately after a first incident of violence, said they were still optimistic about the future of the relationship. Twenty-seven percent of those felt they could still save the marriage. Fifteen percent suggested they stayed because of their affection for the partners.

Understanding how love, affection, and hope relate to women's vulnerability to remain in abusive relationships is a needed undertaking. Women have been identified in our society as being more relationship oriented than men; their socialization and conditioning have reinforced them in this







49

way (Gilligan, 1982). Males have considered their main focus, their life's work, to be their career. Women, traditionally, have viewed their life's work to be wives, mothers, and keepers of the household. Leaving a relationship for a woman is an admission of career failure and, thus, is not taken lightly. An unsuccessful marriage is a subordinate concern for a man compared to a career failure. Self-Esteem

More has been stated about the battered woman's selfesteem than probably any other variable. Researchers and clinicians who have interviewed or provided services to battered women repeatedly comment on the low self-esteem of battered women (Bell, 1977; Bowen, 1982; Brown & Brazzle, 1982; Carlson, 1977; Hilberman & Munson, 1977-78; Pagelow, 1981; Walker, 1977-78; 1978b, 1979, 1981, 1984; Wetzel & Ross, 1983).

From the few studies where instrumentation has been

utilized conflicting results exist. In a comparison of 46 battered women to 12 non-battered women, Star (1978) found lower self-esteem with the battered women. Star, Clark, Goetz, and O'Malia (1979) found, through their study of 57 battered women, that battered women could be described as lacking in self-confidence, low in self-esteem, aloof, anxious, reserved, uneasy in social interactions, and critical or uncompromising.

Hartik (1978) examined 30 battered women compared to 30 non-battered women for personality characteristics and self-






50

concept. She found battered women to be lower in ego strength and self-esteem, with greater identity problems, than non-battered women. Overall the battered women were dissatisfied with themselves physically, morally, socially, and in relationship to their families. Her research indicated that battered women had a very difficult time maintaining even minimal self-esteem. In a comparison of 20 battered women from shelters to 20 non-battered women from the same socioeconomic status, Chan (1979) found the battered women to have lower self-esteem.

Conflicting results come from two studies. Brown and

Brazzle (1982), reporting on the 40 abused women in shelters whom they studied, indicate that 42% of the women had high self-esteem. These authors believed it is inappropriate to assume that battered women have low self-esteem. Mitchell (1980) compared 16 nonabused women who were in therapy for psychological problems to 24 abused women who were in treatment for the abuse. The two groups did not differ in self-esteem, sex role stereotype, locus of control, or anxiety. What she did find, however, when studying within group differences among the battered women was that those who were less educated, unemployed, and severely and/or frequently abused were lower in self-esteem, more anxious, and more externally oriented. These studies not only suggest that battered women may not be lower in self-esteem than their nonbattered counterparts, but also that within group variances may distinguish among battered women.







51

Walker (1984) examined self-esteem in battered women via a semantic differential scale on which they rated themselves, women in general, and men in general. Although the prediction was for low self-esteem, battered women viewed themselves more favorably than they viewed women in general or men in general. The difficulty with this research was that the scale did not appear to be standardized. It was impossible to tell how the battered women's perceptions of themselves and others compared to other group perceptions about themselves, women in general, and men in general. Battered women in this case perceived themselves more favorably than they perceived other groups, but how does this compare with other groups' perceptions about themselves? Other groups' perceptions could be more elevated than battered women's perceptions which would validate a lower self-esteem comparatively for the battered women's group.

Although self-esteem was not be addressed in this

study, more definitive examination of this factor is needed. ResourcesZOccupational History

Martin (1976) was the first to emphasize that financial dependence and unresponsive social systems are very real explanations for why battered women remain with their spouses. At the same time, actual research results (Gelles, 1976) results indicated that the fewer resources and less power the abused woman had, the more likely she was to stay. Resources have been determined to be any of several items,







52

e.g., employment status, having children under 5, income level. In the Gelles study (1976), the resource variable that best distinguished wives who obtain assistance from those who remain with the husband is holding a job. Many authors (Dobash & Dobash, 1977-78; Fleming, 1979; Schechter, 1982; Stacey & Shupe, 1983; Straus, Gelles, & Steinmetz, 1980; Walker, 1979, 1984; Wetzel & Ross, 1983), from their discussions and research with battered women, support the notion that financial dependence does contribute to keeping the battered woman trapped in the relationship. Of Stacey and Shupe's (1983) shelter residents who responded to why they did not leave the situation immediately after the first abuse incident, 30% suggested their economic dependence. Pagelow (1981) retested Gelles' (1976) earlier hypothesis regarding resources, along with the other variables Gelles looked at (severity of abuse, frequency, and family of origin child abuse history), and again found that fewer resources related directly to whether the woman stayed in the relationship.

Occupational underachievement in the husband or higher status in occupation by the wife have been found to increase the risk for spouse abuse to occur (Hornung, McCullough, & Sagimoto, 1981). Conflicting results regarding occupational history suggests that employment status does not have a correlation with whether the abused woman will leave the relationship (Rounsaville, 1978). A battered wife's objective marital dependency (no job, minimal income







53

potential, child-rearing role) is significantly related to severe violence (Kalmuss & Straus, 1982) while her subjective perception of her dependency is not related.

Clearly most research suggests that economic resources are related to whether the abused woman stays. Thus, economic resources was one of the variables studied here in relationship to the battered woman's destination once she leaves the shelter. The method utilized by Kalmuss and Straus (1982) was replicated here. Children

Many researchers have commented on the impact the

children and the parental relationship with the children, have on the decision making for the battering couple (Brown & Brazzle, 1982; Carlson, 1976; Gelles, 1983; Martin, 1976; Snell, Rosenwald & Robey, 1964; Stacey & Shupe, 1983; Straus, 1973).

Snell, Rosenwald, and Robey (1964) questioned wives who charged their husbands with assault and battery regarding their reasons for the action at that particular time, since the marriages were 12-20 years in length and had included other abuse incidents. The women in almost all the cases responded first that the children figured into their decision. Most common were responses describing the presence of an adolescent son, getting older, stronger, and more prone to be either retaliatory towards the father or negatively affected by the domestic disturbances.






54

Brown and Brazzle's (1982) sample indicated that a

reason for not leaving among women was that they feared that the husbands would seek the families out and harm the children or the wives. Twenty-one percent of those in the Stacey and Shupe study (1983) who did not leave the home after the first abuse incident revealed the children and maintaining the family as their reason. The mother's intense attachment to and concern for her children has been noted (Carlson, 1977). Children have been cited as both the reason for leaving and the reason for staying (Cristall, 1978); although in this particular study, twice as often they were the reason for staying.

Where battered women were compared with non-battered

women (Parker & Schumacher, 1977), there was no statistical difference between the number of children in either family. Additionally, when looking at battered women who have left the abuser compared to those who had not (Cristall, 1978), neither the age nor number of children appeared to be statistically different in either group. When the ages of the children are viewed as part of the definition for the women's resources (younger children being a limitation, older children being a potential resource because of their ability to be self-sufficient or assist in supporting the family), a strong correlation exists between a woman's limited resources and her remaining in the relationship (Pagelow, 1981). In another study (Snyder & Sheer, 1981) it was found no sociodemographic variables relating to the






55

children (e.g. age of children, number of children) as singular variables were predictive of the woman's destination upon leaving a shelter. In the study reported herein the researcher incorporated the role of children as a factor in the battered women's destination choice, but subsumed into the resource category as has been done by others (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981). Locus of Control. Learned Helplessness

Walker (1979) theorized, based on her interview and

clinical experience with battered women, that women find it difficult to escape an abusive relationship due to learned helplessness. In later research she reported on her work to test this hypothesis (Walker, 1984).

The learned helplessness phenomenon was identified and most often studied in a laboratory setting with animals and later with humans (cf. Seligman, 1975; Abramson, Seligman, & Teasdale, 1978). Since an operational instrument had not been developed to measure learned helplessness outside the laboratory, Walker (1984) devised several methods to indicate whether the learned helplessness phenomenon was instrumental. A series of questions were combined into scales, e.g., sexual abuse during childhood, childhood health, and combined to form a single measure of childhood contributors to learned helplessness. Also a measure of learned helplessness within the battering relationship was formulated by intercorrelating 15 indices from the interview questionnaire utilized. Walker (1984) noted the reliability







56

of both measures as less than ideal (.57 and .67 respectively) and suggested the need for a better measure of this construct in future studies. In addition, she combined several state instruments as indices of current state related to learned helplessness. Included among them was locus of control (Levenson's Locus of Control Scales). She conducted path analyses to determine if either childhood learned helplessness or relationship learned helplessness was determinant of current state. Both measures appear to influence current state.

Learned helplessness theory has been linked with

depression because of the perception by those experiencing it that they have no power or control over events in their lives (Abramson, Seligman, & Teasdale, 1978; Costello, 1978). A measure of perceived control, i.e., locus of control, was incorporated by Walker'(1984) into her study. She expected battered women would attribute more control to external sources than norm groups would, both on scales measuring powerful others and chance locus of control. She also anticipated lower scores on internal locus of control compared to norms. Battered women, both in and out of battering relationships, did score higher than norm groups on chance locus of control. Battered women still in battering relationships, however, did not score higher than norm groups on the powerful others locus of control scale.

Walker (1984) postulated that a woman still cohabitating with the abuser may not wish to acknowledge the







57

battered's control in her life, nor the unlikelihood of changing him or the environment to prevent further abuse. Battered women did, surprisingly, score significantly higher than norm groups on internal locus of control, attributing themselves as having substantial control over what happens to them. The Levenson Locus of Control Scales were again utilized in this research.

Authoritarianism and Monitoring

Neither authoritarianism nor monitoring have been examined previously for their relationship as potential determinants of battered women's destination. In fact, they had not been investigated in relationship to woman battering in general. In combination they represent salient features of the theory on the roots of violence proposed by A. Miller (1981, 1983, 1986). An examination of the results of research on both factors suggested their fruitfulness for study in this area.

Authoritarianism

As Rokeach (1960) theorized, an individual maintains an open belief system when the need to know dominates and a nonthreatening reality exists. A threatening reality causes the individual to become closed in his or her beliefs, the need to know diminishes, confusion occurs between the information and the informant, and identification with absolute authority takes place. Closing the belief system distances the individual from his or her own anxiety. One with a predominantly open belief system is a low dogmatic







58

(LD); a person having a predominantly closed belief system is a high dogmatic (HD).

High dogmatics indicate greater dependence on absolute authority, a greater vulnerability to be influenced by authority, and less discernment of the value of communications independent of the authority who delivers the communication (Bettinghaus, Miller, & Steinfatt, 1970; C.G. Kemp, 1963; Lazlo & Rosenthal, 1970; Powell, 1962). High dogmatics are more likely to adhere to a position supported by authority (McCarthy & Johnson, 1962). Low dogmatics are not differentially influenced by the authority of a communicator (Harvey & Hays, 1972) and they perceive authority figures more realistically, incorporating both positive and negative characteristics (C.G. Kemp, 1963). Low dogmatics are much less influenced by authority than high dogmatics (Ehrlich & Lee, 1969; Restle, Andrews, & Rokeach, 1964; Vacchiano, Strauss, & Hochman, 1969).

The close-minded person will avoid changing his or her environment or accepting new data and ideas (Vacchiano, Strauss, & Schiffman, 1968). The high dogmatic will focus on the future while minimizing the past or present (Castle, 1971; Jay, 1969; Rokeach, 1954, 1956; Rokeach & Bonier, 1960; Zurcher, Willis, Ikard, & Dohme, 1967). Rokeach (1954, 1956) suggested that this is done in order to create a sense of control regarding the course of life events. There is a direct relationship between authoritarianism and anxiety (cf. Byrne, Blaylock, & Goldberg, 1966; Castle,







59

1971; Norman, 1966; Smithers, 1970) and dogmatism as a defense mechanism (cf. Bernhardson, 1967; Hallenbeck & Lundstedt, 1966; D. Lee & Erhlich, 1971). In general, the high dogmatic can be characterized as having difficulty tolerating frustration and prone to conforming, while the low dogmatic is more tolerant of ambiguities and less accepting of traditional beliefs (Vacchiano, 1977). When information or a situation does not align with the high dogmatic's beliefs, he or she is threatened and avoidant (Hunt & Miller, 1968; G.R. Miller & Rokeach, 1968; Pyron, 1966). In general, close-minded individuals will recall less information when it is inconsistent and positively evaluate consistent information (Kleck & Wheaton, 1967). Low dogmatics are more likely to expose themselves to belief-discrepant information (Donohew, Parker, & McDermott, 1972).

monitoring

Self-monitoring, as conceived by M. Snyder (1974), is observation and control of the self based on the cues one perceives from the situation. The construct involves two aspects: astute sensitivity to the expression and presentation of others and an ability to use the situational cues to monitor and present oneself accordingly (M. Snyder, 1974).

High self-monitoring individuals are skilled at controlling and modifying expressiveness and behavior to align with the situational cues for appropriateness. Low self-







60

monitoring individuals are less able or likely to modify expressiveness and presentation to others (M. Snyder, 1979). The expression of a low self-monitoring individual appears to be more internally guided by his or her own feelings, attitudes, and experiences (M. Snyder, 1974, 1979) rather than by situational or interpersonal cues (M. Snyder & Monson, 1975). Individuals high in self-monitoring notice and accurately remember information about others more than low self-monitoring individuals (Berscheid, Graziano, Monson, & Dermer, 1976). Persons low in self-monitoring spend less time and effort thinking about the contingencies of a prospective date's behavior than those high in selfmonitoring (Berscheid, Graziano, Monson, & Dermer, 1976). High self-monitoring persons keenly attend to the subtleties of behavior and context and utilize the information to infer intentions and predict behaviors of others (Jones & Batumeister, 1976; Kulik & Taylor, 1979). High self-monitors are skilled at accurately judging the intended meaning of verbal expressions (Mill, 1984). Further, those who are high in self-monitoring are especially skilled at reading nonverbal expression and correctly determining underlying affective experience and emotional states of others (Geizer, Rarick, & Soldow, 1977; Krauss, Geller, & Olson, 1976). High self-monitoring individuals have access to very rich dispositional constructs organized around prototypes of others; low self-monitors excel at giving rich, informative descriptions of their own traits and dispositions (M. Snyder







61

& Cantor, 1980). Low self-monitoring individuals focus on information based on their inner states (M. Snyder & Tanke, 1976). A greater reliance on a partner's behavior as a guide to one's own behavior is characteristic of high selfmonitors (Ickes & Barnes, 1977). High self-monitoring individuals connect their identities with their external environment and choose to enter more clearly defined situations rather than ambiguous ones (M. Snyder & Gangestad, 1982; Sampson, 1978).















CHAPTER III

METHODOLOGY

This study was designed to distinguish between those

battered women who, following a shelter stay, did not return to their abusers and those battered women who did return to their abusers in terms of key variables drawn from previous research (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981; Walker 1979, 1984) and the psychodynamic theory of A. Miller (1981, 1983, 1986). Specifically, the relationship was studied between battered women's destination and resources; number of previous separations; internal, powerful others, and chance locus of control, authoritarianism; and monitoring. In Chapter III the research methodology of the investigation is described. The chapter includes the following: population and sample, the sampling procedures, the instrumentation, data collection procedures, and data analysis.

Population and Sample

The population base for this study was battered women who resided in two battered women's shelters located in Florida, the Gainesville and the Ft. Myers shelters, and the three metropolitan Atlanta shelters. The Gainesville shelter mainly served a 15-county area in north central Florida. The Ft. Myers shelter mainly served a six-county


62






63

area in southwest Florida. Two of the Atlanta area shelters were in suburban locations: the Cobb county shelter, approximately 12 miles northwest of downtown Atlanta, and the Clayton County shelter, approximately 12 miles south of downtown Atlanta. These two shelters served mainly residents of their counties plus women and their children from several adjoining counties in the northwestern portion of Georgia. The third metro shelter was located in midtownAtlanta mainly serving residents of urban Fulton and Dekalb counties, as well as occasionally women and their children from the adjoining metro counties. As is common with battered women's shelters throughout the United States, at times each of these shelters operated at capacity and placed women's names on a waiting list for admission. In the case of a woman facing imminent potential for severe violence personnel in each of these shelters worked with those in other shelters for immediate placement and, via law enforcement or other avenues, transported the woman to a distant shelter.

Battered women came to these shelters following

telephone contact with the particular shelter to determine that they met the criteria for admission. The admissions criteria for each shelter was the same: the women must have been recently battered and be in need of safe shelter without available refuge. Battered women were admitted to the shelter at any time of the day or night, any day of the week, depending on when they sought assistance. many women






64

who sought help from the shelters had exhausted financial and familial support.

The collection of demographic information allowed for a comparison of this sample to some demographic information from other known shelter samples. The demographic characteristics of women from these shelters did not differ significantly from those of shelter samples in other studies in the southeast or the nation (Bell, 1977; Carlson, 1977; Label, 1979; Pagelow, 1977c, 1981; Stacey & Shupe, 1983). Although battering occurs across socioeconomic classes, shelter samples are usually more representative of lower and lower middle class socioeconomic classes; women with greater resources have other alternatives for safety (Pagelow, 1981; Stacey & Shupe, 1983). In the study reported herein, basic demographics including age, race or ethnic background, number of children, and residential address before entering shelter (to determine urban, suburban, or rural) were collected (Appendix A). One of the independent variables, number of previous separations, was also included on this portion of the questionnaire.

Many of the women who came to the shelters heard about the services via an intervening police officer, public service advertising, a friend or family, or the advertised battered women's hotlines. Referrals to the shelters could come from the emergency rooms of the hospitals in the area. In addition, other referrals were made by the community crisis hotlines, police, clergy, rape crisis programs,







65

women's health care clinics, and university and medical communities.

To be included in the study, the women must have been battered and must have completed 8 years of schooling according to their response on the shelter intake forms. The 8 years of schooling was deemed appropriate indication of a baseline ability to read and understand the instruments.

Sampling Procedures

The subjects for this study were a volunteer sample of battered women who utilized these shelters during the months of May through December 1987 and met the criteria for the study. Data were not collected every day at each of these shelters, but on the days that the data were collected all women in the shelter, except for those who had already participated, were asked to participate in the study. The sample size was 72 women.

During the days data were collected, all potential subjects admitted to the shelters were given a letter explaining the study, describing the criteria for inclusion, and inviting their participation (see Appendix B). The potential subjects were questioned to determine if they met the criteria for inclusion and, if so, were they willing to participate in the study. This procedure was continued until the pool of individuals who met the inclusion criteria and who agreed to participate totaled 78 women. Completed questionnaires and follow-up destinations were obtained on







66

72 of those 78 women; follow-up was not obtained on 6 women. An additional 6 women who did not participate, either due to choice or education requirements, and their destinations at exit were noted. The reasons for their non-participation were recorded. One woman, although she met the eighth-grade completion requirement, did not possess adequate reading skills to answer the questionnaire; she returned to her abuser. Two women at two different shelters took the questionnaires to complete, were interrupted in the process (one for child rearing, the other to communicate with a social service employee from another agency who was assisting her) and did not complete the questionnaires or make follow-up arrangements. Both left their shelters hastily. one returned to the battered and the other flew to California to move in with her relatives. one woman was incapable of completing the questionnaire due to disorientation and psychological disturbance, and she left the shelter to be psychiatrically hospitalized. One woman did not participate because of a language barrier; she was Spanish speaking and reading. The researcher did not have a Spanish version prepared for her use. She did not return to the abuser.

Instruments

In all, four instruments with a combined total of 92 questions were used to gather the information on authoritarianism, monitoring, resources, and locus of control (number of previous separations was answered by the subject






67

on the demographics questionnaire--Appendix A). This study was designed to use short to moderate length instruments administered by a trained administrator so as not to additionally burden these subjects who were already experiencing a crisis. Battered women usually enter the shelter immediately after what Walker (1979, 1984) has referred to as the acute battering incident. Many battered women have described how difficult it is to leave the home for an unknown refuge environment following such an emotionally charged incident (Martin, 1976; Pagelow, 1981; Walker, 1979, 1984). At that time battered women are fearful of the abusers, as well as afraid of the unknown, embarrassed to make the domestic problem public, and unsure of social support, emotional or otherwise. Knowing that the abuser might perceive the woman's actions of seeking safety and assistance for her and the family as escalation in the current conflict, the woman must muster a great deal of courage to seek shelter help. Battered women battle their own self-blame for the domestic difficulties (Walker, 1979) and often doubt that others will believe they have been abused (Rounsaville, Lifton & Bieber, 1979). Lengthy instrumentation asking the women to respond to many selfstatements at this crisis juncture is thought by many (Dobash & Dobash, 1981, 1981; Schechter, 1982) to be intrusive and to debilitate crisis efforts aimed at supporting her decision to seek safety and get help for the family. with the additional consideration that







68

administering instrumentation to women in crisis is a sensitive issue, the investigator chose the following four measures.

Measure of Resources

For this investigation, the measure for resources was the Resources Index (Appendix C) utilized by Kalmuss and Straus (1982). It is the sum of scores on three dichotomous variables: whether the woman is unemployed or not, whether the woman has children at home of ages 5 or younger or not, and whether her mate earns 75% or more of the total household income or not. The range for the values on this index is 0 (low resources) through 3 (high resources). The internal consistency of the Resources Index in the Kalmuss and Straus study (1982) was .59, measured by Cronbach's alpha coefficient of reliability. Measure of Locus of Control

Locus of control was measured by the Levenson Locus of Control Scales (Levenson, 1973) (Appendix D). The locus of control construct describes individuals' causal beliefs regarding reinforcements which occur to them. They may have a generalized expectancy that reinforcements are contingent upon their own behaviors (internal control). They may also generally believe that reinforcements are contingent upon forces outside their control, e.g., chance, fate, a deity, luck, powerful others (external control). The Levenson Locus of Control Scales are multidimensional, measuring beliefs regarding personal control (Internal Scale),







69

powerful others (Powerful Others Scale), and fate or chance (Chance Scale). A person with a high I Scale score would believe that he or she has great deal of control over his or her life. A person with a high P Scale score generally ascribes powerful others as determining outcomes. A person with a high C Scale score perceives chance or fate as significantly contributory to events.

The Locus of Control Scales are comprised of 24 items,

8 per scale, although they are presented interspersed to subjects in one unified series. Responses are noted on a Likert Scale ranging from strongly disagree (-3) through strongly agree (+3) assigned to each statement. All of the items in the Levenson Scales are stated personally so that the respondent is being measured on his or her perceptions regarding control, not what he or she believes "people in general" think about control (Levenson, 1981).

Kuder-Richardson reliabilities for the I Scale range

from .51 to .67, for the P Scale from .72 to .82, for the C Scale from .73 to .79 (Levenson, 1973, 1974; Wallston, Wallston, & DeVellis, 1978). Split-half reliabilities, utilizing the Spearman-Brown formula, are .62 for the I Scale, .66 for the P Scale, and .64 for the C Scale. Sevenweek test-retest reliabilities are .66, .62, and .73 for the I, P, and C Scales (Lee, 1976).

The P and C Scales are both related to external locus

of control, so it would be expected that in validity studies they would correlate to some extent with each other. In







70

fact, studies of their correlation range from .41 to .60 (Caster & Parsons, 1977; Levenson, 1973; Scanlan, 1979; Wallston, Wallston, & DeVellis, 1978). Only minimal correlation (-.25 to .19) has occurred between the P and C factors with the I factor (Caster & Parsons, 1977; Levenson, 1973; Scanlan, 1979; Wallston, Wallston, & DeVellis, 1978).

The Rotter I-E scale is one of the original locus of

control scales and the most often referenced as the standard in locus of control measurement. Convergent validity is strongest between the C and I Scales with the Rotter I-E Scale, ranging from .43 to .56 for the C factor and from

-.15 to -.41 for the I factor (Donovan & O'Leary, 1975; Hall, Joestring, & Woods, 1977; Levenson, 1972).

There is negligible correlation between the Levenson Scales and the Marlowe-Crowne Social Desirability Scale. In the two studies conducted (Levenson, 1972; Wallston, Wallston, & DeVellis, 1978), the correlations to the Marlowe-Crowne Social Desirability Scale for the I Scale were .04 and .09, for the P Scale .04 and .11, and for the C Scale -.10 and .08.

Borrero-Hernandez (1979) noted several relationships

between the Levenson Scales and other personality variables. The I Scale is positively related to measures of sociability on the California Psychological Inventory (CPI). The C Scale negatively relates to sense of well-being and responsibility on the CPI and positively relates to guilt







71

proneness on the 16 PF. The P Scale correlates significantly with suspiciousness on the 16 PF. Measure of Authoritarianism

Authoritarianism was measured by the Rokeach Dogmatism Scale, Form E (Rokeach, 1956). This is a 40-item scale (see Appendix E). Responses can range from "I agree very much"

(+3) through "I disagree very much" (-3) on a Likert-type Scale. The close-minded or high dogmatic (HD) person believes that authority is absolute, accepts or rejects others based on their agreement or disagreement with authority, and maintains this cognitive system which protects the individual for anxiety. The low dogmatic (LD) person does not hold authority as absolute, accepts or rejects other on different premises than their alignment or disagreement with authority, and has more tolerance for ambiguity and the anxiety that can accompany changing environments (Vacchiano, 1977).

The construct validity of authoritarianism has been supported in numerous studies. A common core of authoritarian factors has been noted between the California F Scale and the Dogmatism Scale (Kerlinger & Rokeach, 1966), although the Dogmatism Scale has been found to be more independent of ideological bias than the California F (Warr, Lee, & Joreskog, 1969). Many studies support the Dogmatism Scale as a measure of authoritarianism (Barker, 1963; Costin, 1965; Hanson, 1968, 1970; Plant, 1960; Rokeach, 1967). High dogmatic persons show a greater dependence on







72

authority than do low dogmatic persons (cf. Bettinghaus, Miller, & Steinfatt, 1970; Kemp, 1963; Powell, 1962; Restle, Andrews, & Rokeach, 1964; Vidulich & Kaiman, 1961). Measure of Monitoringi

Monitoring in this study was measured by the Snyder Self-Monitoring Scale (M. Snyder, 1974) (see Appendix F). It is a 25-item scale which measures whether a person monitors external cues from others to guide his or her own self presentation or if he or she presents and acts in accordance with his or her own affective state (M. Snyder, 1974). Low self-monitoring individuals are more knowledgeable about their own dispositions, attitudes, and traits; high self-monitoring persons are much less dispositionally guided or discerning (M. Snyder & Cantor, 1980).

The KR-20 reliability of the Self-Monitoring scale is .70; test-retest reliability is .83 (M. Snyder, 1974). The Self-Monitoring Scale measures a different construct than those measured by the Marlowe-Crowne Social Desirability Scale, the IIMPI Psychopathic Deviate Scale, the c (Chameleon) scale of the performance style test, the Christie and Geis Machiavellianism scale, and the AlpertHaber Achievement Anxiety Test (M. Snyder, 1974). Numerous studies support the validity of the Self-Monitoring Scale utilizing peer ratings and samples which include college students, insurance and sales personnel, psychiatric patients, and stage actors (Caldwell & O'Reilly, 1982; Gabreyna & Arkin, 1980; Ickes & Barnes, 1977; Lippa, 1978;







73

M. Snyder, 1974; M. Snyder & Monson, 1975; M. Snyder & Swann, 1976; M. Snyder & Tanke, 1976).

Data Collection Procedures

Data were collected by either a trained, full-time

staff person from a shelter, the researcher, or a research associate. At the Ft. Myers, Gainesville, and Cobb shelters data were collected by a trained full-time staff person. At the Clayton shelter and during the latter months of collection at the Cobb shelter, the data were collected by the researcher. At the downtown metro Atlanta shelter the data were collected by either the researcher or the trained research associate. The investigator met individually with each of the selected staff persons and the research associate to train them in the procedures. Training involved the following components: (a) use of the letter of introduction to the subject inviting participation, (b) determination that potential subject meets criteria for inclusion, (c) obtaining informed consent, (d) benefits/risks of participation,

(e) methods to assure confidentiality, (f) instruments and instructions, (g) answering questions during administration,

(h) procedures for follow-up contact, and (i) recording data on the master chart. Each of the individual sessions was 2 to 3 hours long. To determine that each administrator understood and could uniformly perform data collection, the investigator conducted a follow-up session after the participation of at least the first two subjects. The investigator monitored the data collection throughout the







74

period by contact with the trained administrators and the shelter staffs and the review of the data as they were being collected. Additionally, the administrators collecting the data could call the investigator at any time.

Battered women were admitted to these shelters during any of the 24 hours in a day. Given up to 48 hours to emotionally adjust following the acute battering incident, each woman met with a staff member to complete an intake questionnaire and participate in an initial intake interview. Completing the questionnaire was handled as Part II of the shelter's intake procedure.

Following completion of Part I of the intake procedure, which was the shelter's standard intake, the administrator had the potential subject read the letter of introduction (Appendix B). The administrator discussed requisites for participation in the study and potential benefits and risks of participation. Benefits of participation included the following: (a) The subject contributes to research which increases the understanding of the dynamics of battering and adds to theory which can help end conjugal violence against women; (b) the subject assists the participating shelter(s) in increasing the knowledge about their clients so they can provide more effective services to battered women and their children; (c) the subject aids those helping battered women, shelters, legal, criminal, and other social service agencies, to better understand their clients and to dispel stereotypes and myths about battered women; and (d) the subject assists







75

in providing information to battered women which increases self-understanding and decreases feelings of isolation, guilt, and blame. Risks and/or drawbacks of participation were the time required to complete the questionnaire, the adherence to the shelter's follow-up procedure following departure, responding to questions which solicit personal opinions when the subject was in a post-traumatic time period, and reliance on the shelter and principal investigator to assure confidentiality. The administrator recorded on the master chart the subject number, date of administration, and whether the woman qualified and agreed to participate in the study. If the potential subject was not participating, the administrator recorded whether this was due to not meeting the requisites or a refusal to participate.

Each woman volunteering to participate in the study was given a packet which included instructions for the instruments, the instruments themselves, one informed consent form, and a manila envelope for return of the packet (see Appendix G). At this time the subjects completed the demographic questionnaire and the paper and pencil instruments required for this investigation. The demographic questionnaire contained 6 items and took a few moments to complete. The combined instruments contained 92 items total and took from 45 to 90 minutes to complete.

The administrator was present with the subject throughout data collection and trained to answer any questions which







76

arose regarding individual items or the study. The administrator was trained to do this in a way which maintained the integrity of the instrumentation and purpose of the investigation, yet assisted the woman to answer the questionnaire adequately. The administrator collected the completed questionnaire and recorded on the master chart that the following had been totally completed: the informed consent form, the 6 demographic items, the 3 resources items, the 24 locus of control items, the 40 dogmatism items, and the 25 monitoring items.

Subjects' names were not placed on the instruments to

ensure confidentiality. Each packet was given a code number which was paired with the respondent's name on a separate listing.

The trained administrator, following collection of the completed questionnaire, discussed and then recorded the preferred and contingent options for follow-up contact. The woman could select options from the following methods:

(a) telephoning the shelter during a prearranged time,

(b) returning to the shelter for a face-to-face contact (often women return to the shelter to pick up mail, including aid checks, which are addressed to the women at the shelter),

(c) communicating through a third party who agreed to provide the necessary information about the woman to the shelter,

(d) returning a stamped post card addressed to the shelter's post office box with a question to indicate destination on it, (e) responding to a follow-up letter mailed to the woman







77

in a plain envelope including a pre-addressed, postage reply form, and (f) agreeing to receive a telephone call from the shelter.

During the shelter stay, the women participated in peer counseling sessions, met with staff for individual help, and established contact with other social service agencies that provided a more permanent source of food, housing, and transportation when needed. If the women came to the shelter with children, the children also participated in appropriate services, e.g., play therapy. Following a variable length of stay, each woman made the decision regarding her destination following her temporary refuge at the shelter. All shelters had been required by their funding agencies to provide up to 30 days of safe refuge. on occasion, a woman will be planning her own housing away from the abuser and the housing would not be ready for her immediately at the end of the 30 days. In those exceptional cases the shelters granted extensions. Each woman met with a staff member to discuss her departure plans.

At the appropriate time, the trained administrator, in

concert with program staff, initiated the follow-up procedure and recorded the destination as known during the follow-up period (i.e., from 2 to 6 weeks following departure) on the master chart. Working with a battered woman when she leaves a shelter program and returns to live with the abuser necessitates certain precautions. While the battered woman is at the shelter, attempts are made to keep the shelter







78

location a secret and the abuser uninformed about the woman's whereabouts. These efforts are taken to increase the possibility of providing a safe place for women and their children. Increased police surveillance is some- times requested by shelters when the dangerousness and lethality of the abuser is judged to be particularly great or the battered issues threats. Shelter personnel do not release names of women or children housed there. When the woman returns to the abuser, contact from the shelter staff needs to be discreet to continue efforts of safety provision for both the battered woman and the shelter program.

Data Analysis

The researcher posed the following hypotheses.

1. There is no difference in authoritarianism between battered women who do not return to the abuser following a shelter stay and those who do return.

2. There is no difference in monitoring between

battered women who do not return to the abuser following a shelter stay and those who do return.

3. There is no difference in resources between battered women who do not return to the abuser following a shelter stay and those who do return.

4. There is no difference in number of previous

separations between battered women who do not return to the abuser following a shelter stay and those who do return.

5. There is no difference in locus of control (internal, powerful others, chance) between battered women who do






79

not return to the abuser following a shelter stay and those who do return.

6. There is no relationship between authoritarianism monitoring, resources, number of previous separations, and locus of control (internal, powerful others, chance) and the destination of battered woman following a shelter stay.

Hypotheses one through five were analyzed using one-way analyses of variance. Hypothesis six was analyzed using logistic regression. An alpha level for significance was established at .05 as a conventional level of significance. Initially, a Bartlett's test was utilized to determine departures from normality. The results of this test suggested that the sample was not within the range of a normal distribution; therefore, although either discriminant analysis or logistic regression was considered for analysis, logistic regression was chosen since it is less sensitive to departures from normality.

Using logistic regression analysis the researcher can distinguish between two groups and produce a predictive equation. In logistic regression analysis the variables are weighed and linearly combined the variables mathematically to force the optimum distinction between groups.

Logistic regression in a two-group case is the same as a multiple regression where the dependent variable (in this study not returning to the abuser or returning to the abuser) takes on the values of 1 and 0 and maximally discriminates between the two groups (Kerlinger & Pedhazer, 1973).















CHAPTER IV

RESULTS

The results are presented in this chapter. Descriptive statistics were used to characterize the sample group while the research hypotheses were analyzed with one-way analysis of variance (hypothesis one through five) and logistic regression (hypothesis six).

Demographic Information

Demographic information is presented here in order to characterize the overall sample of battered women, as well as subsets of the sample: those who did not return to the batteries and those who did return following the stays at the shelters. The total sample included 72 women. Those who did not return to the batteries included 42 women; those who did return to the abusers numbered 30. In addition to the variable data, general demographic information was gathered on each subject: age of subject, race or ethnic background of subject, number of children, whether subject had left mate before or not, and city of residence before coming to shelter.

A breakdown of the study participants by age appears in Table 4-1. The sample was predominantly in their 20s and 30s (86%), with the mean age being 30.32 years (SD = 7.32); this is quite similar to shelter samples from other


80







81

Table 4-1

Acte of Samvle



Returned Did not Total
Agle to Abuser Return Sample
n M% n M% n M%


17-20 1 (3.3) 3 (7.1) 4 (5.6)

21-25 6 (20.0) 9 (21.4) 15 (20.8)

26-30 12 (40.0) 12 (28.6) 24 (33.3)

31-35 2 (6.7) 7 (16.7) 9 (12.5)

36-40 6 (20.0) 8 (19.0) 14 (19.4)

41-45 2 (6.7) 3 (7.1) 5 (6.9)

46-50 0 0 0

51-55 1 (3.3) 0 1 (1.4)







research (see Appendix H). Mean age for battered women who returned to their abusers was 30.80 years (SD = 7.67); mean age for those who did not was 29.98 years (SD = 7.14).

A breakdown of the study participants by race appears in Table 4-2. The sample had a large percentage of nonwhites (41.7%) which is common for samples including inner city shelters (e.g., D.K. Snyder & Fruchtman, 1981; D.K. Snyder & Scheer, 1981) (see Appendix H). Because there appeared to be an effect due to race, a chi-square analysis was conducted. With a 2 = 16.86, p = 0.000, which is significant for a = .05 and d.f. =1, an effect due to race







82

Table 4-2

Race of amvle



Returned Did Not Total
Race to Abuser Return Sample
n M n M n (-%)



White 26 (86.7) 16 (38.1) 42 (58-3)

Nonwhite 4 (13.3) 26 (61.9) 30 (41.7)









was found i.e., white participants returned to their abusers more often than their nonwhite counterparts. This is discussed in Chapter 5.

The participants' numbers of children are categorized in Table 4-3. The average number of children for the total sample was 1.74 (SD = 1.30). Battered women who returned to the abusers in the study had an average of 1.63 children (SD = 1.45); average number of children for the abused women who did not return to the batteries was 1.81 (SD = 1.19).

A breakdown of the study participants by whether they had left their mates before or not is presented in Table 44. Nearly equal percentages of participants had left their mates before regardless of destination in this study.







83

Table 4-3

Number of Children



Number Returned Did Not Total
of to Abuser Return Sample
Children n (%) n (%) n M



0 8 (26.7) 7 (16.7) 15 (20.8)

1 7 (23.3) 10 (23.8) 17 (23.6)

2 7 (23.3) 12 (26.6) 19 (26.4)

3 6 (20.0) 10 (23.8) 16 (22.2)

4 1 (3.3) 3 (7.1) 4 (5.6)

5 0 0 0

6 1 (3.3) 0 1 (1.4)









Table 4-4

Previous Separations



Previous Returned Did Not Total
Separa- to Abuser Return Sample
tions n (t) n (t) n M



No 9 (30.0) 9 (21.4) 18 (25.0)

Yes 21 (70.0) 33 (78.6) 54 (75.0)







84

Population size of the city in which the study participants lived before coming to the shelter is categorized in Table 4-5. An effect related to population appeared likely, so a chi-square analysis was conducted. A 2 = 10.03, p 0.018, signifies an effect for a = .05 and d.f. = 3; battered women from the small towns were more likely to return to the abuser and as the population of residence increased, battered women were less likely to return to their mates. This result is considered further in Chapter 5.

The Rokeach Dogmatism Scale to measure authoritarianism, Snyder Self-Monitoring Scale was used to measure monitoring, the Kalmuss & Straus (1982) 3-item method to measure resources, a straightforward item on the





Table 4-5

Population of City in Which Subject Resided Before Shelter



Returned Did Not Total
Population to Abuser Return Sample
Size n M n M n M



0-20,000 17 (56.7) 10 (23.8) 27 (37.5)

21,000- 6 (20.0) 9 (21.4) 15 (20.8)
50,000

51,000- 4 (13.3) 8 (19.0) 12 (16.7)
200,000

over 200,000 3 (10.0) 15 (35.7) 18 (42.9)







85

questionnaire to determine number of previous separations, and the Levenson Locus of Control Scales to measure internal

(I), powerful others (P), and chance (C) loci of control. The means and standard deviations for the total sample and the subsets according to destination can be found in Table 4-6. Because a potential effect was likely related to the 75%-or-more-of-income-earned-by-mate item (one of the three from the resource measure) and destination, a chi-square analysis was conducted. Since 2 = 4.46, p = 0.035, which is significant for a = .05 and d.f. = 1, it was found that women whose abusers earn 75% or more of the family income returned to their mates more often than would be expected. Race was not a significant factor impacting this relationship.

Analysis of Hypotheses One Through Five

The first five hypotheses were tested using one-way

ANOVA tests. The first hypothesis was that there would be no difference in the authoritarianism scores between battered women who, following shelter stays, did not return to their abusers and battered women who did return. The one-way ANOVA was not significant at the .05 level (F = 2.44, p = 0.123); therefore, the null hypothesis was not rejected.

The second hypothesis was that there would be no

difference in the monitoring scores between battered women who, following shelter stays, did not return to their batterers and abused women who did return. The one-way






86






IV f-4 H cn im 0 Ln H H N
%0 r, ON qw qT U) H m 0 Ln

H CI4 H
to
4J
0
Ell
00 0 ON 0 OD T-4 CV co m
0 0 CV Iq m U) co N
0; 0; r4 8 8 8 r U C
Ln H H
r-i



4
41 co Ln Ln 0 co CN 0 T-4 fn D
w w r, w U) Iw IV r- co Ln m
N
41
0
z



0 LO co 0 D N Im m
%D N m Iq m ko T-4 0 %D
c4 a; 4 8 8 c; 9
%D rq
ri

:3 4
0 4)
tv to
4) 0
1 0 LO CA 0 0 ON r-4
> N r, m Ln In 1w r-4 %D 10
0 In
41 to C U
4J cq

44 0
ta 41
r. 0)
0 r-i 0 r- o o
r, Ln 1-4 qr Iw (n m %D 0 %D
r U
U) r-4 C9
> H


(1) to
0 0 r-4
to (1) 0 rq w 0
0 >
m 0 41 4)
go 10 0 (d $4
U) rq (1) 4) Z 04 th 4-) 0 :3
$4 V) >1 V) 4J m r. 0 U U r-4
%D m 9 w 0 IU 0 >4 4-4 0 0 0 0
1 C V -4 4) r-4 r. Z.0 0 -rq r-4 m r-4 44 4 m
to m u 0 V 00 V :3 0 V
0 $4 M m tu r, r. 14-4 4) r.
tA to 0 4-) :% '0 V) 0 4) 4) m M 0 $4 w u 0
rq 9: to 4 rq 0 r-i r. r I to '0 u 0 :3 z u
A to 4) V r. IA 4J -r4 0 0 m 04 V t () (a
It 4) Z, :3 0 4) 0 4 0 Ln (0 4) r. 44 03 r. 44
E-4 0 4 Z IX Z U H r, w z M H 0 04 U 0







87

ANOVA was not significant at the .05 level (F = 0.93, p = 0.338); therefore, the null hypothesis was not rejected.

The third hypothesis was that there would be no

difference in the resources scores between battered women who, following shelter stays, did not return to their abusive mates and those battered women who did return to their batteries. The one-way ANOVA was not significant at the .05 level (F = 0.97, p = 0.328); therefore, the null hypothesis was not rejected.

The fourth hypothesis was that there would be no

difference in the number of previous separations between battered women who did not return to their abusers after the shelter stays and those women who did return to the abusive mates. The one-way ANOVA was not significant at the .05 level (F = 1.23, p = 0.271); therefore, the null hypothesis was not rejected.

The fifth hypothesis was that there would be no difference in internal, powerful others, and chance locus of control scores between sheltered battered women who did not return to their batteries and those who did return. Each of the Levenson Locus of Control Scales is independent of the other. The one-way ANOVA for internal locus of control scores was not significant at the .05 level (F = 0.16, p

0.689). The one-way ANOVA for powerful others locus of control scores was not significant at the .05 level (F = 0.07, p = 0.796). The one-way ANOVA for chance locus of control






88

scores was not significant for at the .05 level (F = 3.48, p = 0.066). Since ANOVAs were not significant for all locus of control scores, the null hypothesis was not rejected.

Analysis of Hypothesis Six

The sixth hypothesis was tested using logistic regression. The sixth hypothesis was that there would be no relationship between authoritarianism, monitoring, resources, number of previous separations, locus of control (internal, powerful others, and chance), and the destination of battered women following shelter stays. The logistic regression analysis is presented in Table 4-7. The regression analysis indicated that the variables, in general, were not good predictors of destination following shelter stay. Therefore, the null hypothesis was not rejected.

Summary

No significant differences between the groups occurred on the hypothesized variables of authoritarianism, monitoring, resources, number of previous separations, internal locus of control, powerful others locus of control, and chance locus of control. Some distinctive differences were indicated in this sample between battered women who returned to their abusers and battered women who did not, specifically in the demographic data of race, population of city where subject resided, and husband's percentage of income. No significant relationship was established between these variables and the destination of battered women following a shelter experience.






89






(n Cl cc o %D u L co
-4 IA cl v 0 0 N r, Ln
.q 0 -W O' co en r, % co 0 C; Cl C; I-L C






1- 0 0 0ON0 0 0 0,
0









0

0 co v 0 U 0O
mO 0 m %D m N ~ N

I0 0 0 0 0 0 0

0




4.)
44t C3 r ) ) t



4.)1
0





04 00

4.) 0 4



4.i 019 1.
o 40 (0 cc W 44 N0C
w ~ ~ 04 4J 4 0 04 0 0- 04 v qH 4.) k U V (0 Z
40 Q ) 0 0 0 0 0 0 04 ( 0


to (oR0 4 4
r -'q-4 q4 -f 0 C 4-)

IQ 0p 4 4J C W 04 4JZ 3:U o tv 140 0. 0 0 4) r.0 00 Q
004 >0 Q. 4 z 0x-4E H u 4-4 u u














CHAPTER V

DISCUSSION

Summary

The aim in this study was to determine key variables which would distinguish between battered women who did not return to abusive mates and battered women who did following their stays at refuge shelters. The variables, drawn from previous research (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981; Walker, 1979, 1984) and from the psychodynamic theory of A. Miller (1981, 1983, 1986), were resources, number of previous separations, internal locus of control, powerful others locus of control, chance locus of control, authoritarianism, and monitoring.

A total of 72 battered women who stayed at shelters in Florida and Georgia agreed to participate in the study and completed questionnaires while they were at the shelters. Destination information, i.e., whether they were living with the abuser or not, was recorded during the follow-up period of 2 to 6 weeks after their departures from the shelters.

The demographic information from the questionnaires indicated that battered women who did not return to their abusers and battered women who did were similar to each other in several ways. Most of the women were in their 20s and 30s, had probably left their mates before, and either


90




Full Text
130
Caster, D.U., & Parsons, O.A. (1977). Locus of control in
alcoholics and treatment outcome. Journal of Studies
in Alcohol. 38, 2087-2095.
Castle, T.J. (1971). Temporal correlates of dogmatism.
Journal of Consulting and Clinical Psychology. 36. 70-
81.
Chan, K.B. (1979). Husband-wife violence in Toronto.
Dissertation Abstracts International. 39 6995A-6996A.
Costello, C.G. (1978). A critical review of Seligman's
laboratory experiments on learned helplessness and
depression in humans. Journal of Abnormal Psychology.
87, 21-31.
Costin, F. (1965). Dogmatism and learning: A follow-up of
contradictory findings. Journal of Educational
Research. 59, 186-188.
Cristall, L.M. (1978). A comparison of androgyny and self-
actualization in battered women. Dissertation
Abstracts International. 39, 5039B-5040B. (University
Microfilms No. 79-07826)
Dobash, R.E. (1981). When non-sexist language is sexist.
In B. Warrior (Ed.), Working on wife abuse (pp. 87-88).
Cambridge, MA: Betsy Warrior.
Dobash, R.E., & Dobash, R.P. (1971). Love, honour and
obey: Institutional ideologies and the struggle for
battered women. Contemporary Crises. 1, 403-415.
Dobash, R.E., & Dobash, R.P. (1977-78). Wives: The
'appropriate' victims of marital violence. Victi
mology: An International Journal. 2, 426-442.
Dobash, R.E., & Dobash, R.P. (1979). Violence against
wives: A case against the patriarchy. New York: Free
Press.
Dobash, R.E., & Dobash, R.P. (1981). Social science and
social action: The case of wife beating. Journal of
Family Issues. 2, 439-470.
Dobash, R.E., & Dobash, R.P. (1983). The context specific
approach to researching violence. In D. Finkelhor,
R.J. Gelles, G.T. Hotaling, M.A. Straus (Eds.), The
dark side of families: Current family violence
research (pp. 261-276). Beverly Hills: Sage
Publications.


