The current status of marriage and family therapists' graduate training in the identification, assessment, and treatment...

MISSING IMAGE

Material Information

Title:
The current status of marriage and family therapists' graduate training in the identification, assessment, and treatment of relationship violence
Physical Description:
xi, 187 leaves : ; 29 cm.
Language:
English
Creator:
Bracciale, Marie T
Publication Date:

Subjects

Subjects / Keywords:
Counselor Education thesis, Ph. D
Dissertations, Academic -- Counselor Education -- UF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Marie T. Bracciale.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 022813945
oclc - 847494700
System ID:
AA00025723:00001

Full Text











THE CURRENT STATUS OF MARRIAGE AND FAMILY THERAPISTS'
GRADUATE TRAINING IN THE IDENTIFICATION, ASSESSMENT,
AND TREATMENT OF RELATIONSHIP VIOLENCE














By

MARIE T. BRACCIALE


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


2004
































Copyright 2004

by

Marie T. Bracciale














ACKNOWLEDGMENTS

The author acknowledges the entire faculty in the Counselor Education Department

for preparing me to do this research. I am most appreciative of Dr. Silvia Echevarria-

Doan, chair, and my online editor, Dr. Jerry Byrd, for their many suggestions during

discussions and editing sessions. My thanks go to Dr. Bimrnbaum of California State

University, Fullerton, who developed the SurveyWiz program for the behavioral

sciences, made it accessible to me, and gave me permission to use it to build my

electronic survey; this made it possible for me to reach a nationwide sample. I am deeply

grateful to my committee members: Dr. Barbara Rienzo for her inspiration, Dr. David

Miller for statistical consultation and teaching, Dr. Sondra Smith for her academic

excellence and her role as a mentor for my teaching experience, and Dr. Peter Sherrard

for his continued supervision and coaching and his wonderful ideas that got this research

started in the right direction. I thank each of them for valuable comments, revisions, and

references during the course of my research.

I am indebted to the department chair, Dr. Harry Daniels, for supporting my

education through a Graduate Teaching Assistantship in the department to allow me to

work and to gain university teaching experience that will be valuable for future

endeavors.

I would also like to acknowledge Dr. Max Parker, Ph.D., for being a mentor to me

throughout my program and for being available for consultation whenever I needed him.

The training that I received from Dr. Mary Howard Hamilton in her consultation class








proved helpful in my professional development as a consultant. She also provided

mentorship throughout my program and was always available for consultation.

Personally, I acknowledge my parents, Rose and Roy, for their continued support;

my brothers, Dominick and Daniel, for unconditional love; and especially my partner,

E. J. Updyke, LMFT, for compassion, understanding, and patience through this process.

Others who contributed and whom I wish to thank include Jason Burkhardt for computer

consultation; Rosaria Upchurch, LMFT, for support and wisdom; Rebecca Waumett,

graduate student, who helped me to prepare email lists and databases; Mike Padilla from

educational psychology, who helped with methods selection and data analysis; Amy

Bradshaw, fellow doctoral student, for statistics coding, data entry, and dear friendship;

and Gary Sipe and the entire Stetson University CIT department for their help with the

SurveyWiz program, development of the online survey development, and resolving

glitches along the way.

Another mentor who is acknowledged for her contributions is Dr. Mickie Miller,

who took me under her wing as a graduate assistant at the University of Florida and at the

Brain Institute and helped me to develop courses. I acknowledge the Gainesville Family

Institute and its two faculty members, Dr. Herb Steier and Dr. Gina Early; I am grateful

for their constant consultation and coaching throughout my graduate program. Their

support and mentorship have changed me personally and professionally to be a better

person. Last, I thank all of my supervisees and interns from whom I have learned so

much over the years at the university; it has been a pleasure to work closely with them.














TABLE OF CONTENTS
Page

ACKNOW LEDGM ENTS ................................................................................................. iii

LIST OF TABLES............................................................................................................. ix

ABSTRACT........................................................................................................................ x

CHAPTER

1 INTRODUCTION ........................................................................................................ I

Increase in Incidence of Relationship Violence........................................................... 3
Professional Responsibility: Legal, Ethical, and Therapeutic Issues.......................... 4
Ethical Standards of the Commission on the Accreditation of Marriage
and Fam ily Therapy Education (COAM FTE) ......................................................... 6
Scope of the Problem .................................................................................................... 8
The Extent of the Problem ........................................................................................... 9
Multiforms of Treatment (Metatheoretical, Postmodernism, Feminism,
Sociocultural. and Social Learning) ...................................................................... 11
M etatheoretical Approach................................................................................... 11
Postm odernism .................................................................................................... 12
Sociocultural Theory........................................................................................... 13
Social Learning Theory ...................................................................................... 14
Treatm ent for Battered W om en.................................................................................. 15
Treatm ent for the Batterer.......................................................................................... 18
Duluth M odel...................................................................................................... 19
Society and Culture............................................................................................. 20
The Fam ily.......................................................................................................... 20
The Individual .................................................................................................... 21
Sum m ary ............................................................................................................ 21
Graduate Training in Relationship Violence: Identification of Training Needs ......22
Assessm ent of Violence in Treatm ent........................................................................ 24
Statem ent of the Problem ........................................................................................... 26
Need for the Study...................................................................................................... 27
Purpose of the Study................................................................................................... 29
Rationale for the Study............................................................................................... 31
Definitions of Term s................................................................................................... 32
Organization of the D issertation................................................................................. 34








2 REV IEW OF THE LITERATU RE ............................................................................ 35

Increase in Incidence of Relationship V iolence......................................................... 35
Professional Responsibility: Legal, Ethical, and Therapeutic Issues ........................ 38
Licensing Boards ................................................................................................ 38
Lack of Training in Fam ily Violence ................................................................. 38
M inim um State Requirem ents for Training........................................................ 39
Theoretical Constructs Underlying the Study ............................................................ 40
Fem inist V iew ..................................................................................................... 40
The Study of Power ............................................................................................ 41
Social Learning Theory ...................................................................................... 43
Social Structural Theory..................................................................................... 45
Treatm ent of Relationship Violence........................................................................... 47
Choice to Leave or to Stay in Abusive Relationships ........................................ 47
Gottm an's Batterer O ffender Typologies ........................................................... 49
Hum an Heart Reactivity ..................................................................................... 50
Holtzworth-Munroe and Stuart's Three Major Types of Batterers .................... 51
Identification of Training Needs ................................................................................ 52
Current Status of Training and Need for Additional Curriculum....................... 52
Continuing Education Units................................................................................ 53
Skills in Assessing Relationship Violence and Imminent Danger ..................... 54
Assessing for Risk Factors Associated With Relationship Violence ................. 55
Identifying Risk Factors Associated With Relationship Violence ..................... 55
Prior Relationship A ggression............................................................................ 56
Dem graphic Characteristics.............................................................................. 56
Psychological Characteristics ............................................................................. 57
Specific Psychological Syndrom es..................................................................... 58
Other Risk Factors .............................................................................................. 59
Ability to A ssess V violence Risk ......................................................................... 60
V violence Risk A ssessm ent Study ....................................................................... 60
A ssessing and M managing Risk............................................................................ 61
H CR-20 Item Risk Assessm ent Instrum ent........................................................ 62
Static and D ynam ic Risk..................................................................................... 62
Conclusion.................................................................................................................. 63

3 M ETHODOLOG Y ..................................................................................................... 66

Research M ethod........................................................................................................ 67
Survey Design ............................................................................................................ 69
Rating Scales in the Social Sciences .......................................................................... 70
Research Procedures................................................................................................... 73
Sam ple ................................................................................................................ 73
Sources of Error in Sam ple W eb Surveys .......................................................... 74
Survey Research in Fam ily Therapy .......................................................................... 78
Data Analyses............................................................................................................. 81
The Nature of Factors ......................................................................................... 81
Stages in a Factor Analysis................................................................................. 83


vi





vii


Rotation of Factors ......................................................................................... 85
Validity in Factor Analysis ............................................................................. 86
Reliability in Factor Analysis.......................................................................... 87
Internal Consistency ....................................................................................... 87
M FT Training Program s ................................................................................. 88
Conclusions............................................................................................................ 89

4 RESULTS.............................................................................................................. 91

Participants and Dem graphic Description............................................................. 91
Factor Analysis Results.......................................................................................... 93
Principal Axis Factoring and Oblique Rotation....................................................... 94
Extraction M ethod: Principal Axis Factoring......................................................... 95
Factor Analysis Results ......................................................................................... 96
Identification and Nam ing of Factors...................................................................... 97
Reliability and Validity .......................................................................................... 98
Factor 1 .................................................................................................................. 98
Factor 2..................................................................................................................99
Factor 3................................................................................................................ 104
Sum m ary of Reliability Statistics for Factors 1 to 3.............................................. 106
Summ ary .............................................................................................................. 106

5 DISCUSSION ...................................................................................................... 108

Evaluation and Discussion of the Results.......................................................... 108
Ratings of Importance .................................................................................. 109
Training M ethods ......................................................................................... 112
Clinical Assessm ents ................................................................................... 115
Lim stations of the Study....................................................................................... 117
Respondents ................................................................................................. 117
Potential Confounders .................................................................................. 117
Random Error .............................................................................................. 118
Response Rate ............................................................................................. 120
Possible Nonrespondent Bias ....................................................................... 120
Nonrandomized Samples ............................................................................. 120
Alternative Explanations .............................................................................. 121
Implications and Recommendations for Further Research.................................... 122
Theoretical, Training, and Ethnical Implications.................................................. 123
Conclusion........................................................................................................... 125

APPENDIX

A RELATIONSHIP VIOLENCE TRAINING SURVEY (WEB VERSION)........... 129

B COAMFTE ACCREDITATION STANDARDS.................................................. 136

C AAM FT CODE OF ETHICS .............................................................................. 147








D INVITATION LETTER TO PARTICIPATE IN SURVEY.................................. 155

E FOLLOW-UP LETTER TO NONRESPONDENTS ............................................ 156

F SAMPLE CODED RESPONSE FORM............................................................... 157

G DESCRIPTIVE STATISTICS MEAN RATINGS AND STANDARD
D EV IA TION S ..................................................................................................... 158

H FACTOR CORRELATION MATRIX................................................................. 159

LIST OF REFERENCES............................................................................................. 160

BIOGRAPHICAL SKETCH....................................................................................... 187














LIST OF TABLES

T'Fable pae

1. Variables Used for the Relationship Violence Training Survey (RVTS) ................ 82

2. Characteristics of Respondents................................................................................. 92

3. Eigenvalues and Total Variance Explained ............................................................. 94

4. Factor Loading Matrix: Rotated Factor Loadings of .40 or Greater in the
Principal Axis Factoring, Based Upon N = 171 ................................................... 96

5. Factor 1 (Item s 11-20): Sam ple Size Sum m ary....................................................... 99

6. Factors 1, 2, and 3: Reliability Statistics.................................................................. 99

7. Factor 1: Item Statistics ......................................................................................... 100

8. Factor 1: Summ ary Item Statistics......................................................................... 100

9. Factor 1: Item -Total Statistics............................................................................... 101

10. Factor 1: Scale Statistics........................................................................................ 101

11. Factor 2 (Item s 6-9): Sample Size Summ ary......................................................... 102

12. Factor 2: Item Statistics ......................................................................................... 102

13. Factor 2: Summ ary Item Statistics......................................................................... 103

14. Factor 2: Item -Total Statistics............................................................................... 103

15. Factor 2: Scale Statistics ........................................................................................ 103

16. Factor 3 (Item s 1-5): Sam ple Size Summ ary......................................................... 104

17. Factor 3: Item Statistics ......................................................................................... 104

18. Factor 3: Summ ary Item Statistics......................................................................... 105

19. Factor 3: Item -Total Statistics............................................................................... 105

20. Factor 3: Scale Statistics ........................................................................................ 106














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

THE CURRENT STATUS OF MARRIAGE AND FAMILY THERAPISTS'
GRADUATE TRAINING IN THE IDENTIFICATION, ASSESSMENT,
AND TREATMENT OF RELATIONSHIP VIOLENCE

By

Marie T. Bracciale

August 2004

Chair: Silvia Echevarria-Doan
Major Department: Counselor Education

The incidence of relationship violence (RV) has increased dramatically in the past

decade. Clinical counselors trained in marriage and family therapy (MFT) who interact

with perpetrators and victims of RV must be informed about associated personality

factors, transactional dynamics of perpetrators and victims, structural and environmental

influences leading to RV, and comparative effectiveness of intervention strategies. This

national study focused on graduate MFT training received by clinical members of the

American Association of Marriage and Family Therapists (AAMFT).