77
in a plain envelope including a pre-addressed, postaged reply
form, and (f) agreeing to receive a telephone call from the
shelter.
During the shelter stay, the women participated in peer
counseling sessions, met with staff for individual help, and
established contact with other social service agencies that
provided a more permanent source of food, housing, and
transportation when needed. If the women came to the shelter
with children, the children also participated in appropriate
services, e.g., play therapy. Following a variable length of
stay, each woman made the decision regarding her destination
following her temporary refuge at the shelter. All shelters
had been required by their funding agencies to provide up to
30 days of safe refuge. On occasion, a woman will be
planning her own housing away from the abuser and the housing
would not be ready for her immediately at the end of the 30
days. In those exceptional cases the shelters granted exten
sions. Each woman met with a staff member to discuss her
departure plans.
At the appropriate time, the trained administrator, in
concert with program staff, initiated the follow-up procedure
and recorded the destination as known during the follow-up
period (i.e., from 2 to 6 weeks following departure) on the
master chart. Working with a battered woman when she leaves
a shelter program and returns to live with the abuser
necessitates certain precautions. While the battered woman
is at the shelter, attempts are made to keep the shelter


43
ment of this monitoring skill comes at the cost of under
development of the child's own ability to perceive his or
her own needs.
To summarize, A. Miller (1981, 1983, 1986) proposed
that violence occurs because of the following: Parents have
narcissistic disturbances from their own childhood traumatic
experiences which they have repressed. They treat their own
children cruelly and unsympathetically by splitting off
their "bad selves" on their weaker and more vulnerable
children. Society colludes with the process by mandating
parents to rear obedient and conforming children and viewing
acting out not as children's reactions to their
traumatization but as manifestations of the bad nature of
children who have not been properly reared into subser
vience. Society protects parents, provides avenues histori
cally for socially acceptable violence, e.g., beating wives
and children as appropriate disciplinary measures, striking
students for wrongdoing. Children who are disturbed through
such upbringing will either dull their self-sensing abili
ties, harm others weaker than themselves, become self
destructive, deny their own traumatization and needs, refine
their abilities to monitor and please others in a quest for
love and narcissistic fulfillment, or some combination of
these qualities.
Variables Relating to the Battered Women's Destination
Researchers espousing different theories have examined
a number of variables relating to the question of battered


8
legacy theory does not completely define the determinants of
women's destination.
The most relied upon theoretical explanation regarding
battered women is that some women experience a greater
degree of learned helplessness than others, perceiving
themselves as powerless to alter their situations for the
better, regardless of what they do (Walker, 1979). The
battering cycle of tension building, acute incident, and
loving contrition powerfully reinforces the women's learned
helplessness. Walker (1984) expected that battered women
who remained with abusers would perceive themselves as
having less control over their situations than those who did
not remain with abusers, because they perceived others as
exercising greater control over them. Instead, the findings
indicated that battered women in or out of battering rela
tionships saw themselves as having a great deal of control
over what happens in their lives, more so than a general
population norm. They did not view powerful others, e.g.
the abuser, as having a lot of control over their lives
(Walker, 1984).
Additional investigation of potential determinants of
the destination of domestic violence victims following their
shelter stays is needed. A potential source for additional
variables is A. Miller's (1983) theoretical framework for
violence in society. She has proposed that violence in
society occurs because parents and others, exercising the
prevalent child rearing practices of the past several


132
Gaquin, D.A. (1977-78). Spouse abuse: Data from the
national crime survey. Victimology: An International
Journal. 2, 632-643.
Gayford, J.J. (1975). Battered wives. Medicine. Science
and Law. 15, 237-245.
Geizer, R.S., Rarick, D.L., & Soldow, G.F. (1977).
Deception and judgment accuracy: A study in person
perception. Personality and Social Psychology
Bulletin. 3, 446-449.
Gelles, R.J. (1972) The violent home: A study of
physical aggression between husbands and wives.
Beverly Hills: Sage Publications.
Gelles, R.J. (1975). Violence and pregnancy: A note on
the extent of the problem and needed services. Family
Coordinator. 24. 81-86.
Gelles, R.J. (1976). Abused wivesWhy do they stay?
Journal of Marriage and the Family. 38. 659-668.
Gelles, R.J. (1977). No place to go: The social dynamics
of marital violence. In M. Roy (Ed.), Battered women:
A psvchosociological study of domestic violence
(pp. 46-63). New York: Van Nostrand Reinhold.
Gelles, R.J. (1980). Violence in the family: A review of
research in the seventies. Journal of Marriage and the
Family. 42, 873-884.
Gelles, R.J. (1983). The exchange/social control theory of
violence. In D. Finkelhor, R.J. Gelles, G.T. Hotaling,
M.A. Straus (Eds.), The dark side of families: Current
family violence research (pp. 108-134). Beverly Hills:
Sage Publications.
Gelles, R.J., & Straus, M. (1979). Determinants of
violence in the family: Toward a theoretical integra
tion. In W.R. Burr, R. Hill, F.I. Nye & I.L. Reiss
(Eds.), Contemporary theories about the family (Vol. 1,
pp. 549-581) New York: The Free Press.
Gil, D.G. (1970). Violence against children: Physical
child abuse in the United States. Cambridge, MA:
Harvard University Press.
Giles-Sims, J. (1983). Wife battering: A systems theory
approach. New York: The Guilford Press.
Gill, T., & Coote, A. (1975). Battered women: How to use
the law. Rev. Ed. London: Cobden Trust.


135
Krauss, R.M., Geller, V. & Olson, C. (1976, September).
Modalities and cues in perceiving deception. Paper
presented at the meeting of the American Psychological
Association, Washington, DC.
Kulik, J., & Taylor, S.E. (1979). Self-monitoring and the
use of consensus information. Unpublished manuscript,
Harvard University, Cambridge, MA.
Labell, L.S. (1979). Wife abuse: A sociological study of
battered women and their mates. Victimology: An
International Journal. 4, 258-267.
Langley, R., & Levy, R.C. (1977). Wife beatingThe silent
crisis. New York: E.P. Dutton.
Lazio, J.P., & Rosenthal, R. (1970). Subject dogmatism,
experimenter status, and experimenter expectancy
effects. Personality: An International Journal. 1,
11-23.
Lee, D., & Ehrlich, H.J. (1971). Beliefs about self and
others: A test of the dogmatism theory. Psychological
Reports. 28. 919-922.
Lee, F. (1976). A study of sex differences in locus of
control, tennis, expectancy for success and tennis
achievement. Unpublished doctoral dissertation,
University of Oregon.
Levenson, H. (1972). Distinctions within the concepts of
internal-external control: Development of a new scale.
Proceedings of the 80th Annual Convention of the
American Psychological Association. 7, 261-262.
Levenson, H. (1973). Multidimensional locus of control in
psychiatric patients. Journal of Consulting and
Clinical Psychology. 41, 397-404.
Levenson, H. (1974). Activism and powerful others:
Distinctions within the concept of internal-external
control. Journal of Personality Assessment. 38. 377-
383.
Levenson, H. (1981). Differentiating among internality,
powerful others, and chance. In H.M. Lefcourt (Ed.),
Research with the locus of control construct (Vol. 1,
pp. 15-63). New York: Academic Press.
Levenson, H., & Miller, J. (1976). Multidimensional locus
of control in sociopolitical activists of conservative
and liberal ideologies. Journal of Personality and
Social Psychology. 33. 199-208.


10
proposed by A. Miller (1983, 1986). The previous two
variables already noted in the literature as determinants,
i.e. resources and number of previous separations (Gelles,
1976; Kalmuss & Straus, 1982; Pagelow, 1981), were inves
tigated. Finally, the locus of control determinant which
Walker (1979, 1984) proposed from the learned helplessness
theory was reexamined. Reinvestigation of this was con
sidered important in light of the strong reliance in this
field on her theory and her surprising findings disputing
locus of control as a factor (Walker, 1984).
Research Questions
The specific research questions were
1. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their authoritar
ianism?
2. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their monitoring?
3. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their resources?
4. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their number of
previous separations?


142
Straus, M.A. (1980). Victims and aggressors in marital
violence. American Behavioral Scientist. 1980, 23.
681-704.
Straus, M.A., Gelles, R.J., & Steinmetz, S.K. (1980).
Behind closed doors: Violence in the American family.
Garden City, NY: Doubleday.
Vacchiano, R.B. Dogmatism. (1977). In T. Blass (Ed.),
Personality variables in social behavior (pp. 281-314).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Vacchiano, R.B., Strauss, P.S., & Hochman, L. (1969). The
open and closed mind: A review of dogmatism. Psycho
logical Bulletin. 71. 261-273.
Vacchiano, R.B., Strauss, P.S., & Schiffman, D.C. (1968).
Personality correlates of dogmatism. Journal of
Consulting Psychology. 32., 83-85.
Vidulich, R.N., & Kairman, I.P. (1961). The effects of
information source status and dogmatism upon conformity
behavior. Journal of Abnormal and Social Psychology.
63, 639-642.
Walker, L.E. (1977-78). Battered women and learned
helplessness. Victimology: An International Journal.
2, 525-534.
Walker, L.E. (1978a, March). Feminist psychotherapy with
victims of violence. Paper presented at the Midwinter
Meeting of Division 29 of the American Psychological
Association, Scottsdale, AZ.
Walker. L.E. (1978b). Treatment alternatives for battered
women. In J.R. Chapman & M. Gates (Eds.), The
vVictimization of women (pp. 146-161). Beverly Hills:
Sage Publications.
Walker, L.E. (1979). The battered woman. New York:
Harper & Row.
Walker, L.E. (1980a). Battered women. In A. Brodsky & R.
Hare-Mustin (Eds.), Women and psychotherapy
(pp. 339-363). New York: The Guilford Press.
Walker, L.E. (1980b, October). Clinical aspects of the
battered woman syndrome study. Paper presented at the
Colorado Mental Health Conference, Keystone, CO.
Walker, L.E. (1981). Battered women: Sex roles and
clinical issues. Professional Psychology. 12. 81-91.


APPENDIX B
LETTER OF INTRODUCTION


7
relationship is that a patriarchal society perpetuates abuse
and coercive control of women (Brownmiller, 1975; Dobash &
Dobash, 1979; Schechter 1982). Rights of chastisement
accorded to men as legal property owners of women and
children resulted in the legal and social sanctioning of
male violence against women and children. Patriarchal
values, such as traditional gender identity development
which supports the goals of individuation and separation for
males and attachment for females, reinforce women staying in
relationships unsatisfying to them (Gilligan, 1982) Women
view success of the spousal relationship as their primary
responsibility, and the definition for success is dictated
by the male heads of the household in a traditional family
hierarchy (Dobash & Dobash, 1981). Consequently, battered
women are conditioned within the patriarchal system to
maintain relationships as long as their husbands and society
deem them important.
Although the characteristics of both batterers and
battered women very often include traditional values about
families and stereotypic ideas regarding gender roles, not
all abusers and abused women fit in this category (Walker,
1979). In fact, research to distinguish between women who
remained in abusive relationships via the variable of
traditional ideology has not supported that more traditional
women remain in abusive relationships (Pagelow, 1981).
Although the traditions of a patriarchal society contribute
to an environment "ripe" for wife abuse, the patriarchal


APPENDIX I
COMPARISON NORMS FOR LOCUS OF CONTROL,
AUTHORITARIANISM, AND MONITORING


58
(LD); a person having a predominantly closed belief system
is a high dogmatic (HD).
High dogmatics indicate greater dependence on absolute
authority, a greater vulnerability to be influenced by
authority, and less discernment of the value of communi
cations independent of the authority who delivers the
communication (Bettinghaus, Miller, & Steinfatt, 1970; C.G.
Kemp, 1963; Lazio & Rosenthal, 1970; Powell, 1962). High
dogmatics are more likely to adhere to a position supported
by authority (McCarthy & Johnson, 1962). Low dogmatics are
not differentially influenced by the authority of a communi
cator (Harvey & Hays, 1972) and they perceive authority
figures more realistically, incorporating both positive and
negative characteristics (C.G. Kemp, 1963). Low dogmatics
are much less influenced by authority than high dogmatics
(Ehrlich & Lee, 1969; Restle, Andrews, & Rokeach, 1964;
Vacchiano, Strauss, & Hochman, 1969).
The close-minded person will avoid changing his or her
environment or accepting new data and ideas (Vacchiano,
Strauss, & Schiffman, 1968). The high dogmatic will focus
on the future while minimizing the past or present (Castle,
1971; Jay, 1969; Rokeach, 1954, 1956; Rokeach & Bonier,
1960; Zurcher, Willis, Ikard, & Dohme, 1967). Rokeach
(1954, 1956) suggested that this is done in order to create
a sense of control regarding the course of life events.
There is a direct relationship between authoritarianism and
anxiety (cf. Byrne, Blaylock, & Goldberg, 1966; Castle,


26
wife violence. Where the wife's power was recognized as
high, the wife-to-husband violence was also high.
In studies where college students recalled parental
conflicts, Straus (1974) disproved the popular notion of
"cathartic violence," i.e., if pressure is built up and not
released through some means, it will eventually erupt as
violence. Instead, he found that increased verbal battling
escalated physical aggression, rather than dissipating such
stresses. College students' parents who utilized more
intellectualizing during marital conflicts experienced lower
levels of aggression towards spouses.
Straus, working with others (Allen & Straus, 1975?
Owens & Straus, 1975), found a moderate correlation between
observing, committing, or being victim of violent acts as a
child and adult approval of interpersonal or political
violence. He also found that the greater the husband's
resources, the less likelihood that the husband would
perpetrate violence in the home.
Theoretical Models
These four preceding forces focused on the problem of
domestic violence, building a body of knowledge which
brought multiple facets of wife abuse to light. Theoretical
models of family violence resulted. In some cases, theories
applied to human behavior and aggression in general have
been overlaid on the topic of family violence, e.g., the
theory of learned helplessness (Seligman, 1975) or the
frustration-aggression hypothesis (Dollard, Dobb, Miller,


110
8. Although I might have good ability, I will not be
given leadership responsibility without appealing
to those in positions of power.
9. How many friends I have depends on how nice a
person I am.
10. I have often found that what is going to happen
will happen.
11. My life is chiefly controlled by powerful others.
12. Whether or not I get into a car accident is mostly
a matter of luck.
13. People like myself have very little chance of
protecting our personal interests when they
conflict with those of strong pressure groups.
14. It's not always wise for me to plan too far ahead
because many things turn out to be a matter of
good or bad fortune.
15. Getting what I want requires pleasing those people
above me.
16. Whether or not I get to be a leader depends on
whether I'm lucky enough to be in the right place
at the right time.
17. If important people were to decide they didn't
like me, I probably wouldn't make many friends.
18. I can pretty much determine what will happen in my
life.
19. I am usually able to protect my personal inter
ests .
20. Whether or not I get into a car accident depends
mostly on the other driver.
21. When I get what I want, it's usually because I
worked hard for it.
22. In order to have my plans work, I make sure that
they fit in with the desires of people who have
power over me.
23. My life is determined by my own actions.
It's chiefly a matter of fate whether or not I
have a few friends or many friends.
24.


118
11.
I laugh more when I watch a comedy with
others than when alone.
TRUE/FALSE
12.
In a group of people I am rarely the
center of attention.
TRUE/FALSE
13.
In different situations and with
different people, I often act like very
different persons.
TRUE/FALSE
14.
I am not particularly good at making
other people like me.
TRUE/FALSE
15.
Even if I am not enjoying myself, I
often pretend to be having a good time.
TRUE/FALSE
16.
I'm not always the person I appear to be.
TRUE/FALSE
17.
I would not change my opinions (or the
way I do things) in order to please
someone else or win their favor.
TRUE/FALSE
18.
I have considered being an entertainer.
TRUE/FALSE
19.
In order to get along and be liked, I
tend to be what people expect me to be
rather than anything else.
TRUE/FALSE
20.
I have never been good at games like
charades or improvisational acting.
TRUE/FALSE
21.
I have trouble changing my behavior to
suit different people and different
situations.
TRUE/FALSE
22.
At a party I let others keep the jokes
and stories going.
TRUE/FALSE
23.
I feel a bit awkward in company and do
not show up quite so well as I should.
TRUE/FALSE
24 .
I can look anyone in the eye and tell a
lie with a straight face (if for a right
end.)
TRUE/FALSE
25.
I may deceive people by being friendly
when I really dislike them.
TRUE/FALSE


28
"stay in her place" regardless of what abusive treatment she
receives. All of these socialization imprints provide for a
high level of violence in the family (Straus, 1973) .
Another general systems model has been advanced by
Giles-Sims (1983). She builds on Straus (1973) and Buckley
(1967), as well as Broderick and Smith's (1979) concepts
regarding hierarchies of feedback and control in systems.
The six stage model includes (a) the establishment of the
family system, (b) the first incident of violence,
(c) stabilization of the family system, (d) the choice
point, (e) leaving the system, and (f) resolution or more of
the same. Both the man and woman creating the system come
to it with individual histories and personalities that
influence the way they interact with each other. Precipi
tating events or stressful situations occur in the marriage.
Those events, in combination with prior history relating to
violence, enter into the execution of violence. If the
assault serves to maintain the system or satisfy goals of
the batterer, the next move is by the woman. If her
response to the assault is to ignore, deny, or provide
forgiveness, she colludes in a feedback loop which encour
ages the continuance of such behavior. On the other hand,
if the abused woman gets angry and considers the violence as
a possible pattern, then she may utilize social supports and
examine her alternatives. At this point, the battered woman
may leave. If the assaultive behavior did not satisfy some
of the system achievement or maintenance goals, then the


115
39. Unfortunately, a good many people with whom I have
discussed important social and moral problems
don't really understand what's going on.
40. Most people just don't know what's good for them.


DEMOGRAPHIC INFORMATION QUESTIONNAIRE
1. CODE NUMBER: (assigned by staff)
2. Please list your age:
3A. Please list your race/ethnic background using the codes
below:
A Asian N Native American
B Black W White
H Hispanic 0 Other
3B. If you listed "O" for other in 3A, please specify:
4A. Number of children in your household:
4B. List gender and ages of children: (M/F) (Years Old)
Use other side if additional space is needed.
5A. Have your left your mate before: YES NO
5B. If yes, number of times:
6. City in which you resided before coming to shelter:
103


Levenson Locus of Control Scales
Directions; Below is a series of attitude statements. Each
represents a commonly held opinion and there are no right or
wrong answers. You will probably disagree with some items
and agree with others. We are interested in the extent to
which you agree or disagree with such matters of opinion.
Read each statement carefully. Then indicate the
extent to which you agree or disagree by placing a number in
the left margin next to each statement. The numbers and
their meaning are indicated below:
+1 : I agree slightly -1 : I disagree slightly
+2 : I agree somewhat -2 : I disagree somewhat
+3 : I agree strongly -3 : I disagree strongly
First impressions are usually best in such matters.
Read each statement, decide if you agree or disagree and the
strength of your opinion, and then place a number in the
space provided. Give vour opinion on every statement.
If you find that the numbers to be used in answering do
not adequately indicate your own opinion, use the one which
is closest to the way you feel. Your responses will be kept
confidential.
1. Whether or not I get to be a leader depends mostly
on my ability.
2. To a great extent my life is controlled by
accidental happenings.
3. I feel like what happens in my life is mostly
determined by powerful people.
4. Whether or not I get into a car accident depends
mostly on how good a driver I am.
5. When I make plans, I am almost certain to make
them work.
6. Often there is no chance of protecting my personal
interest from bad luck happenings.
7. When I get what I want, it's usually because I'm
lucky.
109


25
acts that have much more impact than either television
violence or school discipline in producing future genera
tions of assaultive people. In his research, family members
from both violent and nonviolent families approved of
certain violent acts as "normal" in that they served the
purpose of dissipating a husband's tension and/or appeasing
an otherwise hysterical mate. Gelles further supported the
family structure and the family's societal position as more
important factors than individual pathology in the occur
rence of family violence. Gelles estimated that intrafamily
violence occurs in 37% of any general population which
compares similarly to other estimates of this nature
(Levinger, 1966; O'Brien, 1971). Gelles (1975) was the
first researcher to note the higher likelihood of violence
against a wife when she is pregnant.
Straus established the Family Violence Research Program
at the University of New Hampshire in 1970. He found that
the cultural and social signals from a violent society
combine with familial, psychological factors to incite
violence in the home, contradicting the myth of the family
as a warm and loving system (Straus, 1974).
Straus (1973) looked at the balance of power in
families in relationship to the occurrence of violence.
Where equal power existed between the husband and wife, the
husband to wife violence was lowest. High power in either
wife or husband was related to high levels of husband-to-


76
arose regarding individual items or the study. The
administrator was trained to do this in a way which main
tained the integrity of the instrumentation and purpose of
the investigation, yet assisted the woman to answer the
questionnaire adequately. The administrator collected the
completed questionnaire and recorded on the master chart that
the following had been totally completed: the informed
consent form, the 6 demographic items, the 3 resources items,
the 24 locus of control items, the 40 dogmatism items, and
the 25 monitoring items.
Subjects' names were not placed on the instruments to
ensure confidentiality. Each packet was given a code number
which was paired with the respondent's name on a separate
listing.
The trained administrator, following collection of the
completed questionnaire, discussed and then recorded the
preferred and contingent options for follow-up contact. The
woman could select options from the following methods:
(a) telephoning the shelter during a prearranged time,
(b) returning to the shelter for a face-to-face contact
(often women return to the shelter to pick up mail, including
aid checks, which are addressed to the women at the shelter),
(c) communicating through a third party who agreed to provide
the necessary information about the woman to the shelter,
(d) returning a stamped post card addressed to the shelter's
post office box with a question to indicate destination on
it, (e) responding to a follow-up letter mailed to the woman


61
& Cantor, 1980) Low self-monitoring individuals focus on
information based on their inner states (M. Snyder & Tanke,
1976). A greater reliance on a partner's behavior as a
guide to one's own behavior is characteristic of high self
monitors (Ickes & Barnes, 1977). High self-monitoring
individuals connect their identities with their external
environment and choose to enter more clearly defined
situations rather than ambiguous ones (M. Snyder &
Gangestad, 1982; Sampson, 1978).


98
toring would be evidenced by a decreased external moni
toring. In fact, women who monitor others and have
developed the skill to do so may not do so less in order to
add the monitoring of selves more. Therefore, this instru
ment may have been inadequate to gauge the increase in
internal monitoring and whether it distinguishes between the
two groups of battered women.
Two demographic characteristics did distinguish between
abused women not returning to the batterers and abused women
who did. Why have nonwhite battered women not returned to
their abusers more often than their white counterparts?
Perhaps there is greater support from the nonwhite woman's
family of origin and community to leave an abusive relation
ship; that support could be manifested in housing or
economic resources, emotional caring, or less rigid sanc
tions regarding marital dissolution. This finding is
contrary to other studies (Pagelow, 1981; D.K. Snyder &
Scheer, 1981) and requires further investigation.
The other demographic factor of note was related to
population size where the battered woman resided before
coming to the shelter. Battered women from the small towns
may have been more likely to return to the abusers for
several reasons: (a) The mores of small towns are more
conventional and therefore reinforce families staying
together regardless of the conditions. (b) Less safety,
anonymity, and other living options are available in small
towns to protect a battered woman who leaves. (c) Less


91
had no children or, if they had children, had less than
four. Of the 41.7% of the sample who were nonwhites, 86.7%
did not return to their abusers, and this was equally
distributed across shelters and city sizes. Women in this
sample who were from small towns were more likely to return
to their mates (63% returned from towns of 20,000 population
or less) and women who were from larger cities were more
likely not to return to their mates (66% did not return from
cities of 51,000 to 200,000 population and 83.3% did not
return from cities over 200,000 population).
No significant differences (p < .05) were found between
abused women who did not return to their batterers and
abused women who did on any of the variables identified. No
significant relationship was found between all of the
variables and destination.
Limitations of the Study
Several limiting factors impact interpretation of
results from this study. First, to what populations are the
results generalizable? Using five shelters increased the
possibility that the study could be generalized to
sheltered, battered women. Regional or other
characteristics of this sample may reduce generalizability,
however. Additionally, this sample in most cases did not
include women who entered and departed the shelter very
quickly, i.e., within the first 48 hours. A very high
percentage of those women return to their abusive
households, so it is with caution that these results can be


66
72 of those 78 women; follow-up was not obtained on 6 women.
An additional 6 women who did not participate, either due to
choice or education requirements, and their destinations at
exit were noted. The reasons for their non-participation
were recorded. One woman, although she met the eighth-grade
completion requirement, did not possess adequate reading
skills to answer the questionnaire; she returned to her
abuser. Two women at two different shelters took the
questionnaires to complete, were interrupted in the process
(one for child rearing, the other to communicate with a
social service employee from another agency who was
assisting her) and did not complete the questionnaires or
make follow-up arrangements. Both left their shelters
hastily. One returned to the batterer and the other flew to
California to move in with her relatives. One woman was
incapable of completing the questionnaire due to disorienta
tion and psychological disturbance, and she left the shelter
to be psychiatrically hospitalized. One woman did not
participate because of a language barrier; she was Spanish
speaking and reading. The researcher did not have a Spanish
version prepared for her use. She did not return to the
abuser.
Instruments
In all, four instruments with a combined total of 92
questions were used to gather the information on authori
tarianism, monitoring, resources, and locus of control
(number of previous separations was answered by the subject


113
9.It is only natural for a person to be rather
fearful of the future.
10. There is so much to be done and so little time to
do it in.
11. Once I get wound up in a heated discussion I just
can't stop.
12. In a discussion I often find it necessary to
repeat myself several times to make sure I am
being understood.
13. In a heated discussion I generally become so
absorbed in what I am going to say that I forget
to listen to what the others are saying.
14. It is better to be a dead hero than to be a live
coward.
15. While I don't like to admit this even to myself,
my secret ambition is to become a great person,
like Einstein, or Beethoven, or Shakespeare.
16. The main thing in life is for a person to want to
do something important.
17. If given the chance I would do something of great
benefit to the world.
18. In the history of humankind there have probably
been just a handful of really great thinkers.
19. There are a number of persons I have come to hate
because of the things they stand for.
20. A person who does not believe in some great cause
has not really lived.
21. It is only when a person devotes oneself to an
ideal or cause that life becomes meaningful.
22. Of all the different philosophies which exist in
this world there is probably only one which is
correct.
23. A person who gets enthusiastic about too many
causes is likely to be a pretty "wishy-washy" sort
of person.
24. To compromise with our political opponents is
dangerous because it usually leads to the betrayal
of our own side.


29
batterer may choose an alternative behavior that could be
more useful and acceptable. If the woman leaves and will
come back only under the condition that he change, he may
choose some alternative behavior as well. Her other two
options include not coming back at all, or coming back with
no requisite that he change. A weakness of this theory is
that Giles-Sims did not incorporate some of the less
directly observable influences such as the patriarchal
influences or societal gender inequities.
Biological Theory
Elliott (1982) suggested that biological explanations
have been totally ignored in postulating theories about
intrafamilial violence. He detailed how organic and meta
bolic disorders contribute to domestic disturbances; he
specifically focused on episodic dyscontrol syndrome or
"explosive rage." His sample of 286 patients with histories
of violence was a skewed sample because of the biased
selection on the part of the referring physicians, the use
of computerized tomography (the CT scan), and the pain
staking methods utilized to uncover minimal brain dysfunc
tion. Elliott recommended further research with an
unselected group of batterers to uncover neurological
causation factors in domestic violence, and by no means
suggested biology or genetics as an all encompassing factor.
Although neurological dysfunction may be pertinent in a
small number of cases, it is highly unlikely that it is a
major explanatory feature of wife abuse in our society. The


11
5. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their locus of
control (internal, powerful others, chance)?
6. Which combination of the above mentioned variables
provides the best determination of battered women's destina
tion following a shelter stay?
Significance of the Study
This study has implications for those researching
domestic violence, shelter personnel, community agency
personnel, and battered women. A workable theory which
accurately describes the social problem of wife abuse and
which can be used to shape a better family unit and society
has yet to be defined. Whether refinement of current
theories or evolution of a new theory occurs, research is
required on the batterer, the family unit, society, and the
battered woman. Despite the heritage of this problem as a
socially taboo subject, and the paucity of funding provided
to it, the findings from this research effort will add to
what is known about battered women and be a part of the base
for future research efforts on conjugal violence against
women.
If there are identifiable social and psychological
factors that determine the destination of battered women
following a shelter experience, then shelter program
planners could design and administer programs most suited to
the two different populations. Rather than assuming that


34
nonresponsive and arduous for battered women and those
helping them.
The patriarchal tradition today combines with the needs
and realities of a capitalist economy to continue institu
tional and socialization patterns that perpetuate male
dominance and aggression (Schechter, 1982) Battering is
one of the many ways that men maintain control in our
society. Women stay in abusive relationships for two
reasons: (a) Their inferior status under capitalism (lower
wages or wagelessness and unequal division of labor
requiring women to be responsible for maintenance of both
home and family) leaves them financially dependent on men,
and (b) their sex role socialization as wives and mothers
results in a moral obligation to stay and please their
husbands and rear their children properly (Schechter, 1982).
One problem with patriarchal society explanations is
that they are historical and hypothetical in nature, with
minimal empirical work with battered women operationalizing
the specifics of the theory. An example of a researcher who
has captured empirically some of the aspects of patriarchal
society is Pagelow (1981). She has addressed particularly
the question, "why do women stay." The hypothesis is that a
woman of few resources, receiving negative institutional
response (family, law enforcement, social service agencies,
society), and very traditional in her beliefs, is unlikely
to change or believe she can change her battering situation.
Pagelow (1981) supports the proposition that limited


35
resources (as measured by the woman's age, her children's
ages, differences in earnings, husband's earnings, and home
ownership) correlate strongly with the woman's staying in
the relationship. She did not find a strong correlation
between institutional response and traditional ideology with
the women's length of cohabitation while in a battering
relationship.
Social Learning Theory
The most relied upon psychological theory regarding
battered women is that proposed by Walker (cf. 1977-78,
1979, 1980b, 1984). Because of its well established
reference in the battered women's movement and its role as a
significant precedent for this investigation, Walker's
theory will be considered in detail here.
Walker (1977-78, 1979, 1984) utilizes Seligman's (1975)
theory of learned helplessness as a basis to understand the
battering relationship and battered women's responses.
Learned helplessness is combined with a cycle theory of
violence to depict the phenomenon of battering (Walker,
1979). The supporting research (Walker, 1978a, 1978b, 1979,
1980a, 1981) is based on interviews with several hundreds of
battered women and some caregivers to battered women (e.g.,
therapists, shelter workers).
Learned helplessness, according to Seligman (1975),
occurs when animals experience negative reinforcement
noncontingently, i.e., with no apparent link to their own
actions. Eventually they believe that their actions have no


22
involving male violence perpetrated against female victims.
In investigating further, Pagelow (1977b) listed the
parallels: Instances of sexual assault and wife abuse were
both underreported to police; they rarely went to court and
had abysmally low rates of conviction when they did. Both
rapists and wife abusers were likely to be repeat offenders;
few offenses evidenced higher rates of recidivism than these
two. The victims of these crimes were pictured stereo-
typically as masochistic, provoking the crime, reporting it
for hidden reasons, and not dedicated to seeing the perpe
trator prosecuted. Thus, in a "blame the victim" cycle,
women in wife battering cases, as well as rape cases, had to
prove that they were "worthy victims" even though the need
to establish worth was not present in other crimes. Pagelow
reported that these crimes were not predominantly lower
class occurrences as often thought, identifying high
incidence in middle and upper classes also.
Police Involvement
With the addition of computers which stored and
categorized vast amounts of crime scene data, law enforce
ment agencies gained greater awareness of police fatalities
and injuries at "domestic disturbance" response calls
(National Institute of Law Enforcement and Criminal Justice,
1978). Twenty-eight percent of the assaults perpetrated
against police officers have been while the officers were
responding to domestic disturbance complaints (Stephens,
1977) A substantial proportion of all homicides on police


100
1. Tracking battered women's location at later
checkpoints would enhance the accuracy of determining who
maintained their independence. Also, follow-ups in which
information is gathered regarding the women's current
relationship will help to determine if she is still in an
abusive relationship.
2. Specifically develop instrumentation to operation
alize A. Miller's theory more closely. This instrumentation
could then be utilized in any number of areas of research.
What was obvious to this investigator is that there was
little available option for measuring one's ability and
frequency of self monitoring and self empathy.
3. Conduct further investigations related to city
size and race to further understand their impact or connec
tion to destination.
4. A standard measure or measures to assess resources
is needed. The economic factor is judged subjectively to be
important in battering. But researchers using various,
nonstandard methods, without adequate analysis of relia
bility and validity, fail to prove or disprove the impor
tance of this factor.
5. Although neither learned helplessness nor A.
Miller's psychodynamic theory provided assistance in
distinguishing battered women related to destination,
continued study of both with battered women and the phenome
non of battering is recommended. This was the initial
attempt to apply A. Miller's theory to this area, and to


Self-Monitoring Scale
Directions: The statements in the following pages concern
your personal reactions to a number of different situations.
No two statements are exactly alike, so consider each
statement carefully before answering. If a statement is
TRUE or MOSTLY TRUE as applied to you, circle TRUE following
the statement. If a statement is FALSE or NOT USUALLY TRUE
as applied to you, circle FALSE following the statement.
It is important that you answer as frankly and as
honestly as you can. Your answers will be kept in the
strictest confidence.
1. I find it hard to imitate the behavior
of others. TRUE/FALSE
2. My behavior is usually an expression of
my true inner feelings, attitudes, and TRUE/FALSE
beliefs.
3. At parties and social gatherings, I do
not attempt to do or say things that
others will like.
4. I can only argue for ideas which I
already believe.
5. I can make impromptu speeches even on
topics about which I have almost no
information.
6. I guess I put on a show to impress or
entertain people.
7. When I am uncertain how to act in a
social situation, I look to the
behavior of others for cues.
8. I would probably make a good actor.
9. I rarely need the advice of my friends
to choose movies, books, music.
10.I sometimes appear to others to be
experiencing deeper emotions than I
actually am.
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
117


CHAPTER IV
RESULTS
The results are presented in this chapter. Descriptive
statistics were used to characterize the sample group while
the research hypotheses were analyzed with one-way analysis
of variance (hypothesis one through five) and logistic
regression (hypothesis six).
Demographic Information
Demographic information is presented here in order to
characterize the overall sample of battered women, as well
as subsets of the sample: those who did not return to the
batterers and those who did return following the stays at
the shelters. The total sample included 72 women. Those
who did not return to the batterers included 42 women; those
who did return to the abusers numbered 30. In addition to
the variable data, general demographic information was
gathered on each subject: age of subject, race or ethnic
background of subject, number of children, whether subject
had left mate before or not, and city of residence before
coming to shelter.
A breakdown of the study participants by age appears in
Table 4-1. The sample was predominantly in their 20s and
30s (86%), with the mean age being 30.32 years (SD = 7.32);
this is quite similar to shelter samples from other
80


TABLE OF CONTENTS
PAGE
ACKNOWLEDGEMENTS iii
LIST OF TABLES viii
ABSTRACT ix
CHAPTERS
I INTRODUCTION 1
Statement of the Problem 4
Purpose of the Study 4
Rationale for the Study 5
Research Questions 10
Significance of the Study 11
Definition of Terms 13
Battered Woman 13
Battering 14
Shelter 14
Destination 14
Returning to the Abuser 14
Not Returning to the Abuser 15
Locus of Control 15
Resources 16
Number of Previous Separations 16
Authoritarianism 16
Monitoring 17
Organization of Study 17
II REVIEW OF THE LITERATURE 18
Emergence of Conjugal Violence 18
The British Refuge Movement 19
The Women's Rights Movement in the U.S. 21
Police Involvement 22
Study of Societal Violence 24
v


Dear Shelter Resident,
Thank you for agreeing to participate in my study
on the factors related to battered women's destination
following a shelter stay. Enclosed you will find an
informed consent form for your signature and the question
naire for you to complete. Please complete the question
naire and return it to the shelter staff. Your prompt
response will be appreciated. Please ask the designated
staff member if you have any questions.
Thank you again for your interest and participation.
Sincerely,
Joanne F. DeMark
120


LIST OF TABLES
Table PAGE
4-1 Age of Sample 81
4-2 Race of Sample 82
4-3 Number of Children 83
4-4 Previous Separations 83
4-5 Population of City in Which Subject Resided
Before Shelter 84
4-6 Means and Standard Deviations of Study Measures . 86
4-7 Logistics Regression Model of the Relationship
Between Destination and the Research Variables. . 89
H-l Comparison Demographics from Other Battered
Women Studies 12 3
1-1 Locus of Control Comparison Norms 126
1-2 Authoritarianism and Monitoring Comparison Norms 127
viii


3
Bakan, 1971; Gil, 1970; Gillen, 1946; Maurer, 1976; Palmer,
1962; Steele & Pollack, 1974; Welsh, 1976).
Also, domestic violence has significant costs for
society. As examples, public and private sector funds are
used to pay for safe shelters, counseling, police interven
tion, legal avenues, and other resources. Police face
greater injury and death at the scenes of domestic violence
than at any other crime scene at which they intervene (Bard
& Zacker, 1974).
Ratios of wife battering compared to husband battering
range from 11:1 through 13:1 in the United States (Gaquin,
1977-78; Levinger, 1966; Steinmetz, 1977-78). Straus,
Gelles, and Steinmetz (1980) supported social policy,
attention, and treatment specifically on wife abuse for
the following reasons: (a) Wife battering involves more
dangerous and injurious forms of violence, and it results
in greater physical damage and injury; (b) male abusers are
much more often repeaters; (c) wives often act in self-
defense; (d) much abuse by the husband occurs when the wife
is pregnant, posing additional danger to the fetus; and (e)
the mores, laws, and traditions of our society lock women
into marriages in a more substantial way than men.
Wife abuse is a chronic crime which escalates in
severity and poses a serious threat to the safety of the
women involved (Pagelow, 1977a, 1977b, 1981; Walker, 1979).
Battered women commonly report receiving murder threats from
their abusers, as well as perceiving the batterers'


Dogmatism Scale (Form E)
Directions; The following is a study of what the general
public thinks and feels about a number of important social
and personal questions. The best answer to each statement
below is your personal opinion. We have tried to cover many
different and opposing points of view; you may find yourself
agreeing strongly with some of the statements, disagreeing
just as strongly with others, and perhaps uncertain about
others; whether you agree or disagree with any statement,
you can be sure that many people feel the same as you do.
Mark each statement in the left margin according to how
much you agree or disagree with it. Please mark every one.
Write +1, +2, +3, or -1, -2, -3, depending on how you feel
in each case.
+1 : I agree a little -1 : I disagree a little
+2 : I agree on the whole -2 : I disagree on the whole
+3 : I agree very much -3 : I disagree very much
1. The United States and Russia have just about
nothing in common.
2. The highest form of government is a democracy and
the highest form of democracy is a government run
by those who are most intelligent.
3. Even though freedom of speech for all groups is a
worthwhile goal, it is unfortunately necessary to
restrict the freedom of certain political groups.
4. It is only natural that a person should have a
much better acquaintance with ideas one believes
in than with ideas one opposes.
5. A person on one's own is a helpless and miserable
creature.
6. Fundamentally, the world we live in is a pretty
lonesome place.
7. Most people just don't give a "damn" for others.
8. I'd like it if I could find someone who would tell
me how to solve my personal problems.
112


17
Monitoring
Monitoring is defined as the observation of others and
mediating of one's presentation as guided by social cues (M.
Snyder, 1974), as measured by the Snyder Self-Monitoring
Scale (M. Snyder, 1974).
Organization of Study
This study will be presented in five chapters. The
current chapter is a brief introduction of the subject, the
purpose and rationale for the study, and a description of
relevant terms. Chapter II is a review of the related
literature. The research methodology is detailed in the
third chapter. Data analyses and results are presented in
the fourth chapter. Finally, in Chapter V the researcher
offers a discussion and interpretation of the results, a
discussion of the limitations of the study, and further
implications.


55
children (e.g. age of children, number of children) as
singular variables were predictive of the woman's destina
tion upon leaving a shelter. In the study reported herein
the researcher incorporated the role of children as a factor
in the battered women's destination choice, but subsumed
into the resource category as has been done by others
(Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981).
Locus of Control. Learned Helplessness
Walker (1979) theorized, based on her interview and
clinical experience with battered women, that women find it
difficult to escape an abusive relationship due to learned
helplessness. In later research she reported on her work to
test this hypothesis (Walker, 1984).
The learned helplessness phenomenon was identified and
most often studied in a laboratory setting with animals and
later with humans (cf. Seligman, 1975; Abramson, Seligman, &
Teasdale, 1978). Since an operational instrument had not
been developed to measure learned helplessness outside the
laboratory, Walker (1984) devised several methods to
indicate whether the learned helplessness phenomenon was
instrumental. A series of questions were combined into
scales, e.g., sexual abuse during childhood, childhood
health, and combined to form a single measure of childhood
contributors to learned helplessness. Also a measure of
learned helplessness within the battering relationship was
formulated by intercorrelating 15 indices from the interview
questionnaire utilized. Walker (1984) noted the reliability


65
women's health care clinics, and university and medical
communities.
To be included in the study, the women must have been
battered and must have completed 8 years of schooling
according to their response on the shelter intake forms.
The 8 years of schooling was deemed appropriate indication
of a baseline ability to read and understand the instru
ments.
Sampling Procedures
The subjects for this study were a volunteer sample of
battered women who utilized these shelters during the months
of May through December 1987 and met the criteria for the
study. Data were not collected every day at each of these
shelters, but on the days that the data were collected all
women in the shelter, except for those who had already
participated, were asked to participate in the study. The
sample size was 72 women.
During the days data were collected, all potential
subjects admitted to the shelters were given a letter
explaining the study, describing the criteria for inclusion,
and inviting their participation (see Appendix B). The
potential subjects were questioned to determine if they met
the criteria for inclusion and, if so, were they willing to
participate in the study. This procedure was continued
until the pool of individuals who met the inclusion criteria
and who agreed to participate totaled 78 women. Completed
questionnaires and follow-up destinations were obtained on


Theoretical Models 26
General Systems Theory 27
Biological Theory 29
Exchange/Social Control Theory 30
Feminist Theories 31
Social Learning Theory 35
Intrapsychic Theories 38
Variables Relating to the Battered Women's
Destination 43
Frequency and Severity of the Abuse ... 44
Abuse in Family of Origin 44
Length of Marriage, Previous Separations,
Religious Affiliation 45
Love, Affection, Hope 48
Self Esteem 49
Resources/Occupational History 51
Children 53
Locus of Control, Learned Helplessness 55
Authoritarianism and Monitoring 57
Authoritarianism 57
Monitoring 59
III METHODOLOGY 62
Population and Sample 62
Sampling Procedures 65
Instruments 66
Measure of Resources 68
Measure of Locus of Control 68
Measure of Authoritarianism 71
Measure of Monitoring 72
Data Collection Procedures 73
Data Analysis 78
IV RESULTS 80
Demographic Information 80
Analysis of Hypotheses One Through Five ... 85
Analysis of Hypothesis Six 88
Summary 88
V DISCUSSION 90
Summary 90
Limitations of the Study 91
Discussion of Results 93
Recommendations for Future Research 99
vi


134
Hornung, C.A., McCullough, B.C., & Sagimoto, T. (1981).
Status relationship in marriage: Risk factors in
spouse abuse. Journal of Marriage and the Family. 43.,
675-692.
Hunt, M.F., & Miller, G.R. (1968). Open- and close
mindedness, belief-discrepant communication behavior,
and tolerance for cognitive inconsistency. Journal of
Personality and Social Psychology. 8, 35-37.
Ickes, W., & Barnes, R.D. (1977). The role of sex and
self-monitoring in unstructured dyadic interactions.
Journal of Personality and Social Psychology. 35, 315-
330.
Jay, R.L. (1969). Q technique factor analysis of the
Rokeach dogmatism scale. Educational and Psychological
Measurement. 29. 453-459.
Jones, E.E., & Baumeister, R. (1976). The self-monitor
looks at the ingratiator. Journal of Personality. 44.
654-674.
Kalmuss, D.S., & Straus, M.A. (1982). Wife's marital
dependency and wife abuse. Journal of Marriage & the
Family. 44, 277-286.
Kemp, C.G. (1963). Perception of authority in relation to
open and closed belief systems. Science Education. 47.,
482-484.
Kemp, M., Knightly, B., & Norton, M. (1975). Battered
women and the law. London: Cobden Trust.
Kerlinger, F.N., & Pedhazer, E.J. (1973). Multiple
regression in behavioral research. New York: Holt,
Rinehart, & Winston.
Kerlinger, F.N., & Rokeach, M. (1966). The factorial
nature of the F and D scales. Journal of Personality
and Social Psychology. 4, 391-399.
Kernberg, O.F. (1975). Borderline conditions and patho
logical narcissism. New York: Jason Aronson.
Kleck, R.E., & Wheaton, J. (1967). Dogmatism and responses
to opinion-consistent and opinion-inconsistent infor
mation. Journal of Personality and Social Psychology.
5, 249-252.
Klein, J., & Seligman, M.E. (1976). Reversal of
performance deficits and perceptual deficits in learned
helplessness and depression. Journal of Abnormal
Psychology. 85, 11-26.


75
in providing information to battered women which increases
self-understanding and decreases feelings of isolation,
guilt, and blame. Risks and/or drawbacks of participation
were the time required to complete the questionnaire, the
adherence to the shelter's follow-up procedure following
departure, responding to questions which solicit personal
opinions when the subject was in a post-traumatic time
period, and reliance on the shelter and principal investi
gator to assure confidentiality. The administrator recorded
on the master chart the subject number, date of administra
tion, and whether the woman qualified and agreed to parti
cipate in the study. If the potential subject was not
participating, the administrator recorded whether this was
due to not meeting the requisites or a refusal to partici
pate.
Each woman volunteering to participate in the study was
given a packet which included instructions for the instru
ments, the instruments themselves, one informed consent form,
and a manila envelope for return of the packet (see Appendix
G). At this time the subjects completed the demographic
questionnaire and the paper and pencil instruments required
for this investigation. The demographic questionnaire
contained 6 items and took a few moments to complete. The
combined instruments contained 92 items total and took from
45 to 90 minutes to complete.
The administrator was present with the subject through
out data collection and trained to answer any questions which


101
abandon its potential relevance at this point would be
premature.
6. More investigation regarding battered women's
perceptions of authority is warranted. Battered women may
fluctuate between idealizing authority sometimes and, at
others, serving the needs of those in authority while
secretly feeling derisive. This would be consistent with A.
Miller's earliest work (1981) where she addressed the
borderline personality in detail.
7. It is highly possible that certain typologies of
battered women exist as subsets of the destination cate
gories. Of the many studies done related to domestic vio
lence, in only one (D.K. Snyder & Fruchtman, 1981) were
battered women considered heterogenous, not homogenous.
Looking at characteristics of heterogeneity could greatly
increase the potential of actually predicting or determining
destination or when a woman will take abuse no more.


CHAPTER I
INTRODUCTION
In the past decade, domestic violence, particularly
spouse abuse, has been brought increasingly to the public's
attention. What was once considered an unspoken family
issue, condoned and encouraged by society, is now deemed a
social problem (Dobash & Dobash, 1971; Roy, 1977; Stacey &
Shupe, 1983; Straus, 1977-78). Considered as one repre
sentation of violence against women, spouse abuse was
initially addressed through the impetus of the feminist
movement. In an effort to better understand, treat, and
ameliorate the problem of conjugal assault, a host of
disciplines and groups, including helping professionals and
feminists, have undertaken the charge to research the
problem (Fleming, 1979; Hansen & Barnhill, 1982; Hilberman,
1980; NiCarthy, Merriam, & Coffman, 1984; Walker, 1979,
1984). This involvement has led to a focus on providing
safety and treatment for spouse abuse victims: the battered
mates and the children who reside in settings where violence
occurs.
An indication of the extent of the problem is the
number of individuals affected by domestic violence. From
one national survey of violence in American homes,
researchers reported that in one household out of six, a
1


23
are related to domestic disturbance calls (Rochester Police
Bureau, 1974) .
Addressing the American Bar Association, Detroit Police
Chief Bannon (1975) commented that in addition to fearing
injury, police avoid domestic violence for two other
reasons: (a) They lack the skills needed to mediate inter
personal conflicts, and (b) they view domestic conflicts as
private problems, not a public matter. He commented:
Of all the nonathletic occupations, none is as absorbed
with the use of physical coercive force as that of the
police officer. Nor are any more thoroughly socialized
in their masculine role images. This . suggests to
me that traditionally trained and socialized policemen
are the worst possible choices to attempt to intervene
in domestic violence. (p. 8)
The lack of police skills to intervene in conjugal
violence was addressed in 1967 by Bard and the New York City
Police Department. They jointly developed and implemented
the first family crisis intervention program within a police
unit (Bard & Zacker, 1971). Other literature detailed other
early efforts at police training in crisis intervention,
referral, and conflict management (Olsen, 1972; Rochester
Police Bureau, 1974; Spitzner & McGee, 1975).
Examination of police work brought additional domestic
violence factors to light. The incidence and degree of
interpersonal aggression of domestic disturbance calls
cannot be isolated to lower socioeconomic, minority neigh
borhoods as a comparison between Harlem, New York, and
Norwalk, Connecticut, indicated (Bard & Zacker, 1971).
Evidence contradicted the common beliefs that family


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ECI41PHWF_WYV64S INGEST_TIME 2014-10-06T21:42:23Z PACKAGE AA00025735_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


71
proneness on the 16 PF. The P Scale correlates signifi
cantly with suspiciousness on the 16 PF.
Measure of Authoritarianism
Authoritarianism was measured by the Rokeach Dogmatism
Scale, Form E (Rokeach, 1956). This is a 40-item scale (see
Appendix E). Responses can range from "I agree very much"
(+3) through "I disagree very much" (-3) on a Likert-type
Scale. The close-minded or high dogmatic (HD) person
believes that authority is absolute, accepts or rejects
others based on their agreement or disagreement with
authority, and maintains this cognitive system which
protects the individual for anxiety. The low dogmatic (LD)
person does not hold authority as absolute, accepts or
rejects other on different premises than their alignment
or disagreement with authority, and has more tolerance for
ambiguity and the anxiety that can accompany changing
environments (Vacchiano, 1977).
The construct validity of authoritarianism has been
supported in numerous studies. A common core of authori
tarian factors has been noted between the California F Scale
and the Dogmatism Scale (Kerlinger & Rokeach, 1966),
although the Dogmatism Scale has been found to be more
independent of ideological bias than the California F (Warr,
Lee, & Joreskog, 1969). Many studies support the Dogmatism
Scale as a measure of authoritarianism (Barker, 1963;
Costin, 1965; Hanson, 1968, 1970; Plant, 1960; Rokeach,
1967). High dogmatic persons show a greater dependence on


60
monitoring individuals are less able or likely to modify
expressiveness and presentation to others (M. Snyder, 1979).
The expression of a low self-monitoring individual appears
to be more internally guided by his or her own feelings,
attitudes, and experiences (M. Snyder, 1974, 1979) rather
than by situational or interpersonal cues (M. Snyder &
Monson, 1975). Individuals high in self-monitoring notice
and accurately remember information about others more than
low self-monitoring individuals (Berscheid, Graziano,
Monson, & Dermer, 1976). Persons low in self-monitoring
spend less time and effort thinking about the contingencies
of a prospective date's behavior than those high in self
monitoring (Berscheid, Graziano, Monson, & Dermer, 1976).
High self-monitoring persons keenly attend to the subtleties
of behavior and context and utilize the information to infer
intentions and predict behaviors of others (Jones &
Baumeister, 1976; Kulik & Taylor, 1979). High self-monitors
are skilled at accurately judging the intended meaning of
verbal expressions (Mill, 1984). Further, those who are
high in self-monitoring are especially skilled at reading
nonverbal expression and correctly determining underlying
affective experience and emotional states of others (Geizer,
Rarick, & Soldow, 1977; Krauss, Geller, & Olson, 1976).
High self-monitoring individuals have access to very rich
dispositional constructs organized around prototypes of
others; low self-monitors excel at giving rich, informative
descriptions of their own traits and dispositions (M. Snyder


33
There are several theories that represent the feminist
analysis of wife abuse. Dobash and Dobash (1977-78) traced
the history of violence against women by placing it in the
wider context of society, thus supporting that abuse against
wives has been acceptable behavior in a patriarchal society.
Dobash and Dobash (1977-78, 1979) cited marriage laws
throughout history that categorize wives as the possessions
of husbands and dictate the husband's obligations to
chastise and reprimand wives. Many of the laws clearly
reflect double standards. For example, English Common Law
denied wives civil rights and legal status. England,
Europe, and early America had laws supporting the husband's
rights to beat his wife. It was not until the 1700s that
this tradition came into question. In the early 1800s
written laws were adopted in various states that reiterated
the sanctity and privacy of the home unless "permanent
injury or excessive violence" were present (Eisenberg &
Micklow, 1977, p. 149). Finally, not until 1891 does
caselaw reject the husband's legal right to beating or
abusing his wife (Dobash & Dobash, 1977-78). Dobash and
Dobash (1979) believed that it is the tradition of the
patriarchal society that has also impacted on the leaders of
the women's movement in their attempts to address wife abuse
through both traditional institutions and the grass roots
shelter movement. Consequently, nonsupportive laws change
too slowly, funding is inadequate, and institutions remain


6
are demographic, previous separations and Roman Catholic
denomination have proved to be significant. D.K. Snyder and
Scheer (1981) suggested that perhaps religiosity or authori
tarianism would be more revealing than denomination.
Neither previous separations nor denomination is a foolproof
method to determine likelihood of remaining in or leaving
the relationship.
Others (Gelles, 1972, 1976; Martin, 1976; Pagelow,
1981) have proposed that women who are able to leave
battering relationships are those who are economically
independent from the abuser. Several researchers (Gelles,
1976; Kalmuss & Straus, 1982; Pagelow, 1981) have empiri
cally validated this proposition; financial resources
available to a woman have served as predictive indicators of
her remaining in or leaving a battering relationship.
Although the availability of resources is a key factor, it
does not always explain who extricates herself from an
abusive marriage versus who remains in such a marriage.
Some battered women in shelter programs or therapy have
faced tremendous economic and social hardship, at times with
minimal career skills, yet they left their abusers and
established themselves and their children in new locations
on their own (Fleming, 1979; Martin, 1976; Stacey & Shupe,
1983) .
Several theories have been proposed to explain whether
a woman will remain in or leave an abusive relationship.
One theory for the battered woman's entrapment in an abusive


52
e.g., employment status, having children under 5, income
level. In the Gelles study (1976), the resource variable
that best distinguished wives who obtain assistance from
those who remain with the husband is holding a job. Many
authors (Dobash & Dobash, 1977-78; Fleming, 1979; Schechter,
1982; Stacey & Shupe, 1983; Straus, Gelles, & Steinmetz,
1980; Walker, 1979, 1984; Wetzel & Ross, 1983), from their
discussions and research with battered women, support the
notion that financial dependence does contribute to keeping
the battered woman trapped in the relationship. Of Stacey
and Shupe's (1983) shelter residents who responded to why
they did not leave the situation immediately after the first
abuse incident, 30% suggested their economic dependence.
Pagelow (1981) retested Gelles' (1976) earlier hypothesis
regarding resources, along with the other variables Gelles
looked at (severity of abuse, frequency, and family of
origin child abuse history), and again found that fewer
resources related directly to whether the woman stayed in
the relationship.
Occupational underachievement in the husband or higher
status in occupation by the wife have been found to increase
the risk for spouse abuse to occur (Hornung, McCullough, &
Sagimoto, 1981). Conflicting results regarding occupational
history suggests that employment status does not have a
correlation with whether the abused woman will leave the
relationship (Rounsaville, 1978). A battered wife's
objective marital dependency (no job, minimal income


Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE
By
JOANNE F. DeMARK
April 1988
Chair: Dr. Jaquelyn Liss Resnick
Major Department: Counselor Education
The purpose of this study was to examine seven poten
tial determinants of battered women's destination following
shelter experience. The variables were drawn from previous
research and the psychodynamic theory of Alice Miller. It
was hypothesized that battered women who did not return to
their abusers would be characterized as less authoritarian,
less monitoring of others, have less chance and powerful
others locus of control, greater internal locus of control,
greater resources, and have left their mates more often
previous to this shelter visit than battered women who
returned to their abusers.
Questionnaires were completed by 72 women during their
stay at one of five southern U.S. shelters, three in Georgia
and two in Florida. Of the sample, 30 returned to their
abusers and 42 did not.
ix


12
all programs are appropriate for all shelter residents,
better attempts at identifying the appropriate recipients
for certain interventions could be made. The woman who is
setting up independent living arrangements for herself and
her children needs housing assistance, career guidance,
oftentimes school transfers for children, continuing welfare
aid and food stamps, and social networks. The woman
returning to the abuser might benefit from self-defense
training, conflict resolution skills, and strategies to
encourage marital and family counseling. Depending on her
long-range desires and the degree to which her return is
based on economic dependence, the abused woman returning to
the batterer might benefit from strategies for acquiring
career skills or employment opportunities.
Community agency personnel who come in contact with
battered women include those from food and financial
assistance agencies, churches, legal aid, state prosecution
and public defender offices, vocational rehabilitation
programs, health care agencies, school programs, and
sometimes community mental health or crisis counseling
agencies. With more insight into who is likely to return to
the abuser and who is likely to separate longer-term from
the batterer, personnel from each of these agencies can more
appropriately intervene. Oftentimes, staff working with
battered women can get invested in the hope that a client
will decide her destination in a particular direction. A
more definitive profile of who will stay away and who will


27
Mowrer, & Sears, 1939). In other cases theories have
evolved explicitly from the study of domestic violence as in
Walker's (1979, 1984) cycle theory of battering. The
following major theoretical areas are examined: (a) General
systems, (b) biological, (c) exchange/social control, (d)
feminist, (e) social learning, and (f) intrapsychic.
General Systems Theory
Straus (1973) developed a theory to explain family
violence known as the "general systems theory." The family
unit is a social system with a purpose, seeking to meet
goals and adapting to the environmental context in which it
exists (Straus, 1973). A violent family system is described
as follows: A probability for conflict occurs in the family
due to involuntary membership, intensity of relationships,
generation and gender differences, and parental rights of
influence. The probability of violence can be positively
reinforced due to a highly violent society and socialization
of family members in violence vis a vis parental models,
physical punishment, tolerance of sibling assaults, and
macho values, especially for boys. As a result the culture
legitimizes a norm of violence and family members integrate
assaultiveness into their personality characteristics
(Straus, 1973). The sexism of society, through its limited
roles and career opportunities for women, unequal pay,
assumptions regarding wives and mothers, assumptions
regarding males as head of households, and socialization of
women into subordinate roles, reinforces that a woman should


126
Table 1-1
Locus of Control Comparison Norms
This Studv
Walker 1984)
Studv
Levenson
Norms
Returned
to Abuser
Did Not
Return
Total
Sample
Living
With
Abuser
Independent
Total
Sample
Internal
N
30
42
72
94
292
386
48
M
35.63
36.02
35.28
41.87
41.33
41.46
35.46
SD
5.64
6.81
6.31
7.13
6.79
6.88
7.41
Powerful
Others
N
30
42
72
92
291
383
48
M
16.07
17.69
17.43
17.52
18.17
18.01
14.64
SD
9.40
10.53
10.01
10.78
9.39
9.39
9.73
Chance
N
30
42
72
94
289
383
48
M
20.60
16.43
18.17
17.25
17.41
17.37
13.38
SD
9.36
9.36
9.52
9.995
9.32
9.48
9.05


APPENDIX G
LETTER OF INSTRUCTION/INFORMED CONSENT LETTER


88
scores was not significant for at the .05 level (F = 3.48,
p = 0.066). Since ANOVAs were not significant for all locus
of control scores, the null hypothesis was not rejected.
Analysis of Hypothesis Six
The sixth hypothesis was tested using logistic regres
sion. The sixth hypothesis was that there would be no
relationship between authoritarianism, monitoring,
resources, number of previous separations, locus of control
(internal, powerful others, and chance), and the destination
of battered women following shelter stays. The logistic
regression analysis is presented in Table 4-7. The regres
sion analysis indicated that the variables, in general, were
not good predictors of destination following shelter stay.
Therefore, the null hypothesis was not rejected.
Summary
No significant differences between the groups occurred
on the hypothesized variables of authoritarianism, moni
toring, resources, number of previous separations, internal
locus of control, powerful others locus of control, and
chance locus of control. Some distinctive differences were
indicated in this sample between battered women who returned
to their abusers and battered women who did not, specifi
cally in the demographic data of race, population of city
where subject resided, and husband's percentage of income.
No significant relationship was established between these
variables and the destination of battered women following a
shelter experience.


20
that the patriarchal attitudes which fostered husbands
beating wives also frustrated women's groups in society at
large as they sought help for the battering problem. A
contradiction existed in society between protecting the
privacy of home life and preventing the battering of
spouses.
Becoming aware of WA's growth, the psychiatric and
legal communities were next to get involved. In Memorandum
on Battered Wives (1974), the Royal College of Psychiatrists
scholars used case studies to depict the complexity of
factors surrounding spouse abuse. They described wife
battering as an adaptation failure of inadequate acquisition
of appropriate social learning skills by the abusers.
Citing the fact that battered wives use help when available,
the Royal College scholars recommended more services,
research, and education. They proposed that child abuse
often occurs in homes where wife battering exists.
Scott (1974) described persons involved in spouse abuse
as likely to come from many clinical classifications, rather
than just one or two. Common diagnoses he cited were
immature personality, dependence, aggression, jealous
reactions, and drug or alcohol addictions. Scott asserted
that although sadomasochism is often assumed as the reason
for battered wives returning, more often dependency,
ignorance of choice, and fear of loneliness exist as the
actual reasons. By 1975 British pamphlets outlining legal
options and explaining the difficulties and inadequacies of


45
battering each other in her family of origin, the more
willing she was to seek outside intervention when she was
abused by her spouse. Pagelow (1977c, 1980, 1981) was not
able to substantiate the same finding in her research. It
is possible that a similar percentage would occur for non-
battered women's families of origin. Study with families of
college students indicates violence was present in 28% of
the cases.
Length of Marriage, Previous Separations. Religious
Affiliation
Snyder and Scheer (1981) examined a number of variables
for their ability to suggest whether a battered woman would
return to her abuser following a stay at a shelter. They
selected variables from data available at admission to the
shelter so the results would be helpful for shelter workers
in selecting their intervention strategies with the shel
tered women. The variables were not selected to support or
test any particular theory, but to be helpful in theoretical
formulation.
In their study the sampe was comprised of 74 women
admitted to a shelter in Detroit, Michigan. Fifty-five
percent of the women (41 women) were living with the abuser
when follow-up was conducted 6 weeks after they left the
shelter; 45% (33 women) were not. When comparisons were
conducted between the two groups, six variables from
admission were significantly different between the two
groups: two of the reasons for seeking admission (seeks
short-term separation only and seeks conjoint marital


73
M. Snyder, 1974; M. Snyder & Monson, 1975; M. Snyder &
Swann, 1976; M. Snyder & Tanke, 1976).
Data Collection Procedures
Data were collected by either a trained, full-time
staff person from a shelter, the researcher, or a research
associate. At the Ft. Myers, Gainesville, and Cobb shelters
data were collected by a trained full-time staff person. At
the Clayton shelter and during the latter months of collec
tion at the Cobb shelter, the data were collected by the
researcher. At the downtown metro Atlanta shelter the data
were collected by either the researcher or the trained
research associate. The investigator met individually with
each of the selected staff persons and the research associ
ate to train them in the procedures. Training involved the
following components: (a) use of the letter of introduction
to the subject inviting participation, (b) determination that
potential subject meets criteria for inclusion, (c) obtaining
informed consent, (d) benefits/risks of participation,
(e) methods to assure confidentiality, (f) instruments and
instructions, (g) answering questions during administration,
(h) procedures for follow-up contact, and (i) recording data
on the master chart. Each of the individual sessions was
2 to 3 hours long. To determine that each administrator
understood and could uniformly perform data collection,
the investigator conducted a follow-up session after the
participation of at least the first two subjects. The
investigator monitored the data collection throughout the


36
effect on what may happen to them. This theory's appli
cability to humans has also been supported by Seligman and
others (Hiroto, 1974; Klein & Seligman, 1976; W.P. Miller &
Seligman, 1975; W.P. Miller, Seligman, & Kurlander, 1976;
Seligman, 1975; Seligman & Hiroto, 1975). With battered
women, as Walker's theory proposed (1977-78, 1979, 1984),
battering from an abuser occurs at random, often in no
directly rational relationship to the woman's immediately
preceding behavior, and according to no set schedule or
timeframe.
As repeated battering occurs, she comes to believe that
nothing she can do will have any impact on the batterer's
abusive behavior. All of the actions she has previously
tried, e.g., being apologetic and loving, fighting back,
leaving, calling the police, have not resulted in an end to
the unpredictable abuse. Walker proposed that a woman will
more quickly "learn" her powerlessness if she experienced
more rigid sex-role stereotyping as a child which already
reinforces passivity on her part and the importance of
relying on men for help and protection (Dweck, Goetz, &
Straus, 1980; Radloff & Rae, 1979, 1981; Walker, 1977-78,
1980a, 1984).
From Walker's interviews, she determined that the
battering, although neither constant nor occurring in a
consistent manner each time, does occur in a cycle of three
distinct phases (Walker, 1977-78, 1979, 1984). The first
phase is the tension-building stage. During this stage,


68
administering instrumentation to women in crisis is a
sensitive issue, the investigator chose the following four
measures.
Measure of Resources
For this investigation, the measure for resources was
the Resources Index (Appendix C) utilized by Kalmuss and
Straus (1982). It is the sum of scores on three dichotomous
variables: whether the woman is unemployed or not, whether
the woman has children at home of ages 5 or younger or not,
and whether her mate earns 75% or more of the total house
hold income or not. The range for the values on this index
is 0 (low resources) through 3 (high resources). The
internal consistency of the Resources Index in the Kalmuss
and Straus study (1982) was .59, measured by Cronbach's
alpha coefficient of reliability.
Measure of Locus of Control
Locus of control was measured by the Levenson Locus of
Control Scales (Levenson, 1973) (Appendix D). The locus of
control construct describes individuals' causal beliefs
regarding reinforcements which occur to them. They may have
a generalized expectancy that reinforcements are contingent
upon their own behaviors (internal control). They may also
generally believe that reinforcements are contingent upon
forces outside their control, e.g., chance, fate, a deity,
luck, powerful others (external control). The Levenson
Locus of Control Scales are multidimensional, measuring
beliefs regarding personal control (Internal Scale),


APPENDIX D
LEVENSON LOCUS OF CONTROL SCALES


139
Rokeach, M. (1956). Political and religious dogmatism: An
alternative to the authoritarian personality.
Psychological Monographs. 70 (18, whole No. 425).
Rokeach, M. (Ed.). (1960). The open and closed mind. New
York: Basic Books.
Rokeach, M. (1967). Authoritarianism scales and response
bias: Comment on Peabody's paper. Psychological
Bulletin. 67, 349-355.
Rokeach, M., & Bonier, R. (1960). Time perspective,
dogmatism, and anxiety. In M. Rokeach (Ed.), The open
and closed mind (pp. 366-375). New York: Basic Books.
Rotter, J.B. (1966). Generalized expectancies for internal
versus external control of reinforcement. Psycho
logical Monographs. 80, 1.
Rounsaville, B.J. (1978). Battered wives: Barriers to
identification and treatment. American Journal of
Orthopsychiatry. 48, 487-494.
Rounsaville, B.J., Lifton, N., & Bieber, M. (1979). The
natural history of a psychotherapy group for battered
women. Psychiatry. 42., 63-78.
Roy, M. (1977). Battered women: A psvchosociological
study of domestic violence. New York: Van Nostrand
Reinhold.
Royal College of Psychiatrists. (1974). Memorandum on
battered wives. London, England: Chandos House.
Rutschky, K. (1977). Black pedagogy. Berlin, Germany:
Hogarth.
Sampson, E.E. (1978). Personality and the location of
identity. Journal of Personality. 46, 552-568.
Scanlan, T.J. (1979). Self employment as a career option:
An investigation of entrepreneurship from the perspec
tives of Holland's theory of career development and
Levenson's measure of locus of control. Dissertation
Abstracts International. 40, 4561A. (University
Microfilms No. 80-04269)
Schechter, S. (1982). Women and male violence: The
visions and struggles of the battered women's movement.
Boston: South End Press.
Scott, P.D. (1974). Battered wives. British Journal of
Psychiatry. 125. 433.


81
Table 4-1
Age of Sample
Age
Returned
to Abuser
n (%)
Did not
Return
n (%)
Total
Samle
n (%)
17-20
1
(3.3)
3
(7.1)
4
(5.6)
21-25
6
(20.0)
9
(21.4)
15
(20.8)
26-30
12
(40.0)
12
(28.6)
24
(33.3)
31-35
2
(6.7)
7
(16.7)
9
(12.5)
36-40
6
(20.0)
8
(19.0)
14
(19.4)
41-45
2
(6.7)
3
(7.1)
5
(6.9)
46-50
0
0
0
51-55
1
(3.3)
0
1
(1.4)
research (see Appendix H). Mean age for battered women who
returned to their abusers was 30.80 years (SD = 7.67); mean
age for those who did not was 29.98 years (SD = 7.14).
A breakdown of the study participants by race appears
in Table 4-2. The sample had a large percentage of non
whites (41.7%) which is common for samples including inner
city shelters (e.g., D.K. Snyder & Fruchtman, 1981; D.K.
Snyder & Scheer, 1981) (see Appendix H). Because there
appeared to be an effect due to race, a chi-square analysis
was conducted. With a 2 = 16.86, p = 0.000, which is
significant for a = .05 and d.f. = 1, an effect due to race


DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE
By
JOANNE F. DeMARK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988


64
who sought help from the shelters had exhausted financial
and familial support.
The collection of demographic information allowed for a
comparison of this sample to some demographic information
from other known shelter samples. The demographic charac
teristics of women from these shelters did not differ
significantly from those of shelter samples in other studies
in the southeast or the nation (Bell, 1977; Carlson, 1977;
Labell, 1979; Pagelow, 1977c, 1981; Stacey & Shupe, 1983).
Although battering occurs across socioeconomic classes,
shelter samples are usually more representative of lower and
lower middle class socioeconomic classes; women with greater
resources have other alternatives for safety (Pagelow, 1981;
Stacey & Shupe, 1983). In the study reported herein, basic
demographics including age, race or ethnic background,
number of children, and residential address before entering
shelter (to determine urban, suburban, or rural) were
collected (Appendix A). One of the independent variables,
number of previous separations, was also included on this
portion of the questionnaire.
Many of the women who came to the shelters heard about
the services via an intervening police officer, public
service advertising, a friend or family, or the advertised
battered women's hotlines. Referrals to the shelters could
come from the emergency rooms of the hospitals in the area.
In addition, other referrals were made by the community
crisis hotlines, police, clergy, rape crisis programs,


50
concept. She found battered women to be lower in ego
strength and self-esteem, with greater identity problems,
than non-battered women. Overall the battered women were
dissatisfied with themselves physically, morally, socially,
and in relationship to their families. Her research
indicated that battered women had a very difficult time
maintaining even minimal self-esteem. In a comparison of 20
battered women from shelters to 20 non-battered women from
the same socioeconomic status, Chan (1979) found the
battered women to have lower self-esteem.
Conflicting results come from two studies. Brown and
Brazzle (1982), reporting on the 40 abused women in shelters
whom they studied, indicate that 42% of the women had high
self-esteem. These authors believed it is inappropriate to
assume that battered women have low self-esteem. Mitchell
(1980) compared 16 nonabused women who were in therapy for
psychological problems to 24 abused women who were in
treatment for the abuse. The two groups did not differ in
self-esteem, sex role stereotype, locus of control, or
anxiety. What she did find, however, when studying within
group differences among the battered women was that those
who were less educated, unemployed, and severely and/or
frequently abused were lower in self-esteem, more anxious,
and more externally oriented. These studies not only
suggest that battered women may not be lower in self-esteem
than their nonbattered counterparts, but also that within
group variances may distinguish among battered women.


87
ANOVA was not significant at the .05 level (F = 0.93,
p = 0.338); therefore, the null hypothesis was not
rejected.
The third hypothesis was that there would be no
difference in the resources scores between battered women
who, following shelter stays, did not return to their
abusive mates and those battered women who did return to
their batterers. The one-way ANOVA was not significant at
the .05 level (F = 0.97, p = 0.328); therefore, the null
hypothesis was not rejected.
The fourth hypothesis was that there would be no
difference in the number of previous separations between
battered women who did not return to their abusers after the
shelter stays and those women who did return to the abusive
mates. The one-way ANOVA was not significant at the .05
level (F = 1.23, p = 0.271); therefore, the null hypothesis
was not rejected.
The fifth hypothesis was that there would be no differ
ence in internal, powerful others, and chance locus of
control scores between sheltered battered women who did not
return to their batterers and those who did return. Each of
the Levenson Locus of Control Scales is independent of the
other. The one-way ANOVA for internal locus of control
scores was not significant at the .05 level (F = 0.16, p =
0.689). The one-way ANOVA for powerful others locus of con
trol scores was not significant at the .05 level (F = 0.07,
p = 0.796). The one-way ANOVA for chance locus of control


138
Pagelow, M.D. (1980, November). Double victimization of
battered women. Paper presented at the meeting of the
American Society of Criminology, San Francisco.
Pagelow, M.D. (1981). Woman-battering: Victims and their
experiences. Beverly Hills: Sage Publications.
Palmer, S. (1962). The psychology of murder. New York:
Thomas Y. Crowell Company.
Parker, B., & Schumacher, D.N. (1977). The battered wife
syndrome and violence in the nuclear family of origin.
American Journal of Public Health. 67. 760-761.
Pizzey, E. (1974). Scream guietlv or the neighbours will
hear. Middlesex, England: Penguin Books.
Plant, W.T. (1960). Rokeach's dogmatism scale as a measure
of general authoritarianism. Psychological Reports. 6,
164.
Pleck, E., Pleck, J.H., Grossman, M., & Bart, P.B. (1977-
78). The battered data syndrome: A comment on
Steinmetz' article. Victimology: An International
Journal. 2, 680-683.
Powell, F.A. (1962). Open- and closed-mindedness and the
ability to differentiate source and message. Journal
of Abnormal and Social Psychology. 65, 61-64.
Pyron, B.A. (1966). A factor analytical study of simpli
city-complexity of social ordering. Perceptual and
Motor Skills. 2, 259-272.
Radloff, L.S., & Rae, D.S. (1981). Components of the sex
difference in depression. Research in Community Mental
Health. 2, 111-137.
Radloff, L.S., & Rae, D.S. (1979). Susceptibility and
precipitating factors in depression: Sex differences
and similarities. Journal of Abnormal Psychology. 88.
174-181.
Restle, F., Andrews, M., & Rokeach, M. (1964). Difference
between open- and closed-minded subjects on learning-
set and oddity problems. Journal of Abnormal and
Social Psychology. 68, 648-654.
Rochester Police Bureau. (1974). FACITfamily conflict
intervention team experiment: Experimental action
program. Rochester, NY: University of Rochester.
Rokeach, M. (1954). The nature and meaning of dogmatism.
Psychological Review. 61. 194-204.


REFERENCES
Abramson, L.Y., Seligman, M.E., & Teasdale, J.D. (1978).
Learned helplessness in humans: Critique and reformu
lations. Journal of Abnormal Psychology. 87. 49-74.
Allen, M., & Straus, M.A. (1975). Resources, power, and
husband-wife violence. National Council on Family
Relations. 12, 1-29.
Bakan, D. (1971). Slaughter of the innocents: A study of
the battered child phenomenon. Boston: Beacon Press.
Bannon, J. (1975, August). Law enforcement problems with
intra-familv violence. Paper presented to the American
Bar Association, Detroit.
Bard, M., & Zacker, J. (1971). The prevention of family
violence: Dilemmas of community intervention. Journal
of Marriage and the Family. 33. 677-682.
Bard, M., & Zacker, J. (1974). Assaultiveness and alcohol
use in family disputes: Police perceptions.
Criminology. 12, 281-292.
Barker, E.N. (1963). Authoritarianism of the political
right, center, and left. Journal of Social Issues.
19(2) 63-74.
Bell, J.N. (1977, June). Rescuing the battered wife.
Human Behavior, pp. 16-23.
Bernhardson, C.S. (1967). Dogmatism, defense mechanisms,
and social responding. Psychological Reports. 20, 511-
513.
Berscheid, E., Graziano, W., Monson, T., & Dermer, M.
(1976). Outcome dependency: Attention, attribution,
and attraction. Journal of Personality and Social
Psychology. 34, 978-989.
Bettinghaus, E., Miller, G., & Steinfatt, T. (1970).
Source evaluation, syllogistic content, and judgments
of logical validity by high and low-dogmatic persons.
Journal of Personality and Social Psychology. 16, 238-
244.
128


124
Snyder
& Scheer
Studv (1981)
Stacey &
Shupe
Studv (1983)
Star
et al.
Studv (1979)
Walker
Studv (1984)
74
538
57
403
100%
100%
80%

Detroit
Texas
Florida
Colorado &
Surrounding
Region
30.0

32.0
32.2

17-54
18-59


2.02
27 (36%)
343 (64%)
40 (70%)
321 (86%)
47 (64%)
195 (36%)
17 (30%)
51 (14%)
70.3%


*
41 (55%)



33 (45%)



*25% of sample were currently living with abuser.


APPENDIX F
SELF-MONITORING SCALE


67
on the demographics questionnaireAppendix A). This study
was designed to use short to moderate length instruments
administered by a trained administrator so as not to
additionally burden these subjects who were already experi
encing a crisis. Battered women usually enter the shelter
immediately after what Walker (1979, 1984) has referred to
as the acute battering incident. Many battered women have
described how difficult it is to leave the home for an
unknown refuge environment following such an emotionally
charged incident (Martin, 1976; Pagelow, 1981; Walker, 1979,
1984). At that time battered women are fearful of the
abusers, as well as afraid of the unknown, embarrassed to
make the domestic problem public, and unsure of social
support, emotional or otherwise. Knowing that the abuser
might perceive the woman's actions of seeking safety and
assistance for her and the family as escalation in the
current conflict, the woman must muster a great deal of
courage to seek shelter help. Battered women battle their
own self-blame for the domestic difficulties (Walker, 1979)
and often doubt that others will believe they have been
abused (Rounsaville, Lifton & Bieber, 1979). Lengthy
instrumentation asking the women to respond to many self-
statements at this crisis juncture is thought by many
(Dobash & Dobash, 1981, 1981; Schechter, 1982) to be
intrusive and to debilitate crisis efforts aimed at
supporting her decision to seek safety and get help for the
family. With the additional consideration that


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Ja/jufelyn^Liss Resnick, Chair
Professor of Counselor
Education
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
/
j)A[A o
Sandra B. Damico
Professor of Foundations
of Education
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
y/^T-ry 9y
tf&rry A. Grater, Jr.
Professor of Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Phyllis M. Meek
Associate Professor of
Counselor Education


31
sanctioning one type of family violence but not another,
e.g., laws against wife abuse but not child abuse, the
violence will be more frequent in the less sanctioned
category.
The exchange/social control theory for family violence
has application to treatment and policy issues (Gelles,
1983). Treatment interventions and policies that increase
both social controls exercised over the family and the costs
for violent behavior are primary. The resultant policies
should diminish societal norms that glorify violence,
increase financial and gender equality, and create more
responsive criminal justice and social service agencies.
The value in such a theory is that it can be applied across
the board to explain all types of familial abuse.
Feminist Theories
According to feminists, explanations regarding violence
against women are inadequate if concerned with general
explanations regarding overall societal violence or family
violence (Dobash & Dobash, 1979; Schechter, 1982). By
treating all violence between parties as equal, some
significant uniquenesses about domestic violence against
women are masked (Dobash & Dobash, 1981b). Feminist authors
have particularly challenged Straus' (1980) conclusion that
there is nearly sexual equality in perpetuation of violence
in couples and Steinmetz (1977-78) who suggested an equally
prevalent "battered husband syndrome" (p. 501). Pleck,
Pleck, Grossman, and Bart (1977-78) have pointed to inac-


9
centuries, have shown neither validation nor respect for the
feelings, needs, and desires of children. This "spare the
rod, spoil the child" tradition supports the authority of
the parent, whether right or wrong. A. Miller (1981, 1983,
1986) labeled the training of children to meet consistently
the parents' needs and respond without question to the
parents' and society's authority while the children's own
needs go unacknowledged as "poisonous pedagogy." She
suggested that where this tradition is particularly extreme,
children grow up with a greater vulnerability to violence,
either as the perpetrators or the victims. In adulthood,
they are most likely to deny or repress traumatic childhood
experiences, maintain needs to idealize someone else, and
often skillfully monitor and attend to the needs of others
to gain love and acceptance (A. Miller, 1981, 1983, 1986).
This theoretical perspective applied to battered women
suggests that some battered women are more vulnerable to
return to battering relationships than others because of the
following: (a) They monitored and responded more to their
parents' needs than their own, and (b) they strongly
idealized parents and authority in general.
Various factors previously identified and studied had
not satisfactorily accounted for determining whether
battered women's destination following shelter experience
would back with their abusers or not. Additional deter
minants were considered by selecting the variables of
authoritarianism and monitoring from the violence theory


53
potential, child-rearing role) is significantly related to
severe violence (Kalmuss & Straus, 1982) while her subjec
tive perception of her dependency is not related.
Clearly most research suggests that economic resources
are related to whether the abused woman stays. Thus,
economic resources was one of the variables studied here in
relationship to the battered woman's destination once she
leaves the shelter. The method utilized by Kalmuss and
Straus (1982) was replicated here.
Children
Many researchers have commented on the impact the
children and the parental relationship with the children,
have on the decision making for the battering couple (Brown
& Brazzle, 1982; Carlson, 1976; Gelles, 1983; Martin, 1976;
Snell, Rosenwald & Robey, 1964; Stacey & Shupe, 1983;
Straus, 1973).
Snell, Rosenwald, and Robey (1964) questioned wives who
charged their husbands with assault and battery regarding
their reasons for the action at that particular time, since
the marriages were 12-20 years in length and had included
other abuse incidents. The women in almost all the cases
responded first that the children figured into their
decision. Most common were responses describing the
presence of an adolescent son, getting older, stronger, and
more prone to be either retaliatory towards the father or
negatively affected by the domestic disturbances.


CHAPTER V
DISCUSSION
Summary
The aim in this study was to determine key variables
which would distinguish between battered women who did not
return to abusive mates and battered women who did following
their stays at refuge shelters. The variables, drawn from
previous research (Gelles, 1976; Kalmuss & Straus, 1982;
Pagelow, 1981; Walker, 1979, 1984) and from the psycho
dynamic theory of A. Miller (1981, 1983, 1986), were
resources, number of previous separations, internal locus of
control, powerful others locus of control, chance locus of
control, authoritarianism, and monitoring.
A total of 72 battered women who stayed at shelters in
Florida and Georgia agreed to participate in the study and
completed questionnaires while they were at the shelters.
Destination information, i.e., whether they were living with
the abuser or not, was recorded during the follow-up period
of 2 to 6 weeks after their departures from the shelters.
The demographic information from the questionnaires
indicated that battered women who did not return to their
abusers and battered women who did were similar to each
other in several ways. Most of the women were in their 20s
and 30s, had probably left their mates before, and either
90


96
battered women who subsequently returned to batterers, and
so the scores on the Snyder Self-Monitoring Scale would be
lower for battered women who did not return. The results
were not as expected; battered women from the two destina
tions did not score significantly differently on either
measure.
Several plausible explanations are available to illum
inate the lack of significant findings. In this study
battered women who did return to their mates were slightly
less dogmatic than battered women who didn't. However, the
mean scores for both groups would not suggest greater or
lesser dogmatism when compared to a normative sample of
female college students (see Appendix I). It is possible
that battered women, in general, are no more or less rigid
and devoted to authority than other groups in society. This
is compatible with A. Miller's (1983, 1986) beliefs that
most people in today's world are socialized into an extreme
acceptance of parents and authority. So it might be rare to
expect significantly low dogmatism scores from any sizable
percentage of any group in society. Although this resear
cher disagrees with A. Miller's (1986) suggestion that one
would need two in-depth psychoanalytic analyses to "crack"
this idealization of authority, it may be ambitious to
expect that battered women at shelters have come to under
stand the great imperfections of authoritarian rule.
Society has colluded historically in maintaining the near
infallibility of authority in one form or another.


44
women leaving abusive relationships. Researching a variety
of variables is important to build a solid theoretical
understanding of the problem. What follows is a
summarization of that research and, where applicable, a
discussion of how it applies to the investigation.
Frequency and Severity of the Abuse
Gelles (1976) believed that a battered woman would
mobilize on her own behalf the more frequent or severe the
abuse. His research supported this hypothesis. Other
research did not confirm the same finding (Pagelow, 1977c,
1980, 1981). Nor were frequency or severity significant in
defining whether a battered woman would return to the abuser
after a shelter stay (Snyder & Scheer, 1981).
Abuse in Family of Origin
If a battered woman was exposed to more violence in her
family of origin, would she be less likely to stay away from
her husband, the abuser? The families of origin of battered
women have often been violent families. Twenty-three
percent of the family histories of the wives studied by
Gayford (1975) were violent. In other studies (Hilberman &
Munson, 1977-78; Scott, 1974) that percentage was 50% or
more. Parker and Schumacher (1977) found that if the mother
in the wife's family of origin was physically abused, it was
significantly probable that the wife would be battered by
her husband. Half the women in the Hilberman and Munson
study (1977-78) were abused themselves as children. Gelles
(1976) found that the less a woman was exposed to parents


141
Snyder, M., & Monson, T. (1975). Persons, situations, and
the control of social behavior. Journal of Personality
and Social Psychology. 32, 637-644.
Snyder, M., & Swann, W.B., Jr. (1976). When actions
reflect attitudes: The politics of impression manage
ment. Journal of Personality and Social Psychology.
34., 1034-1042.
Snyder, M., & Tanke, E.D. (1976). Behavior and attitude:
Some people are more consistent than others. Journal
of Personality. 44., 501-517.
Spitzner, J.H., & McGee, D.H. (1975). Family-crisis
intervention training, diversion, and the prevention of
crimes of violence. Police Chief. 42., 252-253.
Stacey, W.A., & Shupe, A. (1983). The family secret:
Domestic violence in America. Boston: Beacon Press.
Star, B. (1978). Comparing battered and non-battered
women. Victimology: An International Journal. 2,
32-44.
Star, B., Clark, C.G., Goetz, K.M., & O'Malia, L. (1979).
Psychosocial aspects of wifebattering. Social
Casework. 479-487.
Steele, B.F., & Pollack, C.B. (1974). A psychiatric study
of parents who abuse infants and small children. In
R.E. Heifer & C.H. Kempe (Eds.), The battered child
(pp. 89-133). Chicago: University of Chicago Press.
Steinmetz, K. (1977-78). The battered husband syndrome.
Victimology. 2, 499-509.
Stephens, D.W. (1977). Domestic assault: The police
response. In M. Roy (Ed.), Battered women: A psvcho-
sociological study of domestic violence (pp. 164-172).
New York: Van Nostrand Reinhold.
Straus, M.A. (1973). General systems theory approach to a
theory of violence between family members. Social
Science Information. 12., 105-125.
Straus, M.A. (1974). Cultural and social organization
influences on violence between family members. In R.
Prince & D. Barrier (Eds.), Configurations: Biological
and cultural factors in sexuality and family life
(pp. 53-69). Lexington, MA: D.C. Health.
Straus, M.A. (1977-78). Wife beating: How common and why?
Victimology: An International Journal. 2, 443-458.


APPENDIX E
DOGMATISM SCALE (FORM E)


APPENDICES
A DEMOGRAPHIC INFORMATION QUESTIONNAIRE 103
B LETTER OF INTRODUCTION 105
C MEASURE OF RESOURCES 107
D LEVENSON LOCUS OF CONTROL SCALES 109
E DOGMATISM SCALE (FORM E) 112
F SELF-MONITORING SCALE 117
G LETTER OF INSTRUCTION/INFORMED CONSENT LETTER . 120
H COMPARISON DEMOGRAPHICS 123
I COMPARISON NORMS FOR LOCUS OF CONTROL,
AUTHORITARIANISM, AND MONITORING 126
REFERENCES 128
BIOGRAPHICAL SKETCH 144
vii


114
25. When it comes to differences of opinion in
religion we must be careful not to compromise with
those who believe differently from the way we do.
26. In times like these a person must be pretty
selfish if one considers primarily one's own
happiness.
27. The worst crime a person could commit is to attack
publicly the people who believe in the same thing
one does.
28. In times like these it is often necessary to be
more on guard against ideas put out by people or
groups in one's own camp than by those in the
opposing camp.
29. A group which tolerates too much difference of
opinion among its own members cannot exist for too
long.
30. There are two kinds of people in the world: those
who are for truth and those who are against truth.
31. My blood boils whenever a person stubbornly
refuses to admit one is wrong.
32. A person who thinks primarily of one's own
happiness is beneath contempt.
33. Most of the ideas which get printed nowadays
aren't worth the paper they are printed on.
34. In this complicated world of ours the only way we
can know what is going on is to rely on leaders or
experts who can be trusted.
35. It is often desirable to reserve judgment about
what's going on until one has had a chance to hear
the opinions of those one respects.
36. In the long run the best way to live is to pick
friends and associates whose tastes and beliefs
are the same as one's own.
37. The present is all too often full of unhappiness.
It is only the future that counts.
38. If a person is to accomplish one's mission in
life, it is sometimes necessary to gamble "all or
nothing at all."


4
capabilities to kill them (Walker, 1979). The violence in
wife abuse is often excessive and relentless, with beatings
continuing after the victims are either unconscious or dead
(Okun, 1986; Walker, 1979; Wolfgang, 1958). The severity of
the threat to abused women's lives is substantiated by the
statistics on women who are murdered by their partners
(Dobash & Dobash, 1977-78, 1979). Many more women seek safe
shelter than abusers seek treatment for their problem
(Fleming, 1979). Batterers find it extremely difficult to
acknowledge their behavior as a problem or to take respon
sibility for the outcome of their brutality (Walker, 1979).
The chances are quite slim that battered women who return to
their households will experience an improved conjugal
relationship with less threat of violence (Pagelow, 1981).
Concern for the safety of victims is of utmost importance.
Statement of the Problem
Greater theoretical understanding of wife abuse is
needed. One aspect of this, whether the battered woman will
return to the abuser or not following a shelter stay, is
important, especially with the strong likelihood that
returning to the batterer means continuation of the woman's
abuse. Theoretical explanations have yet to uncover fully
the factors related to battered women's destinations after
shelter stays.
Purpose of the Study
The purpose of this study was to reexamine key factors
from previous empirical and theoretical research, in


16
fate, or luck (Levenson & Miller, 1976). The Levenson Locus
of Control scales (Levenson & Miller, 1976) was be utilized
in this investigation.
Resources
Resources refers to an indication of the battered
woman's objective dependence or independence from her
abuser, especially as it relates to the woman's employment
status, the presence of children age 5 or younger at home,
and 75% or more of the household income being earned by the
mate. In this study it was measured by the Resources Index
(Kalmuss & Straus, 1982).
Number of Previous Separations
Number of previous separations is defined as any time
the woman has left the mate previously. In this study that
was indicated in response to two questions: (a) Have you
left your mate before, and (b) if yes, number of times.
Authoritarianism
Authoritarianism, also known as dogmatism, is defined
by Rokeach (1960) as an organization of belief-disbelief
systems constructed to both satisfy a need for a cognitive
framework for knowing and understanding the world and to
distance threatening aspects of reality. In this study it
was measured by the Dogmatism Scale (Form D), a generalized
measure of authoritarianism, free of political bias
(Rokeach, 1960).


47
ships to leave, despite the abuse. In research which
focused on couples who thought about divorce compared to
those who did not (Booth & White, 1980), results were that
religious affiliation does not have a strong effect on
thinking about divorce. Of the couples who had come from
lengthier marriages with stronger religious belief there
were fewer women who thought about divorce; those couples
with shorter marriages or less religious zeal had more wives
who thought about divorce. They found that the more intense
one's religious belief, the less likely one would think
about divorce. Booth and White (1980), nevertheless, did
find that in religious couples, even where marital satisfac
tion was high, persons experiencing abuse were more likely
to think about divorce.
Stronger religious belief or activity is often tied to
a more traditional way of looking at life (Pagelow, 1981).
Walker (1979, 1984) found in her sample of both battered
women and batterers, as described by the women, a strong
traditional view about the home, family, and sex roles.
Pagelow (1981) felt that the more intense the traditional
ideology of women who have been battered, the more likely
they are to remain in battering relationships and the less
likely they will be to take action to significantly better
the situation. To test this hypothesis, she looked at
whether the women came from a religious family of origin
(i.e., family religious activity), as well as five other
variables she felt reflected traditional ideology. Her


93
currently living with abuser. Although selecting another
point in time for administration of the questionnaire may be
as helpful to pinpoint the relevance of the specific
theories being applied here to the whole problem of woman
abuse, it would not have suited the purpose of establishing
determinants of destination following shelter stay.
Discussion of Results
What might be suggested by the lack of a significant
relationship between the identified variables and battered
women's destinations? Several ideas are presented.
Variables which were drawn from previous studies,
although not couched in a theoretical framework, were
resources and number of previous separations. In this study
even though the overall resources measure did not
distinguish between the two groups, one item from the
measure, whether the mate earned 75% or more of the family
income, did have a significant relationship with
destination. Battered women in this study whose abusers
earned 75% or more of the family income were likely to
return to cohabitation with the batterers. A plethora of
questions across studies indicate researchers' interest in
assessing the economic resource picture which the battered
woman faces and how this impacts her ability to free herself
from a battering relationship. The findings from this
study, and those from others, suggest that understanding
economic impact in a consistent, pertinent way has yet to be
determined.