The Relationship Violence Training Survey (RVTS), designed by the researcher

and measured for content validity by experts in the field, contained two subscales: (a)

assessment of RV, and (b) training/treatment issues in RV. Demographic data included

age, ethnicity, gender, years of therapy experience, and years of supervisory experience.

The RVTS was designed to measure whether graduate MFT programs adequately prepare

therapists in assessment and treatment of RV. Program success was measured by survey








responses regarding required coursework in RV, program endorsement by accreditation

standards, and practitioners' self-reported self-efficacy in assessing and treating RV.

Respondents were 197 clinical members of AAMFT, AAMFT approved supervisors, and

faculty members in MFT training programs, obtained via nationwide random sampling

(response rate of 19.7%).

Statistical analysis of responses to the RTVS indicated three factors: (a)

respondents' rating of the importance of clinical competencies associated with the

identification, assessment, and treatment of RV; (b) respondents' rating of their graduate

training in RV; and (c) respondents' self-rating of their knowledge and skills in

identification, assessment, and treatment of RV. Cronbach alphas for the three factors

were .814, .967 and .812, respectively.

The study results indicated problems in the ability of responding practitioners to

(a) use systematic risk assessments to recognize imminent danger and formulate

appropriate interventions, (b) intervene within violent gay and lesbian relationships, and

(c) obtain Restraining Order Injunctions. The need for improved quality and longer

duration of graduate training in RV and more stringent training requirements by licensing

boards and accreditation standards were supported. The results may be helpful to

researchers, treatment providers, graduate programs, accreditation boards, third-party

payers, and benefits officers.














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

THE CURRENT STATUS OF MARRIAGE AND FAMILY THERAPISTS'
GRADUATE TRAINING IN THE IDENTIFICATION, ASSESSMENT,
AND TREATMENT OF RELATIONSHIP VIOLENCE

By

Marie T. Bracciale

March, 2004

Chair: Silvia Echevarria-Doan
Major Department: Counselor Education

The incidence of relationship violence has increased dramatically in the past

decade. This national study focused on training that clinical members of the American

Association of Marriage and Family Therapists (AAMFT) received during their graduate

marriage and family therapy (MFT) programs.

The Relationship Violence Training Survey was designed to measure whether these

programs prepare therapists in assessment and treatment of relationship violence.

Respondents were 197 clinical members of AAMFT, AAMFT approved supervisors, and

faculty members in MFT training programs, obtained via nationwide random sampling.

Results may be helpful to researchers, treatment providers, graduate programs, licensure

boards, accreditation standards, third-party payers, and benefits officers.













CHAPTER 1
INTRODUCTION

Whether or not human beings are inherently aggressive, as some have asserted

(Lorenz, 1966), it appears that they are most likely to behave aggressively in their most

intimate social relationships. Almost one fourth of all murders occur between relatives,

most often involving spouses killing one another (Straus, 1986). Child abuse, spouse

abuse, and elder abuse have become substantial problems in their own right. Date rape

and courtship violence occur with distressing frequency. Violence between homosexual

couples also has been documented in the literature. No type of interpersonal relationship

seems to be immune (Rosenbaum, Cohen, & Forsstrom-Cohen, 1991). In 2000, the

number of females shot and killed by a husband or intimate acquaintance was nearly 4

times higher than the number murdered by male strangers (Centers for Disease Control

and Prevention, 2001).

Why should one conduct a national survey of practices in assessment and

treatment of relationship violence? Due to the large increase of violence in society (as

the statistics will show), this survey is needed to respond to this incidence. Ethically, we

need to respond as an organization to have an impact on our clients, our research, our

treatment protocols, and our community/courts systems.

Very few national studies have been done on marriage and family therapist

practices. None has directly surveyed practices directly relating to relationship violence.

The results will be information of importance to researchers in the field, treatment








providers, graduate programs, and accreditation boards. It may also be helpful to third-

party payers and benefits officers.

In 1995 Doherty and Simmons conducted the first national survey of marriage and

family therapists on clinical patterns of marriage and family therapists. Their sample of

526 therapists from 15 states gave descriptive information on their training, level of

experience, and professional practices, along with detailed information on recently

completed cases. The findings indicated that marriage and family therapists treat a wide

range of serious mental health and relational problems, that they do so in relatively short-

term fashion, and that they use individual, couple, and family treatment modalities

(Doherty & Simmons, 1996).

The goal of this project is to obtain detailed information about the clinical practices

of clinical members of the American Association of Marriage and Family Therapists

(AAMFT) throughout the United States. Specifically, the Relationship Violence Training

Survey (RVTS) is designed to answer questions about the adequacy and competencies of

marriage and family therapists, supervisors, and faculty members in identification,

assessment, and treatment of relationship violence and to look at the current status of

graduate training in relationship violence (appendix A).

The theoretical basis for this study is drawn from (a) the increase in incidence of

relationship violence as indicated by statistics; (b) professional responsibility: legal,

ethical, and therapeutic issues; (c) multiforms of treatment (metatheoretical,

postmodernism, feminism, sociocultural, social learning); and (d) identification of

training needs.








Increase in Incidence of Relationship Violence

In the United States today, violent crimes occur more frequently within families

than among strangers (Gelles & Straus, 1988; Langan & Innes, 1986; Straus & Gelles,

1990). Government surveys conducted between 1973 and 1981 identified 4.1 million

reports of intrafamilial victimizations (U.S. Department of Justice, 1984). In 1990 alone,

of the 6,008,790 crimes of violence against persons ages 12 and older reported in a

national crime victimization survey (U.S. Department of Justice, 1992), 39% were

perpetrated by a member of the victim's family or by a person in a relationship with the

victim. Less than half of these crimes were reported to the police. Fifty-eight percent of

crimes reported involved the spouse or ex-spouse of the victim.

Although intimate partner violence (IPV) is known to occur among all social

classes, research over the past 30 years has demonstrated a consistent link between low

socioeconomic status (SES) and occurrence of partner violence (Gelles, 1997). In terms

of clinical studies, O'Brien (1971) found that, in a sample of divorcing couples, the

husband's achievement status (measured by job dissatisfaction, education, income, and

occupational status) was much lower among the violent couples than among the

nonviolent couples. Gayford (1975) and Roy (1977) found that husbands of battered

women tended to hold unskilled occupations and have high levels of unemployment.

Hotaling and Sugarman (1986), in their review of 52 case comparison studies, found that

three indicators of the husband's SES (occupational status, income, and educational

level) demonstrated a consistent inverse association with the occurrence of husband-to-

wife violence (Cunradi, Caetano, & Schaefer, 2002).

Data from the 1992-93 National Crime Victimization Survey indicate that young

women (ages 19-29) in low-income families (under $10,000) were more likely than other








women to be victims of partner violence (Bachman & Saltzman, 1995). Longitudinal

research suggests that men characterized by low SES indicators (unemployment, low

educational attainment) are more likely to initiate IPV than men without such

characteristics (Magdol et al., 1977), and these indicators are associated with the

persistence of wife assault over time (Alderondo & Sugarman, 1996). On the other hand,

Quigley and Leonard (1996) found no significant differences in education between

couples whose husbands desisted their marital violence after 2 years of follow-up and

those who did not (Cunradi, Caetano, & Schaefer, 2002).

Professional Responsibility: Legal, Ethical, and Therapeutic Issues

The duty to warn third parties of dangers posed by a client's behavior, as Monahan

(1993) noted, "is now a fact of professional life for nearly all American clinicians" (p.

242). Family therapists, when deciding whether to warn a third party of a threat resulting

from a client's behavior, must consider legal, ethical, and therapeutic issues. On the one

hand, courts in most states have ruled that therapists have a legal duty to warn (or avert

danger to) third parties to whom clients pose a risk. On the other hand, certain sorts of

client confidences are protected by law. Thus, therapists must determine whether they

have a legal duty to warn third parties of the danger posed by their client's behavior and,

if not, whether the client's information is legally protected (thus prohibiting the therapist

from disclosing the information). Ethical and therapeutic issues arise about client

information that state law does not expressly require or forbid therapists to disclose. In

such cases, therapists and clients are generally free to negotiate a mutually agreeable

disclosure policy.

All three factors (legal, ethical, and therapeutic) play a role in formulating the

written disclosure policy statement that each therapist should negotiate with prospective








clients. Some states regulate disclosure statements. In any case, therapists' disclosure

policies must conform to all applicable laws.

The landmark case establishing that therapists, in certain cases, have a legal duty to

warn potential victims of their clients' behavior is Tarasoffv. Regents of University of

California (1976). The details of the case are well known. In the course of therapy,

Prosenjit Poddar threatened to kill Tatiana Tarasoff, a fellow student in his square

dancing class. (Although Tarasoff was not mentioned by name, the therapist was well

aware of the identity of the potential victim.) Poddar was held for observation and then

released. No one notified Tarasoff of the threat that Poddar had made against her. Two

months later, Poddar murdered Tarasoff, and the Tarasoff family filed suit. The court

held that "when a therapist determines or pursuant to the standards of his profession

should determine, that his patient presents a serious danger of violence to another, he

incurs an obligation to use reasonable care to protect the intended victim against such

danger" (p. 346).

A disclosure statement is a written document detailing the policy, negotiated

between therapist and client, concerning therapist disclosure of client information.

(Frequently, the document expresses the standard policy of the therapist and is signed by

the client.) It should inform the client of the therapist's legal responsibilities and indicate

how the therapist will use discretion within the limits of the law (Schlossberger &

Hecker, 1996).

The legal evolution of the "battered women's defense," as it came to be known,

was built on the "rape defense" successfully argued by lawyer Susan Jordan in the second

trial of Inez Garcia and the successful appeal filed by Jordan, Nancy Stearns, and Liz








Schneider for Yvonne Wanrow, the Spokane woman hobbling on crutches who killed a

man for molesting her son. Armed with new, and newsworthy, feminist concepts-

"battered women's syndrome," "unequal combat," and "imminent danger"-their defense

strategies helped to focus a spotlight on domestic violence. After nearly a decade of

feminist agitation in concert with legislative initiatives pioneered by Representative

Lindy Boggs and Senator Barbara Mikulski, Congress passed the Family Violence

Prevention and Services Act in 1984. Today, approximately 1,800 battered women's

shelters, hot lines, and advocacy programs around the country are funded by the federal

program (Brownmiller, 1999).

Ethical Standards of the Commission on the Accreditation
of Marriage and Family Therapy Education (COAMFTE)

The Commission on Accreditation for Marriage and Family Therapy Education

was established by the AAMFT Board of Directors in 1974. In 1978, the Commission

gained official recognition by the U.S. Department of Education as the accrediting

agency for the graduate degree and post-degree training programs in marriage and family

therapy. These training programs are located throughout the United States and Canada.