72
authority than do low dogmatic persons (cf. Bettinghaus,
Miller, & Steinfatt, 1970; Kemp, 1963; Powell, 1962; Restle,
Andrews, & Rokeach, 1964; Vidulich & Kaiman, 1961).
Measure of Monitoring
Monitoring in this study was measured by the Snyder
Self-Monitoring Scale (M. Snyder, 1974) (see Appendix F).
It is a 25-item scale which measures whether a person
monitors external cues from others to guide his or her own
self presentation or if he or she presents and acts in
accordance with his or her own affective state (M. Snyder,
1974). Low self-monitoring individuals are more knowledge
able about their own dispositions, attitudes, and traits;
high self-monitoring persons are much less dispositionally
guided or discerning (M. Snyder & Cantor, 1980).
The KR-20 reliability of the Self-Monitoring Scale is
.70; test-retest reliability is .83 (M. Snyder, 1974). The
Self-Monitoring Scale measures a different construct than
those measured by the Marlowe-Crowne Social Desirability
Scale, the MMPI Psychopathic Deviate Scale, the c
(Chameleon) scale of the performance style test, the
Christie and Geis Machiavellianism scale, and the Alpert-
Haber Achievement Anxiety Test (M. Snyder, 1974). Numerous
studies support the validity of the Self-Monitoring Scale
utilizing peer ratings and samples which include college
students, insurance and sales personnel, psychiatric
patients, and stage actors (Caldwell & O'Reilly, 1982;
Gabreyna & Arkin, 1980; Ickes & Barnes, 1977; Lippa, 1978;


The results indicated no significant differences (p >
.05) between those women who had not returned to their
abusers and those women who had. A logistic regression
analysis identified no significant relationship between the
determinants and battered women's destination.
Selected demographics also were collected and, where
potential effects were suggested, chi-square analyses were
conducted. Nonwhite battered women were less likely to
return to batterers than were their white counterparts.
Also battered women from small towns were more likely to
return to cohabitation with the batterers, and as population
of residence previous to sheltering increased, battered
women were less likely to return to their mates. The
researcher recommended development of specialized instrumen
tation to operationalize Miller's theory and further
investigation of (a) race and city size of battered woman's
residence for relationship to destination, (b) battered
women's perceptions about authority, and (c) battered women
as a heterogenous grouping.
x


21
the law began to proliferate (Gill & Coote, 1975; M. Kemp,
Knightly & Norton, 1975).
The Womens Rights Movement in the U.S.
From 1973-1976, news of what was going on in Great
Britain surfaced in popular magazines in the United States,
many of which had female audiences. On July 9, 1973,
Newsweek ("Britain: Battered," 1973) included a report on
the opening of Chiswick Women's Aid. Ms. Magazine (Search,
1974) and McCall's ("Wife Beaters," 1975) ran similar
stories. Ladies Home Journal (Durbin, 1974) reported on
wife-beating as discussed at the National Organization of
Women's (NOW) national conference. NOW's perspective
suggested that battered wives stayed in destructive rela
tionships due to finances and the positive societal identity
of married females rather than due to masochism.
As coordinator of the NOW Task Force on Battered Women,
Martin wrote Battered Wives (1976), one of the earliest
major works on spouse abuse in the U.S. Martin faulted
society for maintaining gender and marital inequities, the
legal system for being unresponsive, and social service
agencies for poor coordination of services to battered
women.
The organized women's rights movement had already
addressed the plight of rape victims in America, and the
focus on wife abuse was a direct result of the attention of
the women's movement (Hilberman, 1980). The obvious
parallel between the two issues showed that both are actions


137
Miller, W.P., Seligman, M.E., & Kurlander, T.B. (1976).
Learned helplessness, depression, and anxiety. Journal
of Nervous and Mental Diseases. 161. 347-357.
Mitchell, C. (1980). Personality and socioeconomic
differences between abused and nonabused women.
Unpublished master's thesis, Georgia State University,
Atlanta.
Moore, D.M. (1979). Battered women. Beverly Hills,
California: Sage Publications.
National Institute of Law Enforcement and Criminal Justice.
(1978). Spouse abuse: A selected bibliography.
Washington, DC: U.S. General Printing Office.
NiCarthy, G., Merriam, K., & Coffman, S. (1984). Talking
it out: A guide to groups for abused women. Seattle:
The Seal Press.
Norman, R.P. (1966). Dogmatism and psychoneurosis in
college women. Journal of Consulting Psychology. 30,
278.
Nye, F.I. (1979). Choice, exchange, and the family. In
W.R. Burr (Ed.), Contemporary theories about the family
(Vol. 2, pp. 48-60). New York: The Free Press.
O'Brien, J.E. (1971). Violence in divorce prone families.
Journal of Marriage and the Family. 33, 692-698.
Okun, L. (1986). Woman abuse: Facts replacing myths.
Albany: State University of New York Press.
Olsen, A.J. (1972). Volunteers as family counselors. New
York: Home Advisory and Service Council.
Owens, D.M., & Straus, M.A. (1975). The social structure
of violence in childhood and approval of violence as an
adult. Aggressive Behavior. 1, 193-211.
Pagelow, M.D. (1977a). Battered womena new perspective.
Berkeley: University of California.
Pagelow, M.D. (1977b). Blaming the victimParallels in
crimes against women: Rape and battering. Berkeley:
University of California.
Pagelow, M.D. (1977c). Secondary battering: Breaking the
cycle of domestic violence. Berkeley: University of
California.


13
return will aid concerned helpers in setting their expecta
tions while doing what they know to be helpful in each case.
Finally, battered women can gain greater self
understanding and mutual appreciation by more clearly
knowing the challenges they face to establish violence-
free lives. These findings will not provide them with
the causal explanations for why some return to abusers
and some do not. Yet, the findings may provide them with
clues to explore what in their situations are defensive
responses to their abuse and which have roots in both
external and internalized oppression. Greater under
standing can help minimize self-blame and guilt.
Definition of Terms
Operational definitions for terms relevant to the
research are provided to enhance understanding.
Battered Woman
A battered woman is any married or unmarried woman over
the age of 16 who has been physically abused in ways which
caused pain or injury on at least one occasion at the hands
of an intimate male partner.1 Battered wives and wife abuse
-'-A definition for battered women has not been univer
sally agreed upon by researchers. The most notable distinc
tions about this definition are as follows: A battered
woman would not need to present evidence of injury; self-
report of physical battering is sufficient. The battering
need not be a repeated occurrence as preferred by some
authors (Michigan Women's Commission, 1977; Parker &
Schumacher, 1977). Although abuses other than physical may
be just as devastating, and support exists to include
psychological abuse as a component of battering (Moore,
1979; Walker, 1978b, 1979, 1980a,), the definition herein
has been limited to physical harm alone. The purpose is
two-fold: (a) Physical abuse can be documented more


DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE
By
JOANNE F. DeMARK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988

This dissertation is dedicated in loving memory
to my parents, Assunta (Susie) Cini DeMark and
Charles Domenick DeMark, who inspired and nurtured
my educational dreams.

ACKNOWLEDGEMENTS
I extend my appreciation to everyone whose assistance
enabled me to complete this manuscript. The encouragement,
love, support, and assistance from my committee, intimate
friends, and family have been mainstays during my pursuits.
My committee members' attention, direction, and ideas
have been extremely valuable. Dr. Jaquelyn Liss Resnick, my
committee chair, ably guided my process, especially
assisting me with the feminist and academic scholarship
parameters to which I aspired for my work. She was suppor
tive while expecting quality work. Dr. Phyllis Meek
provided consistent encouragement and confidence in my
abilities throughout my university and community work with
her. Dr. Sandra Damico and Dr. Harry Grater offered
excellent review of my research and offered important ideas
and suggestions for the work. I have enjoyed and benefitted
from shared enthusiasms with my committee: feminist theory
and scholarship with Drs. Resnick and Meek; psychodynamic
theory and practice, specifically Alice Miller's ideas with
Dr. Grater; and the joys and intricacies of quantitative and
qualitative research with Dr. Damico. Their time, expertise
and scholarly review of my work have been greatly appre
ciated.
iii

In addition to my doctoral committee, Dr. Joe Wittmer
and John Dixon provided notable academic and consultative
assistance for which I am grateful.
I am thankful to the battered women at the shelters who
volunteered their participation in my study and to the
staffs at the shelters who agreed to assist me, adding to
their already lengthy and stressful responsibilities.
I appreciate my many friends and family, Deb Vingle,
Dayna Buskirk, the late Sally Briegal, Margaret Levings,
Charlene Shackle, Susan DeMark, Rosemary DeMark, Ann McKain,
my group, Lisa Schultz, Ron Remillard, Susan Myers, Kris
Billhart, Sarah Boykin, John Bryant, Carol Samuels, Norma
Calway-Fagan, Gerri Green, the Dunns, Fleury Means, and
others whose kind hearts and nurturing souls have been
sustenance upon which I could draw.
Above all, I wish to thank my closest allies and
intimates, Phyllis DeMark and Kay Leigh Hagan. The
encouragement, support, processing, patience, under
standing, scholarly support, spiritual and emotional
scrutiny, honesty, nurturing, and love from these two
women have been inspiriting for me.
IV

TABLE OF CONTENTS
PAGE
ACKNOWLEDGEMENTS iii
LIST OF TABLES viii
ABSTRACT ix
CHAPTERS
I INTRODUCTION 1
Statement of the Problem 4
Purpose of the Study 4
Rationale for the Study 5
Research Questions 10
Significance of the Study 11
Definition of Terms 13
Battered Woman 13
Battering 14
Shelter 14
Destination 14
Returning to the Abuser 14
Not Returning to the Abuser 15
Locus of Control 15
Resources 16
Number of Previous Separations 16
Authoritarianism 16
Monitoring 17
Organization of Study 17
II REVIEW OF THE LITERATURE 18
Emergence of Conjugal Violence 18
The British Refuge Movement 19
The Women's Rights Movement in the U.S. 21
Police Involvement 22
Study of Societal Violence 24
v

Theoretical Models 26
General Systems Theory 27
Biological Theory 29
Exchange/Social Control Theory 30
Feminist Theories 31
Social Learning Theory 35
Intrapsychic Theories 38
Variables Relating to the Battered Women's
Destination 43
Frequency and Severity of the Abuse ... 44
Abuse in Family of Origin 44
Length of Marriage, Previous Separations,
Religious Affiliation 45
Love, Affection, Hope 48
Self Esteem 49
Resources/Occupational History 51
Children 53
Locus of Control, Learned Helplessness 55
Authoritarianism and Monitoring 57
Authoritarianism 57
Monitoring 59
III METHODOLOGY 62
Population and Sample 62
Sampling Procedures 65
Instruments 66
Measure of Resources 68
Measure of Locus of Control 68
Measure of Authoritarianism 71
Measure of Monitoring 72
Data Collection Procedures 73
Data Analysis 78
IV RESULTS 80
Demographic Information 80
Analysis of Hypotheses One Through Five ... 85
Analysis of Hypothesis Six 88
Summary 88
V DISCUSSION 90
Summary 90
Limitations of the Study 91
Discussion of Results 93
Recommendations for Future Research 99
vi

APPENDICES
A DEMOGRAPHIC INFORMATION QUESTIONNAIRE 103
B LETTER OF INTRODUCTION 105
C MEASURE OF RESOURCES 107
D LEVENSON LOCUS OF CONTROL SCALES 109
E DOGMATISM SCALE (FORM E) 112
F SELF-MONITORING SCALE 117
G LETTER OF INSTRUCTION/INFORMED CONSENT LETTER . 120
H COMPARISON DEMOGRAPHICS 123
I COMPARISON NORMS FOR LOCUS OF CONTROL,
AUTHORITARIANISM, AND MONITORING 126
REFERENCES 128
BIOGRAPHICAL SKETCH 144
vii

LIST OF TABLES
Table PAGE
4-1 Age of Sample 81
4-2 Race of Sample 82
4-3 Number of Children 83
4-4 Previous Separations 83
4-5 Population of City in Which Subject Resided
Before Shelter 84
4-6 Means and Standard Deviations of Study Measures . 86
4-7 Logistics Regression Model of the Relationship
Between Destination and the Research Variables. . 89
H-l Comparison Demographics from Other Battered
Women Studies 12 3
1-1 Locus of Control Comparison Norms 126
1-2 Authoritarianism and Monitoring Comparison Norms 127
viii

Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
DETERMINANTS OF BATTERED WOMEN'S DESTINATION
FOLLOWING A SHELTER EXPERIENCE
By
JOANNE F. DeMARK
April 1988
Chair: Dr. Jaquelyn Liss Resnick
Major Department: Counselor Education
The purpose of this study was to examine seven poten
tial determinants of battered women's destination following
shelter experience. The variables were drawn from previous
research and the psychodynamic theory of Alice Miller. It
was hypothesized that battered women who did not return to
their abusers would be characterized as less authoritarian,
less monitoring of others, have less chance and powerful
others locus of control, greater internal locus of control,
greater resources, and have left their mates more often
previous to this shelter visit than battered women who
returned to their abusers.
Questionnaires were completed by 72 women during their
stay at one of five southern U.S. shelters, three in Georgia
and two in Florida. Of the sample, 30 returned to their
abusers and 42 did not.
ix

The results indicated no significant differences (p >
.05) between those women who had not returned to their
abusers and those women who had. A logistic regression
analysis identified no significant relationship between the
determinants and battered women's destination.
Selected demographics also were collected and, where
potential effects were suggested, chi-square analyses were
conducted. Nonwhite battered women were less likely to
return to batterers than were their white counterparts.
Also battered women from small towns were more likely to
return to cohabitation with the batterers, and as population
of residence previous to sheltering increased, battered
women were less likely to return to their mates. The
researcher recommended development of specialized instrumen
tation to operationalize Miller's theory and further
investigation of (a) race and city size of battered woman's
residence for relationship to destination, (b) battered
women's perceptions about authority, and (c) battered women
as a heterogenous grouping.
x

CHAPTER I
INTRODUCTION
In the past decade, domestic violence, particularly
spouse abuse, has been brought increasingly to the public's
attention. What was once considered an unspoken family
issue, condoned and encouraged by society, is now deemed a
social problem (Dobash & Dobash, 1971; Roy, 1977; Stacey &
Shupe, 1983; Straus, 1977-78). Considered as one repre
sentation of violence against women, spouse abuse was
initially addressed through the impetus of the feminist
movement. In an effort to better understand, treat, and
ameliorate the problem of conjugal assault, a host of
disciplines and groups, including helping professionals and
feminists, have undertaken the charge to research the
problem (Fleming, 1979; Hansen & Barnhill, 1982; Hilberman,
1980; NiCarthy, Merriam, & Coffman, 1984; Walker, 1979,
1984). This involvement has led to a focus on providing
safety and treatment for spouse abuse victims: the battered
mates and the children who reside in settings where violence
occurs.
An indication of the extent of the problem is the
number of individuals affected by domestic violence. From
one national survey of violence in American homes,
researchers reported that in one household out of six, a
1

2
spouse has committed an act of violence against his or her
partner in the past year (Straus, Gelles, & Steinmetz,
1980). The researchers in that survey concluded that an
American's greatest risk of being assaulted, injured, or
murdered occurs in one's own home by a family member. Forty
percent of all female murder victims are killed by their
husbands (Dobash & Dobash, 1977-78, 1979). Beatings
constitute the most prevalent method of wife murder (Fields,
1977-78).
Conjugal abuse affects not only the marital dyad, but
also the children and family unit. When violence occurs in
the family, children are seeing models for how to handle
relationships and disagreement. A cultural tradition exists
to utilize hitting as punishment to curb unwanted behavior.
Theoretically, children not only learn to curb behaviors
through this model; realistically, they also learn that
violence and love are linked, that violence is morally
right, and that violence is justifiable when something is
really important (Gelles, 1977). Thus, an attitude that
violence can be exercised for "the good" of the recipient is
promulgated (A. Miller, 1983). These lessons in childhood
are transferred to the context of other social relation
ships, and violence in families and relationships becomes a
way of life (Gelles, 1977; A. Miller, 1983, 1986). The
intergenerational cycle theory of violence, i.e., that
children who are recipients of violence will grow up to be
perpetrators of violence, has been validated repeatedly (cf.

3
Bakan, 1971; Gil, 1970; Gillen, 1946; Maurer, 1976; Palmer,
1962; Steele & Pollack, 1974; Welsh, 1976).
Also, domestic violence has significant costs for
society. As examples, public and private sector funds are
used to pay for safe shelters, counseling, police interven
tion, legal avenues, and other resources. Police face
greater injury and death at the scenes of domestic violence
than at any other crime scene at which they intervene (Bard
& Zacker, 1974).
Ratios of wife battering compared to husband battering
range from 11:1 through 13:1 in the United States (Gaquin,
1977-78; Levinger, 1966; Steinmetz, 1977-78). Straus,
Gelles, and Steinmetz (1980) supported social policy,
attention, and treatment specifically on wife abuse for
the following reasons: (a) Wife battering involves more
dangerous and injurious forms of violence, and it results
in greater physical damage and injury; (b) male abusers are
much more often repeaters; (c) wives often act in self-
defense; (d) much abuse by the husband occurs when the wife
is pregnant, posing additional danger to the fetus; and (e)
the mores, laws, and traditions of our society lock women
into marriages in a more substantial way than men.
Wife abuse is a chronic crime which escalates in
severity and poses a serious threat to the safety of the
women involved (Pagelow, 1977a, 1977b, 1981; Walker, 1979).
Battered women commonly report receiving murder threats from
their abusers, as well as perceiving the batterers'

4
capabilities to kill them (Walker, 1979). The violence in
wife abuse is often excessive and relentless, with beatings
continuing after the victims are either unconscious or dead
(Okun, 1986; Walker, 1979; Wolfgang, 1958). The severity of
the threat to abused women's lives is substantiated by the
statistics on women who are murdered by their partners
(Dobash & Dobash, 1977-78, 1979). Many more women seek safe
shelter than abusers seek treatment for their problem
(Fleming, 1979). Batterers find it extremely difficult to
acknowledge their behavior as a problem or to take respon
sibility for the outcome of their brutality (Walker, 1979).
The chances are quite slim that battered women who return to
their households will experience an improved conjugal
relationship with less threat of violence (Pagelow, 1981).
Concern for the safety of victims is of utmost importance.
Statement of the Problem
Greater theoretical understanding of wife abuse is
needed. One aspect of this, whether the battered woman will
return to the abuser or not following a shelter stay, is
important, especially with the strong likelihood that
returning to the batterer means continuation of the woman's
abuse. Theoretical explanations have yet to uncover fully
the factors related to battered women's destinations after
shelter stays.
Purpose of the Study
The purpose of this study was to reexamine key factors
from previous empirical and theoretical research, in

5
combination with factors from a new theoretical perspective,
as potential indicators of battered women's destination,
i.e., not returning to the abuser or returning to the
abuser, following a shelter stay.
Rationale for the Study
In several studies researchers have examined various
factors related to women remaining in or leaving a rela
tionship where they are battered, though the researchers did
not formulate global theories which guided their choice of
factors. For example, in early research frequency of abuse
and severity of the abuse were found to be related to the
battered woman leaving the relationship (Gelles, 1976);
however, all investigators since then have failed to
empirically resubstantiate that finding (cf. Pagelow, 1977c,
1980, 1981; D.K. Snyder & Scheer, 1981). Abuse in the
woman's family of origin (Pagelow, 1977c, 1980, 1981), the
age of children, and number of children (Cristall, 1978;
D.K. Snyder & Scheer, 1981) are not related to battered
women's exit from a battering relationship. D.K. Snyder and
Scheer (1981) found that a higher number of previous
separations, shorter length of marriage, and a religious
affiliation other than Roman Catholic were indicative that a
battered woman would not return to the abuser after a
shelter stay (D.K. Snyder & Scheer, 1981). The length of
the relationship was found not to be a significant variable
relating to dissolution of a battering relationship in
another study (Okun, 1986). Of these factors, many of which

6
are demographic, previous separations and Roman Catholic
denomination have proved to be significant. D.K. Snyder and
Scheer (1981) suggested that perhaps religiosity or authori
tarianism would be more revealing than denomination.
Neither previous separations nor denomination is a foolproof
method to determine likelihood of remaining in or leaving
the relationship.
Others (Gelles, 1972, 1976; Martin, 1976; Pagelow,
1981) have proposed that women who are able to leave
battering relationships are those who are economically
independent from the abuser. Several researchers (Gelles,
1976; Kalmuss & Straus, 1982; Pagelow, 1981) have empiri
cally validated this proposition; financial resources
available to a woman have served as predictive indicators of
her remaining in or leaving a battering relationship.
Although the availability of resources is a key factor, it
does not always explain who extricates herself from an
abusive marriage versus who remains in such a marriage.
Some battered women in shelter programs or therapy have
faced tremendous economic and social hardship, at times with
minimal career skills, yet they left their abusers and
established themselves and their children in new locations
on their own (Fleming, 1979; Martin, 1976; Stacey & Shupe,
1983) .
Several theories have been proposed to explain whether
a woman will remain in or leave an abusive relationship.
One theory for the battered woman's entrapment in an abusive

7
relationship is that a patriarchal society perpetuates abuse
and coercive control of women (Brownmiller, 1975; Dobash &
Dobash, 1979; Schechter 1982). Rights of chastisement
accorded to men as legal property owners of women and
children resulted in the legal and social sanctioning of
male violence against women and children. Patriarchal
values, such as traditional gender identity development
which supports the goals of individuation and separation for
males and attachment for females, reinforce women staying in
relationships unsatisfying to them (Gilligan, 1982) Women
view success of the spousal relationship as their primary
responsibility, and the definition for success is dictated
by the male heads of the household in a traditional family
hierarchy (Dobash & Dobash, 1981). Consequently, battered
women are conditioned within the patriarchal system to
maintain relationships as long as their husbands and society
deem them important.
Although the characteristics of both batterers and
battered women very often include traditional values about
families and stereotypic ideas regarding gender roles, not
all abusers and abused women fit in this category (Walker,
1979). In fact, research to distinguish between women who
remained in abusive relationships via the variable of
traditional ideology has not supported that more traditional
women remain in abusive relationships (Pagelow, 1981).
Although the traditions of a patriarchal society contribute
to an environment "ripe" for wife abuse, the patriarchal

8
legacy theory does not completely define the determinants of
women's destination.
The most relied upon theoretical explanation regarding
battered women is that some women experience a greater
degree of learned helplessness than others, perceiving
themselves as powerless to alter their situations for the
better, regardless of what they do (Walker, 1979). The
battering cycle of tension building, acute incident, and
loving contrition powerfully reinforces the women's learned
helplessness. Walker (1984) expected that battered women
who remained with abusers would perceive themselves as
having less control over their situations than those who did
not remain with abusers, because they perceived others as
exercising greater control over them. Instead, the findings
indicated that battered women in or out of battering rela
tionships saw themselves as having a great deal of control
over what happens in their lives, more so than a general
population norm. They did not view powerful others, e.g.
the abuser, as having a lot of control over their lives
(Walker, 1984).
Additional investigation of potential determinants of
the destination of domestic violence victims following their
shelter stays is needed. A potential source for additional
variables is A. Miller's (1983) theoretical framework for
violence in society. She has proposed that violence in
society occurs because parents and others, exercising the
prevalent child rearing practices of the past several

9
centuries, have shown neither validation nor respect for the
feelings, needs, and desires of children. This "spare the
rod, spoil the child" tradition supports the authority of
the parent, whether right or wrong. A. Miller (1981, 1983,
1986) labeled the training of children to meet consistently
the parents' needs and respond without question to the
parents' and society's authority while the children's own
needs go unacknowledged as "poisonous pedagogy." She
suggested that where this tradition is particularly extreme,
children grow up with a greater vulnerability to violence,
either as the perpetrators or the victims. In adulthood,
they are most likely to deny or repress traumatic childhood
experiences, maintain needs to idealize someone else, and
often skillfully monitor and attend to the needs of others
to gain love and acceptance (A. Miller, 1981, 1983, 1986).
This theoretical perspective applied to battered women
suggests that some battered women are more vulnerable to
return to battering relationships than others because of the
following: (a) They monitored and responded more to their
parents' needs than their own, and (b) they strongly
idealized parents and authority in general.
Various factors previously identified and studied had
not satisfactorily accounted for determining whether
battered women's destination following shelter experience
would back with their abusers or not. Additional deter
minants were considered by selecting the variables of
authoritarianism and monitoring from the violence theory

10
proposed by A. Miller (1983, 1986). The previous two
variables already noted in the literature as determinants,
i.e. resources and number of previous separations (Gelles,
1976; Kalmuss & Straus, 1982; Pagelow, 1981), were inves
tigated. Finally, the locus of control determinant which
Walker (1979, 1984) proposed from the learned helplessness
theory was reexamined. Reinvestigation of this was con
sidered important in light of the strong reliance in this
field on her theory and her surprising findings disputing
locus of control as a factor (Walker, 1984).
Research Questions
The specific research questions were
1. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their authoritar
ianism?
2. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their monitoring?
3. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their resources?
4. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their number of
previous separations?

11
5. Do battered women who do not return to their
abusers following shelter stays and battered women who do
return to their abusers differ regarding their locus of
control (internal, powerful others, chance)?
6. Which combination of the above mentioned variables
provides the best determination of battered women's destina
tion following a shelter stay?
Significance of the Study
This study has implications for those researching
domestic violence, shelter personnel, community agency
personnel, and battered women. A workable theory which
accurately describes the social problem of wife abuse and
which can be used to shape a better family unit and society
has yet to be defined. Whether refinement of current
theories or evolution of a new theory occurs, research is
required on the batterer, the family unit, society, and the
battered woman. Despite the heritage of this problem as a
socially taboo subject, and the paucity of funding provided
to it, the findings from this research effort will add to
what is known about battered women and be a part of the base
for future research efforts on conjugal violence against
women.
If there are identifiable social and psychological
factors that determine the destination of battered women
following a shelter experience, then shelter program
planners could design and administer programs most suited to
the two different populations. Rather than assuming that

12
all programs are appropriate for all shelter residents,
better attempts at identifying the appropriate recipients
for certain interventions could be made. The woman who is
setting up independent living arrangements for herself and
her children needs housing assistance, career guidance,
oftentimes school transfers for children, continuing welfare
aid and food stamps, and social networks. The woman
returning to the abuser might benefit from self-defense
training, conflict resolution skills, and strategies to
encourage marital and family counseling. Depending on her
long-range desires and the degree to which her return is
based on economic dependence, the abused woman returning to
the batterer might benefit from strategies for acquiring
career skills or employment opportunities.
Community agency personnel who come in contact with
battered women include those from food and financial
assistance agencies, churches, legal aid, state prosecution
and public defender offices, vocational rehabilitation
programs, health care agencies, school programs, and
sometimes community mental health or crisis counseling
agencies. With more insight into who is likely to return to
the abuser and who is likely to separate longer-term from
the batterer, personnel from each of these agencies can more
appropriately intervene. Oftentimes, staff working with
battered women can get invested in the hope that a client
will decide her destination in a particular direction. A
more definitive profile of who will stay away and who will

13
return will aid concerned helpers in setting their expecta
tions while doing what they know to be helpful in each case.
Finally, battered women can gain greater self
understanding and mutual appreciation by more clearly
knowing the challenges they face to establish violence-
free lives. These findings will not provide them with
the causal explanations for why some return to abusers
and some do not. Yet, the findings may provide them with
clues to explore what in their situations are defensive
responses to their abuse and which have roots in both
external and internalized oppression. Greater under
standing can help minimize self-blame and guilt.
Definition of Terms
Operational definitions for terms relevant to the
research are provided to enhance understanding.
Battered Woman
A battered woman is any married or unmarried woman over
the age of 16 who has been physically abused in ways which
caused pain or injury on at least one occasion at the hands
of an intimate male partner.1 Battered wives and wife abuse
-'-A definition for battered women has not been univer
sally agreed upon by researchers. The most notable distinc
tions about this definition are as follows: A battered
woman would not need to present evidence of injury; self-
report of physical battering is sufficient. The battering
need not be a repeated occurrence as preferred by some
authors (Michigan Women's Commission, 1977; Parker &
Schumacher, 1977). Although abuses other than physical may
be just as devastating, and support exists to include
psychological abuse as a component of battering (Moore,
1979; Walker, 1978b, 1979, 1980a,), the definition herein
has been limited to physical harm alone. The purpose is
two-fold: (a) Physical abuse can be documented more

14
victims will also mean battered women as defined here.
Abused, harmed, or beaten may be substituted for battered.
Battering
Battering refers to intentional physical abuse which
causes pain or injury. Spouse abuse, spousal violence,
conjugal assault, wife beating, or domestic violence will be
used interchangeably with battering.
Shelter
A shelter is an emergency refuge for safety available
for more than one battered woman at a time, operated
generally by women's groups, volunteer organizations, a
governmental or non-profit agency, and other than a safe
place to stay which the woman provides for herself or is
provided for her by family members, in-laws, or friends.
Destination
Destination, in this investigation, refers to either
returning to the abuser or not returning to the abuser
following a shelter stay.
Returning to the Abuser
For the purposes of this study returning to the abuser
means being in residence with the assailant during the
readily; operationalizing a definition for psychological
abuse is, at best, extremely difficult; and (b) due to space
considerations and various other exigencies, shelter
programs for battered women most often provide safe refuge
services utilizing this same delimited definition as an
admissions requirement. Although battered women could also
include those who have been abused by family members
(Pagelow, 1977), those groups have been excluded from this
study. This sample may include women who are cohabitating
with men although not legally married to them, or separated
or divorced partners who are living with each other.

15
follow-up contact which occurs from 2 to 6 weeks following
departure from the shelter, regardless of whether the
battered woman left the shelter to return to him, or in the
interim weeks decided to move back with him.
Not Returning to the Abuser
Women who are in the category of not returning to the
abuser are those who decide to leave the shelter and do not
live with the abuser and are not residing with him as of
follow-up within the time frame of 2 through 6 weeks after
departure from the shelter.
Locus of Control
Locus of control refers to how an individual perceives
contingency reinforcement upon his or her actions(s) and
generally includes an internal and/or an external dimension.
Attributional style may be interchanged with locus of
control. The three types of locus of control in this
investigation are internal, powerful others, and chance.
Internal locus of control refers to when an individual
perceives that a reinforcement has been contingent upon his
or her action(s) (Rotter, 1966). Powerful others locus of
control is defined as when as individual believes that a
reinforcement has not been completely contingent on his or
her action(s), but externally due, to some extent, to the
control of powerful others (Levenson & Miller, 1976) .
Chance locus of control means that an individual perceives
that a reinforcement has not been fully contingent on his or
her action(s), but due, in part, to the result of chance,

16
fate, or luck (Levenson & Miller, 1976). The Levenson Locus
of Control scales (Levenson & Miller, 1976) was be utilized
in this investigation.
Resources
Resources refers to an indication of the battered
woman's objective dependence or independence from her
abuser, especially as it relates to the woman's employment
status, the presence of children age 5 or younger at home,
and 75% or more of the household income being earned by the
mate. In this study it was measured by the Resources Index
(Kalmuss & Straus, 1982).
Number of Previous Separations
Number of previous separations is defined as any time
the woman has left the mate previously. In this study that
was indicated in response to two questions: (a) Have you
left your mate before, and (b) if yes, number of times.
Authoritarianism
Authoritarianism, also known as dogmatism, is defined
by Rokeach (1960) as an organization of belief-disbelief
systems constructed to both satisfy a need for a cognitive
framework for knowing and understanding the world and to
distance threatening aspects of reality. In this study it
was measured by the Dogmatism Scale (Form D), a generalized
measure of authoritarianism, free of political bias
(Rokeach, 1960).

17
Monitoring
Monitoring is defined as the observation of others and
mediating of one's presentation as guided by social cues (M.
Snyder, 1974), as measured by the Snyder Self-Monitoring
Scale (M. Snyder, 1974).
Organization of Study
This study will be presented in five chapters. The
current chapter is a brief introduction of the subject, the
purpose and rationale for the study, and a description of
relevant terms. Chapter II is a review of the related
literature. The research methodology is detailed in the
third chapter. Data analyses and results are presented in
the fourth chapter. Finally, in Chapter V the researcher
offers a discussion and interpretation of the results, a
discussion of the limitations of the study, and further
implications.

CHAPTER II
REVIEW OF THE RELATED LITERATURE
Compared to other psychological topics, the literature
focusing on wife abuse is quite recent, appearing only in
the last 15 years (Hilberman, 1980). Described initially in
this chapter is how conjugal violence emerged as a social
issue and how the beginnings of the public's acknowledgement
of the problem shape the theories which developed. The
socio-psychological theories of violence, particularly those
which have relevance to this investigation, are presented.
Next, the pertinent research in regard to factors related to
the battered woman's destination is detailed. Finally,
previous research findings are given on authoritarianism and
monitoring, the new variables in the current investigation.
Emergence of Conjugal Violence
Woman battering was slow to be recognized as a problem
atic social issue for several reasons: the sanctity and
privacy of the home, the theory of sadomasochism in humans,
and the subordinate role of women to men in both home and
society (Gelles, 1977; Straus, Gelles, & Steinmetz, 1980;
Walker, 1979). In the past two decades, societal issues
have changed substantially resulting in an environment
favorable to examining both the dynamics of conjugal
violence and a society which has supported, excused, and
18

19
tolerated such abuse. From examination of the early
literature several main forces in the early 1970s reflect a
Zeitgeist which has brought wife abuse to public attention:
(a) the British refuge movement, (b) the women's rights
movement in the U.S., (c) examination of police injuries and
fatalities, and (d) the study of societal violence.
The British Refuge Movement
Although the British were the first to open battered
women's shelters, their initial discovery of the social
problem was quite unintentional. In 1971 Erin Pizzey went
to the Chiswick City Council to seek support for estab
lishing a home as an advice center for women. When she
opened the doors to Chiswick Women's Aid (WA), she dis
covered that the majority of women coming for help were
seeking a safe place away from husbands who beat them.
Within a year and a half she collected demographic and
incidence data from the WA program, including 3,000 requests
from battered women seeking refuge. She gave the informa
tion to England's social service administration thereby
documenting the widespread problem of wife abuse and the
lack of support from legal and governmental communities.
Thus, the shelter movement began (Dobash & Dobash, 1971;
Pizzey, 1974).
Dobash and Dobash (1971) recorded the founding and
growth of WA groups throughout Britain. They suggested that
the establishment of shelters was a natural evolution from
the British women's movement. In addition, they suggested

20
that the patriarchal attitudes which fostered husbands
beating wives also frustrated women's groups in society at
large as they sought help for the battering problem. A
contradiction existed in society between protecting the
privacy of home life and preventing the battering of
spouses.
Becoming aware of WA's growth, the psychiatric and
legal communities were next to get involved. In Memorandum
on Battered Wives (1974), the Royal College of Psychiatrists
scholars used case studies to depict the complexity of
factors surrounding spouse abuse. They described wife
battering as an adaptation failure of inadequate acquisition
of appropriate social learning skills by the abusers.
Citing the fact that battered wives use help when available,
the Royal College scholars recommended more services,
research, and education. They proposed that child abuse
often occurs in homes where wife battering exists.
Scott (1974) described persons involved in spouse abuse
as likely to come from many clinical classifications, rather
than just one or two. Common diagnoses he cited were
immature personality, dependence, aggression, jealous
reactions, and drug or alcohol addictions. Scott asserted
that although sadomasochism is often assumed as the reason
for battered wives returning, more often dependency,
ignorance of choice, and fear of loneliness exist as the
actual reasons. By 1975 British pamphlets outlining legal
options and explaining the difficulties and inadequacies of

21
the law began to proliferate (Gill & Coote, 1975; M. Kemp,
Knightly & Norton, 1975).
The Womens Rights Movement in the U.S.
From 1973-1976, news of what was going on in Great
Britain surfaced in popular magazines in the United States,
many of which had female audiences. On July 9, 1973,
Newsweek ("Britain: Battered," 1973) included a report on
the opening of Chiswick Women's Aid. Ms. Magazine (Search,
1974) and McCall's ("Wife Beaters," 1975) ran similar
stories. Ladies Home Journal (Durbin, 1974) reported on
wife-beating as discussed at the National Organization of
Women's (NOW) national conference. NOW's perspective
suggested that battered wives stayed in destructive rela
tionships due to finances and the positive societal identity
of married females rather than due to masochism.
As coordinator of the NOW Task Force on Battered Women,
Martin wrote Battered Wives (1976), one of the earliest
major works on spouse abuse in the U.S. Martin faulted
society for maintaining gender and marital inequities, the
legal system for being unresponsive, and social service
agencies for poor coordination of services to battered
women.
The organized women's rights movement had already
addressed the plight of rape victims in America, and the
focus on wife abuse was a direct result of the attention of
the women's movement (Hilberman, 1980). The obvious
parallel between the two issues showed that both are actions

22
involving male violence perpetrated against female victims.
In investigating further, Pagelow (1977b) listed the
parallels: Instances of sexual assault and wife abuse were
both underreported to police; they rarely went to court and
had abysmally low rates of conviction when they did. Both
rapists and wife abusers were likely to be repeat offenders;
few offenses evidenced higher rates of recidivism than these
two. The victims of these crimes were pictured stereo-
typically as masochistic, provoking the crime, reporting it
for hidden reasons, and not dedicated to seeing the perpe
trator prosecuted. Thus, in a "blame the victim" cycle,
women in wife battering cases, as well as rape cases, had to
prove that they were "worthy victims" even though the need
to establish worth was not present in other crimes. Pagelow
reported that these crimes were not predominantly lower
class occurrences as often thought, identifying high
incidence in middle and upper classes also.
Police Involvement
With the addition of computers which stored and
categorized vast amounts of crime scene data, law enforce
ment agencies gained greater awareness of police fatalities
and injuries at "domestic disturbance" response calls
(National Institute of Law Enforcement and Criminal Justice,
1978). Twenty-eight percent of the assaults perpetrated
against police officers have been while the officers were
responding to domestic disturbance complaints (Stephens,
1977) A substantial proportion of all homicides on police

23
are related to domestic disturbance calls (Rochester Police
Bureau, 1974) .
Addressing the American Bar Association, Detroit Police
Chief Bannon (1975) commented that in addition to fearing
injury, police avoid domestic violence for two other
reasons: (a) They lack the skills needed to mediate inter
personal conflicts, and (b) they view domestic conflicts as
private problems, not a public matter. He commented:
Of all the nonathletic occupations, none is as absorbed
with the use of physical coercive force as that of the
police officer. Nor are any more thoroughly socialized
in their masculine role images. This . suggests to
me that traditionally trained and socialized policemen
are the worst possible choices to attempt to intervene
in domestic violence. (p. 8)
The lack of police skills to intervene in conjugal
violence was addressed in 1967 by Bard and the New York City
Police Department. They jointly developed and implemented
the first family crisis intervention program within a police
unit (Bard & Zacker, 1971). Other literature detailed other
early efforts at police training in crisis intervention,
referral, and conflict management (Olsen, 1972; Rochester
Police Bureau, 1974; Spitzner & McGee, 1975).
Examination of police work brought additional domestic
violence factors to light. The incidence and degree of
interpersonal aggression of domestic disturbance calls
cannot be isolated to lower socioeconomic, minority neigh
borhoods as a comparison between Harlem, New York, and
Norwalk, Connecticut, indicated (Bard & Zacker, 1971).
Evidence contradicted the common beliefs that family

24
disputes are usually associated with alcohol and that
assaultive behaviors by family members usually ceased before
the arrival of police (Bard & Zacker, 1974). Greater
caution was advised for the common practice of diverting
wife beaters from the criminal justice process (Brakel,
1971).
The Study of Societal Violence
By the late 1960s society in general felt more concern
about the rising crime rate, especially of violent crimes.
In 1969 the Eisenhower Commission (Goode, 1969) looked into
the causes and prevention of violence, and in 1971 the U.S.
Surgeon General conducted a similar study (Roy, 1977). The
Surgeon General's examination linked media violence and
violence in real life, suggesting a causal relationship
between the two (Roy, 1977). Goode (1969), a member of the
National Commission on the Causes and Prevention of
Violence, theorized that young children learn violence and
the appropriate use of force through observations of parents
during early childhood. Goode suggested that social class
differences exist for spouse abuse due to lower class
acceptance of violence and conflict in their environment in
general. With more conflict situations arising in lower
class neighborhoods, limited places to escape conflict, and
fewer resources (e.g., money, power) available to get what
counts, violence is a more likely option.
Gelles (1972) coined the family as the "training ground
for violence" (p. 169) with family members modeling violent

25
acts that have much more impact than either television
violence or school discipline in producing future genera
tions of assaultive people. In his research, family members
from both violent and nonviolent families approved of
certain violent acts as "normal" in that they served the
purpose of dissipating a husband's tension and/or appeasing
an otherwise hysterical mate. Gelles further supported the
family structure and the family's societal position as more
important factors than individual pathology in the occur
rence of family violence. Gelles estimated that intrafamily
violence occurs in 37% of any general population which
compares similarly to other estimates of this nature
(Levinger, 1966; O'Brien, 1971). Gelles (1975) was the
first researcher to note the higher likelihood of violence
against a wife when she is pregnant.
Straus established the Family Violence Research Program
at the University of New Hampshire in 1970. He found that
the cultural and social signals from a violent society
combine with familial, psychological factors to incite
violence in the home, contradicting the myth of the family
as a warm and loving system (Straus, 1974).
Straus (1973) looked at the balance of power in
families in relationship to the occurrence of violence.
Where equal power existed between the husband and wife, the
husband to wife violence was lowest. High power in either
wife or husband was related to high levels of husband-to-

26
wife violence. Where the wife's power was recognized as
high, the wife-to-husband violence was also high.
In studies where college students recalled parental
conflicts, Straus (1974) disproved the popular notion of
"cathartic violence," i.e., if pressure is built up and not
released through some means, it will eventually erupt as
violence. Instead, he found that increased verbal battling
escalated physical aggression, rather than dissipating such
stresses. College students' parents who utilized more
intellectualizing during marital conflicts experienced lower
levels of aggression towards spouses.
Straus, working with others (Allen & Straus, 1975?
Owens & Straus, 1975), found a moderate correlation between
observing, committing, or being victim of violent acts as a
child and adult approval of interpersonal or political
violence. He also found that the greater the husband's
resources, the less likelihood that the husband would
perpetrate violence in the home.
Theoretical Models
These four preceding forces focused on the problem of
domestic violence, building a body of knowledge which
brought multiple facets of wife abuse to light. Theoretical
models of family violence resulted. In some cases, theories
applied to human behavior and aggression in general have
been overlaid on the topic of family violence, e.g., the
theory of learned helplessness (Seligman, 1975) or the
frustration-aggression hypothesis (Dollard, Dobb, Miller,

27
Mowrer, & Sears, 1939). In other cases theories have
evolved explicitly from the study of domestic violence as in
Walker's (1979, 1984) cycle theory of battering. The
following major theoretical areas are examined: (a) General
systems, (b) biological, (c) exchange/social control, (d)
feminist, (e) social learning, and (f) intrapsychic.
General Systems Theory
Straus (1973) developed a theory to explain family
violence known as the "general systems theory." The family
unit is a social system with a purpose, seeking to meet
goals and adapting to the environmental context in which it
exists (Straus, 1973). A violent family system is described
as follows: A probability for conflict occurs in the family
due to involuntary membership, intensity of relationships,
generation and gender differences, and parental rights of
influence. The probability of violence can be positively
reinforced due to a highly violent society and socialization
of family members in violence vis a vis parental models,
physical punishment, tolerance of sibling assaults, and
macho values, especially for boys. As a result the culture
legitimizes a norm of violence and family members integrate
assaultiveness into their personality characteristics
(Straus, 1973). The sexism of society, through its limited
roles and career opportunities for women, unequal pay,
assumptions regarding wives and mothers, assumptions
regarding males as head of households, and socialization of
women into subordinate roles, reinforces that a woman should

28
"stay in her place" regardless of what abusive treatment she
receives. All of these socialization imprints provide for a
high level of violence in the family (Straus, 1973) .
Another general systems model has been advanced by
Giles-Sims (1983). She builds on Straus (1973) and Buckley
(1967), as well as Broderick and Smith's (1979) concepts
regarding hierarchies of feedback and control in systems.
The six stage model includes (a) the establishment of the
family system, (b) the first incident of violence,
(c) stabilization of the family system, (d) the choice
point, (e) leaving the system, and (f) resolution or more of
the same. Both the man and woman creating the system come
to it with individual histories and personalities that
influence the way they interact with each other. Precipi
tating events or stressful situations occur in the marriage.
Those events, in combination with prior history relating to
violence, enter into the execution of violence. If the
assault serves to maintain the system or satisfy goals of
the batterer, the next move is by the woman. If her
response to the assault is to ignore, deny, or provide
forgiveness, she colludes in a feedback loop which encour
ages the continuance of such behavior. On the other hand,
if the abused woman gets angry and considers the violence as
a possible pattern, then she may utilize social supports and
examine her alternatives. At this point, the battered woman
may leave. If the assaultive behavior did not satisfy some
of the system achievement or maintenance goals, then the

29
batterer may choose an alternative behavior that could be
more useful and acceptable. If the woman leaves and will
come back only under the condition that he change, he may
choose some alternative behavior as well. Her other two
options include not coming back at all, or coming back with
no requisite that he change. A weakness of this theory is
that Giles-Sims did not incorporate some of the less
directly observable influences such as the patriarchal
influences or societal gender inequities.
Biological Theory
Elliott (1982) suggested that biological explanations
have been totally ignored in postulating theories about
intrafamilial violence. He detailed how organic and meta
bolic disorders contribute to domestic disturbances; he
specifically focused on episodic dyscontrol syndrome or
"explosive rage." His sample of 286 patients with histories
of violence was a skewed sample because of the biased
selection on the part of the referring physicians, the use
of computerized tomography (the CT scan), and the pain
staking methods utilized to uncover minimal brain dysfunc
tion. Elliott recommended further research with an
unselected group of batterers to uncover neurological
causation factors in domestic violence, and by no means
suggested biology or genetics as an all encompassing factor.
Although neurological dysfunction may be pertinent in a
small number of cases, it is highly unlikely that it is a
major explanatory feature of wife abuse in our society. The

30
question would arise as to why so much brain dysfunction
occurs among men only.
Exchanae/Social Control Theory
Exchange and social control apply to all types of
intrafamily violence, including conjugal abuse (Gelles,
1983). Exchange theory suggests that the ways humans
interact with each other is dictated by the rewards they
receive or the negative consequences they avoid (Gelles &
Straus, 1979). The social control theory is the question of
how much control is exerted through society and social
institutions to encourage creation, maintenance, or dimi-
nishment of some belief or act such as violent behavior.
According to this theory, family members batter other family
members simply because they can.
First, the costs of being violent are less than the
rewards; therefore battering occurs. Since effective social
controls aimed at preventing family violence are not in
place, a family member who batters experiences few conse
quences due to these lack of controls. Also, inequality
within the family, the sanctity of the family unit in
society, and the encouraged image of males as macho minimize
social control in the family thereby reducing the conse
quences of violence to the batterer and increasing the
rewards.
When societies have no legal or normative sanctions
regarding family violence, the violence is more frequent
(Nye, 1979). Additionally, when norms or laws exist

31
sanctioning one type of family violence but not another,
e.g., laws against wife abuse but not child abuse, the
violence will be more frequent in the less sanctioned
category.
The exchange/social control theory for family violence
has application to treatment and policy issues (Gelles,
1983). Treatment interventions and policies that increase
both social controls exercised over the family and the costs
for violent behavior are primary. The resultant policies
should diminish societal norms that glorify violence,
increase financial and gender equality, and create more
responsive criminal justice and social service agencies.
The value in such a theory is that it can be applied across
the board to explain all types of familial abuse.
Feminist Theories
According to feminists, explanations regarding violence
against women are inadequate if concerned with general
explanations regarding overall societal violence or family
violence (Dobash & Dobash, 1979; Schechter, 1982). By
treating all violence between parties as equal, some
significant uniquenesses about domestic violence against
women are masked (Dobash & Dobash, 1981b). Feminist authors
have particularly challenged Straus' (1980) conclusion that
there is nearly sexual equality in perpetuation of violence
in couples and Steinmetz (1977-78) who suggested an equally
prevalent "battered husband syndrome" (p. 501). Pleck,
Pleck, Grossman, and Bart (1977-78) have pointed to inac-

32
curate interpretations by Steinmetz of her own data
regarding incidence of husband abuse. Fleming (1979) and
Schechter (1982) challenged these premises, making a strong
case that the more serious violent incidents, often with
more severe physical and psychological repercussions, are
perpetrated by men against women in conjugal relationships.
Feminist researchers pinpoint the reliance on the
social survey as contributing to faulty findings (Dobash &
Dobash, 1983; NiCarthy, Merriam, & Coffman 1984). The data
from survey methods are very informative regarding socio
demographic characteristics, but are of little value for
information regarding the sensitive issues involved in
family violence about which society and individuals have
traditionally kept quiet (Dobash & Dobash, 1983). Conse
quently, many of the findings and conclusions from this
approach have lacked consistency, been proven unfounded, or
are replete with contradiction (Dobash & Dobash, 1983;
Fleming, 1979; Pagelow, 1980; Pleck, Pleck, Grossman & Bart,
1977). To understand this phenomenon in our society, in-
depth systematic interviewing of battered women by experts
on this subject is preferable to the survey technique as a
research method (Dobash & Dobash, 1983; Pagelow, 1981;
Walker, 1984). And for feminist researchers, calling "woman
battering" or "wife abuse" by the term "spouse abuse" is in
itself a sexist act masking the true nature of the problem
(Dobash, 1981).

33
There are several theories that represent the feminist
analysis of wife abuse. Dobash and Dobash (1977-78) traced
the history of violence against women by placing it in the
wider context of society, thus supporting that abuse against
wives has been acceptable behavior in a patriarchal society.
Dobash and Dobash (1977-78, 1979) cited marriage laws
throughout history that categorize wives as the possessions
of husbands and dictate the husband's obligations to
chastise and reprimand wives. Many of the laws clearly
reflect double standards. For example, English Common Law
denied wives civil rights and legal status. England,
Europe, and early America had laws supporting the husband's
rights to beat his wife. It was not until the 1700s that
this tradition came into question. In the early 1800s
written laws were adopted in various states that reiterated
the sanctity and privacy of the home unless "permanent
injury or excessive violence" were present (Eisenberg &
Micklow, 1977, p. 149). Finally, not until 1891 does
caselaw reject the husband's legal right to beating or
abusing his wife (Dobash & Dobash, 1977-78). Dobash and
Dobash (1979) believed that it is the tradition of the
patriarchal society that has also impacted on the leaders of
the women's movement in their attempts to address wife abuse
through both traditional institutions and the grass roots
shelter movement. Consequently, nonsupportive laws change
too slowly, funding is inadequate, and institutions remain

34
nonresponsive and arduous for battered women and those
helping them.
The patriarchal tradition today combines with the needs
and realities of a capitalist economy to continue institu
tional and socialization patterns that perpetuate male
dominance and aggression (Schechter, 1982) Battering is
one of the many ways that men maintain control in our
society. Women stay in abusive relationships for two
reasons: (a) Their inferior status under capitalism (lower
wages or wagelessness and unequal division of labor
requiring women to be responsible for maintenance of both
home and family) leaves them financially dependent on men,
and (b) their sex role socialization as wives and mothers
results in a moral obligation to stay and please their
husbands and rear their children properly (Schechter, 1982).
One problem with patriarchal society explanations is
that they are historical and hypothetical in nature, with
minimal empirical work with battered women operationalizing
the specifics of the theory. An example of a researcher who
has captured empirically some of the aspects of patriarchal
society is Pagelow (1981). She has addressed particularly
the question, "why do women stay." The hypothesis is that a
woman of few resources, receiving negative institutional
response (family, law enforcement, social service agencies,
society), and very traditional in her beliefs, is unlikely
to change or believe she can change her battering situation.
Pagelow (1981) supports the proposition that limited

35
resources (as measured by the woman's age, her children's
ages, differences in earnings, husband's earnings, and home
ownership) correlate strongly with the woman's staying in
the relationship. She did not find a strong correlation
between institutional response and traditional ideology with
the women's length of cohabitation while in a battering
relationship.
Social Learning Theory
The most relied upon psychological theory regarding
battered women is that proposed by Walker (cf. 1977-78,
1979, 1980b, 1984). Because of its well established
reference in the battered women's movement and its role as a
significant precedent for this investigation, Walker's
theory will be considered in detail here.
Walker (1977-78, 1979, 1984) utilizes Seligman's (1975)
theory of learned helplessness as a basis to understand the
battering relationship and battered women's responses.
Learned helplessness is combined with a cycle theory of
violence to depict the phenomenon of battering (Walker,
1979). The supporting research (Walker, 1978a, 1978b, 1979,
1980a, 1981) is based on interviews with several hundreds of
battered women and some caregivers to battered women (e.g.,
therapists, shelter workers).
Learned helplessness, according to Seligman (1975),
occurs when animals experience negative reinforcement
noncontingently, i.e., with no apparent link to their own
actions. Eventually they believe that their actions have no

36
effect on what may happen to them. This theory's appli
cability to humans has also been supported by Seligman and
others (Hiroto, 1974; Klein & Seligman, 1976; W.P. Miller &
Seligman, 1975; W.P. Miller, Seligman, & Kurlander, 1976;
Seligman, 1975; Seligman & Hiroto, 1975). With battered
women, as Walker's theory proposed (1977-78, 1979, 1984),
battering from an abuser occurs at random, often in no
directly rational relationship to the woman's immediately
preceding behavior, and according to no set schedule or
timeframe.
As repeated battering occurs, she comes to believe that
nothing she can do will have any impact on the batterer's
abusive behavior. All of the actions she has previously
tried, e.g., being apologetic and loving, fighting back,
leaving, calling the police, have not resulted in an end to
the unpredictable abuse. Walker proposed that a woman will
more quickly "learn" her powerlessness if she experienced
more rigid sex-role stereotyping as a child which already
reinforces passivity on her part and the importance of
relying on men for help and protection (Dweck, Goetz, &
Straus, 1980; Radloff & Rae, 1979, 1981; Walker, 1977-78,
1980a, 1984).
From Walker's interviews, she determined that the
battering, although neither constant nor occurring in a
consistent manner each time, does occur in a cycle of three
distinct phases (Walker, 1977-78, 1979, 1984). The first
phase is the tension-building stage. During this stage,

37
tension increases in the home with the batterer expressing
anger and dissatisfaction, but without any extreme violent
expression. The woman tries to please the batterer to
minimize the hostility, calm him, and prevent any more
expression of hostility. However, the tension increases,
and her efforts to reduce his angry responses become more
unsuccessful. Eventually, the woman withdraws, thus
eliciting an insecurity and an increase in possessiveness,
jealousy, and oppression from the abuser.
At phase two the tension becomes intolerable and acute
battering occurs. The batterer explodes into physical and
verbal assault. This is the most likely time for the woman
to receive injuries or for external parties, e.g., police,
to become involved. This incident usually dissipates the
tension which had built up, and the violence temporarily
ceases.
In stage three loving contrition ensues. The batterer
may apologize profusely, promise never to repeat the
actions, buy gifts and/or be more helpful to the woman. The
relationship enters into a honeymoon-like appearance. The
batterer may believe he will not be assaultive again, and
the woman wants very much to believe that he will not.
Empirical findings from Walker's (1984) sample support
the three-stage theory. In 65% of the relationships studied
by Walker (1984), evidence supported the tension-building
stage before an acute battering incident. And with 58% of

38
her subjects, a loving contrition stage occurred after the
acute battering incident.
Intrapsvchic Theories
Earliest theoretical explanations of spouse abuse are
based on the concept that abuse occurs because the parti
cipants are individually pathological, i.e., have inadeguate
personality development (Snell, Rosenwald, & Robey, 1964).
This intrapsychic viewpoint suggests that batterers are
passive-aggressive, obsessive-compulsive, paranoid,
sadistic, immature, dependent, excessively jealous, and/or
addicted to drugs or alcohol (Scott, 1974; Shainess, 1977).
The wives contribute to the abuse because they are aggres
sive, masculine, domineering, sexually frigid, and maso
chistic (Snell, Rosenwald, & Robey, 1964; Shainess, 1979).
These theoretical explanations have fallen short because
neither explains persons who have these personality charac
teristics and are not in battering relationships, nor takes
into account a myriad of interactional and social factors
surrounding domestic violence.
The most prevalent Freudian view applied to battered
women is masochism (Shainess, 1979). Yet within the context
of contemporary psychoanalytic theory, Gillman (1980)
proposed an object-relations approach to understanding and
treating battered women, following from the work on border
line personality (cf. Kernberg, 1975). She (Gillman, 1980)
suggested that masochism is not a key factor in battered
women since the defense mechanism of repression, usually

39
present in masochistic individuals, is not often present in
battered women. Battered women endure far too much abuse
for their behavior to be considered "neurotic."
Battered women more often fall into the category of
borderline personality, being neither neurotic nor psy
chotic. The "battered woman personality" is a woman who has
two internal, yet separate, representations of herself: one
good and the other bad. As battering occurs, she is a
worthless person in a destructive relationship to a per
secuting husband (mother). She has difficulty leaving the
relationship in that the good self and other emerges after
the incident is over and as the bad submerges. The split
ting means quite opposing views of self and spouse are
present, yet the battered woman is unaware of the contra
dictory nature of her beliefs (Gillman, 1980).
A source for variables for this investigation was the
theory of violence proposed by A. Miller (1983). Because of
this theory's application to the current study, the theory
is presented in detail.
A. Miller (1983) proposed that violence occurs in
society because of the heritage of strict child rearing
practices which break children's willfulness when they are
very young, and causes them to repress their own traumatic
childhood experiences, and leads to their aggressiveness or
self-destructive behaviors in adolescence and adulthood.
Historically, children have been viewed as mean, demanding,
tyrannical, possessing destructive drives and sexual longing

40
for their parents, and having other negative desires and
behaviors which society and parents have believed are their
responsibilities to ameliorate (Rutschky, 1977). The
child's willfulness must be emphatically curbed and the
child must view the parents as unquestionably right and
superior (Rutschky, 1977). Both physically and emotionally
abusive and humiliating methods are utilized by parents to
accomplish their responsibilities of rearing socially
acceptable children. The parents are unaware of the ways
they are abreacting their own suppressed hatred and anger
from their own childhood by utilizing their offspring to
meet their suppressed needs (A. Miller 1981, 1983, 1986).
In this environment the child's needs are not met but
frustrated. The child experiences hurt and anger, but
cannot have these feelings validated with parents. Without
any validation the child denies the feelings (A. Miller,
1981, 1983, 1986). A proneness to violence is most likely
if the child from such an upbringing does not have the
following: (a) another human being in whom to confide true
feelings, (b) an environment allowing experiencing and
expression of pain (whether at home or elsewhere), (c) other
objects for abreacting hatred, and (d) an education or
intellectual avenue to rationalize hatred (A. Miller, 1983).
Depending, then, on the degree to which the child has been
traumatized and the unavailability of healthy channeling of
negative experiences and feelings which are suppressed,

41
rather than addressed, the likelihood for aggression or
self-destruction can be determined.
What follows is that the child represses any memory of
the traumata and idealizes the parents and the upbringing.
The child, because of love for the parents, takes the
responsibility and the blame for the parents' hurtful
behavior (A. Miller 1983). The child's feelings of anger,
hurt, hatred, helplessness, and pain come forth in other
life circumstances, but disassociated from their origin.
The child is consciously blocked from realizing their
source. In addition to idealizing those who traumatize the
child, the child can strongly identify with the perpetrator
in an effort to distance the self from the identified "bad
child" whom the perpetrator/parent has been disciplining so
sharply. The child has thus split-off, projected his or her
own weakness, feelings of helplessness, and vulnerability by
maintaining a certain derision for others who are younger,
weaker, smaller (A. Miller, 1981).
Since society maintains a lack of sensitivity to the
denied cruelty, children grow up to become parents who
manifest the same types of humiliations and physical and
emotional abuse on their own children (A. Miller, 1983).
This process means that cruelty and violence become passed
on from generation to generation. Only a sensitive dis
covery of one's own traumatization, in a supportive environ
ment, breaks the intergenerational cycle of violence.

42
Children who receive humiliating and cruel treatment
when they need protection, respect, and honesty will grow up
to subject others or themselves to destructive acts. If
they become parents, their own children will be the likely
outlets. Lacking a weaker spouse or children, their
violence may be directed at others in society (A. Miller
1983) Self-cruelty is another possibility. Having
received poor treatment as children, adults can continue to
treat themselves abusively (A. Miller, 1983).
A. Miller (1986) has argued against Freud's theory of
drives in children indicating that his willingness to
consistently interpret neuroses as rooted in the child's
drive conflicts with Freud's own protection and idealization
of the parenting he received. Freud's own narcissistic
wounds, and a lack of an empathic environment for him to
uncover his own repressed traumatization, led him to collude
in blaming the victim and sparing parents' feelings (A.
Miller, 1986).
In addition to being self or other destructive, another
variation of behavior can develop in response to such a
childhood. As the child idealizes the parents and takes
responsibility and blame for traumas, the child comes to
understand performing for the parents and parents' needs as
a means of winning their love. Thus, some children from
this environment have the particular skill of discerning the
needs of others (e.g., the parents) no matter how subtle or
unconscious the signs are (A. Miller, 1981). The develop-

43
ment of this monitoring skill comes at the cost of under
development of the child's own ability to perceive his or
her own needs.
To summarize, A. Miller (1981, 1983, 1986) proposed
that violence occurs because of the following: Parents have
narcissistic disturbances from their own childhood traumatic
experiences which they have repressed. They treat their own
children cruelly and unsympathetically by splitting off
their "bad selves" on their weaker and more vulnerable
children. Society colludes with the process by mandating
parents to rear obedient and conforming children and viewing
acting out not as children's reactions to their
traumatization but as manifestations of the bad nature of
children who have not been properly reared into subser
vience. Society protects parents, provides avenues histori
cally for socially acceptable violence, e.g., beating wives
and children as appropriate disciplinary measures, striking
students for wrongdoing. Children who are disturbed through
such upbringing will either dull their self-sensing abili
ties, harm others weaker than themselves, become self
destructive, deny their own traumatization and needs, refine
their abilities to monitor and please others in a quest for
love and narcissistic fulfillment, or some combination of
these qualities.
Variables Relating to the Battered Women's Destination
Researchers espousing different theories have examined
a number of variables relating to the question of battered

44
women leaving abusive relationships. Researching a variety
of variables is important to build a solid theoretical
understanding of the problem. What follows is a
summarization of that research and, where applicable, a
discussion of how it applies to the investigation.
Frequency and Severity of the Abuse
Gelles (1976) believed that a battered woman would
mobilize on her own behalf the more frequent or severe the
abuse. His research supported this hypothesis. Other
research did not confirm the same finding (Pagelow, 1977c,
1980, 1981). Nor were frequency or severity significant in
defining whether a battered woman would return to the abuser
after a shelter stay (Snyder & Scheer, 1981).
Abuse in Family of Origin
If a battered woman was exposed to more violence in her
family of origin, would she be less likely to stay away from
her husband, the abuser? The families of origin of battered
women have often been violent families. Twenty-three
percent of the family histories of the wives studied by
Gayford (1975) were violent. In other studies (Hilberman &
Munson, 1977-78; Scott, 1974) that percentage was 50% or
more. Parker and Schumacher (1977) found that if the mother
in the wife's family of origin was physically abused, it was
significantly probable that the wife would be battered by
her husband. Half the women in the Hilberman and Munson
study (1977-78) were abused themselves as children. Gelles
(1976) found that the less a woman was exposed to parents

45
battering each other in her family of origin, the more
willing she was to seek outside intervention when she was
abused by her spouse. Pagelow (1977c, 1980, 1981) was not
able to substantiate the same finding in her research. It
is possible that a similar percentage would occur for non-
battered women's families of origin. Study with families of
college students indicates violence was present in 28% of
the cases.
Length of Marriage, Previous Separations. Religious
Affiliation
Snyder and Scheer (1981) examined a number of variables
for their ability to suggest whether a battered woman would
return to her abuser following a stay at a shelter. They
selected variables from data available at admission to the
shelter so the results would be helpful for shelter workers
in selecting their intervention strategies with the shel
tered women. The variables were not selected to support or
test any particular theory, but to be helpful in theoretical
formulation.
In their study the sampe was comprised of 74 women
admitted to a shelter in Detroit, Michigan. Fifty-five
percent of the women (41 women) were living with the abuser
when follow-up was conducted 6 weeks after they left the
shelter; 45% (33 women) were not. When comparisons were
conducted between the two groups, six variables from
admission were significantly different between the two
groups: two of the reasons for seeking admission (seeks
short-term separation only and seeks conjoint marital

46
counseling), relationship to the assailant, length of mar
riage (if married), previous separations, and religious
affiliation (Roman Catholic or not). No significant group
differences were found for any sociodemographic variables or
for measures of the nature or severity of the abuse inci
dent.
These six variables, when entered into a forward
stepwise discriminant function analysis with destination at
follow-up as the criterion variable, yielded three predic
tors. Complete data on all six variables were available for
50 of the 72 abused women; the 50 were subjects for this
portion of the research. The length of the marriage,
occurrence of previous separations, and religious affili
ation best predicted, with an overall classification
accuracy of 79.6%, where the woman would be residing at
follow-up. The authors support cross validation of their
study utilizing independent samples from other settings.
There seems to be no or minimal other research in the
literature for the three variables identified by the Snyder
and Scheer (1981) research as particularly relevant to why
the woman stays or why she returns to an abusive relation
ship. Walker (1979) indicated that abused women who leave
their batterers and return to shelters on multiple occasions
are more likely to ultimately leave the battering relation
ship. Several authors (Martin, 1976; Pagelow, 1981; Walker,
1979, 1984) have reiterated from interviews with battered
women that longer marriages are the more difficult relation-

47
ships to leave, despite the abuse. In research which
focused on couples who thought about divorce compared to
those who did not (Booth & White, 1980), results were that
religious affiliation does not have a strong effect on
thinking about divorce. Of the couples who had come from
lengthier marriages with stronger religious belief there
were fewer women who thought about divorce; those couples
with shorter marriages or less religious zeal had more wives
who thought about divorce. They found that the more intense
one's religious belief, the less likely one would think
about divorce. Booth and White (1980), nevertheless, did
find that in religious couples, even where marital satisfac
tion was high, persons experiencing abuse were more likely
to think about divorce.
Stronger religious belief or activity is often tied to
a more traditional way of looking at life (Pagelow, 1981).
Walker (1979, 1984) found in her sample of both battered
women and batterers, as described by the women, a strong
traditional view about the home, family, and sex roles.
Pagelow (1981) felt that the more intense the traditional
ideology of women who have been battered, the more likely
they are to remain in battering relationships and the less
likely they will be to take action to significantly better
the situation. To test this hypothesis, she looked at
whether the women came from a religious family of origin
(i.e., family religious activity), as well as five other
variables she felt reflected traditional ideology. Her

48
variables in combination did not support her hypothesis.
Her variable about a religious family, however, did have
a statistically significant correlation to the dependent
variable. Wetzel and Ross (1983) supported the idea that
battered women are kept captive by the family and religious
values that they hold among other things.
Love. Affection, and Hope
Research on the impact of love, affection, and hope to
battered women's destination is scant. In a sample of
battered women who were not sheltered women, but battered
women located via social service agencies, legal or medical
sources, advertising, and by word of mouth, the most
commonly cited reason for staying in abusive relationships
was that the women loved their husbands (Cristall, 1978).
Forty percent of 542 shelter residents sampled (Stacey &
Shupe, 1983) when asked why had they remained with the
abusers if they had not left immediately after a first
incident of violence, said they were still optimistic about
the future of the relationship. Twenty-seven percent of
those felt they could still save the marriage. Fifteen
percent suggested they stayed because of their affection for
the partners.
Understanding how love, affection, and hope relate to
women's vulnerability to remain in abusive relationships is
a needed undertaking. Women have been identified in our
society as being more relationship oriented than men; their
socialization and conditioning have reinforced them in this

49
way (Gilligan, 1982). Males have considered their main
focus, their life's work, to be their career. Women,
traditionally, have viewed their life's work to be wives,
mothers, and keepers of the household. Leaving a relation
ship for a woman is an admission of career failure and,
thus, is not taken lightly. An unsuccessful marriage is a
subordinate concern for a man compared to a career failure.
Self-Esteem
More has been stated about the battered woman's self
esteem than probably any other variable. Researchers and
clinicians who have interviewed or provided services to
battered women repeatedly comment on the low self-esteem of
battered women (Bell, 1977; Bowen, 1982; Brown & Brazzle,
1982; Carlson, 1977; Hilberman & Munson, 1977-78; Pagelow,
1981; Walker, 1977-78; 1978b, 1979, 1981, 1984; Wetzel &
Ross, 1983) .
From the few studies where instrumentation has been
utilized conflicting results exist. In a comparison of 46
battered women to 12 non-battered women, Star (1978) found
lower self-esteem with the battered women. Star, Clark,
Goetz, and O'Malia (1979) found, through their study of 57
battered women, that battered women could be described as
lacking in self-confidence, low in self-esteem, aloof,
anxious, reserved, uneasy in social interactions, and
critical or uncompromising.
Hartik (1978) examined 30 battered women compared to 30
non-battered women for personality characteristics and self-

50
concept. She found battered women to be lower in ego
strength and self-esteem, with greater identity problems,
than non-battered women. Overall the battered women were
dissatisfied with themselves physically, morally, socially,
and in relationship to their families. Her research
indicated that battered women had a very difficult time
maintaining even minimal self-esteem. In a comparison of 20
battered women from shelters to 20 non-battered women from
the same socioeconomic status, Chan (1979) found the
battered women to have lower self-esteem.
Conflicting results come from two studies. Brown and
Brazzle (1982), reporting on the 40 abused women in shelters
whom they studied, indicate that 42% of the women had high
self-esteem. These authors believed it is inappropriate to
assume that battered women have low self-esteem. Mitchell
(1980) compared 16 nonabused women who were in therapy for
psychological problems to 24 abused women who were in
treatment for the abuse. The two groups did not differ in
self-esteem, sex role stereotype, locus of control, or
anxiety. What she did find, however, when studying within
group differences among the battered women was that those
who were less educated, unemployed, and severely and/or
frequently abused were lower in self-esteem, more anxious,
and more externally oriented. These studies not only
suggest that battered women may not be lower in self-esteem
than their nonbattered counterparts, but also that within
group variances may distinguish among battered women.

51
Walker (1984) examined self-esteem in battered women
via a semantic differential scale on which they rated
themselves, women in general, and men in general. Although
the prediction was for low self-esteem, battered women
viewed themselves more favorably than they viewed women in
general or men in general. The difficulty with this
research was that the scale did not appear to be standar
dized. It was impossible to tell how the battered women's
perceptions of themselves and others compared to other group
perceptions about themselves, women in general, and men in
general. Battered women in this case perceived themselves
more favorably than they perceived other groups, but how
does this compare with other groups' perceptions about them
selves? Other groups' perceptions could be more elevated
than battered women's perceptions which would validate a
lower self-esteem comparatively for the battered women's
group.
Although self-esteem was not be addressed in this
study, more definitive examination of this factor is needed.
Resources/Occupational History
Martin (1976) was the first to emphasize that financial
dependence and unresponsive social systems are very real
explanations for why battered women remain with their
spouses. At the same time, actual research results (Gelles,
1976) results indicated that the fewer resources and less
power the abused woman had, the more likely she was to stay.
Resources have been determined to be any of several items,

52
e.g., employment status, having children under 5, income
level. In the Gelles study (1976), the resource variable
that best distinguished wives who obtain assistance from
those who remain with the husband is holding a job. Many
authors (Dobash & Dobash, 1977-78; Fleming, 1979; Schechter,
1982; Stacey & Shupe, 1983; Straus, Gelles, & Steinmetz,
1980; Walker, 1979, 1984; Wetzel & Ross, 1983), from their
discussions and research with battered women, support the
notion that financial dependence does contribute to keeping
the battered woman trapped in the relationship. Of Stacey
and Shupe's (1983) shelter residents who responded to why
they did not leave the situation immediately after the first
abuse incident, 30% suggested their economic dependence.
Pagelow (1981) retested Gelles' (1976) earlier hypothesis
regarding resources, along with the other variables Gelles
looked at (severity of abuse, frequency, and family of
origin child abuse history), and again found that fewer
resources related directly to whether the woman stayed in
the relationship.
Occupational underachievement in the husband or higher
status in occupation by the wife have been found to increase
the risk for spouse abuse to occur (Hornung, McCullough, &
Sagimoto, 1981). Conflicting results regarding occupational
history suggests that employment status does not have a
correlation with whether the abused woman will leave the
relationship (Rounsaville, 1978). A battered wife's
objective marital dependency (no job, minimal income

53
potential, child-rearing role) is significantly related to
severe violence (Kalmuss & Straus, 1982) while her subjec
tive perception of her dependency is not related.
Clearly most research suggests that economic resources
are related to whether the abused woman stays. Thus,
economic resources was one of the variables studied here in
relationship to the battered woman's destination once she
leaves the shelter. The method utilized by Kalmuss and
Straus (1982) was replicated here.
Children
Many researchers have commented on the impact the
children and the parental relationship with the children,
have on the decision making for the battering couple (Brown
& Brazzle, 1982; Carlson, 1976; Gelles, 1983; Martin, 1976;
Snell, Rosenwald & Robey, 1964; Stacey & Shupe, 1983;
Straus, 1973).
Snell, Rosenwald, and Robey (1964) questioned wives who
charged their husbands with assault and battery regarding
their reasons for the action at that particular time, since
the marriages were 12-20 years in length and had included
other abuse incidents. The women in almost all the cases
responded first that the children figured into their
decision. Most common were responses describing the
presence of an adolescent son, getting older, stronger, and
more prone to be either retaliatory towards the father or
negatively affected by the domestic disturbances.

54
Brown and Brazzle's (1982) sample indicated that a
reason for not leaving among women was that they feared that
the husbands would seek the families out and harm the
children or the wives. Twenty-one percent of those in the
Stacey and Shupe study (1983) who did not leave the home
after the first abuse incident revealed the children and
maintaining the family as their reason. The mother's
intense attachment to and concern for her children has been
noted (Carlson, 1977). Children have been cited as both the
reason for leaving and the reason for staying (Cristall,
1978) ; although in this particular study, twice as often
they were the reason for staying.
Where battered women were compared with non-battered
women (Parker & Schumacher, 1977), there was no statistical
difference between the number of children in either family.
Additionally, when looking at battered women who have left
the abuser compared to those who had not (Cristall, 1978),
neither the age nor number of children appeared to be
statistically different in either group. When the ages of
the children are viewed as part of the definition for the
women's resources (younger children being a limitation,
older children being a potential resource because of their
ability to be self-sufficient or assist in supporting the
family), a strong correlation exists between a woman's
limited resources and her remaining in the relationship
(Pagelow, 1981). In another study (Snyder & Sheer, 1981) it
was found no sociodemographic variables relating to the

55
children (e.g. age of children, number of children) as
singular variables were predictive of the woman's destina
tion upon leaving a shelter. In the study reported herein
the researcher incorporated the role of children as a factor
in the battered women's destination choice, but subsumed
into the resource category as has been done by others
(Gelles, 1976; Kalmuss & Straus, 1982; Pagelow, 1981).
Locus of Control. Learned Helplessness
Walker (1979) theorized, based on her interview and
clinical experience with battered women, that women find it
difficult to escape an abusive relationship due to learned
helplessness. In later research she reported on her work to
test this hypothesis (Walker, 1984).
The learned helplessness phenomenon was identified and
most often studied in a laboratory setting with animals and
later with humans (cf. Seligman, 1975; Abramson, Seligman, &
Teasdale, 1978). Since an operational instrument had not
been developed to measure learned helplessness outside the
laboratory, Walker (1984) devised several methods to
indicate whether the learned helplessness phenomenon was
instrumental. A series of questions were combined into
scales, e.g., sexual abuse during childhood, childhood
health, and combined to form a single measure of childhood
contributors to learned helplessness. Also a measure of
learned helplessness within the battering relationship was
formulated by intercorrelating 15 indices from the interview
questionnaire utilized. Walker (1984) noted the reliability

56
of both measures as less than ideal (.57 and .67 respec
tively) and suggested the need for a better measure of this
construct in future studies. In addition, she combined
several state instruments as indices of current state
related to learned helplessness. Included among them was
locus of control (Levenson's Locus of Control Scales). She
conducted path analyses to determine if either childhood
learned helplessness or relationship learned helplessness
was determinant of current state. Both measures appear to
influence current state.
Learned helplessness theory has been linked with
depression because of the perception by those experiencing
it that they have no power or control over events in their
lives (Abramson, Seligman, & Teasdale, 1978; Costello,
1978). A measure of perceived control, i.e., locus of
control, was incorporated by Walker (1984) into her study.
She expected battered women would attribute more control to
external sources than norm groups would, both on scales
measuring powerful others and chance locus of control. She
also anticipated lower scores on internal locus of control
compared to norms. Battered women, both in and out of
battering relationships, did score higher than norm groups
on chance locus of control. Battered women still in
battering relationships, however, did not score higher than
norm groups on the powerful others locus of control scale.
Walker (1984) postulated that a woman still cohabi-
tating with the abuser may not wish to acknowledge the

57
batterer's control in her life, nor the unlikelihood of
changing him or the environment to prevent further abuse.
Battered women did, surprisingly, score significantly higher
than norm groups on internal locus of control, attributing
themselves as having substantial control over what happens
to them. The Levenson Locus of Control Scales were again
utilized in this research.
Authoritarianism and Monitoring
Neither authoritarianism nor monitoring have been
examined previously for their relationship as potential
determinants of battered women's destination. In fact, they
had not been investigated in relationship to woman battering
in general. In combination they represent salient features
of the theory on the roots of violence proposed by A. Miller
(1981, 1983, 1986). An examination of the results of
research on both factors suggested their fruitfulness for
study in this area.
Authoritarianism
As Rokeach (1960) theorized, an individual maintains an
open belief system when the need to know dominates and a
nonthreatening reality exists. A threatening reality causes
the individual to become closed in his or her beliefs, the
need to know diminishes, confusion occurs between the
information and the informant, and identification with
absolute authority takes place. Closing the belief system
distances the individual from his or her own anxiety. One
with a predominantly open belief system is a low dogmatic

58
(LD); a person having a predominantly closed belief system
is a high dogmatic (HD).
High dogmatics indicate greater dependence on absolute
authority, a greater vulnerability to be influenced by
authority, and less discernment of the value of communi
cations independent of the authority who delivers the
communication (Bettinghaus, Miller, & Steinfatt, 1970; C.G.
Kemp, 1963; Lazio & Rosenthal, 1970; Powell, 1962). High
dogmatics are more likely to adhere to a position supported
by authority (McCarthy & Johnson, 1962). Low dogmatics are
not differentially influenced by the authority of a communi
cator (Harvey & Hays, 1972) and they perceive authority
figures more realistically, incorporating both positive and
negative characteristics (C.G. Kemp, 1963). Low dogmatics
are much less influenced by authority than high dogmatics
(Ehrlich & Lee, 1969; Restle, Andrews, & Rokeach, 1964;
Vacchiano, Strauss, & Hochman, 1969).
The close-minded person will avoid changing his or her
environment or accepting new data and ideas (Vacchiano,
Strauss, & Schiffman, 1968). The high dogmatic will focus
on the future while minimizing the past or present (Castle,
1971; Jay, 1969; Rokeach, 1954, 1956; Rokeach & Bonier,
1960; Zurcher, Willis, Ikard, & Dohme, 1967). Rokeach
(1954, 1956) suggested that this is done in order to create
a sense of control regarding the course of life events.
There is a direct relationship between authoritarianism and
anxiety (cf. Byrne, Blaylock, & Goldberg, 1966; Castle,

59
1971; Norman, 1966; Smithers, 1970) and dogmatism as a
defense mechanism (cf. Bernhardson, 1967; Hallenbeck &
Lundstedt, 1966; D. Lee & Erhlich, 1971). In general, the
high dogmatic can be characterized as having difficulty
tolerating frustration and prone to conforming, while the
low dogmatic is more tolerant of ambiguities and less
accepting of traditional beliefs (Vacchiano, 1977) When
information or a situation does not align with the high
dogmatic's beliefs, he or she is threatened and avoidant
(Hunt & Miller, 1968; G.R. Miller & Rokeach, 1968; Pyron,
1966). In general, close-minded individuals will recall
less information when it is inconsistent and positively
evaluate consistent information (Kleck & Wheaton, 1967).
Low dogmatics are more likely to expose themselves to
belief-discrepant information (Donohew, Parker, & McDermott,
1972) .
Monitoring
Self-monitoring, as conceived by M. Snyder (1974), is
observation and control of the self based on the cues one
perceives from the situation. The construct involves two
aspects: astute sensitivity to the expression and presen
tation of others and an ability to use the situational cues
to monitor and present oneself accordingly (M. Snyder,
1974) .
High self-monitoring individuals are skilled at con
trolling and modifying expressiveness and behavior to align
with the situational cues for appropriateness. Low self-

60
monitoring individuals are less able or likely to modify
expressiveness and presentation to others (M. Snyder, 1979).
The expression of a low self-monitoring individual appears
to be more internally guided by his or her own feelings,
attitudes, and experiences (M. Snyder, 1974, 1979) rather
than by situational or interpersonal cues (M. Snyder &
Monson, 1975). Individuals high in self-monitoring notice
and accurately remember information about others more than
low self-monitoring individuals (Berscheid, Graziano,
Monson, & Dermer, 1976). Persons low in self-monitoring
spend less time and effort thinking about the contingencies
of a prospective date's behavior than those high in self
monitoring (Berscheid, Graziano, Monson, & Dermer, 1976).
High self-monitoring persons keenly attend to the subtleties
of behavior and context and utilize the information to infer
intentions and predict behaviors of others (Jones &
Baumeister, 1976; Kulik & Taylor, 1979). High self-monitors
are skilled at accurately judging the intended meaning of
verbal expressions (Mill, 1984). Further, those who are
high in self-monitoring are especially skilled at reading
nonverbal expression and correctly determining underlying
affective experience and emotional states of others (Geizer,
Rarick, & Soldow, 1977; Krauss, Geller, & Olson, 1976).
High self-monitoring individuals have access to very rich
dispositional constructs organized around prototypes of
others; low self-monitors excel at giving rich, informative
descriptions of their own traits and dispositions (M. Snyder

61
& Cantor, 1980) Low self-monitoring individuals focus on
information based on their inner states (M. Snyder & Tanke,
1976). A greater reliance on a partner's behavior as a
guide to one's own behavior is characteristic of high self
monitors (Ickes & Barnes, 1977). High self-monitoring
individuals connect their identities with their external
environment and choose to enter more clearly defined
situations rather than ambiguous ones (M. Snyder &
Gangestad, 1982; Sampson, 1978).

CHAPTER III
METHODOLOGY
This study was designed to distinguish between those
battered women who, following a shelter stay, did not return
to their abusers and those battered women who did return to
their abusers in terms of key variables drawn from previous
research (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow,
1981; Walker 1979, 1984) and the psychodynamic theory of A.
Miller (1981, 1983, 1986). Specifically, the relationship
was studied between battered women's destination and
resources; number of previous separations; internal,
powerful others, and chance locus of control, authori
tarianism; and monitoring. In Chapter III the research
methodology of the investigation is described. The chapter
includes the following: population and sample, the sampling
procedures, the instrumentation, data collection procedures,
and data analysis.
Population and Sample
The population base for this study was battered women
who resided in two battered women's shelters located in
Florida, the Gainesville and the Ft. Myers shelters, and the
three metropolitan Atlanta shelters. The Gainesville
shelter mainly served a 15-county area in north central
Florida. The Ft. Myers shelter mainly served a six-county
62

63
area in southwest Florida. Two of the Atlanta area shelters
were in suburban locations: the Cobb county shelter,
approximately 12 miles northwest of downtown Atlanta, and
the Clayton County shelter, approximately 12 miles south of
downtown Atlanta. These two shelters served mainly resi
dents of their counties plus women and their children from
several adjoining counties in the northwestern portion of
Georgia. The third metro shelter was located in midtown-
Atlanta mainly serving residents of urban Fulton and Dekalb
counties, as well as occasionally women and their children
from the adjoining metro counties. As is common with
battered women's shelters throughout the United States, at
times each of these shelters operated at capacity and placed
women's names on a waiting list for admission. In the case
of a woman facing imminent potential for severe violence
personnel in each of these shelters worked with those in
other shelters for immediate placement and, via law enforce
ment or other avenues, transported the woman to a distant
shelter.
Battered women came to these shelters following
telephone contact with the particular shelter to determine
that they met the criteria for admission. The admissions
criteria for each shelter was the same: the women must have
been recently battered and be in need of safe shelter
without available refuge. Battered women were admitted to
the shelter at any time of the day or night, any day of the
week, depending on when they sought assistance. Many women

64
who sought help from the shelters had exhausted financial
and familial support.
The collection of demographic information allowed for a
comparison of this sample to some demographic information
from other known shelter samples. The demographic charac
teristics of women from these shelters did not differ
significantly from those of shelter samples in other studies
in the southeast or the nation (Bell, 1977; Carlson, 1977;
Labell, 1979; Pagelow, 1977c, 1981; Stacey & Shupe, 1983).
Although battering occurs across socioeconomic classes,
shelter samples are usually more representative of lower and
lower middle class socioeconomic classes; women with greater
resources have other alternatives for safety (Pagelow, 1981;
Stacey & Shupe, 1983). In the study reported herein, basic
demographics including age, race or ethnic background,
number of children, and residential address before entering
shelter (to determine urban, suburban, or rural) were
collected (Appendix A). One of the independent variables,
number of previous separations, was also included on this
portion of the questionnaire.
Many of the women who came to the shelters heard about
the services via an intervening police officer, public
service advertising, a friend or family, or the advertised
battered women's hotlines. Referrals to the shelters could
come from the emergency rooms of the hospitals in the area.
In addition, other referrals were made by the community
crisis hotlines, police, clergy, rape crisis programs,

65
women's health care clinics, and university and medical
communities.
To be included in the study, the women must have been
battered and must have completed 8 years of schooling
according to their response on the shelter intake forms.
The 8 years of schooling was deemed appropriate indication
of a baseline ability to read and understand the instru
ments.
Sampling Procedures
The subjects for this study were a volunteer sample of
battered women who utilized these shelters during the months
of May through December 1987 and met the criteria for the
study. Data were not collected every day at each of these
shelters, but on the days that the data were collected all
women in the shelter, except for those who had already
participated, were asked to participate in the study. The
sample size was 72 women.
During the days data were collected, all potential
subjects admitted to the shelters were given a letter
explaining the study, describing the criteria for inclusion,
and inviting their participation (see Appendix B). The
potential subjects were questioned to determine if they met
the criteria for inclusion and, if so, were they willing to
participate in the study. This procedure was continued
until the pool of individuals who met the inclusion criteria
and who agreed to participate totaled 78 women. Completed
questionnaires and follow-up destinations were obtained on

66
72 of those 78 women; follow-up was not obtained on 6 women.
An additional 6 women who did not participate, either due to
choice or education requirements, and their destinations at
exit were noted. The reasons for their non-participation
were recorded. One woman, although she met the eighth-grade
completion requirement, did not possess adequate reading
skills to answer the questionnaire; she returned to her
abuser. Two women at two different shelters took the
questionnaires to complete, were interrupted in the process
(one for child rearing, the other to communicate with a
social service employee from another agency who was
assisting her) and did not complete the questionnaires or
make follow-up arrangements. Both left their shelters
hastily. One returned to the batterer and the other flew to
California to move in with her relatives. One woman was
incapable of completing the questionnaire due to disorienta
tion and psychological disturbance, and she left the shelter
to be psychiatrically hospitalized. One woman did not
participate because of a language barrier; she was Spanish
speaking and reading. The researcher did not have a Spanish
version prepared for her use. She did not return to the
abuser.
Instruments
In all, four instruments with a combined total of 92
questions were used to gather the information on authori
tarianism, monitoring, resources, and locus of control
(number of previous separations was answered by the subject

67
on the demographics questionnaireAppendix A). This study
was designed to use short to moderate length instruments
administered by a trained administrator so as not to
additionally burden these subjects who were already experi
encing a crisis. Battered women usually enter the shelter
immediately after what Walker (1979, 1984) has referred to
as the acute battering incident. Many battered women have
described how difficult it is to leave the home for an
unknown refuge environment following such an emotionally
charged incident (Martin, 1976; Pagelow, 1981; Walker, 1979,
1984). At that time battered women are fearful of the
abusers, as well as afraid of the unknown, embarrassed to
make the domestic problem public, and unsure of social
support, emotional or otherwise. Knowing that the abuser
might perceive the woman's actions of seeking safety and
assistance for her and the family as escalation in the
current conflict, the woman must muster a great deal of
courage to seek shelter help. Battered women battle their
own self-blame for the domestic difficulties (Walker, 1979)
and often doubt that others will believe they have been
abused (Rounsaville, Lifton & Bieber, 1979). Lengthy
instrumentation asking the women to respond to many self-
statements at this crisis juncture is thought by many
(Dobash & Dobash, 1981, 1981; Schechter, 1982) to be
intrusive and to debilitate crisis efforts aimed at
supporting her decision to seek safety and get help for the
family. With the additional consideration that

68
administering instrumentation to women in crisis is a
sensitive issue, the investigator chose the following four
measures.
Measure of Resources
For this investigation, the measure for resources was
the Resources Index (Appendix C) utilized by Kalmuss and
Straus (1982). It is the sum of scores on three dichotomous
variables: whether the woman is unemployed or not, whether
the woman has children at home of ages 5 or younger or not,
and whether her mate earns 75% or more of the total house
hold income or not. The range for the values on this index
is 0 (low resources) through 3 (high resources). The
internal consistency of the Resources Index in the Kalmuss
and Straus study (1982) was .59, measured by Cronbach's
alpha coefficient of reliability.
Measure of Locus of Control
Locus of control was measured by the Levenson Locus of
Control Scales (Levenson, 1973) (Appendix D). The locus of
control construct describes individuals' causal beliefs
regarding reinforcements which occur to them. They may have
a generalized expectancy that reinforcements are contingent
upon their own behaviors (internal control). They may also
generally believe that reinforcements are contingent upon
forces outside their control, e.g., chance, fate, a deity,
luck, powerful others (external control). The Levenson
Locus of Control Scales are multidimensional, measuring
beliefs regarding personal control (Internal Scale),

69
powerful others (Powerful Others Scale), and fate or chance
(Chance Scale). A person with a high I Scale score would
believe that he or she has great deal of control over his or
her life. A person with a high P Scale score generally
ascribes powerful others as determining outcomes. A person
with a high C Scale score perceives chance or fate as
significantly contributory to events.
The Locus of Control Scales are comprised of 24 items,
8 per scale, although they are presented interspersed to
subjects in one unified series. Responses are noted on a
Likert Scale ranging from strongly disagree (-3) through
strongly agree (+3) assigned to each statement. All of the
items in the Levenson Scales are stated personally so that
the respondent is being measured on his or her perceptions
regarding control, not what he or she believes "people in
general" think about control (Levenson, 1981).
Kuder-Richardson reliabilities for the I Scale range
from .51 to .67, for the P Scale from .72 to .82, for the C
Scale from .73 to .79 (Levenson, 1973, 1974; Wallston,
Wallston, & DeVellis, 1978). Split-half reliabilities,
utilizing the Spearman-Brown formula, are .62 for the I
Scale, .66 for the P Scale, and .64 for the C Scale. Seven-
week test-retest reliabilities are .66, .62, and .73 for the
I, P, and C Scales (Lee, 1976).
The P and C Scales are both related to external locus
of control, so it would be expected that in validity studies
they would correlate to some extent with each other. In

70
fact, studies of their correlation range from .41 to .60
(Caster & Parsons, 1977; Levenson, 1973; Scanlan, 1979;
Wallston, Wallston, & DeVellis, 1978). Only minimal
correlation (-.25 to .19) has occurred between the P and C
factors with the I factor (Caster & Parsons, 1977; Levenson,
1973; Scanlan, 1979; Wallston, Wallston, & DeVellis, 1978).
The Rotter I-E scale is one of the original locus of
control scales and the most often referenced as the standard
in locus of control measurement. Convergent validity is
strongest between the C and I Scales with the Rotter I-E
Scale, ranging from .43 to .56 for the C factor and from
-.15 to -.41 for the I factor (Donovan & O'Leary, 1975;
Hall, Joestring, & Woods, 1977; Levenson, 1972).
There is negligible correlation between the Levenson
Scales and the Marlowe-Crowne Social Desirability Scale.
In the two studies conducted (Levenson, 1972; Wallston,
Wallston, & DeVellis, 1978), the correlations to the
Marlowe-Crowne Social Desirability Scale for the I Scale
were .04 and .09, for the P Scale .04 and .11, and for the
C Scale -.10 and .08.
Borrero-Hernandez (1979) noted several relationships
between the Levenson Scales and other personality variables.
The I Scale is positively related to measures of sociability
on the California Psychological Inventory (CPI). The C
Scale negatively relates to sense of well-being and
responsibility on the CPI and positively relates to guilt

71
proneness on the 16 PF. The P Scale correlates signifi
cantly with suspiciousness on the 16 PF.
Measure of Authoritarianism
Authoritarianism was measured by the Rokeach Dogmatism
Scale, Form E (Rokeach, 1956). This is a 40-item scale (see
Appendix E). Responses can range from "I agree very much"
(+3) through "I disagree very much" (-3) on a Likert-type
Scale. The close-minded or high dogmatic (HD) person
believes that authority is absolute, accepts or rejects
others based on their agreement or disagreement with
authority, and maintains this cognitive system which
protects the individual for anxiety. The low dogmatic (LD)
person does not hold authority as absolute, accepts or
rejects other on different premises than their alignment
or disagreement with authority, and has more tolerance for
ambiguity and the anxiety that can accompany changing
environments (Vacchiano, 1977).
The construct validity of authoritarianism has been
supported in numerous studies. A common core of authori
tarian factors has been noted between the California F Scale
and the Dogmatism Scale (Kerlinger & Rokeach, 1966),
although the Dogmatism Scale has been found to be more
independent of ideological bias than the California F (Warr,
Lee, & Joreskog, 1969). Many studies support the Dogmatism
Scale as a measure of authoritarianism (Barker, 1963;
Costin, 1965; Hanson, 1968, 1970; Plant, 1960; Rokeach,
1967). High dogmatic persons show a greater dependence on

72
authority than do low dogmatic persons (cf. Bettinghaus,
Miller, & Steinfatt, 1970; Kemp, 1963; Powell, 1962; Restle,
Andrews, & Rokeach, 1964; Vidulich & Kaiman, 1961).
Measure of Monitoring
Monitoring in this study was measured by the Snyder
Self-Monitoring Scale (M. Snyder, 1974) (see Appendix F).
It is a 25-item scale which measures whether a person
monitors external cues from others to guide his or her own
self presentation or if he or she presents and acts in
accordance with his or her own affective state (M. Snyder,
1974). Low self-monitoring individuals are more knowledge
able about their own dispositions, attitudes, and traits;
high self-monitoring persons are much less dispositionally
guided or discerning (M. Snyder & Cantor, 1980).
The KR-20 reliability of the Self-Monitoring Scale is
.70; test-retest reliability is .83 (M. Snyder, 1974). The
Self-Monitoring Scale measures a different construct than
those measured by the Marlowe-Crowne Social Desirability
Scale, the MMPI Psychopathic Deviate Scale, the c
(Chameleon) scale of the performance style test, the
Christie and Geis Machiavellianism scale, and the Alpert-
Haber Achievement Anxiety Test (M. Snyder, 1974). Numerous
studies support the validity of the Self-Monitoring Scale
utilizing peer ratings and samples which include college
students, insurance and sales personnel, psychiatric
patients, and stage actors (Caldwell & O'Reilly, 1982;
Gabreyna & Arkin, 1980; Ickes & Barnes, 1977; Lippa, 1978;