In 1978, in recognition of its increased level of activities and responsibilities, the

Commission was restructured and renamed the COAMFTE.

The Commission on Recognition of Post-secondary Accreditation (CORPA)

officially granted recognition to the COAMFTE in 1994. CORPA is a nongovernmental

organization that works to foster and facilitate the role of accrediting bodies in promoting

and ensuring the quality and diversity of American postsecondary education.








COAMFTE serves under a broad mandate from the AAMFT Board of Directors to

set standards for and accredit master's, doctoral, and post-degree clinical training

programs in marriage and family therapy (COAMFTE, 1994).

Specific standards developed by the COAMFTE (appendix B) outlining the

importance of this study include the following:

320: Area II: Clinical Knowledge

320.02 Area II content will address contemporary issues, which include but are
not limited to gender, violence, addictions, and abuse, in the treatment of
individuals, couples, and families from a relational/systemic perspective.

340.04 Area IV will address the AAMFT Code of Ethics [appendix C],
confidentiality issues, the legal responsibilities and liabilities of clinical practice
and research, family law, record keeping, reimbursement, and the business aspects
of practice. Area IV content will inform students about the interface between
therapist responsibility and the professional, social, and political context of
treatment.

Accreditation Bodies

COAMFTE Preamble to the Standards on Accreditation, Version 10.2

Accreditation is a voluntary process on the part of the program whose major
purpose is to ensure quality in a marriage and family therapy program. All
accredited programs are expected to meet or exceed all standards of accreditation
throughout their period of accreditation. The integrity of an institution and the
program is fundamental and critical to the process of accreditation. Accreditation
standards are usually regarded as minimal requirements for quality training. All
accredited programs are free to include other requirements, which they deem
necessary and contribute to the overall quality of the program. Programs must
continually evaluate their programs in relation to their institution's mission and
their own program mission, goals and educational objectives. Accreditation
standards, like other aspects of accreditation, are part of a slowly evolving,
continuous process. In the long view, there are continuing conversations among
accreditors, training programs, trainees, trained professionals, employers, and
consumers from which the standards and other aspects of accreditation evolve.

The objective of these standards is to assure, as much as possible that individuals
trained in accredited programs are competently trained to become marriage and
family therapists at the entry and doctoral levels. The standards are designed to be
unique to the practice and supervision of marriage and family therapy. Some
standards apply to training programs in general, including elements such as








organizational stability, faculty accessibility, appropriate student selection
processes, and fairness to students and employees. Some standards apply to all
psychotherapy training, including elements such as adequate numbers of client
contact hours and supervision hours.

Graduates from COAMFTE accredited marriage and family therapy programs are
trained to be clinical mental health practitioners. COAMFTE adopts the Standard
Occupational Classification of the Bureau of Labor and Statistics which states that
Mats are qualified to "[d]diagnose and treat mental and emotional disorders,
whether cognitive, affective, or behavioral, within the context of marriage and
family systems. [They] Apply psychotherapeutic and family systems theories and
techniques in the delivery of professional services to individuals, couples, and
families for the purpose of treating such diagnosed nervous and mental disorders."
All persons properly trained in marriage and family therapy are to be competent in
working with individuals. (American Association for Marriage and Family
Therapy, 2003)

Marriage and family therapists are bound ethically by the standards of accreditation

and clinical ethical guidelines to understand the importance of assessing for abuse and

violence with each case. In some cases, the reasons that abuse and violence are present

may be difficult to detect.

Scope of the Problem

Current researchers who report on the treatment of relationship violence have only

recently written that this type of treatment is evolving into a specialized field. Although

for many years there have been specialized treatment programs available in a few states,

it is only recently that the need has been recognized to develop and fund specialized

treatment programs in every state. This recognition has been spurred on by two factors:

(a) increasing and widespread acknowledgement of the extent of the problem, and (b)

increasing acceptance that traditional forms of therapy are not effective with the

relationship violence population (Crawford, 1981).

Two decades of empirical research on child abuse, wife beating, and domestic

violence are conclusive on one point: The causes of violence are multidimensional (Gil,








1971; Straus, Gelles, & Steinmetz, 1980). There is no one cause of this violence-not

poverty, not stress, not mental illness or psychopathology, not being raised in a violent

home, and not alcohol and/or drugs (Gelles & Maynard, 1987).

It is apparent that the treatment for a multidimensional problem would include

varied approaches, depending on the context. Straus's (1973) general systems model of

violence between family members was one of the first theoretical applications of a

systems perspective to family violence. Another presentation of a research systems

model is Giles-Sims's (1985) examination of wife battering. Giles-Sims's systems

models looked at the factors influencing a battered woman's decision to stay, flee, and/or

return to violent relationships.

The Extent of the Problem

Domestic violence is one of the most common crimes. Many of us know someone

in our close family or among our friends to whom it has happened, or we have

experienced it ourselves, but we tend to think that we are different or alone, not realizing

perhaps just how widespread and enduring domestic abuse is. Domestic violence occurs

in almost all cultures and countries, across all known divisions of wealth, race, caste, and

social class. There may never have been a time when it did not exist; it certainly

stretches back deep into history. Centuries, indeed millennia, are filled with millions of

assaults, attacks, rapes, violations, psychological abuses, maimings, killings of women in

their homes by men (Hague & Malos, 1998).

Although it is true that approximately 6 million women in the United States are

beaten in their homes each year, there is more to the domestic violence picture. For

example, 70% of male partners (batterers) also abuse children in the home. More tragic,

boys often attempt to protect their mothers from battering and are themselves injured or








killed. Incredibly, over 60% of males ages 15 to 20 who were incarcerated for homicide

had killed their mother's batterer (Summers & Hoffman, 2002).

Domestic violence is not a new phenomenon. It has been a common occurrence

throughout recorded history. In many societies, women were traditionally considered the

property of the man; his duty was to discipline her and the children (and slaves) with

thorough beatings. The only concerns about this related to the thickness of the stick that

the law allowed for the beatings. Although there were some earlier unenforced laws

against spousal abuse, it was only as recently as the 1970s in the United States that the

justice system began to view the problem seriously and consider domestic violence as a

crime. Up until that time, social services for victims of domestic violence were almost

nonexistent.

There are many contributing factors to domestic violence, usually associated with

differing views of what the problem is based on. Some of these views are learned

behavior, gender socialization, patriarchy or power and privilege, and risk factors,

including criminal or psychological profiles. Even countries view the nature of the

problem differently. For example, some of the risk factors for domestic violence in

England and Wales are gender inequality, poverty, social exclusion, having a criminal

background, and having experienced abuse as a child. The families are classified as

patriarchal (male dominated), and women have a subordinate status. This also seems to

be the case in Italy, where the view is held that religion keeps domestic violence as an

isolated and personal problem rather than a serious social ill. Both the Catholic church

and the state view domestic violence as a personal and private matter. Battered women

are encouraged to return home to their abusing partners. In their zeal to preserve the








family, domestic violence has often been overlooked and even considered "normal." In

Australia, most citizens see domestic violence as rooted in the aggressive nature of men.

Yet the indigenous population sees it as learned behavior. In Jamaica, the problem is

regarded as stemming from a lack of education, abject poverty, drug abuse, and the

mythology surrounding the traditional role of the male. In Russia, Germany, and

Slovenia, major political change is identified as another contributing factor (Summers &

Hoffman, 2002).

Overall, domestic violence presents a profoundly disturbing and distressing picture.

Some people choose to turn away from that picture. Partner abuse is an epidemic with

potentially dire consequences for individuals, families, and society. Family therapists

must be competent to assess for and intervene in abuse situations (Haddock, 2002).

Multiforms of Treatment (Metatheoretical, Postmodernism,
Feminism, Sociocultural, and Social Learning)

Various theoretical and treatment approaches are being used within the relationship

violence field. Four of these approaches are reviewed in this section.

Metatheoretical Approach

Metatheoretical, as a general term, refers to the philosophical (e.g., epistemic,

ontological, metaphysical) assumptions that influence or form the basic structure of

various disciplines in science and the social sciences (Hoshmand, 1996). When working

with batterers, the most therapeutic interventions in cases of relationship violence are still

based on behavioral approaches (Bagarozzi, 1983; Taylor & Gunn, 1984). Cook and

Franz-Cook (1984) presented a systematic treatment approach to wife battering.

Margolin (1979) and Taylor and Gunn proposed conjoint therapy for spouse abuse cases.

The radical feminist perspective and the systemic view of family violence are not








mutually exclusive, and Cook and Franz-Cook stated that treatment based on both views

is necessary and important.

Archer (2000), in a meta-analysis, compared samples selected for male violence

(from battered women's shelters) with community samples to assess whether the couple

violence looked different across these populations. Very high levels of male aggression

were reported in shelter samples, whereas in community samples women were slightly

more physically aggressive. Archer also examined studies of couples undergoing

treatment for marital problems and found that men were slightly more likely than women

to be physically aggressive. However, in contrast to shelter samples, the level of male

aggression was much lower. This suggests that couples receiving counseling, even for

problems specifically related to male violence, do not have nearly the same kind of

imbalance in physical aggression as might be found in couples in which the woman has

sought shelter from abuse (Greene & Bogo, 2002). This study will describe current

empirical research to support a conceptual framework for helping marriage and family

therapists to assess and treat relationship violence. Depending on the context, flexibility

in searching for approaches when working with the needs of couples may be indicated. A

broader lens that takes into account the various faces of intimate violence may expand

alternatives for assessing and treating these couples (Greene & Bogo, 2002).

Postmodernism

Postmodern perspectives have had considerable impact on the field of couple

therapy in the 1990s. Focused on self-organizing and proactive features of human

knowing, they emphasize that reality is constructed, reflecting language, culture, and

social context (Anderson, 1997; Neimeyer, 1993). Meaning and knowledge are seen as

being created through social communication with others. The most radical forms of








constructionism see every case as unique and suggest that no single version of reality or

problem formulation is better than another. Problems are viewed as "interpretations" that

can be "dis-solved" in language (Anderson).

Specific techniques have been developed in solution-focused and narrative

therapies to help clients to "deconstruct" the problematic aspects of their relationship and

allow new possibilities to emerge. However, there are many ways to help clients to

create new meanings and many ways to access and work with aspects of experience that

have gone "unstoried." More generally, this perspective may be viewed as an "attitude"

or philosophical stance for therapy rather than as a model for intervention or a set of

techniques.

From a respectful, collaborative stance, therapists regard clients as experts on their

own reality and discover with clients how they construct that reality. Therapists show

sensitivity to each individual and enlarge the frame to include larger contextual issues,

such as gender, class, and culture. Therapists also focus more on a couple's strengths and

competencies, striving to honor and validate clients' wisdom and strengths in dealing

with difficult realities. Social-constructionist ideas can also be integrated with more

traditional research if certain guidelines are followed (Myers Avis, 1996); for example, if

researchers recognize and reveal their own values and beliefs with the research context.

Sociocultural Theory

The search for the causes of domestic aggression has focused largely on

sociocultural and psychological factors. It has been a short search, the primary strategy

of which has been to identify characteristics of participants that distinguish them from

their nonaggressive counterparts. It has been an theoretical search in which theory is

occasionally invoked, post hoc, to explain one or another research finding. Social








learning theory, for example, is used as an explanation for the intergenerational

transmission of aggression, and female masochism is sometimes employed to account for

the battered woman's reluctance to leave an abusive mate (Rosenbaum & O'Leary,

1981).