73
M. Snyder, 1974; M. Snyder & Monson, 1975; M. Snyder &
Swann, 1976; M. Snyder & Tanke, 1976).
Data Collection Procedures
Data were collected by either a trained, full-time
staff person from a shelter, the researcher, or a research
associate. At the Ft. Myers, Gainesville, and Cobb shelters
data were collected by a trained full-time staff person. At
the Clayton shelter and during the latter months of collec
tion at the Cobb shelter, the data were collected by the
researcher. At the downtown metro Atlanta shelter the data
were collected by either the researcher or the trained
research associate. The investigator met individually with
each of the selected staff persons and the research associ
ate to train them in the procedures. Training involved the
following components: (a) use of the letter of introduction
to the subject inviting participation, (b) determination that
potential subject meets criteria for inclusion, (c) obtaining
informed consent, (d) benefits/risks of participation,
(e) methods to assure confidentiality, (f) instruments and
instructions, (g) answering questions during administration,
(h) procedures for follow-up contact, and (i) recording data
on the master chart. Each of the individual sessions was
2 to 3 hours long. To determine that each administrator
understood and could uniformly perform data collection,
the investigator conducted a follow-up session after the
participation of at least the first two subjects. The
investigator monitored the data collection throughout the

74
period by contact with the trained administrators and the
shelter staffs and the review of the data as they were being
collected. Additionally, the administrators collecting the
data could call the investigator at any time.
Battered women were admitted to these shelters during
any of the 24 hours in a day. Given up to 48 hours to
emotionally adjust following the acute battering incident,
each woman met with a staff member to complete an intake
questionnaire and participate in an initial intake inter
view. Completing the questionnaire was handled as Part II
of the shelter's intake procedure.
Following completion of Part I of the intake procedure,
which was the shelter's standard intake, the administrator
had the potential subject read the letter of introduction
(Appendix B). The administrator discussed requisites for
participation in the study and potential benefits and risks
of participation. Benefits of participation included the
following: (a) The subject contributes to research which
increases the understanding of the dynamics of battering and
adds to theory which can help end conjugal violence against
women; (b) the subject assists the participating shelter(s)
in increasing the knowledge about their clients so they can
provide more effective services to battered women and their
children; (c) the subject aids those helping battered women,
shelters, legal, criminal, and other social service agencies,
to better understand their clients and to dispel stereotypes
and myths about battered women; and (d) the subject assists

75
in providing information to battered women which increases
self-understanding and decreases feelings of isolation,
guilt, and blame. Risks and/or drawbacks of participation
were the time required to complete the questionnaire, the
adherence to the shelter's follow-up procedure following
departure, responding to questions which solicit personal
opinions when the subject was in a post-traumatic time
period, and reliance on the shelter and principal investi
gator to assure confidentiality. The administrator recorded
on the master chart the subject number, date of administra
tion, and whether the woman qualified and agreed to parti
cipate in the study. If the potential subject was not
participating, the administrator recorded whether this was
due to not meeting the requisites or a refusal to partici
pate.
Each woman volunteering to participate in the study was
given a packet which included instructions for the instru
ments, the instruments themselves, one informed consent form,
and a manila envelope for return of the packet (see Appendix
G). At this time the subjects completed the demographic
questionnaire and the paper and pencil instruments required
for this investigation. The demographic questionnaire
contained 6 items and took a few moments to complete. The
combined instruments contained 92 items total and took from
45 to 90 minutes to complete.
The administrator was present with the subject through
out data collection and trained to answer any questions which

76
arose regarding individual items or the study. The
administrator was trained to do this in a way which main
tained the integrity of the instrumentation and purpose of
the investigation, yet assisted the woman to answer the
questionnaire adequately. The administrator collected the
completed questionnaire and recorded on the master chart that
the following had been totally completed: the informed
consent form, the 6 demographic items, the 3 resources items,
the 24 locus of control items, the 40 dogmatism items, and
the 25 monitoring items.
Subjects' names were not placed on the instruments to
ensure confidentiality. Each packet was given a code number
which was paired with the respondent's name on a separate
listing.
The trained administrator, following collection of the
completed questionnaire, discussed and then recorded the
preferred and contingent options for follow-up contact. The
woman could select options from the following methods:
(a) telephoning the shelter during a prearranged time,
(b) returning to the shelter for a face-to-face contact
(often women return to the shelter to pick up mail, including
aid checks, which are addressed to the women at the shelter),
(c) communicating through a third party who agreed to provide
the necessary information about the woman to the shelter,
(d) returning a stamped post card addressed to the shelter's
post office box with a question to indicate destination on
it, (e) responding to a follow-up letter mailed to the woman

77
in a plain envelope including a pre-addressed, postaged reply
form, and (f) agreeing to receive a telephone call from the
shelter.
During the shelter stay, the women participated in peer
counseling sessions, met with staff for individual help, and
established contact with other social service agencies that
provided a more permanent source of food, housing, and
transportation when needed. If the women came to the shelter
with children, the children also participated in appropriate
services, e.g., play therapy. Following a variable length of
stay, each woman made the decision regarding her destination
following her temporary refuge at the shelter. All shelters
had been required by their funding agencies to provide up to
30 days of safe refuge. On occasion, a woman will be
planning her own housing away from the abuser and the housing
would not be ready for her immediately at the end of the 30
days. In those exceptional cases the shelters granted exten
sions. Each woman met with a staff member to discuss her
departure plans.
At the appropriate time, the trained administrator, in
concert with program staff, initiated the follow-up procedure
and recorded the destination as known during the follow-up
period (i.e., from 2 to 6 weeks following departure) on the
master chart. Working with a battered woman when she leaves
a shelter program and returns to live with the abuser
necessitates certain precautions. While the battered woman
is at the shelter, attempts are made to keep the shelter

78
location a secret and the abuser uninformed about the woman's
whereabouts. These efforts are taken to increase the
possibility of providing a safe place for women and their
children. Increased police surveillance is some- times
requested by shelters when the dangerousness and lethality of
the abuser is judged to be particularly great or the batterer
issues threats. Shelter personnel do not release names of
women or children housed there. When the woman returns to
the abuser, contact from the shelter staff needs to be
discreet to continue efforts of safety provision for both the
battered woman and the shelter program.
Data Analysis
The researcher posed the following hypotheses.
1. There is no difference in authoritarianism between
battered women who do not return to the abuser following a
shelter stay and those who do return.
2. There is no difference in monitoring between
battered women who do not return to the abuser following a
shelter stay and those who do return.
3. There is no difference in resources between battered
women who do not return to the abuser following a shelter
stay and those who do return.
4. There is no difference in number of previous
separations between battered women who do not return to the
abuser following a shelter stay and those who do return.
5. There is no difference in locus of control (inter
nal, powerful others, chance) between battered women who do

79
not return to the abuser following a shelter stay and those
who do return.
6. There is no relationship between authoritarianism
monitoring, resources, number of previous separations, and
locus of control (internal, powerful others, chance) and the
destination of battered woman following a shelter stay.
Hypotheses one through five were analyzed using one-way
analyses of variance. Hypothesis six was analyzed using
logistic regression. An alpha level for significance was
established at .05 as a conventional level of significance.
Initially, a Bartlett's test was utilized to determine
departures from normality. The results of this test sug
gested that the sample was not within the range of a normal
distribution; therefore, although either discriminant
analysis or logistic regression was considered for analysis,
logistic regression was chosen since it is less sensitive to
departures from normality.
Using logistic regression analysis the researcher can
distinguish between two groups and produce a predictive
equation. In logistic regression analysis the variables are
weighed and linearly combined the variables mathematically to
force the optimum distinction between groups.
Logistic regression in a two-group case is the same as a
multiple regression where the dependent variable (in this
study not returning to the abuser or returning to the
abuser) takes on the values of 1 and 0 and maximally discrim
inates between the two groups (Kerlinger & Pedhazer, 1973).

CHAPTER IV
RESULTS
The results are presented in this chapter. Descriptive
statistics were used to characterize the sample group while
the research hypotheses were analyzed with one-way analysis
of variance (hypothesis one through five) and logistic
regression (hypothesis six).
Demographic Information
Demographic information is presented here in order to
characterize the overall sample of battered women, as well
as subsets of the sample: those who did not return to the
batterers and those who did return following the stays at
the shelters. The total sample included 72 women. Those
who did not return to the batterers included 42 women; those
who did return to the abusers numbered 30. In addition to
the variable data, general demographic information was
gathered on each subject: age of subject, race or ethnic
background of subject, number of children, whether subject
had left mate before or not, and city of residence before
coming to shelter.
A breakdown of the study participants by age appears in
Table 4-1. The sample was predominantly in their 20s and
30s (86%), with the mean age being 30.32 years (SD = 7.32);
this is quite similar to shelter samples from other
80

81
Table 4-1
Age of Sample
Age
Returned
to Abuser
n (%)
Did not
Return
n (%)
Total
Samle
n (%)
17-20
1
(3.3)
3
(7.1)
4
(5.6)
21-25
6
(20.0)
9
(21.4)
15
(20.8)
26-30
12
(40.0)
12
(28.6)
24
(33.3)
31-35
2
(6.7)
7
(16.7)
9
(12.5)
36-40
6
(20.0)
8
(19.0)
14
(19.4)
41-45
2
(6.7)
3
(7.1)
5
(6.9)
46-50
0
0
0
51-55
1
(3.3)
0
1
(1.4)
research (see Appendix H). Mean age for battered women who
returned to their abusers was 30.80 years (SD = 7.67); mean
age for those who did not was 29.98 years (SD = 7.14).
A breakdown of the study participants by race appears
in Table 4-2. The sample had a large percentage of non
whites (41.7%) which is common for samples including inner
city shelters (e.g., D.K. Snyder & Fruchtman, 1981; D.K.
Snyder & Scheer, 1981) (see Appendix H). Because there
appeared to be an effect due to race, a chi-square analysis
was conducted. With a 2 = 16.86, p = 0.000, which is
significant for a = .05 and d.f. = 1, an effect due to race

82
Table 4-2
Race of Sample
Returned
Did Not
Total
Race
to Abuser
Return
Samle
n (%)
n (%)
n (%)
White
26 (86.7)
16 (38.1)
42 (58.3)
Nonwhite
4 (13.3)
26 (61.9)
30 (41.7)
was found i.e., white participants returned to their abusers
more often than their nonwhite counterparts. This is
discussed in Chapter 5.
The participants' numbers of children are categorized
in Table 4-3. The average number of children for the total
sample was 1.74 (SD = 1.30). Battered women who returned to
the abusers in the study had an average of 1.63 children
(SD = 1.45); average number of children for the abused women
who did not return to the batterers was 1.81 (SD = 1.19).
A breakdown of the study participants by whether they
had left their mates before or not is presented in Table 4-
4. Nearly equal percentages of participants had left their
mates before regardless of destination in this study.

83
Table 4-3
Number of Children
Number
of
Children
Returned
to Abuser
n (%)
Did Not
Return
n (%)
Total
Sample
n (%)
0
8
(26.7)
7
(16.7)
15
(20.8)
1
7
(23.3)
10
(23.8)
17
(23.6)
2
7
(23.3)
12
(26.6)
19
(26.4)
3
6
(20.0)
10
(23.8)
16
(22.2)
4
1
(3.3)
3
(7.1)
4
(5.6)
5
0
0
0
6
1
(3.3)
0
1
(1.4)
Table 4-4
Previous
Separations
Previous
Returned
Did Not
Total
Separa-
to Abuser
Return
Sample
tions
n (%)
n (%)
n (%)
No
9 (30.0)
9 (21.4)
18 (25.0)
Yes
21 (70.0)
33 (78.6)
54 (75.0)

84
Population size of the city in which the study partici
pants lived before coining to the shelter is categorized in
Table 4-5. An effect related to population appeared likely,
so a chi-square analysis was conducted. A 2 = 10.03, p =
0.018, signifies an effect for a = .05 and d.f. = 3; bat
tered women from the small towns were more likely to return
to the abuser and as the population of residence increased,
battered women were less likely to return to their mates.
This result is considered further in Chapter 5.
The Rokeach Dogmatism Scale to measure authori
tarianism, Snyder Self-Monitoring Scale was used to measure
monitoring, the Kalmuss & Straus (1982) 3-item method to
measure resources, a straightforward item on the
Table 4-5
Population of City in Which Subject Resided Before Shelter
Returned
Population to Abuser
Size n (%)
Did Not Total
Return Sample
n (%) n (%)
0-20,000
21,000-
50,000
51,000-
200,000
17 (56.7)
6 (20.0)
4 (13.3)
over 200,000 3 (10.0)
10 (23.8)
9 (21.4)
8 (19.0)
15 (35.7)
27 (37.5)
15 (20.8)
12 (16.7)
18 (42.9)

85
questionnaire to determine number of previous separations,
and the Levenson Locus of Control Scales to measure internal
(I), powerful others (P), and chance (C) loci of control.
The means and standard deviations for the total sample and
the subsets according to destination can be found in Table
4-6. Because a potential effect was likely related to the
75%-or-more-of-income-earned-by-mate item (one of the three
from the resource measure) and destination, a chi-square
analysis was conducted. Since 2 = 4.46, p = 0.035, which
is significant for a = .05 and d.f. = 1, it was found that
women whose abusers earn 75% or more of the family income
returned to their mates more often than would be expected.
Race was not a significant factor impacting this relation
ship.
Analysis of Hypotheses One Through Five
The first five hypotheses were tested using one-way
ANOVA tests. The first hypothesis was that there would be
no difference in the authoritarianism scores between
battered women who, following shelter stays, did not return
to their abusers and battered women who did return. The
one-way ANOVA was not significant at the .05 level (F =
2.44, p = 0.123); therefore, the null hypothesis was not
rejected.
The second hypothesis was that there would be no
difference in the monitoring scores between battered women
who, following shelter stays, did not return to their
batterers and abused women who did return. The one-way

Table 4-6
Means and Standard Deviations of Study Measures
Measures
Returned
to Abuser
Did Not Return
Total Samle
M
SD
M
SD
M
SD
Authoritarianism
152.72
23.20
163.60
32.88
159.08
29.64
Monitoring
8.50
3.75
9.25
3.75
9.00
3.71
Resources
1.17
0.99
1.38
0.85
1.29
0.91
Not Employed
0.40
0.50
0.40
0.50
0.40
0.49
Children 5 or
Younger at Home
0.40
0.50
0.36
0.48
0.38
0.49
75% or More Income
Earned by Mate
0.37
0.49
0.62
0.49
0.51
0.50
Number of Previous
Separations
2.33
2.11
3.17
3.70
2.82
3.15
Internal Locus
of Control
35.63
5.64
36.02
6.81
35.28
6.31
Powerful Others
Locus of Control
16.07
9.40
17.69
10.53
17.43
10.01
Chance Locus
of Control
20.60
9.36
16.43
9.36
18.17
9.52

87
ANOVA was not significant at the .05 level (F = 0.93,
p = 0.338); therefore, the null hypothesis was not
rejected.
The third hypothesis was that there would be no
difference in the resources scores between battered women
who, following shelter stays, did not return to their
abusive mates and those battered women who did return to
their batterers. The one-way ANOVA was not significant at
the .05 level (F = 0.97, p = 0.328); therefore, the null
hypothesis was not rejected.
The fourth hypothesis was that there would be no
difference in the number of previous separations between
battered women who did not return to their abusers after the
shelter stays and those women who did return to the abusive
mates. The one-way ANOVA was not significant at the .05
level (F = 1.23, p = 0.271); therefore, the null hypothesis
was not rejected.
The fifth hypothesis was that there would be no differ
ence in internal, powerful others, and chance locus of
control scores between sheltered battered women who did not
return to their batterers and those who did return. Each of
the Levenson Locus of Control Scales is independent of the
other. The one-way ANOVA for internal locus of control
scores was not significant at the .05 level (F = 0.16, p =
0.689). The one-way ANOVA for powerful others locus of con
trol scores was not significant at the .05 level (F = 0.07,
p = 0.796). The one-way ANOVA for chance locus of control

88
scores was not significant for at the .05 level (F = 3.48,
p = 0.066). Since ANOVAs were not significant for all locus
of control scores, the null hypothesis was not rejected.
Analysis of Hypothesis Six
The sixth hypothesis was tested using logistic regres
sion. The sixth hypothesis was that there would be no
relationship between authoritarianism, monitoring,
resources, number of previous separations, locus of control
(internal, powerful others, and chance), and the destination
of battered women following shelter stays. The logistic
regression analysis is presented in Table 4-7. The regres
sion analysis indicated that the variables, in general, were
not good predictors of destination following shelter stay.
Therefore, the null hypothesis was not rejected.
Summary
No significant differences between the groups occurred
on the hypothesized variables of authoritarianism, moni
toring, resources, number of previous separations, internal
locus of control, powerful others locus of control, and
chance locus of control. Some distinctive differences were
indicated in this sample between battered women who returned
to their abusers and battered women who did not, specifi
cally in the demographic data of race, population of city
where subject resided, and husband's percentage of income.
No significant relationship was established between these
variables and the destination of battered women following a
shelter experience.

Table 4-7
Logistic Regression Model of the Relationship Between Destination and the Research
Variables
Parameter
Estimate
Standard Error
Chi-Square
Probability
Intercept
0.200
0.811
0.06
0.8053
Authoritarianism
-0.332
0.350
0.90
0.3431
Monitoring
-0.587
1.498
0.45
0.6948
Resources
-0.062
0.234
0.07
0.7896
Number of Previous
Separations
-0.023
0.068
0.12
0.7309
Internal Locus of
Control
0.020
0.035
0.32
0.5721
Powerful Others
Locus of Control
-0.005
0.025
0.04
0.8475
Chance Locus of
Control
0.037
0.028
1.75
0.1858

CHAPTER V
DISCUSSION
Summary
The aim in this study was to determine key variables
which would distinguish between battered women who did not
return to abusive mates and battered women who did following
their stays at refuge shelters. The variables, drawn from
previous research (Gelles, 1976; Kalmuss & Straus, 1982;
Pagelow, 1981; Walker, 1979, 1984) and from the psycho
dynamic theory of A. Miller (1981, 1983, 1986), were
resources, number of previous separations, internal locus of
control, powerful others locus of control, chance locus of
control, authoritarianism, and monitoring.
A total of 72 battered women who stayed at shelters in
Florida and Georgia agreed to participate in the study and
completed questionnaires while they were at the shelters.
Destination information, i.e., whether they were living with
the abuser or not, was recorded during the follow-up period
of 2 to 6 weeks after their departures from the shelters.
The demographic information from the questionnaires
indicated that battered women who did not return to their
abusers and battered women who did were similar to each
other in several ways. Most of the women were in their 20s
and 30s, had probably left their mates before, and either
90

91
had no children or, if they had children, had less than
four. Of the 41.7% of the sample who were nonwhites, 86.7%
did not return to their abusers, and this was equally
distributed across shelters and city sizes. Women in this
sample who were from small towns were more likely to return
to their mates (63% returned from towns of 20,000 population
or less) and women who were from larger cities were more
likely not to return to their mates (66% did not return from
cities of 51,000 to 200,000 population and 83.3% did not
return from cities over 200,000 population).
No significant differences (p < .05) were found between
abused women who did not return to their batterers and
abused women who did on any of the variables identified. No
significant relationship was found between all of the
variables and destination.
Limitations of the Study
Several limiting factors impact interpretation of
results from this study. First, to what populations are the
results generalizable? Using five shelters increased the
possibility that the study could be generalized to
sheltered, battered women. Regional or other
characteristics of this sample may reduce generalizability,
however. Additionally, this sample in most cases did not
include women who entered and departed the shelter very
quickly, i.e., within the first 48 hours. A very high
percentage of those women return to their abusive
households, so it is with caution that these results can be

92
said to reflect that unique group of battered women using
shelters. It is possible that those women could be
characterized quite differently from even those participants
in this study who returned to batterers after the first 48
hours. Including responses from the "quick exit" women
might have impacted findings as well as generalizability.
The volunteer nature of the sample and the self-report, par
ticularly for destination, may also present limitations.
Secondly, destination of battered women is assumed, in
this study, to reflect two divergent realities: living with
an abuser or not living with an abuser. Some women may go
back to spouses and, although rare, not receive abuse again.
And some women who do not return to their mates may select,
again though rare, to form relationships with other abusive
partners. In fact, this research was not designed to
include differing typologies of women within either of the
destination categories.
A third limitation exists in the operational definition
for destination in this study being 2 to 6 weeks following
the shelter experience. Some women categorized as not
returning to their abusers by this definition will indeed
move back in with their mates sometime later.
The researcher selected the time while battered women are at
shelters as an appropriate time to perceive psychological
differences between the two groups. It could be that the
differences between the two groups would be more profound
during other times, e.g., 6 months away from abuser or while

93
currently living with abuser. Although selecting another
point in time for administration of the questionnaire may be
as helpful to pinpoint the relevance of the specific
theories being applied here to the whole problem of woman
abuse, it would not have suited the purpose of establishing
determinants of destination following shelter stay.
Discussion of Results
What might be suggested by the lack of a significant
relationship between the identified variables and battered
women's destinations? Several ideas are presented.
Variables which were drawn from previous studies,
although not couched in a theoretical framework, were
resources and number of previous separations. In this study
even though the overall resources measure did not
distinguish between the two groups, one item from the
measure, whether the mate earned 75% or more of the family
income, did have a significant relationship with
destination. Battered women in this study whose abusers
earned 75% or more of the family income were likely to
return to cohabitation with the batterers. A plethora of
questions across studies indicate researchers' interest in
assessing the economic resource picture which the battered
woman faces and how this impacts her ability to free herself
from a battering relationship. The findings from this
study, and those from others, suggest that understanding
economic impact in a consistent, pertinent way has yet to be
determined.

94
Number of previous separations was not a distinguishing
variable between the groups in this study. This contrasts
with the findings by D.K. Snyder and Scheer (1981) that
previous separations were a predictor of destination. It is
possible that their findings are not generalizable since
different findings occurred here, or that they have opera
tionalized previous separations differently than this
researcher. Okun (1986) and Walker (1979) expressed beliefs
that repeated separations were likely before the woman would
finally terminate the relationship. This may be so, but the
quality and nature of these separations, preceding and
subsequent events, may be so unique, that a simple quantita
tive tally will rarely be predictive. Data gathering on
frequency and severity of abuse have been partial attempts
to define the complex context of the battering relationship
and the separations; researchers, through simple quantita
tive analysis of these factors, have failed to distinguish
between women with abusers or women who have ended battering
relationships (Pagelow, 1977c, 1980, 1981; D.K. Snyder &
Scheer, 1981).
The locus of control variables were included to inves
tigate a component of the "learned helplessness" model
applied to battered women by Walker (1979). Her research
(1984) included, "contrary to expectations" (p. 80), the
findings that battered women saw themselves as having a
great deal of control over their lives, more so than how
women in general viewed their internal control.

95
Additionally, the battered women in her sample who were
living with abusers did not believe, to a significantly
greater extent, that powerful others exercised a lot of
control over them compared to the beliefs of a norm group of
women or battered women who were no longer with batterers.
If learned helplessness were operant, one would expect that
battered women who returned from shelters to abusive
partners would have significantly less internal locus of
control and significantly greater powerful others locus of
control than battered women who did not return to batterers.
Findings from this investigation were not significant in
distinguishing between the two groups. The means and
standard deviations from this study can be compared to
results from the Walker study and Levenson's norms in
Appendix I.
The relationship of A. Miller's (1981, 1983, 1986)
psychodynamic theory to battered women's destination was
accomplished by utilizing measures of authoritarianism and
monitoring. If Miller's theory, as operationalized here,
were applicable, the expectations would be twofold. First,
battered women who did not return to their abusers would
have received lower scores on the Rokeach Dogmatism Scale as
evidence that they had shed their idealized view of author
ity (i.e., the male head of household, the marriage institu
tion) Secondly, battered women who did not return to
batterers would be more likely to monitor their own feelings
and needs and monitor others much less when compared to

96
battered women who subsequently returned to batterers, and
so the scores on the Snyder Self-Monitoring Scale would be
lower for battered women who did not return. The results
were not as expected; battered women from the two destina
tions did not score significantly differently on either
measure.
Several plausible explanations are available to illum
inate the lack of significant findings. In this study
battered women who did return to their mates were slightly
less dogmatic than battered women who didn't. However, the
mean scores for both groups would not suggest greater or
lesser dogmatism when compared to a normative sample of
female college students (see Appendix I). It is possible
that battered women, in general, are no more or less rigid
and devoted to authority than other groups in society. This
is compatible with A. Miller's (1983, 1986) beliefs that
most people in today's world are socialized into an extreme
acceptance of parents and authority. So it might be rare to
expect significantly low dogmatism scores from any sizable
percentage of any group in society. Although this resear
cher disagrees with A. Miller's (1986) suggestion that one
would need two in-depth psychoanalytic analyses to "crack"
this idealization of authority, it may be ambitious to
expect that battered women at shelters have come to under
stand the great imperfections of authoritarian rule.
Society has colluded historically in maintaining the near
infallibility of authority in one form or another.

97
Battered women who come to shelters and establish
themselves independently from their abusers may be trans
ferring their allegiances from one form of authority to
another. Loyalty to the authority of marriage and the
husband may become loyalty to the authority of the shelter
program, the social service network, and/or, at times,
family and friends, who encourage separation from and
dissolution of the abusive conjugal relationship. Thus,
separation from the abuser may not be concomitant with any
lessening of a rigid, authoritarian belief system.
The battered women with differing destinations after
shelter experience could not be distinguished based on their
monitoring of others and mediating of their own behavior
according to external cues. In fact, battered women
appeared to have very similar levels of monitoring compared
to a female norm sample (see Appendix I). A. Miller's
theory specifically addresses the excessive need to achieve
and perform for others, conditioned and reinforced in
childhood as the right thing to do, at the cost of under
development of one's abilities to perceive one's own needs
and feelings. The self monitoring scores suggest that
battered women in either group had no less orientation
towards the monitoring of others. This instrument may be
inadeguate, however, in measuring any increase in self-
directedness, self empathy, monitoring of the self, and
mediating one's behavior due to internal cues. It was the
researcher's assumption that an increased internal moni-

98
toring would be evidenced by a decreased external moni
toring. In fact, women who monitor others and have
developed the skill to do so may not do so less in order to
add the monitoring of selves more. Therefore, this instru
ment may have been inadequate to gauge the increase in
internal monitoring and whether it distinguishes between the
two groups of battered women.
Two demographic characteristics did distinguish between
abused women not returning to the batterers and abused women
who did. Why have nonwhite battered women not returned to
their abusers more often than their white counterparts?
Perhaps there is greater support from the nonwhite woman's
family of origin and community to leave an abusive relation
ship; that support could be manifested in housing or
economic resources, emotional caring, or less rigid sanc
tions regarding marital dissolution. This finding is
contrary to other studies (Pagelow, 1981; D.K. Snyder &
Scheer, 1981) and requires further investigation.
The other demographic factor of note was related to
population size where the battered woman resided before
coming to the shelter. Battered women from the small towns
may have been more likely to return to the abusers for
several reasons: (a) The mores of small towns are more
conventional and therefore reinforce families staying
together regardless of the conditions. (b) Less safety,
anonymity, and other living options are available in small
towns to protect a battered woman who leaves. (c) Less

99
income and employment opportunities may be available to the
woman to assist her in supporting herself and a family.
(d) The woman may perceive fewer options for other meaning
ful intimate relationships. In addition, battered women in
urban settings have greater access to subsidized housing or
additional sheltering, other women who are in the same or
similar situations, and urban amenities such as more pre
valent mass transit and child care to assist them in gaining
autonomy. Little previous research has been done investi
gating this aspect.
The findings from this research support the beliefs
(Dobash & Dobash, 1979; Hilberman, 1980; Hilberman & Munson,
1977-78; Martin, 1976; Schechter, 1982) that the person
alities of women who are in or remain in abusive relation
ships are not somehow abnormal or different than other
women. The scores on authoritarianism, locus of control,
and monitoring, three personality measures, did not distin
guish between battered women who (potentially) were ending
their abusive relationships and those who were continuing in
them. Nor did their scores appear to be different from
sample comparison norms (Appendix I).
Recommendations for Future Research
The results of this investigation have implications for
future research. Given both the limitations and factors
cited during the discussion, all presented in this chapter,
the following suggestions are made to guide future study in
this area:

100
1. Tracking battered women's location at later
checkpoints would enhance the accuracy of determining who
maintained their independence. Also, follow-ups in which
information is gathered regarding the women's current
relationship will help to determine if she is still in an
abusive relationship.
2. Specifically develop instrumentation to operation
alize A. Miller's theory more closely. This instrumentation
could then be utilized in any number of areas of research.
What was obvious to this investigator is that there was
little available option for measuring one's ability and
frequency of self monitoring and self empathy.
3. Conduct further investigations related to city
size and race to further understand their impact or connec
tion to destination.
4. A standard measure or measures to assess resources
is needed. The economic factor is judged subjectively to be
important in battering. But researchers using various,
nonstandard methods, without adequate analysis of relia
bility and validity, fail to prove or disprove the impor
tance of this factor.
5. Although neither learned helplessness nor A.
Miller's psychodynamic theory provided assistance in
distinguishing battered women related to destination,
continued study of both with battered women and the phenome
non of battering is recommended. This was the initial
attempt to apply A. Miller's theory to this area, and to

101
abandon its potential relevance at this point would be
premature.
6. More investigation regarding battered women's
perceptions of authority is warranted. Battered women may
fluctuate between idealizing authority sometimes and, at
others, serving the needs of those in authority while
secretly feeling derisive. This would be consistent with A.
Miller's earliest work (1981) where she addressed the
borderline personality in detail.
7. It is highly possible that certain typologies of
battered women exist as subsets of the destination cate
gories. Of the many studies done related to domestic vio
lence, in only one (D.K. Snyder & Fruchtman, 1981) were
battered women considered heterogenous, not homogenous.
Looking at characteristics of heterogeneity could greatly
increase the potential of actually predicting or determining
destination or when a woman will take abuse no more.

APPENDIX A
DEMOGRAPHIC INFORMATION QUESTIONNAIRE

DEMOGRAPHIC INFORMATION QUESTIONNAIRE
1. CODE NUMBER: (assigned by staff)
2. Please list your age:
3A. Please list your race/ethnic background using the codes
below:
A Asian N Native American
B Black W White
H Hispanic 0 Other
3B. If you listed "O" for other in 3A, please specify:
4A. Number of children in your household:
4B. List gender and ages of children: (M/F) (Years Old)
Use other side if additional space is needed.
5A. Have your left your mate before: YES NO
5B. If yes, number of times:
6. City in which you resided before coming to shelter:
103

APPENDIX B
LETTER OF INTRODUCTION

Dear Shelter Resident,
I would like to request your participation in a
research study I am conducting. I am a doctoral student in
the Counseling Psychology program at the University of
Florida in Gainesville, Florida. I am conducting a study on
women who stay at battered women's shelters. Specifically,
my purpose is to examine the following factors: level of
resources, beliefs about control and authority, and moni
toring skills through a questionnaire.
I am attempting to locate women who have been battered
by their mates (husband, boyfriend, ex-husband), are staying
at a battered women's shelter, and have lived with the mate
before coming to the shelter. In order to participate,
women must have completed an eighth grade education.
Participation in this study involves two things:
1) Completion of a questionnaire containing 92 items during
your stay at the shelter, and 2) participation in the
shelter's follow-up procedure at six weeks after you leave
the shelter. All responses are anonymous with each indi
vidual receiving a code number.
As much as I would like to compensate each woman for
her participation, the most I can offer is my gratitude for
helping in a study that I hope will contribute to the safety
of battered women. To those who are interested, a copy of
the results of the study will be available through your
shelter.
A designated shelter staff member is available to
answer any questions you might have and to find out if you
are interested in participating.
Thank you for taking the time to read this letter. I
hope that you will agree to participate. I am eager to have
your involvement in this study.
Sincerely,
Joanne F. DeMark
105

APPENDIX C
MEASURE OF RESOURCES

Measure of Resources
1.
Are you currently employed?
No
Yes
2.
Do you have children of age five or
younger at home?
No
Yes
3.
What percentage of the total family
income is earned by your mate?
%
107

APPENDIX D
LEVENSON LOCUS OF CONTROL SCALES

Levenson Locus of Control Scales
Directions; Below is a series of attitude statements. Each
represents a commonly held opinion and there are no right or
wrong answers. You will probably disagree with some items
and agree with others. We are interested in the extent to
which you agree or disagree with such matters of opinion.
Read each statement carefully. Then indicate the
extent to which you agree or disagree by placing a number in
the left margin next to each statement. The numbers and
their meaning are indicated below:
+1 : I agree slightly -1 : I disagree slightly
+2 : I agree somewhat -2 : I disagree somewhat
+3 : I agree strongly -3 : I disagree strongly
First impressions are usually best in such matters.
Read each statement, decide if you agree or disagree and the
strength of your opinion, and then place a number in the
space provided. Give vour opinion on every statement.
If you find that the numbers to be used in answering do
not adequately indicate your own opinion, use the one which
is closest to the way you feel. Your responses will be kept
confidential.
1. Whether or not I get to be a leader depends mostly
on my ability.
2. To a great extent my life is controlled by
accidental happenings.
3. I feel like what happens in my life is mostly
determined by powerful people.
4. Whether or not I get into a car accident depends
mostly on how good a driver I am.
5. When I make plans, I am almost certain to make
them work.
6. Often there is no chance of protecting my personal
interest from bad luck happenings.
7. When I get what I want, it's usually because I'm
lucky.
109

110
8. Although I might have good ability, I will not be
given leadership responsibility without appealing
to those in positions of power.
9. How many friends I have depends on how nice a
person I am.
10. I have often found that what is going to happen
will happen.
11. My life is chiefly controlled by powerful others.
12. Whether or not I get into a car accident is mostly
a matter of luck.
13. People like myself have very little chance of
protecting our personal interests when they
conflict with those of strong pressure groups.
14. It's not always wise for me to plan too far ahead
because many things turn out to be a matter of
good or bad fortune.
15. Getting what I want requires pleasing those people
above me.
16. Whether or not I get to be a leader depends on
whether I'm lucky enough to be in the right place
at the right time.
17. If important people were to decide they didn't
like me, I probably wouldn't make many friends.
18. I can pretty much determine what will happen in my
life.
19. I am usually able to protect my personal inter
ests .
20. Whether or not I get into a car accident depends
mostly on the other driver.
21. When I get what I want, it's usually because I
worked hard for it.
22. In order to have my plans work, I make sure that
they fit in with the desires of people who have
power over me.
23. My life is determined by my own actions.
It's chiefly a matter of fate whether or not I
have a few friends or many friends.
24.

APPENDIX E
DOGMATISM SCALE (FORM E)

Dogmatism Scale (Form E)
Directions; The following is a study of what the general
public thinks and feels about a number of important social
and personal questions. The best answer to each statement
below is your personal opinion. We have tried to cover many
different and opposing points of view; you may find yourself
agreeing strongly with some of the statements, disagreeing
just as strongly with others, and perhaps uncertain about
others; whether you agree or disagree with any statement,
you can be sure that many people feel the same as you do.
Mark each statement in the left margin according to how
much you agree or disagree with it. Please mark every one.
Write +1, +2, +3, or -1, -2, -3, depending on how you feel
in each case.
+1 : I agree a little -1 : I disagree a little
+2 : I agree on the whole -2 : I disagree on the whole
+3 : I agree very much -3 : I disagree very much
1. The United States and Russia have just about
nothing in common.
2. The highest form of government is a democracy and
the highest form of democracy is a government run
by those who are most intelligent.
3. Even though freedom of speech for all groups is a
worthwhile goal, it is unfortunately necessary to
restrict the freedom of certain political groups.
4. It is only natural that a person should have a
much better acquaintance with ideas one believes
in than with ideas one opposes.
5. A person on one's own is a helpless and miserable
creature.
6. Fundamentally, the world we live in is a pretty
lonesome place.
7. Most people just don't give a "damn" for others.
8. I'd like it if I could find someone who would tell
me how to solve my personal problems.
112

113
9.It is only natural for a person to be rather
fearful of the future.
10. There is so much to be done and so little time to
do it in.
11. Once I get wound up in a heated discussion I just
can't stop.
12. In a discussion I often find it necessary to
repeat myself several times to make sure I am
being understood.
13. In a heated discussion I generally become so
absorbed in what I am going to say that I forget
to listen to what the others are saying.
14. It is better to be a dead hero than to be a live
coward.
15. While I don't like to admit this even to myself,
my secret ambition is to become a great person,
like Einstein, or Beethoven, or Shakespeare.
16. The main thing in life is for a person to want to
do something important.
17. If given the chance I would do something of great
benefit to the world.
18. In the history of humankind there have probably
been just a handful of really great thinkers.
19. There are a number of persons I have come to hate
because of the things they stand for.
20. A person who does not believe in some great cause
has not really lived.
21. It is only when a person devotes oneself to an
ideal or cause that life becomes meaningful.
22. Of all the different philosophies which exist in
this world there is probably only one which is
correct.
23. A person who gets enthusiastic about too many
causes is likely to be a pretty "wishy-washy" sort
of person.
24. To compromise with our political opponents is
dangerous because it usually leads to the betrayal
of our own side.

114
25. When it comes to differences of opinion in
religion we must be careful not to compromise with
those who believe differently from the way we do.
26. In times like these a person must be pretty
selfish if one considers primarily one's own
happiness.
27. The worst crime a person could commit is to attack
publicly the people who believe in the same thing
one does.
28. In times like these it is often necessary to be
more on guard against ideas put out by people or
groups in one's own camp than by those in the
opposing camp.
29. A group which tolerates too much difference of
opinion among its own members cannot exist for too
long.
30. There are two kinds of people in the world: those
who are for truth and those who are against truth.
31. My blood boils whenever a person stubbornly
refuses to admit one is wrong.
32. A person who thinks primarily of one's own
happiness is beneath contempt.
33. Most of the ideas which get printed nowadays
aren't worth the paper they are printed on.
34. In this complicated world of ours the only way we
can know what is going on is to rely on leaders or
experts who can be trusted.
35. It is often desirable to reserve judgment about
what's going on until one has had a chance to hear
the opinions of those one respects.
36. In the long run the best way to live is to pick
friends and associates whose tastes and beliefs
are the same as one's own.
37. The present is all too often full of unhappiness.
It is only the future that counts.
38. If a person is to accomplish one's mission in
life, it is sometimes necessary to gamble "all or
nothing at all."

115
39. Unfortunately, a good many people with whom I have
discussed important social and moral problems
don't really understand what's going on.
40. Most people just don't know what's good for them.

APPENDIX F
SELF-MONITORING SCALE

Self-Monitoring Scale
Directions: The statements in the following pages concern
your personal reactions to a number of different situations.
No two statements are exactly alike, so consider each
statement carefully before answering. If a statement is
TRUE or MOSTLY TRUE as applied to you, circle TRUE following
the statement. If a statement is FALSE or NOT USUALLY TRUE
as applied to you, circle FALSE following the statement.
It is important that you answer as frankly and as
honestly as you can. Your answers will be kept in the
strictest confidence.
1. I find it hard to imitate the behavior
of others. TRUE/FALSE
2. My behavior is usually an expression of
my true inner feelings, attitudes, and TRUE/FALSE
beliefs.
3. At parties and social gatherings, I do
not attempt to do or say things that
others will like.
4. I can only argue for ideas which I
already believe.
5. I can make impromptu speeches even on
topics about which I have almost no
information.
6. I guess I put on a show to impress or
entertain people.
7. When I am uncertain how to act in a
social situation, I look to the
behavior of others for cues.
8. I would probably make a good actor.
9. I rarely need the advice of my friends
to choose movies, books, music.
10.I sometimes appear to others to be
experiencing deeper emotions than I
actually am.
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
TRUE/FALSE
117

118
11.
I laugh more when I watch a comedy with
others than when alone.
TRUE/FALSE
12.
In a group of people I am rarely the
center of attention.
TRUE/FALSE
13.
In different situations and with
different people, I often act like very
different persons.
TRUE/FALSE
14.
I am not particularly good at making
other people like me.
TRUE/FALSE
15.
Even if I am not enjoying myself, I
often pretend to be having a good time.
TRUE/FALSE
16.
I'm not always the person I appear to be.
TRUE/FALSE
17.
I would not change my opinions (or the
way I do things) in order to please
someone else or win their favor.
TRUE/FALSE
18.
I have considered being an entertainer.
TRUE/FALSE
19.
In order to get along and be liked, I
tend to be what people expect me to be
rather than anything else.
TRUE/FALSE
20.
I have never been good at games like
charades or improvisational acting.
TRUE/FALSE
21.
I have trouble changing my behavior to
suit different people and different
situations.
TRUE/FALSE
22.
At a party I let others keep the jokes
and stories going.
TRUE/FALSE
23.
I feel a bit awkward in company and do
not show up quite so well as I should.
TRUE/FALSE
24 .
I can look anyone in the eye and tell a
lie with a straight face (if for a right
end.)
TRUE/FALSE
25.
I may deceive people by being friendly
when I really dislike them.
TRUE/FALSE

APPENDIX G
LETTER OF INSTRUCTION/INFORMED CONSENT LETTER

Dear Shelter Resident,
Thank you for agreeing to participate in my study
on the factors related to battered women's destination
following a shelter stay. Enclosed you will find an
informed consent form for your signature and the question
naire for you to complete. Please complete the question
naire and return it to the shelter staff. Your prompt
response will be appreciated. Please ask the designated
staff member if you have any questions.
Thank you again for your interest and participation.
Sincerely,
Joanne F. DeMark
120

121
Informed Consent Form
University of Florida
Department of Counselor Education
Subject's Name:
Project Title: Determinants of Battered Women's Destination
Following a Shelter Stay
Principal Investigator: Joanne F. DeMark Date:
I agree to participate in the research as explained below:
The aim of this study is to examine some factors
related to the battering of women. Specifically,
I am looking at level of resources, beliefs about
control and authority, and monitoring skills. The
women will need to complete a questionnaire and
participate in a shelter follow-up procedure. The
time required to complete the questionnaire is
approximately 45-90 minutes.
There is no monetary remuneration for participating in
this study.
Please feel free to ask any questions which you may
have at any time.
The above stated nature and purpose of this research,
including discomforts and risks involved (if any) have been
explained to me. Furthermore, I understand that this
investigation may be used for educational purposes which may
include publication. I also understand that subjects will
not be identified by name in any reporting or publication.
I understand that I may withdraw my consent at any time
without prejudice.
This information will be kept confidential within legal
limits (or to the extent provided by law).
I have read and understand the procedures described
above. I agree to participate in the procedure and I have
received a copy of this description.
Signed
I have defined and fully explained this research to the
participant whose signature appears above.
Signed.

APPENDIX H
COMPARISON DEMOGRAPHICS

Table H-l
Comparison Demographics from Other Battered Women Studies
Characteristic
Okun
Studv 1986^
Pagelow
Studv (1981}
Snyder &
Fruchtman
Studv f1981^
N
300
350
89
Shelter
Sample
100%
90.6%
100%
Location
Michigan
California
& Florida
Detriot
Mean Age
27.7
29.9
29.2
Age Range
16-55
17-68
17-58
Average Number
of Children
1.84
2.3
White
237 (79%)
271 (78%)
30 (34%)
Nonwhite
63 (21%)
77 (22%)
59 (66%)
Left Mate
Before
59.7%
80.0%
68.0%
Returned
to Abuser



Did Not
Return



123

124
Snyder
& Scheer
Studv (1981)
Stacey &
Shupe
Studv (1983)
Star
et al.
Studv (1979)
Walker
Studv (1984)
74
538
57
403
100%
100%
80%

Detroit
Texas
Florida
Colorado &
Surrounding
Region
30.0

32.0
32.2

17-54
18-59


2.02
27 (36%)
343 (64%)
40 (70%)
321 (86%)
47 (64%)
195 (36%)
17 (30%)
51 (14%)
70.3%


*
41 (55%)



33 (45%)



*25% of sample were currently living with abuser.

APPENDIX I
COMPARISON NORMS FOR LOCUS OF CONTROL,
AUTHORITARIANISM, AND MONITORING

126
Table 1-1
Locus of Control Comparison Norms
This Studv
Walker 1984)
Studv
Levenson
Norms
Returned
to Abuser
Did Not
Return
Total
Sample
Living
With
Abuser
Independent
Total
Sample
Internal
N
30
42
72
94
292
386
48
M
35.63
36.02
35.28
41.87
41.33
41.46
35.46
SD
5.64
6.81
6.31
7.13
6.79
6.88
7.41
Powerful
Others
N
30
42
72
92
291
383
48
M
16.07
17.69
17.43
17.52
18.17
18.01
14.64
SD
9.40
10.53
10.01
10.78
9.39
9.39
9.73
Chance
N
30
42
72
94
289
383
48
M
20.60
16.43
18.17
17.25
17.41
17.37
13.38
SD
9.36
9.36
9.52
9.995
9.32
9.48
9.05

Table 1-2
Authoritarianism and Monitoring Comparison Norms
Comparison
This Studv
Returned Did Not
to Abuser Return
Total
Sample
Norms
Authoritarianism
N
30
42
70
1310
M
152.72
163.60
159.08
163.56
SD
23.20
32.88
29.64
25.47
Monitorina
N
30
42
70
128
M
8.50
9.25
9.00
8.44
SD
3.75
3.75
3.71
6.89
127

REFERENCES
Abramson, L.Y., Seligman, M.E., & Teasdale, J.D. (1978).
Learned helplessness in humans: Critique and reformu
lations. Journal of Abnormal Psychology. 87. 49-74.
Allen, M., & Straus, M.A. (1975). Resources, power, and
husband-wife violence. National Council on Family
Relations. 12, 1-29.
Bakan, D. (1971). Slaughter of the innocents: A study of
the battered child phenomenon. Boston: Beacon Press.
Bannon, J. (1975, August). Law enforcement problems with
intra-familv violence. Paper presented to the American
Bar Association, Detroit.
Bard, M., & Zacker, J. (1971). The prevention of family
violence: Dilemmas of community intervention. Journal
of Marriage and the Family. 33. 677-682.
Bard, M., & Zacker, J. (1974). Assaultiveness and alcohol
use in family disputes: Police perceptions.
Criminology. 12, 281-292.
Barker, E.N. (1963). Authoritarianism of the political
right, center, and left. Journal of Social Issues.
19(2) 63-74.
Bell, J.N. (1977, June). Rescuing the battered wife.
Human Behavior, pp. 16-23.
Bernhardson, C.S. (1967). Dogmatism, defense mechanisms,
and social responding. Psychological Reports. 20, 511-
513.
Berscheid, E., Graziano, W., Monson, T., & Dermer, M.
(1976). Outcome dependency: Attention, attribution,
and attraction. Journal of Personality and Social
Psychology. 34, 978-989.
Bettinghaus, E., Miller, G., & Steinfatt, T. (1970).
Source evaluation, syllogistic content, and judgments
of logical validity by high and low-dogmatic persons.
Journal of Personality and Social Psychology. 16, 238-
244.
128

129
Booth, A., & White, L. (1980). Thinking about divorce.
Journal of Marriage and the Family. 42. 605-616.
Borrero-Hernandez, A. (1979). Unidimensional and multi
dimensional measures of locus of control and their
relationship to selected personality variables.
Dissertation Abstracts International. 40, 1328B-1329B.
(University Microfilms No. 79-19052)
Bourdouris, J. Homicide and the family. (1971). Journal
of Marriage and the Family. 33. 667-676.
Bowen, N.H. (1982). Guidelines for career counseling with
abused women. Vocational Guidance Quarterly. 31. 123-
127.
Brakel, S.J. (1971). Diversion from the criminal process:
Informal discretion, motivation, and formalization.
Denver Law Journal. 48, 211-238.
Briggs, S.R., Cheek, J.M., & Buss, A.H. (1980). An
analysis of the self-monitoring scale. Journal of
Personality and Social Psychology. 34. 978-989.
Britain, Battered Wives. (1973, July 9). Newsweek, p. 39.
Broderick, C.B., & Smith, J. (1979). The general systems
approaches to the family. In W.R. Burr, R. Hill, F. I.
Nye, & I. L. Reiss (Eds.), Contemporary theories about
the family (Vol. 1, pp. 227-256). New York: The Free
Press.
Brown, B.A., & Brazzle, G.R. (1982). The self concept and
marital satisfaction of women in metroplex shelters.
Unpublished master's thesis, University of Texas,
Arlington.
Brownmiller, S. (1975). Against our will: Men, women, and
rape. New York: Simon and Schuster.
Buckley, W. (1967). Sociology and modern systems theory.
Englewood Cliffs, NJ: Prentice-Hall.
Byrne, D., Blaylock, B., & Goldberg, J. (1966). Dogmatism
and defense mechanisms. Psychological Reports. 18.
739-742.
Caldwell, D. F., & O'Reilly, C.A. (1982). Boundary
spanning and individual performance: The impact of
self-monitoring. Journal of Applied Psychology. 67.,
124-127.
Carlson, B.E. (1977). Battered women and their assailants.
Social Work. 22., 455-460.

130
Caster, D.U., & Parsons, O.A. (1977). Locus of control in
alcoholics and treatment outcome. Journal of Studies
in Alcohol. 38, 2087-2095.
Castle, T.J. (1971). Temporal correlates of dogmatism.
Journal of Consulting and Clinical Psychology. 36. 70-
81.
Chan, K.B. (1979). Husband-wife violence in Toronto.
Dissertation Abstracts International. 39 6995A-6996A.
Costello, C.G. (1978). A critical review of Seligman's
laboratory experiments on learned helplessness and
depression in humans. Journal of Abnormal Psychology.
87, 21-31.
Costin, F. (1965). Dogmatism and learning: A follow-up of
contradictory findings. Journal of Educational
Research. 59, 186-188.
Cristall, L.M. (1978). A comparison of androgyny and self-
actualization in battered women. Dissertation
Abstracts International. 39, 5039B-5040B. (University
Microfilms No. 79-07826)
Dobash, R.E. (1981). When non-sexist language is sexist.
In B. Warrior (Ed.), Working on wife abuse (pp. 87-88).
Cambridge, MA: Betsy Warrior.
Dobash, R.E., & Dobash, R.P. (1971). Love, honour and
obey: Institutional ideologies and the struggle for
battered women. Contemporary Crises. 1, 403-415.
Dobash, R.E., & Dobash, R.P. (1977-78). Wives: The
'appropriate' victims of marital violence. Victi
mology: An International Journal. 2, 426-442.
Dobash, R.E., & Dobash, R.P. (1979). Violence against
wives: A case against the patriarchy. New York: Free
Press.
Dobash, R.E., & Dobash, R.P. (1981). Social science and
social action: The case of wife beating. Journal of
Family Issues. 2, 439-470.
Dobash, R.E., & Dobash, R.P. (1983). The context specific
approach to researching violence. In D. Finkelhor,
R.J. Gelles, G.T. Hotaling, M.A. Straus (Eds.), The
dark side of families: Current family violence
research (pp. 261-276). Beverly Hills: Sage
Publications.

131
Dollard, J., Dobb, L.W., Miller, N.E., Mowrer, O.H., &
Sears, R.R. (1939). Frustration and aggression. New
York: Yale University Press.
Donohew, L., Parker, J.M., & McDermott, V. (1972).
Psychophysiological measurements of information
selection: Two studies. Journal of Communications.
22, 54-63.
Donovan, D.M., & O'Leary, M.R. (1975). Comparison of
perceived and experienced control among alcoholics and
nonalcoholics. Journal of Abnormal Psychology. 84.
726-728.
Dweck, C.C., Goetz, T.E., & Straus, N.L. (1980). Sex
difference in learned helplessness: IV. An experi
mental and naturalistic study of failure generalization
and its mediators. Journal of Personality and Social
Psychology. 38. 441-452.
Dubinsky, A.J., Hartley, S.W., & Yammarino, F.J. (1985).
Boundary spanners and self-monitoring: An extended
view. Psychological Reports. 57, 287-294.
Durbin, K. (1974, June). Wife-beating. Ladies Home
Journal, pp. 62-67, 72.
Ehrlich, H.J., & Lee, D. (1969). Dogmatism, learning, and
resistance to change: A review and a new paradigm.
Psychological Bulletin. 71, 249-259.
Eisenberg, S., & Micklow, P. (1977). The assaulted wife:
'Catch 22' revisited. Women's Rights Law Reporter. 3-
4, 138-161.
Elliott, F.A. (1982). Biological contributions to family
violence. In J. Hansen & L. Barnhill (Eds.), Clinical
approaches to family violence (pp. 35-58). Rockville,
MD: Aspen.
Fields, M. (1977-78). Wife-beating: Facts and figures.
Victimology; An International Journal. 2, 643-647.
Fleming, J.B. (1979). Stopping wife abuse: A guide to the
emotional, psychological, and legal implications for
the abused woman and those helping her. Garden City,
NY: Anchor Press/Doubleday.
Gabrenya, W.K., & Arkin, R.M. (1980). Self-monitoring
scale: Factor structure and correlates. Personality
and Social Psychology Bulletin. 6, 13-22.

132
Gaquin, D.A. (1977-78). Spouse abuse: Data from the
national crime survey. Victimology: An International
Journal. 2, 632-643.
Gayford, J.J. (1975). Battered wives. Medicine. Science
and Law. 15, 237-245.
Geizer, R.S., Rarick, D.L., & Soldow, G.F. (1977).
Deception and judgment accuracy: A study in person
perception. Personality and Social Psychology
Bulletin. 3, 446-449.
Gelles, R.J. (1972) The violent home: A study of
physical aggression between husbands and wives.
Beverly Hills: Sage Publications.
Gelles, R.J. (1975). Violence and pregnancy: A note on
the extent of the problem and needed services. Family
Coordinator. 24. 81-86.
Gelles, R.J. (1976). Abused wivesWhy do they stay?
Journal of Marriage and the Family. 38. 659-668.
Gelles, R.J. (1977). No place to go: The social dynamics
of marital violence. In M. Roy (Ed.), Battered women:
A psvchosociological study of domestic violence
(pp. 46-63). New York: Van Nostrand Reinhold.
Gelles, R.J. (1980). Violence in the family: A review of
research in the seventies. Journal of Marriage and the
Family. 42, 873-884.
Gelles, R.J. (1983). The exchange/social control theory of
violence. In D. Finkelhor, R.J. Gelles, G.T. Hotaling,
M.A. Straus (Eds.), The dark side of families: Current
family violence research (pp. 108-134). Beverly Hills:
Sage Publications.
Gelles, R.J., & Straus, M. (1979). Determinants of
violence in the family: Toward a theoretical integra
tion. In W.R. Burr, R. Hill, F.I. Nye & I.L. Reiss
(Eds.), Contemporary theories about the family (Vol. 1,
pp. 549-581) New York: The Free Press.
Gil, D.G. (1970). Violence against children: Physical
child abuse in the United States. Cambridge, MA:
Harvard University Press.
Giles-Sims, J. (1983). Wife battering: A systems theory
approach. New York: The Guilford Press.
Gill, T., & Coote, A. (1975). Battered women: How to use
the law. Rev. Ed. London: Cobden Trust.

133
Gillen, J.L. (1946). The Wisconsin prisoner: Studies in
crimoaenesis. Madison: University of Wisconsin Press.
Gilligan, C. (1982). In a different voice: Psychological
theory and women's development. Cambridge, MA:
Harvard University Press.
Gillman, I.S. (1980). An object-relations approach to the
phenomenon and treatment of battered women.
Psychiatry. 43, 346-358.
Goode, W.J. (1969). Violence among intimates. In D.
Mulvihill & M. Tumin (Eds.), Crimes of Violence (pp.
116-141). Washington, DC: U.S. Government Printing
Office.
Hall, E., Joestring, J., & Woods, M.J. (1977). Relation
ships among measures of locus of control for black and
white students. Psychological Reports. 40, 59-62.
Hallenbeck, P.N., & Lundstedt, S. (1966). Some relations
between dogmatism, denial and depression. Journal of
Social Psychology. 70, 53-58.
Hansen, J.C., & Barnhill, L.R. (1982). Clinical approaches
to family violence. Rockville, MD: Aspen Publica
tions.
Hanson, D.J. (1968). Dogmatism and authoritarianism.
Journal of Social Psychology. 76, 89-95.
Hanson, D.J. (1970). Validity test of the dogmatism scale.
Psychological Reports. 26, 585-586.
Hartik, L.M. (1978). Identification of personality
characteristics and self-concept factors of battered
wives. Dissertation Abstracts International. 40. 893B-
894B. (University Microfilms No. 79-18190)
Harvey, J., & Hays, D.G. (1972). Effect of dogmatism and
authority of the sources of communication upon persua
sion. Psychological Reports. 30, 119-122.
Hilberman, E. (1980). The wife beater's wife reconsidered.
American Journal of Psychiatry. 137. 1336-1347.
Hilberman, E., & Munson, K. (1977-78). Sixty battered
women. Victimology; An International Journal. 2,
460-470.
Hiroto, D.S. (1974). Locus of control and learned help
lessness. Journal of Experimental psychology. 102.
187-193.

134
Hornung, C.A., McCullough, B.C., & Sagimoto, T. (1981).
Status relationship in marriage: Risk factors in
spouse abuse. Journal of Marriage and the Family. 43.,
675-692.
Hunt, M.F., & Miller, G.R. (1968). Open- and close
mindedness, belief-discrepant communication behavior,
and tolerance for cognitive inconsistency. Journal of
Personality and Social Psychology. 8, 35-37.
Ickes, W., & Barnes, R.D. (1977). The role of sex and
self-monitoring in unstructured dyadic interactions.
Journal of Personality and Social Psychology. 35, 315-
330.
Jay, R.L. (1969). Q technique factor analysis of the
Rokeach dogmatism scale. Educational and Psychological
Measurement. 29. 453-459.
Jones, E.E., & Baumeister, R. (1976). The self-monitor
looks at the ingratiator. Journal of Personality. 44.
654-674.
Kalmuss, D.S., & Straus, M.A. (1982). Wife's marital
dependency and wife abuse. Journal of Marriage & the
Family. 44, 277-286.
Kemp, C.G. (1963). Perception of authority in relation to
open and closed belief systems. Science Education. 47.,
482-484.
Kemp, M., Knightly, B., & Norton, M. (1975). Battered
women and the law. London: Cobden Trust.
Kerlinger, F.N., & Pedhazer, E.J. (1973). Multiple
regression in behavioral research. New York: Holt,
Rinehart, & Winston.
Kerlinger, F.N., & Rokeach, M. (1966). The factorial
nature of the F and D scales. Journal of Personality
and Social Psychology. 4, 391-399.
Kernberg, O.F. (1975). Borderline conditions and patho
logical narcissism. New York: Jason Aronson.
Kleck, R.E., & Wheaton, J. (1967). Dogmatism and responses
to opinion-consistent and opinion-inconsistent infor
mation. Journal of Personality and Social Psychology.
5, 249-252.
Klein, J., & Seligman, M.E. (1976). Reversal of
performance deficits and perceptual deficits in learned
helplessness and depression. Journal of Abnormal
Psychology. 85, 11-26.

135
Krauss, R.M., Geller, V. & Olson, C. (1976, September).
Modalities and cues in perceiving deception. Paper
presented at the meeting of the American Psychological
Association, Washington, DC.
Kulik, J., & Taylor, S.E. (1979). Self-monitoring and the
use of consensus information. Unpublished manuscript,
Harvard University, Cambridge, MA.
Labell, L.S. (1979). Wife abuse: A sociological study of
battered women and their mates. Victimology: An
International Journal. 4, 258-267.
Langley, R., & Levy, R.C. (1977). Wife beatingThe silent
crisis. New York: E.P. Dutton.
Lazio, J.P., & Rosenthal, R. (1970). Subject dogmatism,
experimenter status, and experimenter expectancy
effects. Personality: An International Journal. 1,
11-23.
Lee, D., & Ehrlich, H.J. (1971). Beliefs about self and
others: A test of the dogmatism theory. Psychological
Reports. 28. 919-922.
Lee, F. (1976). A study of sex differences in locus of
control, tennis, expectancy for success and tennis
achievement. Unpublished doctoral dissertation,
University of Oregon.
Levenson, H. (1972). Distinctions within the concepts of
internal-external control: Development of a new scale.
Proceedings of the 80th Annual Convention of the
American Psychological Association. 7, 261-262.
Levenson, H. (1973). Multidimensional locus of control in
psychiatric patients. Journal of Consulting and
Clinical Psychology. 41, 397-404.
Levenson, H. (1974). Activism and powerful others:
Distinctions within the concept of internal-external
control. Journal of Personality Assessment. 38. 377-
383.
Levenson, H. (1981). Differentiating among internality,
powerful others, and chance. In H.M. Lefcourt (Ed.),
Research with the locus of control construct (Vol. 1,
pp. 15-63). New York: Academic Press.
Levenson, H., & Miller, J. (1976). Multidimensional locus
of control in sociopolitical activists of conservative
and liberal ideologies. Journal of Personality and
Social Psychology. 33. 199-208.

136
Levinger, G. (1966). Source of marital dissatisfaction
among applicants for divorce. American Journal of
Orthopsychiatry. 2, 804-806.
Lippa, R. (1978). Expressive control, expressive consis
tency and the correspondence between expressive
behavior and personality. Journal of Personality. 46.
438-461.
Lobel, K. (1986). Naming the violence: Speaking out about
lesbian battering. Seattle: The Seal Press.
Martin, D. (1976). Battered wives. San Francisco, CA:
Glide Publications.
Maurer, A. (1976, June). Physical punishment of children.
Paper presented at the California State Psychological
Convention, Anaheim.
McCarthy, J., & Johnson, R.C. (1962). Interpretation of
the "City Hall Riots" as a function of general
dogmatism. Psychological Reports. 11. 243-245.
Michigan's Women's Commission. (1977). Domestic assault:
A report on family violence in Michigan. Lansing:
State of Michigan.
Mill, J. (1984). High and low self-monitoring individuals:
Their decoding skills and empathetic expression.
Journal of Personality. 52. 372-388.
Miller, A. (1981). The drama of the gifted child. New
York: Basic Books.
Miller, A. (1983). For vour own good: Hidden cruelty in
child-rearing and the roots of violence. New York:
Farrar, Straus, Giroux.
Miller, A. (1986). Thou shalt not be aware: Society's
betrayal of the child. New York: Meridian.
Miller, G.R., & Rokeach, M. (1968). Individual differences
and tolerance for inconsistency. In R.P. Abelson, E.
Aronson, W.J. McGuire, T.N. Newcomb, M.J. Rosenberg,
and P.H. Tannenbaum (Eds.), Theories of cognitive
consistency: A sourcebook (pp. 599-623). Chicago:
Rand McNally.
Miller, W.P., & Seligman, M.E. (1975). Depression and
learned helplessness in man. Journal of Abnormal
Psychology. 84. 228-238.

137
Miller, W.P., Seligman, M.E., & Kurlander, T.B. (1976).
Learned helplessness, depression, and anxiety. Journal
of Nervous and Mental Diseases. 161. 347-357.
Mitchell, C. (1980). Personality and socioeconomic
differences between abused and nonabused women.
Unpublished master's thesis, Georgia State University,
Atlanta.
Moore, D.M. (1979). Battered women. Beverly Hills,
California: Sage Publications.
National Institute of Law Enforcement and Criminal Justice.
(1978). Spouse abuse: A selected bibliography.
Washington, DC: U.S. General Printing Office.
NiCarthy, G., Merriam, K., & Coffman, S. (1984). Talking
it out: A guide to groups for abused women. Seattle:
The Seal Press.
Norman, R.P. (1966). Dogmatism and psychoneurosis in
college women. Journal of Consulting Psychology. 30,
278.
Nye, F.I. (1979). Choice, exchange, and the family. In
W.R. Burr (Ed.), Contemporary theories about the family
(Vol. 2, pp. 48-60). New York: The Free Press.
O'Brien, J.E. (1971). Violence in divorce prone families.
Journal of Marriage and the Family. 33, 692-698.
Okun, L. (1986). Woman abuse: Facts replacing myths.
Albany: State University of New York Press.
Olsen, A.J. (1972). Volunteers as family counselors. New
York: Home Advisory and Service Council.
Owens, D.M., & Straus, M.A. (1975). The social structure
of violence in childhood and approval of violence as an
adult. Aggressive Behavior. 1, 193-211.
Pagelow, M.D. (1977a). Battered womena new perspective.
Berkeley: University of California.
Pagelow, M.D. (1977b). Blaming the victimParallels in
crimes against women: Rape and battering. Berkeley:
University of California.
Pagelow, M.D. (1977c). Secondary battering: Breaking the
cycle of domestic violence. Berkeley: University of
California.

138
Pagelow, M.D. (1980, November). Double victimization of
battered women. Paper presented at the meeting of the
American Society of Criminology, San Francisco.
Pagelow, M.D. (1981). Woman-battering: Victims and their
experiences. Beverly Hills: Sage Publications.
Palmer, S. (1962). The psychology of murder. New York:
Thomas Y. Crowell Company.
Parker, B., & Schumacher, D.N. (1977). The battered wife
syndrome and violence in the nuclear family of origin.
American Journal of Public Health. 67. 760-761.
Pizzey, E. (1974). Scream guietlv or the neighbours will
hear. Middlesex, England: Penguin Books.
Plant, W.T. (1960). Rokeach's dogmatism scale as a measure
of general authoritarianism. Psychological Reports. 6,
164.
Pleck, E., Pleck, J.H., Grossman, M., & Bart, P.B. (1977-
78). The battered data syndrome: A comment on
Steinmetz' article. Victimology: An International
Journal. 2, 680-683.
Powell, F.A. (1962). Open- and closed-mindedness and the
ability to differentiate source and message. Journal
of Abnormal and Social Psychology. 65, 61-64.
Pyron, B.A. (1966). A factor analytical study of simpli
city-complexity of social ordering. Perceptual and
Motor Skills. 2, 259-272.
Radloff, L.S., & Rae, D.S. (1981). Components of the sex
difference in depression. Research in Community Mental
Health. 2, 111-137.
Radloff, L.S., & Rae, D.S. (1979). Susceptibility and
precipitating factors in depression: Sex differences
and similarities. Journal of Abnormal Psychology. 88.
174-181.
Restle, F., Andrews, M., & Rokeach, M. (1964). Difference
between open- and closed-minded subjects on learning-
set and oddity problems. Journal of Abnormal and
Social Psychology. 68, 648-654.
Rochester Police Bureau. (1974). FACITfamily conflict
intervention team experiment: Experimental action
program. Rochester, NY: University of Rochester.
Rokeach, M. (1954). The nature and meaning of dogmatism.
Psychological Review. 61. 194-204.

139
Rokeach, M. (1956). Political and religious dogmatism: An
alternative to the authoritarian personality.
Psychological Monographs. 70 (18, whole No. 425).
Rokeach, M. (Ed.). (1960). The open and closed mind. New
York: Basic Books.
Rokeach, M. (1967). Authoritarianism scales and response
bias: Comment on Peabody's paper. Psychological
Bulletin. 67, 349-355.
Rokeach, M., & Bonier, R. (1960). Time perspective,
dogmatism, and anxiety. In M. Rokeach (Ed.), The open
and closed mind (pp. 366-375). New York: Basic Books.
Rotter, J.B. (1966). Generalized expectancies for internal
versus external control of reinforcement. Psycho
logical Monographs. 80, 1.
Rounsaville, B.J. (1978). Battered wives: Barriers to
identification and treatment. American Journal of
Orthopsychiatry. 48, 487-494.
Rounsaville, B.J., Lifton, N., & Bieber, M. (1979). The
natural history of a psychotherapy group for battered
women. Psychiatry. 42., 63-78.
Roy, M. (1977). Battered women: A psvchosociological
study of domestic violence. New York: Van Nostrand
Reinhold.
Royal College of Psychiatrists. (1974). Memorandum on
battered wives. London, England: Chandos House.
Rutschky, K. (1977). Black pedagogy. Berlin, Germany:
Hogarth.
Sampson, E.E. (1978). Personality and the location of
identity. Journal of Personality. 46, 552-568.
Scanlan, T.J. (1979). Self employment as a career option:
An investigation of entrepreneurship from the perspec
tives of Holland's theory of career development and
Levenson's measure of locus of control. Dissertation
Abstracts International. 40, 4561A. (University
Microfilms No. 80-04269)
Schechter, S. (1982). Women and male violence: The
visions and struggles of the battered women's movement.
Boston: South End Press.
Scott, P.D. (1974). Battered wives. British Journal of
Psychiatry. 125. 433.

140
Search, G. (1974, June). Notes from abroad: London-
Battered wives. Ms. Magazine, pp. 24-26.
Seligman, M.E. (1975) Helplessness: On depression.
development, and death. San Francisco: W.H. Freeman.
Seligman, M.E., & Hiroto, D.S. (1974). Generality of
learned helplessness in man. Journal of Personality
and Social Psychology. 31. 311-327.
Shainess, N. (1977). Psychological aspects of wife
battering. In M. Roy, (Ed.), Battered women: A
psvchosocioloaical study of domestic violence
(pp. 111-119). New York: Van Nostrand Reinhold.
Shainess, N. (1979). Vulnerability to violence: Masochism
as process. American Journal of Psychotherapy. 33.
174-189.
Smithers, A. (1970). Personality patterns and levels of
dogmatism. British Journal of Social and Clinical
Psychology. 9, 183-184.
Snell, J., Rosenwald, R., & Robey, A. (1964). The wife
beater's wife: A study of family interaction.
Archives of General Psychiatry. 11. 107-112.
Snyder, D.K., & Fruchtman, L.A. (1981). Differential
patterns of wife abuse: A data-based typology.
Journal of Consulting & Clinical Psychology. 49.
878-885.
Snyder, D.K., & Scheer, N.S. (1981). Predicting disposi
tion following brief residence at a shelter for women.
American Journal of Community Psychology. 9, 559-566.
Snyder, M. (1974). Self-monitoring of expressive behavior.
Journal of Personality and Social Psychology. 30, 526-
537.
Snyder, M. (1979). Self-monitoring processes. In L.
Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 12, pp. 86-124). New York: Academic
Press.
Snyder, M., & Cantor, N. (1980). Thinking about ourselves
and others: Self-monitoring and social knowledge.
Journal of Personality and Social Psychology. 39. 222-
234.
Snyder, M., & Gangestad, S. (1982). Choosing social
situations: Two investigations of self-monitoring
processes. Journal of Personality and Social
Psychology. 43. 123-135.

141
Snyder, M., & Monson, T. (1975). Persons, situations, and
the control of social behavior. Journal of Personality
and Social Psychology. 32, 637-644.
Snyder, M., & Swann, W.B., Jr. (1976). When actions
reflect attitudes: The politics of impression manage
ment. Journal of Personality and Social Psychology.
34., 1034-1042.
Snyder, M., & Tanke, E.D. (1976). Behavior and attitude:
Some people are more consistent than others. Journal
of Personality. 44., 501-517.
Spitzner, J.H., & McGee, D.H. (1975). Family-crisis
intervention training, diversion, and the prevention of
crimes of violence. Police Chief. 42., 252-253.
Stacey, W.A., & Shupe, A. (1983). The family secret:
Domestic violence in America. Boston: Beacon Press.
Star, B. (1978). Comparing battered and non-battered
women. Victimology: An International Journal. 2,
32-44.
Star, B., Clark, C.G., Goetz, K.M., & O'Malia, L. (1979).
Psychosocial aspects of wifebattering. Social
Casework. 479-487.
Steele, B.F., & Pollack, C.B. (1974). A psychiatric study
of parents who abuse infants and small children. In
R.E. Heifer & C.H. Kempe (Eds.), The battered child
(pp. 89-133). Chicago: University of Chicago Press.
Steinmetz, K. (1977-78). The battered husband syndrome.
Victimology. 2, 499-509.
Stephens, D.W. (1977). Domestic assault: The police
response. In M. Roy (Ed.), Battered women: A psvcho-
sociological study of domestic violence (pp. 164-172).
New York: Van Nostrand Reinhold.
Straus, M.A. (1973). General systems theory approach to a
theory of violence between family members. Social
Science Information. 12., 105-125.
Straus, M.A. (1974). Cultural and social organization
influences on violence between family members. In R.
Prince & D. Barrier (Eds.), Configurations: Biological
and cultural factors in sexuality and family life
(pp. 53-69). Lexington, MA: D.C. Health.
Straus, M.A. (1977-78). Wife beating: How common and why?
Victimology: An International Journal. 2, 443-458.

142
Straus, M.A. (1980). Victims and aggressors in marital
violence. American Behavioral Scientist. 1980, 23.
681-704.
Straus, M.A., Gelles, R.J., & Steinmetz, S.K. (1980).
Behind closed doors: Violence in the American family.
Garden City, NY: Doubleday.
Vacchiano, R.B. Dogmatism. (1977). In T. Blass (Ed.),
Personality variables in social behavior (pp. 281-314).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Vacchiano, R.B., Strauss, P.S., & Hochman, L. (1969). The
open and closed mind: A review of dogmatism. Psycho
logical Bulletin. 71. 261-273.
Vacchiano, R.B., Strauss, P.S., & Schiffman, D.C. (1968).
Personality correlates of dogmatism. Journal of
Consulting Psychology. 32., 83-85.
Vidulich, R.N., & Kairman, I.P. (1961). The effects of
information source status and dogmatism upon conformity
behavior. Journal of Abnormal and Social Psychology.
63, 639-642.
Walker, L.E. (1977-78). Battered women and learned
helplessness. Victimology: An International Journal.
2, 525-534.
Walker, L.E. (1978a, March). Feminist psychotherapy with
victims of violence. Paper presented at the Midwinter
Meeting of Division 29 of the American Psychological
Association, Scottsdale, AZ.
Walker. L.E. (1978b). Treatment alternatives for battered
women. In J.R. Chapman & M. Gates (Eds.), The
vVictimization of women (pp. 146-161). Beverly Hills:
Sage Publications.
Walker, L.E. (1979). The battered woman. New York:
Harper & Row.
Walker, L.E. (1980a). Battered women. In A. Brodsky & R.
Hare-Mustin (Eds.), Women and psychotherapy
(pp. 339-363). New York: The Guilford Press.
Walker, L.E. (1980b, October). Clinical aspects of the
battered woman syndrome study. Paper presented at the
Colorado Mental Health Conference, Keystone, CO.
Walker, L.E. (1981). Battered women: Sex roles and
clinical issues. Professional Psychology. 12. 81-91.

143
Walker, L.E. (1984). The battered woman syndrome. New
York: Springer Publishing.
Wallston, K.A., Wallston, B.S., & DeVellis, R. (1978).
Development of the multidimensional health locus of
control (MHLC) scales. Health Education Monographs. 6,
160-170.
Warr, P.B., Lee, R.E., & Joreskog, K.G. (1969). A note on
the factorial nature of the F and D scales. British
Journal of Psychology. 60, 119-123.
Welsh, R.S. (1976). Severe parental punishment and
delinquency: A developmental theory. Journal of Child
Clinical Psychology. 35, 17-21.
Wetzel, L., & Ross, M.A. (1983). Psychological and social
ramifications of battering: Observations leading to a
counseling methodology for victims of domestic
violence. Personnel & Guidance Journal. 61, 423-428.
Wife beaters. (1975, June). McCall/s. p. 35.
Wolfgang, M.E. (1958). Patterns of criminal homicide.
Philadelphia: University of Pennsylvania Press.
Zurcher, L.A., Willis, J.E., Ikard, F., & Dohme, J.A.
(1967). Dogmatism, future orientation, and perception
of time. Journal of Social Psychology. 63. 205-209.

BIOGRAPHICAL SKETCH
Joanne F. DeMark was born on March 4, 1949, in New
Castle, Pennsylvania. She attended schools in western
Pennsylvania, graduating from Lincoln High School in Ellwood
City, Pennsylvania, in 1967. She received her Bachelor of
Science degree in English education from Indiana University
of Pennsylvania in 1971.
She taught high school and was a social worker in
Broward County, Florida, for 2 years until 1974. She then
moved to Gainesville, Florida, where she spent the next
several years working in social service administrative,
advocacy and treatment roles with juvenile delinquents and
criminally committed patients with psychotic disorders. She
was the first Rape Victim Advocate for the Eighth Judicial
Circuit and was an employee and board member for the Sexual
and Physical Abuse Resource Center. She completed her M.Ed.
and Ed.S. at the University of Florida in counselor educa
tion, community agency track, in 1979.
Ms. DeMark began her doctoral studies in counseling
psychology in 1980. She completed her APA-approved doctoral
internship at the Counseling Center at Georgia, State
University in Atlanta, Georgia in 1982. Since 1982, she has
been employed with HBO & Company, a healthcare information
systems and services company where she conducts management
144

145
and organizational development for the company, as well as
for their hospital clients. Over the years Ms. DeMark has
lectured in or taught undergraduate and graduate courses at
the University of Florida, Santa Fe Community College,
Dekalb Community College, Emory University, and Georgia
State University. Additionally, she has conducted public
and media presentations and consulted and led numerous work
shops on sexual assault, woman battering, women's liberation
issues, leadership development, and organizational culture
and change.

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Ja/jufelyn^Liss Resnick, Chair
Professor of Counselor
Education
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
/
j)A[A o
Sandra B. Damico
Professor of Foundations
of Education
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
y/^T-ry 9y
tf&rry A. Grater, Jr.
Professor of Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Phyllis M. Meek
Associate Professor of
Counselor Education

This dissertation was submitted to the Graduate Faculty
of the College of Education and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
April 1988
Dean, College of Education
Dean, Graduate School



39
present in masochistic individuals, is not often present in
battered women. Battered women endure far too much abuse
for their behavior to be considered "neurotic."
Battered women more often fall into the category of
borderline personality, being neither neurotic nor psy
chotic. The "battered woman personality" is a woman who has
two internal, yet separate, representations of herself: one
good and the other bad. As battering occurs, she is a
worthless person in a destructive relationship to a per
secuting husband (mother). She has difficulty leaving the
relationship in that the good self and other emerges after
the incident is over and as the bad submerges. The split
ting means quite opposing views of self and spouse are
present, yet the battered woman is unaware of the contra
dictory nature of her beliefs (Gillman, 1980).
A source for variables for this investigation was the
theory of violence proposed by A. Miller (1983). Because of
this theory's application to the current study, the theory
is presented in detail.
A. Miller (1983) proposed that violence occurs in
society because of the heritage of strict child rearing
practices which break children's willfulness when they are
very young, and causes them to repress their own traumatic
childhood experiences, and leads to their aggressiveness or
self-destructive behaviors in adolescence and adulthood.
Historically, children have been viewed as mean, demanding,
tyrannical, possessing destructive drives and sexual longing


Measure of Resources
1.
Are you currently employed?
No
Yes
2.
Do you have children of age five or
younger at home?
No
Yes
3.
What percentage of the total family
income is earned by your mate?
%
107


94
Number of previous separations was not a distinguishing
variable between the groups in this study. This contrasts
with the findings by D.K. Snyder and Scheer (1981) that
previous separations were a predictor of destination. It is
possible that their findings are not generalizable since
different findings occurred here, or that they have opera
tionalized previous separations differently than this
researcher. Okun (1986) and Walker (1979) expressed beliefs
that repeated separations were likely before the woman would
finally terminate the relationship. This may be so, but the
quality and nature of these separations, preceding and
subsequent events, may be so unique, that a simple quantita
tive tally will rarely be predictive. Data gathering on
frequency and severity of abuse have been partial attempts
to define the complex context of the battering relationship
and the separations; researchers, through simple quantita
tive analysis of these factors, have failed to distinguish
between women with abusers or women who have ended battering
relationships (Pagelow, 1977c, 1980, 1981; D.K. Snyder &
Scheer, 1981).
The locus of control variables were included to inves
tigate a component of the "learned helplessness" model
applied to battered women by Walker (1979). Her research
(1984) included, "contrary to expectations" (p. 80), the
findings that battered women saw themselves as having a
great deal of control over their lives, more so than how
women in general viewed their internal control.


15
follow-up contact which occurs from 2 to 6 weeks following
departure from the shelter, regardless of whether the
battered woman left the shelter to return to him, or in the
interim weeks decided to move back with him.
Not Returning to the Abuser
Women who are in the category of not returning to the
abuser are those who decide to leave the shelter and do not
live with the abuser and are not residing with him as of
follow-up within the time frame of 2 through 6 weeks after
departure from the shelter.
Locus of Control
Locus of control refers to how an individual perceives
contingency reinforcement upon his or her actions(s) and
generally includes an internal and/or an external dimension.
Attributional style may be interchanged with locus of
control. The three types of locus of control in this
investigation are internal, powerful others, and chance.
Internal locus of control refers to when an individual
perceives that a reinforcement has been contingent upon his
or her action(s) (Rotter, 1966). Powerful others locus of
control is defined as when as individual believes that a
reinforcement has not been completely contingent on his or
her action(s), but externally due, to some extent, to the
control of powerful others (Levenson & Miller, 1976) .
Chance locus of control means that an individual perceives
that a reinforcement has not been fully contingent on his or
her action(s), but due, in part, to the result of chance,


42
Children who receive humiliating and cruel treatment
when they need protection, respect, and honesty will grow up
to subject others or themselves to destructive acts. If
they become parents, their own children will be the likely
outlets. Lacking a weaker spouse or children, their
violence may be directed at others in society (A. Miller
1983) Self-cruelty is another possibility. Having
received poor treatment as children, adults can continue to
treat themselves abusively (A. Miller, 1983).
A. Miller (1986) has argued against Freud's theory of
drives in children indicating that his willingness to
consistently interpret neuroses as rooted in the child's
drive conflicts with Freud's own protection and idealization
of the parenting he received. Freud's own narcissistic
wounds, and a lack of an empathic environment for him to
uncover his own repressed traumatization, led him to collude
in blaming the victim and sparing parents' feelings (A.
Miller, 1986).
In addition to being self or other destructive, another
variation of behavior can develop in response to such a
childhood. As the child idealizes the parents and takes
responsibility and blame for traumas, the child comes to
understand performing for the parents and parents' needs as
a means of winning their love. Thus, some children from
this environment have the particular skill of discerning the
needs of others (e.g., the parents) no matter how subtle or
unconscious the signs are (A. Miller, 1981). The develop-


ACKNOWLEDGEMENTS
I extend my appreciation to everyone whose assistance
enabled me to complete this manuscript. The encouragement,
love, support, and assistance from my committee, intimate
friends, and family have been mainstays during my pursuits.
My committee members' attention, direction, and ideas
have been extremely valuable. Dr. Jaquelyn Liss Resnick, my
committee chair, ably guided my process, especially
assisting me with the feminist and academic scholarship
parameters to which I aspired for my work. She was suppor
tive while expecting quality work. Dr. Phyllis Meek
provided consistent encouragement and confidence in my
abilities throughout my university and community work with
her. Dr. Sandra Damico and Dr. Harry Grater offered
excellent review of my research and offered important ideas
and suggestions for the work. I have enjoyed and benefitted
from shared enthusiasms with my committee: feminist theory
and scholarship with Drs. Resnick and Meek; psychodynamic
theory and practice, specifically Alice Miller's ideas with
Dr. Grater; and the joys and intricacies of quantitative and
qualitative research with Dr. Damico. Their time, expertise
and scholarly review of my work have been greatly appre
ciated.
iii


Table 1-2
Authoritarianism and Monitoring Comparison Norms
Comparison
This Studv
Returned Did Not
to Abuser Return
Total
Sample
Norms
Authoritarianism
N
30
42
70
1310
M
152.72
163.60
159.08
163.56
SD
23.20
32.88
29.64
25.47
Monitorina
N
30
42
70
128
M
8.50
9.25
9.00
8.44
SD
3.75
3.75
3.71
6.89
127


133
Gillen, J.L. (1946). The Wisconsin prisoner: Studies in
crimoaenesis. Madison: University of Wisconsin Press.
Gilligan, C. (1982). In a different voice: Psychological
theory and women's development. Cambridge, MA:
Harvard University Press.
Gillman, I.S. (1980). An object-relations approach to the
phenomenon and treatment of battered women.
Psychiatry. 43, 346-358.
Goode, W.J. (1969). Violence among intimates. In D.
Mulvihill & M. Tumin (Eds.), Crimes of Violence (pp.
116-141). Washington, DC: U.S. Government Printing
Office.
Hall, E., Joestring, J., & Woods, M.J. (1977). Relation
ships among measures of locus of control for black and
white students. Psychological Reports. 40, 59-62.
Hallenbeck, P.N., & Lundstedt, S. (1966). Some relations
between dogmatism, denial and depression. Journal of
Social Psychology. 70, 53-58.
Hansen, J.C., & Barnhill, L.R. (1982). Clinical approaches
to family violence. Rockville, MD: Aspen Publica
tions.
Hanson, D.J. (1968). Dogmatism and authoritarianism.
Journal of Social Psychology. 76, 89-95.
Hanson, D.J. (1970). Validity test of the dogmatism scale.
Psychological Reports. 26, 585-586.
Hartik, L.M. (1978). Identification of personality
characteristics and self-concept factors of battered
wives. Dissertation Abstracts International. 40. 893B-
894B. (University Microfilms No. 79-18190)
Harvey, J., & Hays, D.G. (1972). Effect of dogmatism and
authority of the sources of communication upon persua
sion. Psychological Reports. 30, 119-122.
Hilberman, E. (1980). The wife beater's wife reconsidered.
American Journal of Psychiatry. 137. 1336-1347.
Hilberman, E., & Munson, K. (1977-78). Sixty battered
women. Victimology; An International Journal. 2,
460-470.
Hiroto, D.S. (1974). Locus of control and learned help
lessness. Journal of Experimental psychology. 102.
187-193.


79
not return to the abuser following a shelter stay and those
who do return.
6. There is no relationship between authoritarianism
monitoring, resources, number of previous separations, and
locus of control (internal, powerful others, chance) and the
destination of battered woman following a shelter stay.
Hypotheses one through five were analyzed using one-way
analyses of variance. Hypothesis six was analyzed using
logistic regression. An alpha level for significance was
established at .05 as a conventional level of significance.
Initially, a Bartlett's test was utilized to determine
departures from normality. The results of this test sug
gested that the sample was not within the range of a normal
distribution; therefore, although either discriminant
analysis or logistic regression was considered for analysis,
logistic regression was chosen since it is less sensitive to
departures from normality.
Using logistic regression analysis the researcher can
distinguish between two groups and produce a predictive
equation. In logistic regression analysis the variables are
weighed and linearly combined the variables mathematically to
force the optimum distinction between groups.
Logistic regression in a two-group case is the same as a
multiple regression where the dependent variable (in this
study not returning to the abuser or returning to the
abuser) takes on the values of 1 and 0 and maximally discrim
inates between the two groups (Kerlinger & Pedhazer, 1973).


Table H-l
Comparison Demographics from Other Battered Women Studies
Characteristic
Okun
Studv 1986^
Pagelow
Studv (1981}
Snyder &
Fruchtman
Studv f1981^
N
300
350
89
Shelter
Sample
100%
90.6%
100%
Location
Michigan
California
& Florida
Detriot
Mean Age
27.7
29.9
29.2
Age Range
16-55
17-68
17-58
Average Number
of Children
1.84
2.3
White
237 (79%)
271 (78%)
30 (34%)
Nonwhite
63 (21%)
77 (22%)
59 (66%)
Left Mate
Before
59.7%
80.0%
68.0%
Returned
to Abuser



Did Not
Return



123


56
of both measures as less than ideal (.57 and .67 respec
tively) and suggested the need for a better measure of this
construct in future studies. In addition, she combined
several state instruments as indices of current state
related to learned helplessness. Included among them was
locus of control (Levenson's Locus of Control Scales). She
conducted path analyses to determine if either childhood
learned helplessness or relationship learned helplessness
was determinant of current state. Both measures appear to
influence current state.
Learned helplessness theory has been linked with
depression because of the perception by those experiencing
it that they have no power or control over events in their
lives (Abramson, Seligman, & Teasdale, 1978; Costello,
1978). A measure of perceived control, i.e., locus of
control, was incorporated by Walker (1984) into her study.
She expected battered women would attribute more control to
external sources than norm groups would, both on scales
measuring powerful others and chance locus of control. She
also anticipated lower scores on internal locus of control
compared to norms. Battered women, both in and out of
battering relationships, did score higher than norm groups
on chance locus of control. Battered women still in
battering relationships, however, did not score higher than
norm groups on the powerful others locus of control scale.
Walker (1984) postulated that a woman still cohabi-
tating with the abuser may not wish to acknowledge the


57
batterer's control in her life, nor the unlikelihood of
changing him or the environment to prevent further abuse.
Battered women did, surprisingly, score significantly higher
than norm groups on internal locus of control, attributing
themselves as having substantial control over what happens
to them. The Levenson Locus of Control Scales were again
utilized in this research.
Authoritarianism and Monitoring
Neither authoritarianism nor monitoring have been
examined previously for their relationship as potential
determinants of battered women's destination. In fact, they
had not been investigated in relationship to woman battering
in general. In combination they represent salient features
of the theory on the roots of violence proposed by A. Miller
(1981, 1983, 1986). An examination of the results of
research on both factors suggested their fruitfulness for
study in this area.
Authoritarianism
As Rokeach (1960) theorized, an individual maintains an
open belief system when the need to know dominates and a
nonthreatening reality exists. A threatening reality causes
the individual to become closed in his or her beliefs, the
need to know diminishes, confusion occurs between the
information and the informant, and identification with
absolute authority takes place. Closing the belief system
distances the individual from his or her own anxiety. One
with a predominantly open belief system is a low dogmatic


38
her subjects, a loving contrition stage occurred after the
acute battering incident.
Intrapsvchic Theories
Earliest theoretical explanations of spouse abuse are
based on the concept that abuse occurs because the parti
cipants are individually pathological, i.e., have inadeguate
personality development (Snell, Rosenwald, & Robey, 1964).
This intrapsychic viewpoint suggests that batterers are
passive-aggressive, obsessive-compulsive, paranoid,
sadistic, immature, dependent, excessively jealous, and/or
addicted to drugs or alcohol (Scott, 1974; Shainess, 1977).
The wives contribute to the abuse because they are aggres
sive, masculine, domineering, sexually frigid, and maso
chistic (Snell, Rosenwald, & Robey, 1964; Shainess, 1979).
These theoretical explanations have fallen short because
neither explains persons who have these personality charac
teristics and are not in battering relationships, nor takes
into account a myriad of interactional and social factors
surrounding domestic violence.
The most prevalent Freudian view applied to battered
women is masochism (Shainess, 1979). Yet within the context
of contemporary psychoanalytic theory, Gillman (1980)
proposed an object-relations approach to understanding and
treating battered women, following from the work on border
line personality (cf. Kernberg, 1975). She (Gillman, 1980)
suggested that masochism is not a key factor in battered
women since the defense mechanism of repression, usually


CHAPTER III
METHODOLOGY
This study was designed to distinguish between those
battered women who, following a shelter stay, did not return
to their abusers and those battered women who did return to
their abusers in terms of key variables drawn from previous
research (Gelles, 1976; Kalmuss & Straus, 1982; Pagelow,
1981; Walker 1979, 1984) and the psychodynamic theory of A.
Miller (1981, 1983, 1986). Specifically, the relationship
was studied between battered women's destination and
resources; number of previous separations; internal,
powerful others, and chance locus of control, authori
tarianism; and monitoring. In Chapter III the research
methodology of the investigation is described. The chapter
includes the following: population and sample, the sampling
procedures, the instrumentation, data collection procedures,
and data analysis.
Population and Sample
The population base for this study was battered women
who resided in two battered women's shelters located in
Florida, the Gainesville and the Ft. Myers shelters, and the
three metropolitan Atlanta shelters. The Gainesville
shelter mainly served a 15-county area in north central
Florida. The Ft. Myers shelter mainly served a six-county
62


BIOGRAPHICAL SKETCH
Joanne F. DeMark was born on March 4, 1949, in New
Castle, Pennsylvania. She attended schools in western
Pennsylvania, graduating from Lincoln High School in Ellwood
City, Pennsylvania, in 1967. She received her Bachelor of
Science degree in English education from Indiana University
of Pennsylvania in 1971.
She taught high school and was a social worker in
Broward County, Florida, for 2 years until 1974. She then
moved to Gainesville, Florida, where she spent the next
several years working in social service administrative,
advocacy and treatment roles with juvenile delinquents and
criminally committed patients with psychotic disorders. She
was the first Rape Victim Advocate for the Eighth Judicial
Circuit and was an employee and board member for the Sexual
and Physical Abuse Resource Center. She completed her M.Ed.
and Ed.S. at the University of Florida in counselor educa
tion, community agency track, in 1979.
Ms. DeMark began her doctoral studies in counseling
psychology in 1980. She completed her APA-approved doctoral
internship at the Counseling Center at Georgia, State
University in Atlanta, Georgia in 1982. Since 1982, she has
been employed with HBO & Company, a healthcare information
systems and services company where she conducts management
144


95
Additionally, the battered women in her sample who were
living with abusers did not believe, to a significantly
greater extent, that powerful others exercised a lot of
control over them compared to the beliefs of a norm group of
women or battered women who were no longer with batterers.
If learned helplessness were operant, one would expect that
battered women who returned from shelters to abusive
partners would have significantly less internal locus of
control and significantly greater powerful others locus of
control than battered women who did not return to batterers.
Findings from this investigation were not significant in
distinguishing between the two groups. The means and
standard deviations from this study can be compared to
results from the Walker study and Levenson's norms in
Appendix I.
The relationship of A. Miller's (1981, 1983, 1986)
psychodynamic theory to battered women's destination was
accomplished by utilizing measures of authoritarianism and
monitoring. If Miller's theory, as operationalized here,
were applicable, the expectations would be twofold. First,
battered women who did not return to their abusers would
have received lower scores on the Rokeach Dogmatism Scale as
evidence that they had shed their idealized view of author
ity (i.e., the male head of household, the marriage institu
tion) Secondly, battered women who did not return to
batterers would be more likely to monitor their own feelings
and needs and monitor others much less when compared to


19
tolerated such abuse. From examination of the early
literature several main forces in the early 1970s reflect a
Zeitgeist which has brought wife abuse to public attention:
(a) the British refuge movement, (b) the women's rights
movement in the U.S., (c) examination of police injuries and
fatalities, and (d) the study of societal violence.
The British Refuge Movement
Although the British were the first to open battered
women's shelters, their initial discovery of the social
problem was quite unintentional. In 1971 Erin Pizzey went
to the Chiswick City Council to seek support for estab
lishing a home as an advice center for women. When she
opened the doors to Chiswick Women's Aid (WA), she dis
covered that the majority of women coming for help were
seeking a safe place away from husbands who beat them.
Within a year and a half she collected demographic and
incidence data from the WA program, including 3,000 requests
from battered women seeking refuge. She gave the informa
tion to England's social service administration thereby
documenting the widespread problem of wife abuse and the
lack of support from legal and governmental communities.
Thus, the shelter movement began (Dobash & Dobash, 1971;
Pizzey, 1974).
Dobash and Dobash (1971) recorded the founding and
growth of WA groups throughout Britain. They suggested that
the establishment of shelters was a natural evolution from
the British women's movement. In addition, they suggested


APPENDIX H
COMPARISON DEMOGRAPHICS


14
victims will also mean battered women as defined here.
Abused, harmed, or beaten may be substituted for battered.
Battering
Battering refers to intentional physical abuse which
causes pain or injury. Spouse abuse, spousal violence,
conjugal assault, wife beating, or domestic violence will be
used interchangeably with battering.
Shelter
A shelter is an emergency refuge for safety available
for more than one battered woman at a time, operated
generally by women's groups, volunteer organizations, a
governmental or non-profit agency, and other than a safe
place to stay which the woman provides for herself or is
provided for her by family members, in-laws, or friends.
Destination
Destination, in this investigation, refers to either
returning to the abuser or not returning to the abuser
following a shelter stay.
Returning to the Abuser
For the purposes of this study returning to the abuser
means being in residence with the assailant during the
readily; operationalizing a definition for psychological
abuse is, at best, extremely difficult; and (b) due to space
considerations and various other exigencies, shelter
programs for battered women most often provide safe refuge
services utilizing this same delimited definition as an
admissions requirement. Although battered women could also
include those who have been abused by family members
(Pagelow, 1977), those groups have been excluded from this
study. This sample may include women who are cohabitating
with men although not legally married to them, or separated
or divorced partners who are living with each other.