Sex-role socialization, in general, tends to support the notion that the success or

failure of intimate relationships is the woman's responsibility, and this may lead some

women to make great efforts to stay in intimate relationships, even after episodes of

abuse, to show their commitment to their partner and to weathering the "difficult times"

together (Dutton & Painter, 1981; Strube, 1988). In addition, when an abusive event

occurs, the woman may presume that it will not recur, and so will "try to make the

relationship work under the belief that, if she tries hard enough, her efforts will succeed"

(Strube, p. 240).

Sociological and sociocultural theory assumes that social structures affect people

and their behavior. The major social structural influences on family violence are age,

gender, position in the socioeconomic structure, and race and ethnicity (Gelles, 1983).

Social Learning Theory

According to Bandura's (1973, 1977) social learning analysis of aggression,

witnessing interparental violence may predispose some young males to abusive behaviors

in their adult intimate relationships with women. Social learning theory maintains that

violence (in the form of a learned response) is transmitted from the family of origin to the

adult intimate dyad through the vicarious reinforcement of interpersonal violence as a

method of conflict resolution and a means to the maintenance of power and control in

intimate relationships.








In addition to external reward/punishment contingencies, Bandura (1973)

proposed that the following self-regulatory mechanisms modulate self-recrimination

processes by "neutralizing" aggressive behaviors: (a) justification of the behavior on the

basis of some higher authority (e.g., scripture); (b) comparison of the behavior with more

serious violence; (c) projection of responsibility for the behavior onto drugs, alcohol,

work stressors, or the provocation of the victim; (d) normalization of the behaviors as a

common and socially acceptable occurrence; (e) depersonalization of the victim through

the use of disparaging labels; and (f) minimization of the consequences of the behavior.

One or more of these neutralizing tactics have been observed among batterers in

treatment (Carden, 1994).

Treatment for Battered Women

Varieties of counseling approaches have been proposed for battered women in

recent years, including grieving, existential, and shame therapy (Turner & Shapiro,

1986). Some of these approaches are reviewed in this section.

Empowerment and safety-based interventions have been found useful. In her text

Counseling Female Offenders: A Strength-Restorative Approach, Katherine Van

Wormer (2001) established a link between the crimes of female offenders and

environmental factors such as substance abuse and sexual abuse. Combining strategies

from the fields of criminal justice and social work, she showed how to empower female

offenders and how to rehabilitate them to society by building on their personal strengths.

From her unique "strengths-restorative" approach, the author presented strategies for

anger management, substance abuse treatment, and domestic violence counseling.

In 2000, the University of Northern Iowa applied for and received a federal grant

to combat gender-based violence in a comprehensive manner. The planning and








implementation of the grant used an interdisciplinary approach, linking pre-existing law

enforcement, prevention programs, and victim services while adding a variety of new

tools and efforts. Below are some of the new and expanded programs offered as part of

the Violence Against Women grant. Some of the programs that they have developed are

RAD-Rape Aggression Defense Class and a Blue Light program, with five blue light

safety phones spread throughout the campus. These are some examples of safety

prevention programs that are effective and work well to decrease violence on campuses

(University of Northern Iowa, 2003).

Fundamental to the counseling efforts should be a design to move the battered

woman from status as a "victim" to that of a "survivor" (Rieker & Carmen, 1986). It is

this shift in self-perception that is most associated with safety and recovery (Gondolf &

Fisher, 1988).

The Duluth Model (Minnesota Program Development, 2003) contacts partners of

offending men and offers advocacy, community resources, and a women's group. The

model uses a curriculum called In Our Best Interest: A Process for Personal and Social

Change for their battered women's group. Women who have been arrested for using

violence are also ordered to attend nonviolence classes.

According to several victimization studies, battered women tend to move through

several phases in response to abuse (Ferraro & Johnson, 1983; Mills, 1985). These

phases are distinguished by an attributional shift on the woman's part. In essence, she

begins to perceive that the battering was not "all her fault" but was largely due to her

husband's behavior. It is not up to her to change the batterer; in fact, it is not likely that








he will change. Instead, she is capable of taking care of herself, with the support and

assistance from others that she deserves (Gondolf & Fisher, 1991).

Therefore, the objective in counseling might be to reinforce and encourage this

realization. Many shelters subscribe to an "empowerment" mode of counseling to

achieve this end. The feminist approach is directed toward helping the woman to realize

her options and choices and to begin to make decisions that assure her worth, integrity,

and determination (Bograd, 1988). One study of formerly battered women rated this sort

of counseling to be the most effective in stopping violence (Bowker, L., 1983).

Some clinicians (e.g., Almeida & Durkin, 1999) who are sensitive to issues of

power, abuse, and trauma have recommended the use of gender-specific support groups

for the victim as the most appropriate treatment strategy. It can be noted that a

combination of individual and group treatment is often desirable.

Some theorists have argued against the use of couples therapy in situations of abuse

(Avis, 1992; Bograd, 1992; Dutton; 1992). However, feminist-identified family

therapists have begun to experiment cautiously with the use of couples treatment in

situations of abuse (Goldner, Penn, Sheinberg, & Walker, 1990; Jenkins, 1990; Jory &

Anderson, 2000). Some therapists are using proposed criteria to determine situations in

which couples therapy may be appropriate. These indicators, listed by Bograd and

Mederos (1999), rule out the appropriate use of couples therapy. Since the issues of

partner abuse situations are complex, therapists should resist formulaic approaches to

treatment planning. Relevant variables should be carefully considered in making

treatment decisions, such as the power differential between the partners, the nature and

extent of the abuse, lethality indicators, the effects of the abuse on the victim, the ability








of the victim to stay safe, the resources of the victim, the responsibility taken by the

perpetrator, and the commitment to change demonstrated by the perpetrator (Haddock,

2002).

Treatment for the Batterer

The treatment of batterers has similarly seen a proliferation of approaches and, with

it, increased debate. The leading programs are characterized by group process that

prompts men to take responsibility for their abuse, to exercise alternatives to the violence,

and to restructure their sex-role perceptions (Gondolf, 1987b). However, there is a

questionable trend toward short-term anger control treatment that unwittingly reinforces

the batterer's penchant for control (Gondolf & Russell, 1986).

The research on cessation suggests that batterers who reform their behavior pass

through a series of developmental stages (Fagan, 1987; Gondolf, 1987a). The change

process begins with "realization." The egocentric batterer acknowledges the

consequences of his abuse and that it may be in his own self-interest to contain the anger

that led to the abuse. Gradually, the batterer becomes more "other oriented" and begins

to make "behavioral changes" to improve relationships, or at least to avoid totally

destroying them. Some men eventually begin to think more in terms of values and

principles and integrate these into a change of self-concept. Consequently, a number of

leading batterer programs employ a phased approach that moves batterers from didactic

sessions of accountability and consequence to social support groups with a focus on

service (Gondolf, 1985).

When treating batterers, marriage and family therapists should be aware of the

heterogeneity among batterers across several dimensions, including the severity of the

violence and the psychopathology and physiological responses of the batterer








(Holtzworth-Munroe, Smutzler, Bates, & Sandin, 1996; Jacobson & Gottman, 1998).

Understanding the differences between two types of batterers-Type 1 ("cobras") and

Type 2 ("pit bulls")-can also be useful (Gottman et al., 1995; Jacobson & Gottman).

These typologies are described in the literature review section of this dissertation.

According to Haddock (2000), novice therapists should be advised against treating

most perpetrators of abuse; indeed, for court-ordered batterers, most states mandate

certain kinds of treatment by certified professionals. Therapists should be familiar with

local agencies that serve batterers and the methods for making referrals to local batterer-

specific treatment programs.

Duluth Model

In 1981 nine city, county, and private agencies in Duluth, Minnesota, adopted

policies and procedures that coordinated their intervention in domestic assault cases.

These measures focused on protecting victims from continued acts of violence by

combining legal sanctions, nonviolence classes, and, when necessary, incarceration to

end the violence. Consistently applied, their message to offenders is clear: "Your use of

violence is unacceptable."

The Domestic Abuse Intervention Project (DAIP) was the coordinating agency for

this effort. An additional component of the DAIP was the nonviolence program, which

consisted of classes for offenders who were court-ordered to the programs. The programs

used the curriculum "Power and Control: Tactics of Men Who Batter," a 24-week

educational curriculum (Pence & Paymar, 1993).

Batterer intervention programs, which seek to educate or rehabilitate known

perpetrators of IPV to be nonviolent, have proliferated since the 1980s under the auspices

of both the criminal justice system and the mental health system. Three theoretical








approaches to the conduct of these programs have been consistently documented (Healey,

Smith, & O'Sullivan, 1997): society and culture, the family, and the individual. These

theories influence the content and delivery of interventions.

Society and Culture

Feminist theorists attribute battering to social and cultural norms and values that

endorse or tolerate the use of violence by men against their women partners. The

feminist model of intervention educates men concerning the impact of these social norms

and values and attempts to resocialize men through education, emphasizing nonviolence

and equality in relationships.

It has been well documented by feminist researchers that gender is a central

organizing principle for both individuals and couple relationships and therefore must be

an integral feature in family therapy (Goldner, 1985; Hare-Mustin, 1986). More research,

analysis, and understanding are needed regarding how intersecting factors such as gender,

class, race, and ethnicity operate in cases of couple violence.

We must acknowledge the limits of generalizations that can be made on the basis of

populations included in research to date. Most couple therapy clients are White and

middle class. The considerable influence of cultural diversity and changing gender roles

is, as yet, largely unexamined (Johnson & Lebow, 2000). The adaptation of couple and

family therapy to consider the impact of culture will be a vital concern of future research

in the field.

The Family

Family-based theories of IPV focus on the structure and social isolation of families.

The family systems model of intervention focuses on communication skills, with the goal

of family preservation, and may use couples counseling/conjoint therapy. Wife battering,








like other forms of family violence, raises a variety of family issues. But, unlike child

and elder abuse, it threatens the very foundation of the family structure-the marriage (or

partnership). Therefore, the most crucial family issue is whether the family is to

continue. Given the tendency of wife battering to escalate and denial of the problem to

persist, most practitioners in the field have, in the past, strongly recommend separating

the batterer from the battered women and children (Gondolf & Fisher, 1991).

The Individual

Psychological theories attribute perpetration of IPV to personality disorders, the

batterer's social environment during childhood, or biological predispositions.

Psychotherapeutic interventions target individual problems and/or build cognitive skills

to help the batterer to control violent behaviors.

Summary

Currently, there is little evidence to suggest the effectiveness of one approach over

another or of the differential effectiveness of different programs with different "types" of

batterers, although one study has suggested that process-psychodynamic groups may

function better for men with dependent personalities and cognitive-behavioral groups

may be more effective for those with antisocial traits (Saunders, 1996).

The most widely evaluated intervention model for men who batter is that of group

interventions using cognitive-behavioral techniques, often in combination with feminist

content. One review of these studies reported that rates of successful outcomes (i.e.,

reduced or no reassault) from these programs varied from 53% to 85% (Tollman &

Bennett, 1990). However, other reviews have pointed out those methodological problems

in the studies limit conclusions about the effectiveness of such programs.








Graduate Training in Relationship Violence:
Identification of Training Needs

The importance of intervening in cases of so-called "minor" spousal violence is

underscored by the assumption by many researchers that minor violence, if left untreated,

can escalate into severe or life-threatening violence (Rosenberg, 1985). In most graduate

programs, this area of assessment is poorly presented, unless graduate students

specifically discuss relationship violence during intake. Research suggests that, unlike

graduate students, emergency personnel are well trained in this area of assessment, since

these personnel are primarily working the "front lines" with respect to relationship

violence victims. These personnel have incorporated violence assessments into their

intake procedures.

According to Wolf-Smith & LaRossa (1992), professional counselors, therapists,

and social workers have an obligation to help victims to gain insight into their abuse.

They also have an obligation to be nonjudgmental of whatever decision a woman makes

about her abusive relationship. Professionals provide varied services to victims and their

families to help heal the effects of violent relationships. Treatment providers may

recommend many different approaches, including individual, group, and/or family

therapy. However, the specific treatment approaches that are currently being taught in

graduate school training for counselors are unknown. Furthermore, it is unknown

whether the treatment recommended varies depending on the context of the case. Are

professionals being trained to treat the victim? the batterer? the couple? the individual?

the family? Is there coursework from a relationship-violence context or through a

relational violence lens? This information would allow the professional counselor to gain

knowledge in treating clients who present with issues of relationship violence.








"Victims of male battering face difficult choices-choices about what to say to

their abusers, choices about whether to stay. Respecting the choices that women make is

an integral part of the counseling/therapeutic process" (Wolf-Smith & LaRossa, 1992,

p. 324).

Understanding the tendency to reduce the inherent complexity of partner abuse

cases in response to concerns about potential lethality, therapists would benefit by

learning to conceptualize each case within its own unique and multifaceted context

(Dutton, 1992; Goldner, 1999), while attending to the intersections of gender, race, class,

religion, and sexual orientation (Bograd, 1999).

Goldner's (1992, 1999) "both/and" stance can be introduced as a way to manage

these complexities.

There is an enormous pressure to "get it right" immediately and, as a result, the
impulse is to lapses into extremes: to side with one partner against the other, to
refuse to ever take sides at all, to exaggerate or minimize danger, to insist on a
particular paradigm and argue against all others-in other words, to polarize
everything. (Goldner, 1992, p. 56)

Part of the difficulty for faculty and marriage and family supervisors in training

students effectively has been that the domestic violence literature includes disparate and

controversial findings. On the one hand, studies of community samples find generally

low levels of violence perpetrated by both males and females. On the other hand, studies

of clinical samples drawn from courts, hospitals, and shelters find severe violence,

mainly perpetrated by men (Archer, 2000; Johnson, 1995). For instance, feminist

researchers have studied primarily clinical samples and have concluded that intimate

violence is the result of patriarchy and, thus, is primarily perpetrated by men as a means

to maintain power and control (Dobash & Dobash, 1979; Pagelow, 1984). Family

conflict researchers have studied mainly representative community samples and have








concluded that intimate violence between partners results from individual, relational, and

societal variables that tend to be more gender neutral (Berkowitz, 1993; Straus & Smith,

1990). These two different perspectives have led to a longstanding debate about the

veracity of each position, which impacts training in terms of approaches taken to address

the identified problem.

The majority of studies on couple violence have limitations, notably largely using

samples involving only severely violent men. Conclusions and conceptualizations about

violence and appropriate clinical interventions have been generalized from these samples

to all couples in which there is aggression (Johnson & Ferraro, 2000). Comparably little

research has involved couples voluntarily seeking conjoint treatment for intimate

violence (Brown & O'Leary, 1997). Despite growing evidence of difference between

these populations, distinctions have yet to be included in assessment (Greene & Bogo,

2002).

Assessment of Violence in Treatment

For those professionals using traditional assessment methods, rather than

assessments geared toward gleaning information about relationship violence, clients may

be allowed to "maintain the silence," since appropriate questions are not asked initially.

Without vital skills in assessing relationship violence specifically, counselors are more

likely colluding with the system to maintain the violence and thus may be putting their

clients at risk. It is unclear why the majority of counselors are not trained in this

important area of assessment.

Previous research establishes a precedent for needed improvements in the education

and training of psychotherapists. For example, Hansen, Harway, and Cervantes (1991)

surveyed the national membership of the AAMFT. Therapists participating in the survey








were asked to read two vignettes with proven therapeutic interventions for domestic

violence cases portraying female victims and domestic violence. The results indicated

that most of the counselors did not attend to the seriousness of the violence and many did

not attend to it at all. Indeed, one of the vignettes was based on an actual case study in

which a family member had been killed. In this survey of 362 members of the AAMFT,

respondents were asked to conceptualize the case and to describe how they would

intervene. Forty-one percent of those surveyed indicated no recognition of domestic

violence.

The use of various written assessments for detecting abuse and violence may prove

beneficial. One such instrument is the HCR-20: Assessing Risk for Violence (Version 2)

by Webster, Douglas, Eaves, and Hart (1997). Another assessment tool is the Clinical

Guidelines on Routine Screening published by the Family Violence Prevention Fund and

available at no cost from their Web site. This includes screening questions, history intake

form, abuse assessment screen, domestic violence screening/documentation form,

assessment of patient safety, referrals, reporting procedures, and photographs that might

be taken (Family Violence Prevention Fund, 2003)

To underscore the importance of assuming abuse with each case until ruled out,

Bograd and Mederos (1999) developed a comprehensive protocol for screening for abuse.

Therapists should also be informed of other written instruments, including multimodal

assessments of partner abuse (Alderondo, 1998; Alderondo & Strauss, 1994; Gottman,

1999).

In her 2002 text Seeking Safety: A Treatment Manual for PTSD and Substance

Abuse Najavits discussed safety plans as well as many other resources, including








individual and group therapy guidelines. Seeking Safety also provides clinicians with a

session format, including a check-in and check-out procedure that helps the client to

commit to safe coping strategies.

Dunford (2000), along with other researchers (Holtzworth-Munroe, Meehan,

Herron, Rehman, & Stuart, 2000; Saunders, 1996; Waltz, Babcock, Jacobson, &

Gottman, 2000), has suggested that treating physically aggressive men as one

homogenous group, rather than tailoring interventions according to the different

motivations and needs of physically aggressive men, could be responsible for the

ineffectiveness of treatment. Dunford urged therapists to give "full and preferential

attention" to the possibility that one-size-fits-all approaches to treatment may not meet

the needs of these couples (p. 475). Examining the effectiveness of distinguishing

between couples in various types of violent relationships and tailoring treatment

interventions accordingly constitutes a promising area for future clinical exploration and

empirical research (Greene & Bogo, 2002)

In the cases of common couple violence, intervention should maintain a dual and

simultaneous focus on both anger management and relationship building. Gottman's

(1999) research has also highlighted the importance of addressing issues beyond conflict,

such as strengthening the marital bond.

Statement of the Problem

Partner abuse is an epidemic with potentially dire consequences for individuals,

families, and society. As part of responding to this epidemic, researchers and clinicians

suggest that therapists should develop competence in the areas of assessing and treating

violence.








Intrafamilial violence has been documented in relationships of every race, religion,

social class, and educational level (Straus & Gelles, 1986; Straus et al., 1980; U.S.

Department of Justice, 1984, 1992). In response, theorists and practitioners have created

specialized treatment methods and programs for recovery from relationship violence

(Bagarozzi, 1983; Cook & Franz-Cook, 1984; Dobash & Dobash, 1979; Giles-Sims,

1981; Margolin, 1979; Straus, 1973; Taylor, & Gunn, 1984; Walker, 1979). However,

battered victims often enter counseling with a marriage and family therapist without

having knowledge about what type of treatment may be recommended.

Need for the Study

Data on the prevalence and seriousness of intimate partner violence vary, but they

unequivocally establish partner abuse as a widespread and serious problem among

heterosexual, gay, and lesbian couples (Bograd & Mederos, 1999; Carillo & Tello, 1998;

Renzetti, 1997). Haddock (2002) suggested that it is imperative that family therapists

possess the knowledge and skills to manage partner abuse cases effectively.

The most significant theories on treatment approaches are discussed and

highlighted in this dissertation. The results of this study will have broad implications for

curriculum development on the graduate level of counseling programs and graduate

institutes, mental health policy and codes of ethics, social policy, and, most important,

counselor efficacy in treatment of relationship violence.

If specific treatments are recommended more often because they are believed to

produce positive outcomes, clinicians may be able to utilize these available interventions.

Since family violence is multifaceted and multidimensional, this research may help

clinicians to determine alternative treatment modalities that will prove to be successful

when working with a relationship violence population.








Clearly, this information will help graduate programs to provide better and more

informed, theoretically based training on the topic of relationship violence, which in turn

will produce clinicians who can effectively assess risk factors for relationship violence,

effectively treatment plan, create safety plans with clients, and provide valuable services

to clients with problems in the area of relationship violence.

Data will be collected to answer the following questions:

1. How do MFTs rate themselves on their knowledge in the identification,
assessment, and treatment of relationship violence?

2. How do MFTs rate their graduate training in the identification, assessment, and
treatment of relationship violence?

3. How do MFTs rate the importance of specified competencies and skills in the
identification, assessment, and treatment of relationship violence?

Future credentialing boards may have to expand requirements regarding training

hours required in domestic violence training. By surveying professionals in marriage and

family therapy, this study will provide information that may assist credentialing boards,

graduate and institute training programs, and third party payers for this specific client

group.

Implications of this study are not limited to graduate training programs and

counseling professionals. Since this treatment issue is embedded in the context of legal,

educational, medical, emergency, social, and family services, all are affected by advances

in prevention and treatment programs for relationship violence.

The results of this study and its external validity will help to inform theoretical

constructs in the area, as well as the current status of graduate training in relationship

violence. Future research may focus on specific methodology and identification of which

techniques are most often recommended in these cases.








As a result of this research, graduate schools, program developers, and managed

care companies may want to create assessments that include questions addressing

relationship violence to better serve their client base. These providers may find it more

cost effective to screen thoroughly clients who may be in current danger for violence.

Mental health professionals across disciplines may be able to use the results of this study

to facilitate the prevention and treatment of relationship violence.

Purpose of the Study

The specific type of method of inquiry for this study is a Web-based survey

developed by the researcher to gain information on factors related to the current status of

graduate-level preparation on relationship violence within four main subtopics regarding

relationship violence: identification, assessment, treatment, and training.

The primary purpose of this study is to determine the factors most frequently

reported by MFTs to be related to the identification, assessment and treatment of violence

via (a) a self-rating scale of their knowledge in the identification, assessment, and

treatment of relationship violence, (b) a self-rating scale of their graduate training in the

identification, assessment, and treatment of relationship violence and (c) their rating of

the importance of competencies and skills in the identification, assessment, and treatment

of relationship violence. The populations to be sampled are licensed marriage and family

therapists, approved marriage and family therapist supervisors, and marriage and family

therapy faculty.

The results of this survey would be helpful to the field of marriage and family

therapy, as it will help to inform researchers regarding the current status of theoretical

constructs currently being used in the field. The results of this survey would also be

helpful in looking at clinical patterns that may be generalizable to the larger population of








marriage and family therapists, accreditation boards, and marriage and family therapist

faculty and training programs.

The variables in this study are (a) the quality of graduate training in the

identification, assessment, and treatment of relationship violence as reported by MFTs,

and (b) competencies in the identification, assessment, and treatment of relationship

violence as reported by MFTs. Additional variables are (a) the MFTs' self-rating of their

knowledge in the identification, assessment, and treatment of relationship violence;

(b) their ratings of the adequacy of their graduate training in relationship violence; and

(c) their ratings of the importance of specified competencies and skills in the

identification, assessment, and treating of relationships violence. Specific competencies

and skills to be rated by the respondents include assessing for relationship violence,

assessing imminent danger, performing assessments quickly, understanding batterers'

typologies, identifying clients' strengths, and recognizing the signs and symptoms of

relationship violence, their ability to adhere to ethical standards, their use of multicultural

mode, their ability to include identification of risk factors in their assessments of clients,

and their ability to recommend risk assessment manuals. Intervening or extraneous

variables are demographics items such as gender, age, marital status, sexual orientation,

race, discipline, years of experience, and accreditation.

The study will examine the following question: What factors are most frequently

reported by MFTs to be related to the identification, assessment, and treatment of

relationship violence?