This dissertation is dedicated in loving memory
to my parents, Assunta (Susie) Cini DeMark and
Charles Domenick DeMark, who inspired and nurtured
my educational dreams.


In addition to my doctoral committee, Dr. Joe Wittmer
and John Dixon provided notable academic and consultative
assistance for which I am grateful.
I am thankful to the battered women at the shelters who
volunteered their participation in my study and to the
staffs at the shelters who agreed to assist me, adding to
their already lengthy and stressful responsibilities.
I appreciate my many friends and family, Deb Vingle,
Dayna Buskirk, the late Sally Briegal, Margaret Levings,
Charlene Shackle, Susan DeMark, Rosemary DeMark, Ann McKain,
my group, Lisa Schultz, Ron Remillard, Susan Myers, Kris
Billhart, Sarah Boykin, John Bryant, Carol Samuels, Norma
Calway-Fagan, Gerri Green, the Dunns, Fleury Means, and
others whose kind hearts and nurturing souls have been
sustenance upon which I could draw.
Above all, I wish to thank my closest allies and
intimates, Phyllis DeMark and Kay Leigh Hagan. The
encouragement, support, processing, patience, under
standing, scholarly support, spiritual and emotional
scrutiny, honesty, nurturing, and love from these two
women have been inspiriting for me.
IV


145
and organizational development for the company, as well as
for their hospital clients. Over the years Ms. DeMark has
lectured in or taught undergraduate and graduate courses at
the University of Florida, Santa Fe Community College,
Dekalb Community College, Emory University, and Georgia
State University. Additionally, she has conducted public
and media presentations and consulted and led numerous work
shops on sexual assault, woman battering, women's liberation
issues, leadership development, and organizational culture
and change.


Table 4-7
Logistic Regression Model of the Relationship Between Destination and the Research
Variables
Parameter
Estimate
Standard Error
Chi-Square
Probability
Intercept
0.200
0.811
0.06
0.8053
Authoritarianism
-0.332
0.350
0.90
0.3431
Monitoring
-0.587
1.498
0.45
0.6948
Resources
-0.062
0.234
0.07
0.7896
Number of Previous
Separations
-0.023
0.068
0.12
0.7309
Internal Locus of
Control
0.020
0.035
0.32
0.5721
Powerful Others
Locus of Control
-0.005
0.025
0.04
0.8475
Chance Locus of
Control
0.037
0.028
1.75
0.1858


84
Population size of the city in which the study partici
pants lived before coining to the shelter is categorized in
Table 4-5. An effect related to population appeared likely,
so a chi-square analysis was conducted. A 2 = 10.03, p =
0.018, signifies an effect for a = .05 and d.f. = 3; bat
tered women from the small towns were more likely to return
to the abuser and as the population of residence increased,
battered women were less likely to return to their mates.
This result is considered further in Chapter 5.
The Rokeach Dogmatism Scale to measure authori
tarianism, Snyder Self-Monitoring Scale was used to measure
monitoring, the Kalmuss & Straus (1982) 3-item method to
measure resources, a straightforward item on the
Table 4-5
Population of City in Which Subject Resided Before Shelter
Returned
Population to Abuser
Size n (%)
Did Not Total
Return Sample
n (%) n (%)
0-20,000
21,000-
50,000
51,000-
200,000
17 (56.7)
6 (20.0)
4 (13.3)
over 200,000 3 (10.0)
10 (23.8)
9 (21.4)
8 (19.0)
15 (35.7)
27 (37.5)
15 (20.8)
12 (16.7)
18 (42.9)


92
said to reflect that unique group of battered women using
shelters. It is possible that those women could be
characterized quite differently from even those participants
in this study who returned to batterers after the first 48
hours. Including responses from the "quick exit" women
might have impacted findings as well as generalizability.
The volunteer nature of the sample and the self-report, par
ticularly for destination, may also present limitations.
Secondly, destination of battered women is assumed, in
this study, to reflect two divergent realities: living with
an abuser or not living with an abuser. Some women may go
back to spouses and, although rare, not receive abuse again.
And some women who do not return to their mates may select,
again though rare, to form relationships with other abusive
partners. In fact, this research was not designed to
include differing typologies of women within either of the
destination categories.
A third limitation exists in the operational definition
for destination in this study being 2 to 6 weeks following
the shelter experience. Some women categorized as not
returning to their abusers by this definition will indeed
move back in with their mates sometime later.
The researcher selected the time while battered women are at
shelters as an appropriate time to perceive psychological
differences between the two groups. It could be that the
differences between the two groups would be more profound
during other times, e.g., 6 months away from abuser or while


121
Informed Consent Form
University of Florida
Department of Counselor Education
Subject's Name:
Project Title: Determinants of Battered Women's Destination
Following a Shelter Stay
Principal Investigator: Joanne F. DeMark Date:
I agree to participate in the research as explained below:
The aim of this study is to examine some factors
related to the battering of women. Specifically,
I am looking at level of resources, beliefs about
control and authority, and monitoring skills. The
women will need to complete a questionnaire and
participate in a shelter follow-up procedure. The
time required to complete the questionnaire is
approximately 45-90 minutes.
There is no monetary remuneration for participating in
this study.
Please feel free to ask any questions which you may
have at any time.
The above stated nature and purpose of this research,
including discomforts and risks involved (if any) have been
explained to me. Furthermore, I understand that this
investigation may be used for educational purposes which may
include publication. I also understand that subjects will
not be identified by name in any reporting or publication.
I understand that I may withdraw my consent at any time
without prejudice.
This information will be kept confidential within legal
limits (or to the extent provided by law).
I have read and understand the procedures described
above. I agree to participate in the procedure and I have
received a copy of this description.
Signed
I have defined and fully explained this research to the
participant whose signature appears above.
Signed.


51
Walker (1984) examined self-esteem in battered women
via a semantic differential scale on which they rated
themselves, women in general, and men in general. Although
the prediction was for low self-esteem, battered women
viewed themselves more favorably than they viewed women in
general or men in general. The difficulty with this
research was that the scale did not appear to be standar
dized. It was impossible to tell how the battered women's
perceptions of themselves and others compared to other group
perceptions about themselves, women in general, and men in
general. Battered women in this case perceived themselves
more favorably than they perceived other groups, but how
does this compare with other groups' perceptions about them
selves? Other groups' perceptions could be more elevated
than battered women's perceptions which would validate a
lower self-esteem comparatively for the battered women's
group.
Although self-esteem was not be addressed in this
study, more definitive examination of this factor is needed.
Resources/Occupational History
Martin (1976) was the first to emphasize that financial
dependence and unresponsive social systems are very real
explanations for why battered women remain with their
spouses. At the same time, actual research results (Gelles,
1976) results indicated that the fewer resources and less
power the abused woman had, the more likely she was to stay.
Resources have been determined to be any of several items,


Dear Shelter Resident,
I would like to request your participation in a
research study I am conducting. I am a doctoral student in
the Counseling Psychology program at the University of
Florida in Gainesville, Florida. I am conducting a study on
women who stay at battered women's shelters. Specifically,
my purpose is to examine the following factors: level of
resources, beliefs about control and authority, and moni
toring skills through a questionnaire.
I am attempting to locate women who have been battered
by their mates (husband, boyfriend, ex-husband), are staying
at a battered women's shelter, and have lived with the mate
before coming to the shelter. In order to participate,
women must have completed an eighth grade education.
Participation in this study involves two things:
1) Completion of a questionnaire containing 92 items during
your stay at the shelter, and 2) participation in the
shelter's follow-up procedure at six weeks after you leave
the shelter. All responses are anonymous with each indi
vidual receiving a code number.
As much as I would like to compensate each woman for
her participation, the most I can offer is my gratitude for
helping in a study that I hope will contribute to the safety
of battered women. To those who are interested, a copy of
the results of the study will be available through your
shelter.
A designated shelter staff member is available to
answer any questions you might have and to find out if you
are interested in participating.
Thank you for taking the time to read this letter. I
hope that you will agree to participate. I am eager to have
your involvement in this study.
Sincerely,
Joanne F. DeMark
105


40
for their parents, and having other negative desires and
behaviors which society and parents have believed are their
responsibilities to ameliorate (Rutschky, 1977). The
child's willfulness must be emphatically curbed and the
child must view the parents as unquestionably right and
superior (Rutschky, 1977). Both physically and emotionally
abusive and humiliating methods are utilized by parents to
accomplish their responsibilities of rearing socially
acceptable children. The parents are unaware of the ways
they are abreacting their own suppressed hatred and anger
from their own childhood by utilizing their offspring to
meet their suppressed needs (A. Miller 1981, 1983, 1986).
In this environment the child's needs are not met but
frustrated. The child experiences hurt and anger, but
cannot have these feelings validated with parents. Without
any validation the child denies the feelings (A. Miller,
1981, 1983, 1986). A proneness to violence is most likely
if the child from such an upbringing does not have the
following: (a) another human being in whom to confide true
feelings, (b) an environment allowing experiencing and
expression of pain (whether at home or elsewhere), (c) other
objects for abreacting hatred, and (d) an education or
intellectual avenue to rationalize hatred (A. Miller, 1983).
Depending, then, on the degree to which the child has been
traumatized and the unavailability of healthy channeling of
negative experiences and feelings which are suppressed,


70
fact, studies of their correlation range from .41 to .60
(Caster & Parsons, 1977; Levenson, 1973; Scanlan, 1979;
Wallston, Wallston, & DeVellis, 1978). Only minimal
correlation (-.25 to .19) has occurred between the P and C
factors with the I factor (Caster & Parsons, 1977; Levenson,
1973; Scanlan, 1979; Wallston, Wallston, & DeVellis, 1978).
The Rotter I-E scale is one of the original locus of
control scales and the most often referenced as the standard
in locus of control measurement. Convergent validity is
strongest between the C and I Scales with the Rotter I-E
Scale, ranging from .43 to .56 for the C factor and from
-.15 to -.41 for the I factor (Donovan & O'Leary, 1975;
Hall, Joestring, & Woods, 1977; Levenson, 1972).
There is negligible correlation between the Levenson
Scales and the Marlowe-Crowne Social Desirability Scale.
In the two studies conducted (Levenson, 1972; Wallston,
Wallston, & DeVellis, 1978), the correlations to the
Marlowe-Crowne Social Desirability Scale for the I Scale
were .04 and .09, for the P Scale .04 and .11, and for the
C Scale -.10 and .08.
Borrero-Hernandez (1979) noted several relationships
between the Levenson Scales and other personality variables.
The I Scale is positively related to measures of sociability
on the California Psychological Inventory (CPI). The C
Scale negatively relates to sense of well-being and
responsibility on the CPI and positively relates to guilt


143
Walker, L.E. (1984). The battered woman syndrome. New
York: Springer Publishing.
Wallston, K.A., Wallston, B.S., & DeVellis, R. (1978).
Development of the multidimensional health locus of
control (MHLC) scales. Health Education Monographs. 6,
160-170.
Warr, P.B., Lee, R.E., & Joreskog, K.G. (1969). A note on
the factorial nature of the F and D scales. British
Journal of Psychology. 60, 119-123.
Welsh, R.S. (1976). Severe parental punishment and
delinquency: A developmental theory. Journal of Child
Clinical Psychology. 35, 17-21.
Wetzel, L., & Ross, M.A. (1983). Psychological and social
ramifications of battering: Observations leading to a
counseling methodology for victims of domestic
violence. Personnel & Guidance Journal. 61, 423-428.
Wife beaters. (1975, June). McCall/s. p. 35.
Wolfgang, M.E. (1958). Patterns of criminal homicide.
Philadelphia: University of Pennsylvania Press.
Zurcher, L.A., Willis, J.E., Ikard, F., & Dohme, J.A.
(1967). Dogmatism, future orientation, and perception
of time. Journal of Social Psychology. 63. 205-209.


82
Table 4-2
Race of Sample
Returned
Did Not
Total
Race
to Abuser
Return
Samle
n (%)
n (%)
n (%)
White
26 (86.7)
16 (38.1)
42 (58.3)
Nonwhite
4 (13.3)
26 (61.9)
30 (41.7)
was found i.e., white participants returned to their abusers
more often than their nonwhite counterparts. This is
discussed in Chapter 5.
The participants' numbers of children are categorized
in Table 4-3. The average number of children for the total
sample was 1.74 (SD = 1.30). Battered women who returned to
the abusers in the study had an average of 1.63 children
(SD = 1.45); average number of children for the abused women
who did not return to the batterers was 1.81 (SD = 1.19).
A breakdown of the study participants by whether they
had left their mates before or not is presented in Table 4-
4. Nearly equal percentages of participants had left their
mates before regardless of destination in this study.


63
area in southwest Florida. Two of the Atlanta area shelters
were in suburban locations: the Cobb county shelter,
approximately 12 miles northwest of downtown Atlanta, and
the Clayton County shelter, approximately 12 miles south of
downtown Atlanta. These two shelters served mainly resi
dents of their counties plus women and their children from
several adjoining counties in the northwestern portion of
Georgia. The third metro shelter was located in midtown-
Atlanta mainly serving residents of urban Fulton and Dekalb
counties, as well as occasionally women and their children
from the adjoining metro counties. As is common with
battered women's shelters throughout the United States, at
times each of these shelters operated at capacity and placed
women's names on a waiting list for admission. In the case
of a woman facing imminent potential for severe violence
personnel in each of these shelters worked with those in
other shelters for immediate placement and, via law enforce
ment or other avenues, transported the woman to a distant
shelter.
Battered women came to these shelters following
telephone contact with the particular shelter to determine
that they met the criteria for admission. The admissions
criteria for each shelter was the same: the women must have
been recently battered and be in need of safe shelter
without available refuge. Battered women were admitted to
the shelter at any time of the day or night, any day of the
week, depending on when they sought assistance. Many women


41
rather than addressed, the likelihood for aggression or
self-destruction can be determined.
What follows is that the child represses any memory of
the traumata and idealizes the parents and the upbringing.
The child, because of love for the parents, takes the
responsibility and the blame for the parents' hurtful
behavior (A. Miller 1983). The child's feelings of anger,
hurt, hatred, helplessness, and pain come forth in other
life circumstances, but disassociated from their origin.
The child is consciously blocked from realizing their
source. In addition to idealizing those who traumatize the
child, the child can strongly identify with the perpetrator
in an effort to distance the self from the identified "bad
child" whom the perpetrator/parent has been disciplining so
sharply. The child has thus split-off, projected his or her
own weakness, feelings of helplessness, and vulnerability by
maintaining a certain derision for others who are younger,
weaker, smaller (A. Miller, 1981).
Since society maintains a lack of sensitivity to the
denied cruelty, children grow up to become parents who
manifest the same types of humiliations and physical and
emotional abuse on their own children (A. Miller, 1983).
This process means that cruelty and violence become passed
on from generation to generation. Only a sensitive dis
covery of one's own traumatization, in a supportive environ
ment, breaks the intergenerational cycle of violence.


37
tension increases in the home with the batterer expressing
anger and dissatisfaction, but without any extreme violent
expression. The woman tries to please the batterer to
minimize the hostility, calm him, and prevent any more
expression of hostility. However, the tension increases,
and her efforts to reduce his angry responses become more
unsuccessful. Eventually, the woman withdraws, thus
eliciting an insecurity and an increase in possessiveness,
jealousy, and oppression from the abuser.
At phase two the tension becomes intolerable and acute
battering occurs. The batterer explodes into physical and
verbal assault. This is the most likely time for the woman
to receive injuries or for external parties, e.g., police,
to become involved. This incident usually dissipates the
tension which had built up, and the violence temporarily
ceases.
In stage three loving contrition ensues. The batterer
may apologize profusely, promise never to repeat the
actions, buy gifts and/or be more helpful to the woman. The
relationship enters into a honeymoon-like appearance. The
batterer may believe he will not be assaultive again, and
the woman wants very much to believe that he will not.
Empirical findings from Walker's (1984) sample support
the three-stage theory. In 65% of the relationships studied
by Walker (1984), evidence supported the tension-building
stage before an acute battering incident. And with 58% of


140
Search, G. (1974, June). Notes from abroad: London-
Battered wives. Ms. Magazine, pp. 24-26.
Seligman, M.E. (1975) Helplessness: On depression.
development, and death. San Francisco: W.H. Freeman.
Seligman, M.E., & Hiroto, D.S. (1974). Generality of
learned helplessness in man. Journal of Personality
and Social Psychology. 31. 311-327.
Shainess, N. (1977). Psychological aspects of wife
battering. In M. Roy, (Ed.), Battered women: A
psvchosocioloaical study of domestic violence
(pp. 111-119). New York: Van Nostrand Reinhold.
Shainess, N. (1979). Vulnerability to violence: Masochism
as process. American Journal of Psychotherapy. 33.
174-189.
Smithers, A. (1970). Personality patterns and levels of
dogmatism. British Journal of Social and Clinical
Psychology. 9, 183-184.
Snell, J., Rosenwald, R., & Robey, A. (1964). The wife
beater's wife: A study of family interaction.
Archives of General Psychiatry. 11. 107-112.
Snyder, D.K., & Fruchtman, L.A. (1981). Differential
patterns of wife abuse: A data-based typology.
Journal of Consulting & Clinical Psychology. 49.
878-885.
Snyder, D.K., & Scheer, N.S. (1981). Predicting disposi
tion following brief residence at a shelter for women.
American Journal of Community Psychology. 9, 559-566.
Snyder, M. (1974). Self-monitoring of expressive behavior.
Journal of Personality and Social Psychology. 30, 526-
537.
Snyder, M. (1979). Self-monitoring processes. In L.
Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 12, pp. 86-124). New York: Academic
Press.
Snyder, M., & Cantor, N. (1980). Thinking about ourselves
and others: Self-monitoring and social knowledge.
Journal of Personality and Social Psychology. 39. 222-
234.
Snyder, M., & Gangestad, S. (1982). Choosing social
situations: Two investigations of self-monitoring
processes. Journal of Personality and Social
Psychology. 43. 123-135.


99
income and employment opportunities may be available to the
woman to assist her in supporting herself and a family.
(d) The woman may perceive fewer options for other meaning
ful intimate relationships. In addition, battered women in
urban settings have greater access to subsidized housing or
additional sheltering, other women who are in the same or
similar situations, and urban amenities such as more pre
valent mass transit and child care to assist them in gaining
autonomy. Little previous research has been done investi
gating this aspect.
The findings from this research support the beliefs
(Dobash & Dobash, 1979; Hilberman, 1980; Hilberman & Munson,
1977-78; Martin, 1976; Schechter, 1982) that the person
alities of women who are in or remain in abusive relation
ships are not somehow abnormal or different than other
women. The scores on authoritarianism, locus of control,
and monitoring, three personality measures, did not distin
guish between battered women who (potentially) were ending
their abusive relationships and those who were continuing in
them. Nor did their scores appear to be different from
sample comparison norms (Appendix I).
Recommendations for Future Research
The results of this investigation have implications for
future research. Given both the limitations and factors
cited during the discussion, all presented in this chapter,
the following suggestions are made to guide future study in
this area:


69
powerful others (Powerful Others Scale), and fate or chance
(Chance Scale). A person with a high I Scale score would
believe that he or she has great deal of control over his or
her life. A person with a high P Scale score generally
ascribes powerful others as determining outcomes. A person
with a high C Scale score perceives chance or fate as
significantly contributory to events.
The Locus of Control Scales are comprised of 24 items,
8 per scale, although they are presented interspersed to
subjects in one unified series. Responses are noted on a
Likert Scale ranging from strongly disagree (-3) through
strongly agree (+3) assigned to each statement. All of the
items in the Levenson Scales are stated personally so that
the respondent is being measured on his or her perceptions
regarding control, not what he or she believes "people in
general" think about control (Levenson, 1981).
Kuder-Richardson reliabilities for the I Scale range
from .51 to .67, for the P Scale from .72 to .82, for the C
Scale from .73 to .79 (Levenson, 1973, 1974; Wallston,
Wallston, & DeVellis, 1978). Split-half reliabilities,
utilizing the Spearman-Brown formula, are .62 for the I
Scale, .66 for the P Scale, and .64 for the C Scale. Seven-
week test-retest reliabilities are .66, .62, and .73 for the
I, P, and C Scales (Lee, 1976).
The P and C Scales are both related to external locus
of control, so it would be expected that in validity studies
they would correlate to some extent with each other. In


59
1971; Norman, 1966; Smithers, 1970) and dogmatism as a
defense mechanism (cf. Bernhardson, 1967; Hallenbeck &
Lundstedt, 1966; D. Lee & Erhlich, 1971). In general, the
high dogmatic can be characterized as having difficulty
tolerating frustration and prone to conforming, while the
low dogmatic is more tolerant of ambiguities and less
accepting of traditional beliefs (Vacchiano, 1977) When
information or a situation does not align with the high
dogmatic's beliefs, he or she is threatened and avoidant
(Hunt & Miller, 1968; G.R. Miller & Rokeach, 1968; Pyron,
1966). In general, close-minded individuals will recall
less information when it is inconsistent and positively
evaluate consistent information (Kleck & Wheaton, 1967).
Low dogmatics are more likely to expose themselves to
belief-discrepant information (Donohew, Parker, & McDermott,
1972) .
Monitoring
Self-monitoring, as conceived by M. Snyder (1974), is
observation and control of the self based on the cues one
perceives from the situation. The construct involves two
aspects: astute sensitivity to the expression and presen
tation of others and an ability to use the situational cues
to monitor and present oneself accordingly (M. Snyder,
1974) .
High self-monitoring individuals are skilled at con
trolling and modifying expressiveness and behavior to align
with the situational cues for appropriateness. Low self-


CHAPTER II
REVIEW OF THE RELATED LITERATURE
Compared to other psychological topics, the literature
focusing on wife abuse is quite recent, appearing only in
the last 15 years (Hilberman, 1980). Described initially in
this chapter is how conjugal violence emerged as a social
issue and how the beginnings of the public's acknowledgement
of the problem shape the theories which developed. The
socio-psychological theories of violence, particularly those
which have relevance to this investigation, are presented.
Next, the pertinent research in regard to factors related to
the battered woman's destination is detailed. Finally,
previous research findings are given on authoritarianism and
monitoring, the new variables in the current investigation.
Emergence of Conjugal Violence
Woman battering was slow to be recognized as a problem
atic social issue for several reasons: the sanctity and
privacy of the home, the theory of sadomasochism in humans,
and the subordinate role of women to men in both home and
society (Gelles, 1977; Straus, Gelles, & Steinmetz, 1980;
Walker, 1979). In the past two decades, societal issues
have changed substantially resulting in an environment
favorable to examining both the dynamics of conjugal
violence and a society which has supported, excused, and
18


136
Levinger, G. (1966). Source of marital dissatisfaction
among applicants for divorce. American Journal of
Orthopsychiatry. 2, 804-806.
Lippa, R. (1978). Expressive control, expressive consis
tency and the correspondence between expressive
behavior and personality. Journal of Personality. 46.
438-461.
Lobel, K. (1986). Naming the violence: Speaking out about
lesbian battering. Seattle: The Seal Press.
Martin, D. (1976). Battered wives. San Francisco, CA:
Glide Publications.
Maurer, A. (1976, June). Physical punishment of children.
Paper presented at the California State Psychological
Convention, Anaheim.
McCarthy, J., & Johnson, R.C. (1962). Interpretation of
the "City Hall Riots" as a function of general
dogmatism. Psychological Reports. 11. 243-245.
Michigan's Women's Commission. (1977). Domestic assault:
A report on family violence in Michigan. Lansing:
State of Michigan.
Mill, J. (1984). High and low self-monitoring individuals:
Their decoding skills and empathetic expression.
Journal of Personality. 52. 372-388.
Miller, A. (1981). The drama of the gifted child. New
York: Basic Books.
Miller, A. (1983). For vour own good: Hidden cruelty in
child-rearing and the roots of violence. New York:
Farrar, Straus, Giroux.
Miller, A. (1986). Thou shalt not be aware: Society's
betrayal of the child. New York: Meridian.
Miller, G.R., & Rokeach, M. (1968). Individual differences
and tolerance for inconsistency. In R.P. Abelson, E.
Aronson, W.J. McGuire, T.N. Newcomb, M.J. Rosenberg,
and P.H. Tannenbaum (Eds.), Theories of cognitive
consistency: A sourcebook (pp. 599-623). Chicago:
Rand McNally.
Miller, W.P., & Seligman, M.E. (1975). Depression and
learned helplessness in man. Journal of Abnormal
Psychology. 84. 228-238.


78
location a secret and the abuser uninformed about the woman's
whereabouts. These efforts are taken to increase the
possibility of providing a safe place for women and their
children. Increased police surveillance is some- times
requested by shelters when the dangerousness and lethality of
the abuser is judged to be particularly great or the batterer
issues threats. Shelter personnel do not release names of
women or children housed there. When the woman returns to
the abuser, contact from the shelter staff needs to be
discreet to continue efforts of safety provision for both the
battered woman and the shelter program.
Data Analysis
The researcher posed the following hypotheses.
1. There is no difference in authoritarianism between
battered women who do not return to the abuser following a
shelter stay and those who do return.
2. There is no difference in monitoring between
battered women who do not return to the abuser following a
shelter stay and those who do return.
3. There is no difference in resources between battered
women who do not return to the abuser following a shelter
stay and those who do return.
4. There is no difference in number of previous
separations between battered women who do not return to the
abuser following a shelter stay and those who do return.
5. There is no difference in locus of control (inter
nal, powerful others, chance) between battered women who do


129
Booth, A., & White, L. (1980). Thinking about divorce.
Journal of Marriage and the Family. 42. 605-616.
Borrero-Hernandez, A. (1979). Unidimensional and multi
dimensional measures of locus of control and their
relationship to selected personality variables.
Dissertation Abstracts International. 40, 1328B-1329B.
(University Microfilms No. 79-19052)
Bourdouris, J. Homicide and the family. (1971). Journal
of Marriage and the Family. 33. 667-676.
Bowen, N.H. (1982). Guidelines for career counseling with
abused women. Vocational Guidance Quarterly. 31. 123-
127.
Brakel, S.J. (1971). Diversion from the criminal process:
Informal discretion, motivation, and formalization.
Denver Law Journal. 48, 211-238.
Briggs, S.R., Cheek, J.M., & Buss, A.H. (1980). An
analysis of the self-monitoring scale. Journal of
Personality and Social Psychology. 34. 978-989.
Britain, Battered Wives. (1973, July 9). Newsweek, p. 39.
Broderick, C.B., & Smith, J. (1979). The general systems
approaches to the family. In W.R. Burr, R. Hill, F. I.
Nye, & I. L. Reiss (Eds.), Contemporary theories about
the family (Vol. 1, pp. 227-256). New York: The Free
Press.
Brown, B.A., & Brazzle, G.R. (1982). The self concept and
marital satisfaction of women in metroplex shelters.
Unpublished master's thesis, University of Texas,
Arlington.
Brownmiller, S. (1975). Against our will: Men, women, and
rape. New York: Simon and Schuster.
Buckley, W. (1967). Sociology and modern systems theory.
Englewood Cliffs, NJ: Prentice-Hall.
Byrne, D., Blaylock, B., & Goldberg, J. (1966). Dogmatism
and defense mechanisms. Psychological Reports. 18.
739-742.
Caldwell, D. F., & O'Reilly, C.A. (1982). Boundary
spanning and individual performance: The impact of
self-monitoring. Journal of Applied Psychology. 67.,
124-127.
Carlson, B.E. (1977). Battered women and their assailants.
Social Work. 22., 455-460.


5
combination with factors from a new theoretical perspective,
as potential indicators of battered women's destination,
i.e., not returning to the abuser or returning to the
abuser, following a shelter stay.
Rationale for the Study
In several studies researchers have examined various
factors related to women remaining in or leaving a rela
tionship where they are battered, though the researchers did
not formulate global theories which guided their choice of
factors. For example, in early research frequency of abuse
and severity of the abuse were found to be related to the
battered woman leaving the relationship (Gelles, 1976);
however, all investigators since then have failed to
empirically resubstantiate that finding (cf. Pagelow, 1977c,
1980, 1981; D.K. Snyder & Scheer, 1981). Abuse in the
woman's family of origin (Pagelow, 1977c, 1980, 1981), the
age of children, and number of children (Cristall, 1978;
D.K. Snyder & Scheer, 1981) are not related to battered
women's exit from a battering relationship. D.K. Snyder and
Scheer (1981) found that a higher number of previous
separations, shorter length of marriage, and a religious
affiliation other than Roman Catholic were indicative that a
battered woman would not return to the abuser after a
shelter stay (D.K. Snyder & Scheer, 1981). The length of
the relationship was found not to be a significant variable
relating to dissolution of a battering relationship in
another study (Okun, 1986). Of these factors, many of which


2
spouse has committed an act of violence against his or her
partner in the past year (Straus, Gelles, & Steinmetz,
1980). The researchers in that survey concluded that an
American's greatest risk of being assaulted, injured, or
murdered occurs in one's own home by a family member. Forty
percent of all female murder victims are killed by their
husbands (Dobash & Dobash, 1977-78, 1979). Beatings
constitute the most prevalent method of wife murder (Fields,
1977-78).
Conjugal abuse affects not only the marital dyad, but
also the children and family unit. When violence occurs in
the family, children are seeing models for how to handle
relationships and disagreement. A cultural tradition exists
to utilize hitting as punishment to curb unwanted behavior.
Theoretically, children not only learn to curb behaviors
through this model; realistically, they also learn that
violence and love are linked, that violence is morally
right, and that violence is justifiable when something is
really important (Gelles, 1977). Thus, an attitude that
violence can be exercised for "the good" of the recipient is
promulgated (A. Miller, 1983). These lessons in childhood
are transferred to the context of other social relation
ships, and violence in families and relationships becomes a
way of life (Gelles, 1977; A. Miller, 1983, 1986). The
intergenerational cycle theory of violence, i.e., that
children who are recipients of violence will grow up to be
perpetrators of violence, has been validated repeatedly (cf.


APPENDIX C
MEASURE OF RESOURCES


48
variables in combination did not support her hypothesis.
Her variable about a religious family, however, did have
a statistically significant correlation to the dependent
variable. Wetzel and Ross (1983) supported the idea that
battered women are kept captive by the family and religious
values that they hold among other things.
Love. Affection, and Hope
Research on the impact of love, affection, and hope to
battered women's destination is scant. In a sample of
battered women who were not sheltered women, but battered
women located via social service agencies, legal or medical
sources, advertising, and by word of mouth, the most
commonly cited reason for staying in abusive relationships
was that the women loved their husbands (Cristall, 1978).
Forty percent of 542 shelter residents sampled (Stacey &
Shupe, 1983) when asked why had they remained with the
abusers if they had not left immediately after a first
incident of violence, said they were still optimistic about
the future of the relationship. Twenty-seven percent of
those felt they could still save the marriage. Fifteen
percent suggested they stayed because of their affection for
the partners.
Understanding how love, affection, and hope relate to
women's vulnerability to remain in abusive relationships is
a needed undertaking. Women have been identified in our
society as being more relationship oriented than men; their
socialization and conditioning have reinforced them in this


This dissertation was submitted to the Graduate Faculty
of the College of Education and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
April 1988
Dean, College of Education
Dean, Graduate School


85
questionnaire to determine number of previous separations,
and the Levenson Locus of Control Scales to measure internal
(I), powerful others (P), and chance (C) loci of control.
The means and standard deviations for the total sample and
the subsets according to destination can be found in Table
4-6. Because a potential effect was likely related to the
75%-or-more-of-income-earned-by-mate item (one of the three
from the resource measure) and destination, a chi-square
analysis was conducted. Since 2 = 4.46, p = 0.035, which
is significant for a = .05 and d.f. = 1, it was found that
women whose abusers earn 75% or more of the family income
returned to their mates more often than would be expected.
Race was not a significant factor impacting this relation
ship.
Analysis of Hypotheses One Through Five
The first five hypotheses were tested using one-way
ANOVA tests. The first hypothesis was that there would be
no difference in the authoritarianism scores between
battered women who, following shelter stays, did not return
to their abusers and battered women who did return. The
one-way ANOVA was not significant at the .05 level (F =
2.44, p = 0.123); therefore, the null hypothesis was not
rejected.
The second hypothesis was that there would be no
difference in the monitoring scores between battered women
who, following shelter stays, did not return to their
batterers and abused women who did return. The one-way


74
period by contact with the trained administrators and the
shelter staffs and the review of the data as they were being
collected. Additionally, the administrators collecting the
data could call the investigator at any time.
Battered women were admitted to these shelters during
any of the 24 hours in a day. Given up to 48 hours to
emotionally adjust following the acute battering incident,
each woman met with a staff member to complete an intake
questionnaire and participate in an initial intake inter
view. Completing the questionnaire was handled as Part II
of the shelter's intake procedure.
Following completion of Part I of the intake procedure,
which was the shelter's standard intake, the administrator
had the potential subject read the letter of introduction
(Appendix B). The administrator discussed requisites for
participation in the study and potential benefits and risks
of participation. Benefits of participation included the
following: (a) The subject contributes to research which
increases the understanding of the dynamics of battering and
adds to theory which can help end conjugal violence against
women; (b) the subject assists the participating shelter(s)
in increasing the knowledge about their clients so they can
provide more effective services to battered women and their
children; (c) the subject aids those helping battered women,
shelters, legal, criminal, and other social service agencies,
to better understand their clients and to dispel stereotypes
and myths about battered women; and (d) the subject assists


32
curate interpretations by Steinmetz of her own data
regarding incidence of husband abuse. Fleming (1979) and
Schechter (1982) challenged these premises, making a strong
case that the more serious violent incidents, often with
more severe physical and psychological repercussions, are
perpetrated by men against women in conjugal relationships.
Feminist researchers pinpoint the reliance on the
social survey as contributing to faulty findings (Dobash &
Dobash, 1983; NiCarthy, Merriam, & Coffman 1984). The data
from survey methods are very informative regarding socio
demographic characteristics, but are of little value for
information regarding the sensitive issues involved in
family violence about which society and individuals have
traditionally kept quiet (Dobash & Dobash, 1983). Conse
quently, many of the findings and conclusions from this
approach have lacked consistency, been proven unfounded, or
are replete with contradiction (Dobash & Dobash, 1983;
Fleming, 1979; Pagelow, 1980; Pleck, Pleck, Grossman & Bart,
1977). To understand this phenomenon in our society, in-
depth systematic interviewing of battered women by experts
on this subject is preferable to the survey technique as a
research method (Dobash & Dobash, 1983; Pagelow, 1981;
Walker, 1984). And for feminist researchers, calling "woman
battering" or "wife abuse" by the term "spouse abuse" is in
itself a sexist act masking the true nature of the problem
(Dobash, 1981).


131
Dollard, J., Dobb, L.W., Miller, N.E., Mowrer, O.H., &
Sears, R.R. (1939). Frustration and aggression. New
York: Yale University Press.
Donohew, L., Parker, J.M., & McDermott, V. (1972).
Psychophysiological measurements of information
selection: Two studies. Journal of Communications.
22, 54-63.
Donovan, D.M., & O'Leary, M.R. (1975). Comparison of
perceived and experienced control among alcoholics and
nonalcoholics. Journal of Abnormal Psychology. 84.
726-728.
Dweck, C.C., Goetz, T.E., & Straus, N.L. (1980). Sex
difference in learned helplessness: IV. An experi
mental and naturalistic study of failure generalization
and its mediators. Journal of Personality and Social
Psychology. 38. 441-452.
Dubinsky, A.J., Hartley, S.W., & Yammarino, F.J. (1985).
Boundary spanners and self-monitoring: An extended
view. Psychological Reports. 57, 287-294.
Durbin, K. (1974, June). Wife-beating. Ladies Home
Journal, pp. 62-67, 72.
Ehrlich, H.J., & Lee, D. (1969). Dogmatism, learning, and
resistance to change: A review and a new paradigm.
Psychological Bulletin. 71, 249-259.
Eisenberg, S., & Micklow, P. (1977). The assaulted wife:
'Catch 22' revisited. Women's Rights Law Reporter. 3-
4, 138-161.
Elliott, F.A. (1982). Biological contributions to family
violence. In J. Hansen & L. Barnhill (Eds.), Clinical
approaches to family violence (pp. 35-58). Rockville,
MD: Aspen.
Fields, M. (1977-78). Wife-beating: Facts and figures.
Victimology; An International Journal. 2, 643-647.
Fleming, J.B. (1979). Stopping wife abuse: A guide to the
emotional, psychological, and legal implications for
the abused woman and those helping her. Garden City,
NY: Anchor Press/Doubleday.
Gabrenya, W.K., & Arkin, R.M. (1980). Self-monitoring
scale: Factor structure and correlates. Personality
and Social Psychology Bulletin. 6, 13-22.


54
Brown and Brazzle's (1982) sample indicated that a
reason for not leaving among women was that they feared that
the husbands would seek the families out and harm the
children or the wives. Twenty-one percent of those in the
Stacey and Shupe study (1983) who did not leave the home
after the first abuse incident revealed the children and
maintaining the family as their reason. The mother's
intense attachment to and concern for her children has been
noted (Carlson, 1977). Children have been cited as both the
reason for leaving and the reason for staying (Cristall,
1978) ; although in this particular study, twice as often
they were the reason for staying.
Where battered women were compared with non-battered
women (Parker & Schumacher, 1977), there was no statistical
difference between the number of children in either family.
Additionally, when looking at battered women who have left
the abuser compared to those who had not (Cristall, 1978),
neither the age nor number of children appeared to be
statistically different in either group. When the ages of
the children are viewed as part of the definition for the
women's resources (younger children being a limitation,
older children being a potential resource because of their
ability to be self-sufficient or assist in supporting the
family), a strong correlation exists between a woman's
limited resources and her remaining in the relationship
(Pagelow, 1981). In another study (Snyder & Sheer, 1981) it
was found no sociodemographic variables relating to the


49
way (Gilligan, 1982). Males have considered their main
focus, their life's work, to be their career. Women,
traditionally, have viewed their life's work to be wives,
mothers, and keepers of the household. Leaving a relation
ship for a woman is an admission of career failure and,
thus, is not taken lightly. An unsuccessful marriage is a
subordinate concern for a man compared to a career failure.
Self-Esteem
More has been stated about the battered woman's self
esteem than probably any other variable. Researchers and
clinicians who have interviewed or provided services to
battered women repeatedly comment on the low self-esteem of
battered women (Bell, 1977; Bowen, 1982; Brown & Brazzle,
1982; Carlson, 1977; Hilberman & Munson, 1977-78; Pagelow,
1981; Walker, 1977-78; 1978b, 1979, 1981, 1984; Wetzel &
Ross, 1983) .
From the few studies where instrumentation has been
utilized conflicting results exist. In a comparison of 46
battered women to 12 non-battered women, Star (1978) found
lower self-esteem with the battered women. Star, Clark,
Goetz, and O'Malia (1979) found, through their study of 57
battered women, that battered women could be described as
lacking in self-confidence, low in self-esteem, aloof,
anxious, reserved, uneasy in social interactions, and
critical or uncompromising.
Hartik (1978) examined 30 battered women compared to 30
non-battered women for personality characteristics and self-


24
disputes are usually associated with alcohol and that
assaultive behaviors by family members usually ceased before
the arrival of police (Bard & Zacker, 1974). Greater
caution was advised for the common practice of diverting
wife beaters from the criminal justice process (Brakel,
1971).
The Study of Societal Violence
By the late 1960s society in general felt more concern
about the rising crime rate, especially of violent crimes.
In 1969 the Eisenhower Commission (Goode, 1969) looked into
the causes and prevention of violence, and in 1971 the U.S.
Surgeon General conducted a similar study (Roy, 1977). The
Surgeon General's examination linked media violence and
violence in real life, suggesting a causal relationship
between the two (Roy, 1977). Goode (1969), a member of the
National Commission on the Causes and Prevention of
Violence, theorized that young children learn violence and
the appropriate use of force through observations of parents
during early childhood. Goode suggested that social class
differences exist for spouse abuse due to lower class
acceptance of violence and conflict in their environment in
general. With more conflict situations arising in lower
class neighborhoods, limited places to escape conflict, and
fewer resources (e.g., money, power) available to get what
counts, violence is a more likely option.
Gelles (1972) coined the family as the "training ground
for violence" (p. 169) with family members modeling violent


46
counseling), relationship to the assailant, length of mar
riage (if married), previous separations, and religious
affiliation (Roman Catholic or not). No significant group
differences were found for any sociodemographic variables or
for measures of the nature or severity of the abuse inci
dent.
These six variables, when entered into a forward
stepwise discriminant function analysis with destination at
follow-up as the criterion variable, yielded three predic
tors. Complete data on all six variables were available for
50 of the 72 abused women; the 50 were subjects for this
portion of the research. The length of the marriage,
occurrence of previous separations, and religious affili
ation best predicted, with an overall classification
accuracy of 79.6%, where the woman would be residing at
follow-up. The authors support cross validation of their
study utilizing independent samples from other settings.
There seems to be no or minimal other research in the
literature for the three variables identified by the Snyder
and Scheer (1981) research as particularly relevant to why
the woman stays or why she returns to an abusive relation
ship. Walker (1979) indicated that abused women who leave
their batterers and return to shelters on multiple occasions
are more likely to ultimately leave the battering relation
ship. Several authors (Martin, 1976; Pagelow, 1981; Walker,
1979, 1984) have reiterated from interviews with battered
women that longer marriages are the more difficult relation-


97
Battered women who come to shelters and establish
themselves independently from their abusers may be trans
ferring their allegiances from one form of authority to
another. Loyalty to the authority of marriage and the
husband may become loyalty to the authority of the shelter
program, the social service network, and/or, at times,
family and friends, who encourage separation from and
dissolution of the abusive conjugal relationship. Thus,
separation from the abuser may not be concomitant with any
lessening of a rigid, authoritarian belief system.
The battered women with differing destinations after
shelter experience could not be distinguished based on their
monitoring of others and mediating of their own behavior
according to external cues. In fact, battered women
appeared to have very similar levels of monitoring compared
to a female norm sample (see Appendix I). A. Miller's
theory specifically addresses the excessive need to achieve
and perform for others, conditioned and reinforced in
childhood as the right thing to do, at the cost of under
development of one's abilities to perceive one's own needs
and feelings. The self monitoring scores suggest that
battered women in either group had no less orientation
towards the monitoring of others. This instrument may be
inadeguate, however, in measuring any increase in self-
directedness, self empathy, monitoring of the self, and
mediating one's behavior due to internal cues. It was the
researcher's assumption that an increased internal moni-


Table 4-6
Means and Standard Deviations of Study Measures
Measures
Returned
to Abuser
Did Not Return
Total Samle
M
SD
M
SD
M
SD
Authoritarianism
152.72
23.20
163.60
32.88
159.08
29.64
Monitoring
8.50
3.75
9.25
3.75
9.00
3.71
Resources
1.17
0.99
1.38
0.85
1.29
0.91
Not Employed
0.40
0.50
0.40
0.50
0.40
0.49
Children 5 or
Younger at Home
0.40
0.50
0.36
0.48
0.38
0.49
75% or More Income
Earned by Mate
0.37
0.49
0.62
0.49
0.51
0.50
Number of Previous
Separations
2.33
2.11
3.17
3.70
2.82
3.15
Internal Locus
of Control
35.63
5.64
36.02
6.81
35.28
6.31
Powerful Others
Locus of Control
16.07
9.40
17.69
10.53
17.43
10.01
Chance Locus
of Control
20.60
9.36
16.43
9.36
18.17
9.52


30
question would arise as to why so much brain dysfunction
occurs among men only.
Exchanae/Social Control Theory
Exchange and social control apply to all types of
intrafamily violence, including conjugal abuse (Gelles,
1983). Exchange theory suggests that the ways humans
interact with each other is dictated by the rewards they
receive or the negative consequences they avoid (Gelles &
Straus, 1979). The social control theory is the question of
how much control is exerted through society and social
institutions to encourage creation, maintenance, or dimi-
nishment of some belief or act such as violent behavior.
According to this theory, family members batter other family
members simply because they can.
First, the costs of being violent are less than the
rewards; therefore battering occurs. Since effective social
controls aimed at preventing family violence are not in
place, a family member who batters experiences few conse
quences due to these lack of controls. Also, inequality
within the family, the sanctity of the family unit in
society, and the encouraged image of males as macho minimize
social control in the family thereby reducing the conse
quences of violence to the batterer and increasing the
rewards.
When societies have no legal or normative sanctions
regarding family violence, the violence is more frequent
(Nye, 1979). Additionally, when norms or laws exist


APPENDIX A
DEMOGRAPHIC INFORMATION QUESTIONNAIRE


83
Table 4-3
Number of Children
Number
of
Children
Returned
to Abuser
n (%)
Did Not
Return
n (%)
Total
Sample
n (%)
0
8
(26.7)
7
(16.7)
15
(20.8)
1
7
(23.3)
10
(23.8)
17
(23.6)
2
7
(23.3)
12
(26.6)
19
(26.4)
3
6
(20.0)
10
(23.8)
16
(22.2)
4
1
(3.3)
3
(7.1)
4
(5.6)
5
0
0
0
6
1
(3.3)
0
1
(1.4)
Table 4-4
Previous
Separations
Previous
Returned
Did Not
Total
Separa-
to Abuser
Return
Sample
tions
n (%)
n (%)
n (%)
No
9 (30.0)
9 (21.4)
18 (25.0)
Yes
21 (70.0)
33 (78.6)
54 (75.0)