Rationale for the Study

Given the magnitude of the problem of intrafamilial violence, it is reasonable to

assume that clinical counselors in hospitals, colleges, corporations, agencies, and private








settings already interact on a regular basis with perpetrators and victims. This especially

includes those working in the field of alcohol and other drug addictions. To treat

relationship violence effectively, it is essential that these practitioners be informed about

personality factors associated with spouse abusers; transactional dynamics of perpetrators

and victims; and the structure and comparative effectiveness of existing intervention

strategies. Counselors and researchers, as experts in psychosocial development,

personality dynamics, and change processes, have much to contribute in the way of

theoretically and empirically derived answers to the questions, "Why does he do it?"

"What will stop him?" and "Why does she stay?" (Carden, 1994).

Although some models derived from certain perspectives (systemic or feminist)

seem to "fit" better with certain types of violence, there is a growing recognition in the

family therapy field of the need to integrate various theoretical perspectives and practice

models for effective practice. Johnson and Lebow (2000) saw the trend toward

integration as a "sign of a maturing field that general principles and interventions become

delineated and applied in varying formats and context" (p. 32).

Lebow (1997) believed that integrative approaches have the potential to offer

greater flexibility, an increased repertoire of interventions, higher treatment efficacy, and

greater acceptability among clients. The last of these is particularly important in

domestic violence cases, given that many clinicians have noted that women and men

often want to be seen together (Goldner, 1999; Lipchick & Kubicki, 1996; Shamai,

1996). Shamai noted how the categorical dismissal of systemic principles in the treatment

of domestic violence, may be akin to "throwing the baby out with the bath water" (p.








202) and serves to detract from efforts to develop a more effective, broader range of

interventions.

Goldner and her colleagues at the Ackerman Institute (Goldner, 1998, 1999;

Goldner et al., 1990) spent 10 years developing an integrative treatment model for

intimate violence in which systemic and feminist perspectives inform and enrich one

another. They have articulated how each of these perspectives in isolation from the other

serves as an insufficient explanatory framework, and they highlight the need to move

from an either/or orientation to a both/and position. Over the years, a complex and

sophisticated "multisystemic" approach to treatment, consisting of several different

approaches-feminist, systemic, psychodynamic, narrative, neurobiological, and

behavioral-has been integrated under the guiding principle that one level of description

or explanation does not have to exclude another (Greenspun, 2000).

This research and its results will advance the existing knowledge in thie relationship

violence field as it is part of a programmatic research effort. That is, when the results of

the study are considered in relation to other concurrent and/or sequential studies, there

may be theoretical and/or practical applications to the marriage and family field of

research.

Definitions of Terms

For purposes of this study, relationship violence is defined as any unwanted

physical, sexual, emotional, or financial control or power between two partners who are

in an intimate relationship. Verbal aggression would be included in the definition of

relationship violence.

Several terms are used in the literature to identify and describe family members

affected by relationship violence. In particular, offender or perpetrator refers to the








person who has committed the crime of domestic violence. The terms survivor or victim

are utilized in this study to identify a male or female adult who has experienced violence

in an intimate relationship.

According to the Florida Statute 741.28-741.31,

Domestic violence refers to any assault, aggravated assault, battery, aggravated
battery, sexual assault, sexual battery, stalking, aggravated stalking, kidnapping,
false imprisonment, or any criminal offense resulting in physical injury or death of
one family or household member by another who is or was residing in the same
single dwelling unit.

Family or household member means spouses, former spouses, persons related by
blood or marriage, persons who are presently residing together as if a family or
who have resided together in the past as if a family, and persons who have a child
in common, regardless of whether they have been married or have resided together
at any time.

In feminist writing on the patriarchal structure and content of language, writers

stress that women's experience is silenced and made invisible by the lack of words with

which to name it (Daly, 1978; Spender, 1980). A major contribution of feminist social

action around sexual violence has been to provide or create new terms with which to

describe and name the experience. For example, the terms battered woman and sexual

harassment did not exist 20 years ago. Even if a name exists and is known, the way it is

understood can vary greatly. For example, feminists have challenged the limited

traditional definitions of forms of sexual violence by expanding the definition of rape to

include unwanted and/or forced intercourse between husband and wife and by including

psychological abuse and coercive sex in the definition of domestic violence. Limited

definitions tend to draw on stereotypes of forms of sexual violence, stressing particular

features and ignoring others (Kelly, 1988).

In this study, individual therapy is used to describe sessions with an individual

client alone, using individual techniques and theories of counseling. Conjoint therapy (or








couples therapy) is used to describe sessions with a couple or dyad together. Group

therapy is used to describe therapy involving three or more clients and a counselor.

Family therapy is used to describe a variety of counseling approaches that work with

family members together in a session with a counselor.

The term survey is used to include cross-sectional and longitudinal studies using

questionnaires or structured interviews for data collection with the intent of generalizing

from a sample to a population (Babbie, 1990). A variable is a discrete phenomenon that

can be measured or observed in two or more categories (Kerlinger, 1979). Psychologists

use the term construct interchangeably with the term variable, according to Kerlinger.

Variables could be gender, age, SES, or attitudes or behaviors such as racism, social

control, political power, and socialization. Because the phenomena vary (in two or more

categories), they are called variables (Creswell, 1994).

Organization of the Dissertation

Presented in chapter 2 is a review of the related literature, including an overview of

relationship violence and the training of family therapists. Individual, group, conjoint,

and family therapy treatment modalities are discussed as they apply to the population of

this study. Presented in chapter 3 is a description of the methodology for the study,

including the research design, sample, Internet survey, instruments, and data analysis.

Chapter 4 presents a demographic description of the participants and results of factor

analysis of the data. Chapter 5 presents an evaluation and discussion of the results,

identifies the limitations of the study, discusses implications of the findings, and presents

a conclusion to the study.













CHAPTER 2
REVIEW OF THE LITERATURE

To review what was stated in Chapter 1, this study will be organized around four

points of research and literature review. They are (a) increase in incidence of relationship

violence as indicated by statistics; (b) professional responsibility: legal, ethical, and

therapeutic issues; (c) multiforms of treatment (metatheoretical, postmodernism,

feminism, sociocultural, and social learning); and (d) identification cf training needs.

Increase in Incidence of Relationship Violence

Violence in America has reached epidemic proportions and is exceeding the

capacity and the responsibility of law enforcement alone to curtail it. Although still

unacceptably high, overall criminal violence statistics have declined in recent years.

However, the frequency and severity of violence against children, women, and the elderly

are increasing at alarming rates. Between 1986 and 1993, reported cases of child abuse

and neglect increased by 98%, while reported cases of elder abuse increased by 106%. It

has been estimated that one woman in the United States is physically abused by her

husband every 9 seconds (Heise, Ellsberg, & Gottenmoeller, 1999). Every day in the

United States, four women murdered by male partners. This horrific fact is made worse

by the realization that there are more women killed in acts of domestic violence in any 5-

year period than all of the Americans killed in the Vietnam War (Berry, 1998).

Preventing violence and providing appropriate treatment for the victims of family

violence are important concerns for the health care system and society. Training

professional counselors in the current approaches to treating family violence is an area








that has recently drawn much attention. Incorporating family violence coursework into

graduate training programs is an ethical decision with which each department is faced.

The incidence of graduating students who are ill prepared to handle complex cases

involving family violence does not serve future clients, the clinician, or the community in

which the violence is occurring.

Interpersonal relationship violence has a long history as a deep-seated social

phenomenon. Several social historians have documented the informal and formal

sanctions that have encouraged wife battering (Davidson, 1978; Martin, 1976; Pleck,

1987). In early 19th-century America, a husband was permitted to discipline his wife

physically without prosecution for assault and battery. The legendary "rule of thumb"

law derived from English common law eventually restricted the instrument of wife

beatings to a stick no thicker than the man's thumb. Only in the past 15 years have courts

finally considered wife battering to be a criminal offense. These historical circumstances

led several social scientists to explain that men batter women basically because they are

permitted and encouraged to do so (Gelles, 1983).

Such "selective inattention," as it has been called (Pleck, Pleck, Grossman, & Bart,

1978), has important social implications. It was not until the women's movement in the

1970s identified and responded to wife battering that it emerged as a "social problem"

(Tiemey, 1982). Prior to this time, social scientists, physicians, social workers,

psychologists, and clergy had virtually overlooked and even denied that wife battering

existed. This markedly contrasted the extensive professional involvement in the issue of

child battering (Finkelhor, 1983).








Feminists concluded that such negligence was a symptom of the sexist attitudes

that pervade our society and contribute to relationship violence (Martin, 1976).

According to the feminist analysis, wife battering is the rape, sexual harassment, incest,

and pornography to which women are disproportionately subjected. To address this

problem and compensate for what other social services and the criminal justice system

have largely shunned, nearly 1,000 women's shelters were established across the country,

largely through the grass roots efforts of the women's movement of the 1970s (Schechter,

1982).

While differences remain over the definition of battering and its dynamics, there is

consensus that social services have inadequately responded to the problem. Clergy have

been accused of promoting compliance and submission to the abusive man (Horton,

1988); physicians have tended to identify battered women as "troublesome" (Kurz,

1987); police have, for the most part, taken a "hands off" approach (Dolon, Hendricks, &

Meagher, 1986); psychiatric staff are inclined to overmedicate battered women and return

them to their spouses (Gondolf, 1990).

These deficiencies are reflected in the observations of battered women. A survey

of formerly battered women rated women's shelters to be the most effective avenue in

helping to end the violence. Lawyers were the next most helpful. Other forms of social

service were, on the whole, rated as less than satisfactory (Bowker, L., 1983, 1986).

Partner violence has many causes. This is one of the reasons that there is no single

solution. The major points discussed in this chapter are viewed through a multitude of

lenses, including looking at power, sociocultural factors, worldwide prevalence of

relationship violence, feminist views, social learning theory, and the social structure of








our society. Treatment philosophies and batterer's typologies are also discussed. Current

literature is reviewed to identify risk factors for relationship violence.

Professional Responsibility: Legal, Ethical, and Therapeutic Issues

Licensing Boards

A skill deficit of the magnitude described above poses legal problems for licensing

boards, accreditation boards, and academic and training institutions overseeing therapists.

The various boards and institutions would theoretically be ethically and morally

responsible to victims of family violence if therapists were not properly trained to assess

and intervene in these specialized cases. Therefore, more training and specialized

programs should be developed to meet the growing demand of family violence cases that

professionals treat. Counselors are well trained to initiate screening for suicide but not as

likely to screen for safety issues related to domestic violence (Jansinski & Williams,

1998).

Lack of Training in Family Violence

One reason for the lack of appropriate training in prevention and treatment of

family violence is thought to be that professionals are uncomfortable in asking clients

about possible violence. This discomfort is common and should be explored with

counseling students during their training and supervision (Jansinski & Williams, 1998).

A supervision practicum in family violence issues would make the training component

more comprehensive; individual issues that face students regarding this issue could be

discussed.

Counselor education programs should include in their curriculum a course in family

violence or should address this theoretical approach in one or more core classes.

Students should be informed regarding methods of family violence assessment and








should be given vignettes to test their level of mastery in this skill area. Such training

should take place before students are made eligible for practicums or internships.

Requirements such as these would better prepare students to treat family violence cases

effectively.

Minimum State Requirements for Training

Currently, most states require a minimum of 2 hours in domestic violence training

for licensure or professional credentialing. In 2003 California has led the way by

increasing the minimum state requirement to 6 hours in domestic violence training.

However, some state and profession licensing and credentialing bodies have no required

training in the area of relationship violence.

In her article on training family therapists to assess for and intervene in partner

abuse, Haddock (2002) summarized core assumptions regarding curriculum development

in this area. First, feminist theoretical perspectives and treatment approaches are

essential to the effective treatment of partner abuse; however, privileging feminist

explanatory theories and treatment protocols do not disallow the inclusion of

complementary theoretical explanations and approaches (Goldner, 1992, 1999). The

second core assumption presented by Haddock was that therapists should be exposed to

current developments, controversies, and inconsistencies in the theoretical literature;

however, they also should be provided with specific principles, protocols, and skills to

guide their clinical work. The third core assumption presented by Haddock was that,

given therapists' tendency to reduce the inherent complexity of partner abuse cases in

response to concerns about potential lethality, therapists should learn to conceptualize

each case within its in own unique and multifaceted context (Dutton, 1992; Goldner,

1999), while attending to the intersections of gender, race, class, religion, and sexual








orientation (Bograd, 1999). The fourth core assumption presented by Haddock was that

therapists should be cautioned to utilize the most conservative assessment and treatment

approaches and to obtain direct and close supervision on cases involving partner abuse.

Theoretical Constructs Underlying the Study

Feminist View

Many different approaches are used when counseling battered women and

perpetrators of relationship violence. One approach is the feminist approach, in which

the batterer is viewed as being completely responsible for the battering. Feminists view

the battering as a criminal act that is used to control, intimidate, and inflict harm

(Goldner, 1999).

Walker (1995) particularly broadened the playing field by bringing in anecdotes

about how to match certain types of batterers with specific psychotherapies. Given the

sociopolitical context in which battering occurs, if psychotherapy is to be successful in

the treatment of battering, it must be integrated within a community-wide response. This

would involve coordination among therapists, police officers, probation officers,

prosecutors, judges, and advocates. Walker (1995) claimed that the Duluth model (Pence

& Paymar, 1993) is particularly promising, since treatment includes attention to the entire

social context of battering through community organizing.

Feminist theory has also been used as a conceptual framework for explaining the

presence of violence between males and females. In feminist theory, violence is viewed

as a manifestation of the patriarchal structure in our culture, which is reflected in the

patterns of behaviors and attitudes of individuals (Gentemann, 1984; Kalmuss, 1984). As

part of Lenore Walker's (1999) theorizing, violence against women in general has been

conceptualized as gender-based; interventions are stressed by using advocacy, victim








services, and educational efforts. The feminist perspective on domestic violence is

accepted worldwide. Where women and girls are primary targets of male abuse, violence

cannot be eradicated without looking carefully at gender socialization issues that

maintain and possibly facilitate such violence in the home.

The incorporation of social justice perspectives into family therapy training,

practice, and research has been a recent focus of family therapy scholarship (Bograd,

1999; Haddock, Zimmerman, & MacPhee, 2000; McGoldrick et al., 1999). These

authors have focused on intersections of race, class, sexual orientation, and gender as

they relate to the training and practice of family therapy. Goldner (1988) recognized that

both age and gender organize families, and she implored family therapists to reduce

power differentials based on gender within family therapy.

Feminist therapists claim that a therapist who does not directly challenge power

and control based on traditional gender hierarchies contributes to the maintenance of

unjust power differentials between men and women (Avis, 1996; Goldner, 1988; Hare-

Mustin, 1987; Walters, Carter, Papp, & Silverstein, 1988). These feminist scholars urge

therapists to challenge powerful acts by men over women and to expose gender

hierarchies by calling attention to them in therapy. As Hardy (2000) pointed out, family

therapists are often guilty of allowing the most powerful members of society to continue

speaking while inadvertently silencing those with less power. To heal strained

relationships, we must let the oppressed have an opportunity to speak during therapy that

takes the needs of all family members into account.

The Study of Power

In his 1993 book Understanding Family Process, Broderick organized family

process literature, which he called "relational space," into three major areas: (a)








regulation of interpersonal distance, (b) regulation of interpersonal transactions, and (c)

regulation of "vertical space," by which he meant power. The idea of "regulation"

implied a homeostatic set point theory. In a therapy context, these three areas were

respectively discussed as positivity/caring, responsiveness, and status/influence

(Gottman, Notarius, Gonso, & Markman, 1976).

Historically, the regulation of interpersonal distance was first explored by

examining the clarity of communication. Hypotheses were advanced to explore the role

of unclear communication in dysfunctional families and family distress. More specific

hypotheses were advanced that unclear communication was responsible for

psychopathology (Bateson, Jackson, Haley, & Weakland, 1956; Watzlawick, Beavin, &

Jackson, 1967), and the cybernetic model or the systems approach to family process was

born.

In their decade review of observing marital interactions, Gottman and Notarius

(2000) stated that historical research was complex but included the result that balance in

husband-wife power was related to marital quality; however, self-report and

observational measures did not show a high level of agreement in classifying couples.

These researchers stated that the issues of blending the study of affect and power are

central to the integration of psychological and sociological approaches to marriage.

Power is now being studied more precisely, using coding of the couple's influence

patterns during a discussion of the Inventory of Marital Conflicts (Olson & Ryder, 1970)

used by Gray-Little, Baucom, and Hamby (1996). They found that egalitarian couples

had the highest "Time 1" marital satisfaction and fewer negative marital inventory

conflicts scale behaviors. Power is also being explored in the context of gender and








relational hierarchy. Feminist writers have pointed to the central role that power must

play in understanding marriages. Quantitative observational research has now begun to

explore these ideas (Gottman & Notatius, 2001).

Equipped with a sound theoretical basis for addressing intimate violence in a

couple format, feminist family therapists have continued to develop couple approaches

that are sensitive to issues of power and gender (Almeida & Durkin, 1999; Bograd &

Mederos, 1999; Greenspun, 2000; Jory & Anderson, 2000; Lipchik & Kubicki, 1996;

Shamai, 1996). Hardy (2000) cautioned that the most powerful members of society

continue to speak while inadvertently silencing those with less power.

Social Learning Theory

Research by Alexander, Moore, and Alexander (1991) applied social learning

theory to investigate the intergenerational transmission of violence among dating

partners. This perspective explains that a man's involvement in dating violence is best

predicted by a personal history of severe abuse by his father. Although his behavior is

not directly predicted by having witnessed violence between his parents, his attitude

toward women apparently is thus predicted. His own attitudes are relevant to the

perpetration of violence only as they interact with those of his partners.

Scientists have traditionally believed that experiencing physical abuse as a child

increases the risk of later aggressive behaviors, based on social learning theory (Bandura,

1973). Social learning theory proposes that people acquire novel behaviors and expand

personal behavior repertoires by observing others' behaviors as well as through classical

and operant conditioning (Bandura, 1965, 1973). They observe other individuals and use

imitation to perform novel behaviors.








To imitate a behavior, the person must have some motivation or incentive for doing

so. Incentives are what the person expects to obtain once the behavior is performed.

Incentives act as reinforcers. When incentives are available, observation is more quickly

translated into action. Incentives also influence the attention and retention processes.

Children pay attention when given incentives to do so and, with more attention, more

information is retained. In the Bobo doll experiment, some children witnessed the adult

being rewarded for aggression. Therefore, these children performed the same act to

achieve the same rewards (Bandura, 1977).

Bandura asserted that people can also learn new behaviors without practice and

without reinforcement. Simply stated, an observer may copy a model's behavior long

after he or she saw the action performed, even without any immediate reinforcement

being earned by the model or the observer. Researchers have begun to consider the role

of previous abuse and violence on propensity for victimization, although only very

limited data have supported this theory; indeed, myriad unanswered questions remain.

Social learning theory not only deals with learning but also seeks to describe how a

group of social and personal competencies (i.e., personality) could evolve out of social

conditions within which learning occurs. It also addresses techniques of personality

assessment and behavior modification in clinical and educational settings (Bandura,

1977).

Social learning theory not only deals with learning but also seeks to describe how a

group of social and personal competencies (i.e., personality) could evolve out of social

conditions within which learning occurs. It also addresses techniques of personality








assessment and behavior modification in clinical and educational settings (Bandura,

1977).

Further, the principles of social learning theory have been applied to a wide range

of social behaviors, including competitiveness, aggressiveness, sex roles, deviance, and

pathological behavior (Bandura & Walters, 1963). Currently, it is impossible to ascertain

the exact social cognitive processes at work or how the observation of violence translates

into victim versus perpetrator status. The specific effect of social learning theory remains

an unresolved issue (Lewis & Fremouw, 2001).

Lenore Walker studied some intriguing animal behavior theories based on

laboratory research. She began to see startling connections between the coping

mechanisms of battered women and the behavior of caged dogs subjected to random and

variable electric shocks. "Learned helplessness" became one the linchpins of The

Battered Woman, published by Harper and Row in 1979. Coining the phrase "battered

women's syndrome," Walker delineated a common cycle of violence: a honeymoon

period followed by a buildup of tension, followed by an explosion and battery, followed

by regrets and apologies, followed by another honeymoon period, and so forth. By

featuring stories of several professional women who had endured physical abuse in

marriage, Walker put to rest the myth that battery was strictly a lower-class problem

(Brownmiller, 1999).

Social Structural Theory

The social structural approach examines patterns first applied by Goode (1971),

which applied Blood and Wolfe's (1960) resource theory of power to explain a husband's

use of physical force against his wife. Goode maintained that violence is a resource,

similar to money or personal attributes, that can be used to deter unwanted actions or to








induce desired behaviors. The use of violence thus can be seen as the most overt and

effective means of husbands' social control over wives (Yllo & Bograd, 1988) in that it is

used when other and more subtle methods of control do not lead to submission.

In a study by Allen and Straus (1980), key propositions of Goode's (1971) resource

theory of violence were tested using occupational prestige, educational level, income, and

satisfaction with income as measures of extrinsic resources. This study also used

variables for assessing interpersonal, intrinsic resources. The researchers found a strong,

positive correlation between the low resources/working class variable and the husband's

use of physical force. This is consistent with other studies that have shown that husbands

who experienced resource deprivation were more likely to physically abuse their wives

(Bowker, L., 1983; Pagelow, 1981).

Around the world, state-sanctioned violence, such as civil and interstate wars, often

increases the amount of violence against women; rape and brutal physical beatings of the

enemy's women have been considered just the spoils of war. Better understanding of the

relationships among civil war, domestic violence, and women's mental health is

important to help countries to provide both prevention and rehabilitation strategies

(Walker, 1999).

Strong cultural traditions tying women to small communities with few resources

(and power), state-sponsored conflicts, and greater acceptance of gender inequities all

contribute to the greater risk for a woman to be battered in her home (Heise, 1994; Koss

et al., 1994; Root, 1992; Walker, 1994).

An association in the marital literature between spouse abuse and self-esteem is

well documented (Goldstein, 1985; Hotaling, 1988), with the level of self-esteem








negatively correlated with frequency and severity of violence (Cascardi & O'Leary,

1992). However, it is often unclear whether low self-esteem precipitates violence or is

the result of chronic battering. Empirical support exists that low self-esteem contributes

to the difficulty that victims experience in disengaging from an abusive marriage (Aguilar

& Nightingale, 1994). As a result, low self-esteem may be associated with increased

tolerance for interpersonal violence.

Treatment of Relationship Violence

Choice to Leave or to Stay in Abusive Relationships

The cognitive process that a victim undergoes when choosing to disengage from an

abusive relationship provides important information with regard to victimology. Rosen

and Stith (1995), employing a multiple-case qualitative research design, identified a

progression of hierarchical cognitive steps utilized by individuals who eventually

disengaged from abusers.

The process included a five-step progression of disengagement: (a) seeds of doubt,

such as fleeting thoughts, often not identified until the dissolution of the relationship;

(b) turning points, or events that significantly impacted the intimate relationship;

(c) reappraisals, occurring when the victim re-evaluated the relationship; (d) paradigmatic

shifts, which included a shift in perspective about the relationship; and (e) last straw

events, or events providing the impetus to terminate the abusive relationship.

The explanations of why individuals remain in abusive relationships has

traditionally focused on dispositional characteristics such as low self-esteem, learned

helplessness, or masochistic personalities (Rusbult & Martz, 1995). A new model shifts

the focus away from blaming the victim and examines the interdependent nature of

ongoing relationships. This model proposes that victims apply a two-step model,








considering their resources and level of satisfaction when deciding to stay or leave an

abusive relationship (Choice & Lamrnke, 1997). According to this model, an abused

woman may stay in the relationship for several reasons. For example, she may

experience feelings of satisfaction, believing her best available alternative to staying is

not attractive enough to terminate the relationship. Second, she may believe that she has

a strong emotional investment and, thus, does not feel ready to leave the relationship.

This model is not proposing that victims want to be abused, but rather that some women

choose to remain in a relationship despite the abuse (Lewis & Fremouw, 2001).

Rusbult and Martz (1995) provided additional support for the supposition that

victims' decisions to stay in abusive relationships are influenced by environmental as

well as intrapersonal variables. Researchers found that a victim's choice to remain in an

abusive relationship was strongly related to level of satisfaction in the relationship,

quality of alternatives, and size of the investment. These models (Choice & Lamnke,

1997; Rusbult & Martz) represent a shift in conceptual understanding of the subtle

nuances and dynamics of victimization. Additional research is needed to examine the

conditions that trigger victims to leave abusive relationships. As described above, a

variety of factors may predispose a couple to relationship violence.

Most likely, it is a combination of these factors that motivates a perpetrator to

offend. A contemporary study of American couples conservatively documented that 1 in

8 husbands had committed a violent act against his wife during the preceding year. A

comprehensive review of studies using probability samples revealed that the reported rate

of wife abuse in the United States was between 11% and 22% (Straus & Gelles, 1990).








Gottman's Batterer Offender Typologies

Two distinct typologies are described by Gottman et al. (1995). In the 200

seriously violent couples studied, these researchers found at least two kinds of batterers:

Type 1 ("cobras") and Type 2 ("pit bulls"). This research dealt primarily with the

physiological response of male batterers to a high-conflict marital discussion. In this

study, Type 1 men's heart rates lowered from baseline in response to the high-conflict

marital discussion, while Type 2 men's heart rates increased from baseline. Type 2 refers

to men whose emotions quickly boil over, whereas Type 1 refers to men who remain cool

and methodical as they inflict pain and humiliation on their partners. This was evident in

physiological data taken from heart monitors while participants had angry responses to

their partners.

Type 2 men were referred to as "pit bulls." These men are characterized as having

a quick temper and are physiologically aroused when behaving aggressively. Using the

metaphor of the "pit bull," this type of dog is used in dog fights and is known as a vicious

fighter that will often fight until death. These dogs become aroused and aggressive when

other dogs are being aggressive. The "pit bulls" or Type 2 men scrutinize their wives and

display excessive need for approval and self-fulfillment.

According to Gottman et al. (1995), the Type 1 batterer is referred to as the

"cobra." Most striking is the swift escalation of anger and violence displayed by these

men during a disagreement, again relating to the analogy of the cobra snake, which is

swift and dangerous.

The "cobra" men enjoy shocking and scaring people. They are opposite to the "pit

bulls" in that they are not worried or jealous, nor are they emotionally dependent. The

significant finding of the Gottman et al. (1995) research was that "the cobra" offender's








heart rate actually got lower during an argument. Internally, these men remain calm, yet

they are externally more violent and severe in their violence than the "pit bulls."

Jacobson and Gottman (1998) discussed the pit bulls and cobras further. They

concluded that these two batterer types resemble dysphoric/borderline batterers and

generally violent/antisocial batterers, respectively (Holtzworth-Munroe et al., 2000).

Although Jacobson and Gottman focused their study on severely violent men, they also

discovered what they called a "low-level violent" group of couples, which they followed

with the expectation of tracking the development of violence from minor to more severe

forms. Unexpectedly, however, this group almost never escalated their use of violence,

and they were described by Jacobson and Gottman as a "stable group of couples who

periodically have arguments that escalate into pushing and shoving, but never reach the

point where we could call the men batterers" (p. 25). This description coincides with

Johnson's (1995) description of common couple violence.

Overall, findings across research by Gottman et al. (1995), Meehan, Holtzworth-

Munroe, and Herron (2001), and Babcock, Yerrington, Green, and Miller (2001) do not

lend strong support to the Type 1-Type 2 batterer typology. Although severely violent

men can be split into two groups on the basis of heart rate reactivity, so can both less

violent and nonviolent men, raising questions about the theoretical meaning of this

distinction. In addition, consistent differences between Type 1 and Type 2 severely

violent men have not been found across studies.

Human Heart Reactivity

A review of previous research suggests that the use of human heart reactivity

(HRR) may prove useful for differentiating male batterers on the dimension of








anger/hostility. There is less evidence that this approach will be effective for studying

antisocial personality in these men.

Meehan et al. (2001) failed to replicate the batterer typology proposed by the

Gottman et al. (1995) article. Therefore, caution is advised when discussing possible

implications of these studies for public policy and clinical application. To be

appropriately cautious, it should be noted that the Gottman et al. (1995) typology was

supported in one study (i.e., their initial study) but was not supported in another study

(i.e., the Meehan et al. study). Such an even scoreboard suggests that further attempts at

replication are necessary before definitive conclusions may be drawn about the validity of

the Gottman et al. typology.

It is possible that the Gottman et al. (1995) typology will remain a valid one after

further scrutiny; therefore, clinicians and public policy makers should not abandon the

idea of batterer typologies altogether. Indeed, across multiple research laboratories, there

is increasing convergence regarding the descriptive and theoretical dimensions that can

be used to meaningfully categorize variability among samples of batterers.

Holtzworth-Munroe and Stuart's Three Major Types of Batterers

Other research on men who batter women has suggested that there are three major

types of batterers: (a) those who use violence as a strategy to gain power and control

within their family, (b) those who use violence as a strategy and are also mentally ill, and

(c) those who have serious personality flaws that permit them to use violence to commit

other criminal acts as well as to abuse their partners (Dutton, 1995; Holtzworth-Munroe

& Stuart, 1994; Jacobson & Gottman, 1998; Meloy, 1998; Saunders, 1992; Walker &

Meloy, 1998).








Although other typologies for differentiating male batterers have been proposed,

research on female perpetrators is much less extensive. Holtzworth-Munroe and Stuart

(1994) proposed three types of male batterers: family-only, dysphoric/borderline, and

generally violent/antisocial. Batterers can be identified along three descriptive

dimensions: severity/frequency of violence, generality of violence, and psychopathology

or personality disorders, as well as by risk factors correlated with the development of

violent behavior (such as witnessing violence in the family of origin).

Dysphoric/borderline and generally violent/antisocial batterers engage in moderate to

severe levels of violence, and the latter are most likely to be involved in criminal

behavior and use violence both within and outside the home. The dysphoric/borderline

batterers tend to confine their violence to the intimate relationship. Family-only batterers

engage in the least amount of violence, show little or no psychopathology, and have very

low levels of risk factors. Empirical testing of the model has supported this batterer

typology (Hamberger, Lohr, Bonge, & Tolin, 1996; Holtzworth-Munroe et al., 2000;

Tweed & Dutton, 1998; Waltz et al., 2000).

Identification of Training Needs

Current Status of Training and Need for Additional Curriculum

There was a time when learning one particular school of therapy was deemed

sufficient empowerment to treat all manner of clients---or at least sufficient to make one

feel prepared to treat all types of clients. With time, specific treatments began to be

developed for particular problems. Alcoholism became widely recognized as a problem

that necessitated a particular type of therapy. Phobias were found to respond best to

behavioral treatment. Treatment began to be problem specific rather than school

oriented. Added to this is a new pressure: the necessity of keeping up in a field where,








seemingly, every new day reveals a new type of therapy for a new type of problem

(Salter, 1988). The treatment of family violence as well as child sexual abuse offenders

and victims has only recently evolved into a specialized field. Although for many years

there have been specialized treatment programs available in a few states, the need to set

up more specialized treatment programs in every state has only recently been recognized.

This recognition has been spurred by two factors: (a) the increasingly widespread

acknowledgement of the extent of the problem, and (b) the increasing acceptance that

traditional forms of therapy are not effective with this population (Crawford, 1981).

However, specialized treatment does appear to have an impact (Knopp, 1984).

Continuing Education Units

Academic institutions, institutes, workshops, and continuing education credits are

responsible for the successful dissemination of academic knowledge pertinent to the

profession. These educational areas must provide an integrated course of study as well as

appropriate professional training. An ethical education program with quality is one that

uses traditional and managed care approaches of training emphasizing the current

theories of practice and those presented in the Diagnostic and Statistical Manual of

Mental Disorders (DSM-IV) of the American Psychiatric Association (1994). It should

be a program that develops competency in people working with diverse aspects of

society, resulting in effective, cost-efficient treatment plans and service delivery. Only

recently (in 1995), California began specialized spousal abuse training for marriage and

family therapists. Psychologists are required to have coursework in human sexuality,

chemical and substance abuse, and child abuse.

However, spousal abuse training is not currently a requirement for licensing in

California or in many other states. Both trainees and interns are permitted to perform








counseling in a governmental entity, school, college, university, nonprofit agency,

charitable corporation, or licensed health facility under the direct supervision of a

licensed supervisor, regardless of whether that supervisor has received specialized

training in the area of spousal abuse (Board of Behavioral Science Examiners [BBSE],

2000).

Skills in Assessing Relationship Violence and Imminent Danger

In previous research there is a precedent for needed improvement in the education

and training of psychotherapists. Hansen et al. (1991) surveyed members of a marriage

and family therapy organization. The therapists in this survey were asked to read two

vignettes with proven therapeutic interventions for domestic violence cases. These cases

portrayed female victims and domestic violence. The results indicated that the

counselors did not attend to the seriousness of the violence portrayed in the vignettes, and

many did not attend to the violence at all. As a group, the therapists did not attend to the

crisis nature of the cases portrayed and failed to intervene for needed protection of female

victims from their batterers. Forty-one percent of the therapists surveyed indicated no

recognition of domestic violence. Interventions provided by therapists in this study were

also negligent with regard to violence potential. For example, 55% of respondents

reported that they would not intervene even when the violence portrayed required

immediate action. Only 2% reported a potential for lethality in these cases. A scant 11%

of the respondents indicated that they would obtain protection for the wife by helping her

to develop a safety plan, obtain shelter, or obtain a restraining order.

Psychologists were significantly less likely than other respondents to conceptualize

the case in terms of conflict or to describe the problem as violence. Because few

counseling or clinical psychology graduate programs provide academic or clinical








exposure to the problem of domestic violence presently (and even fewer have done so in

the past), there is reason to expect that knowledge and skill deficits similar to those

observed among physicians and family therapists in the Hansen et al. study might be

found among practicing psychologists as well (Carden, 1994).

Assessing for Risk Factors Associated With Relationship Violence

Riggs, Caulfield, and Street (2000) discussed the need for identifying risk factors

for domestic violence. They stated that the extent and potential dangerousness of the

problem of domestic violence warrant systematic screening and assessment in all mental

health settings. Few empirical studies have approached the question of domestic violence

with the aim of identifying risk markers, making it impossible to identify persons at risk

for perpetrating or becoming victims of domestic violence.

Identifying Risk Factors Associated With Relationship Violence

A number of factors have been identified as correlates of domestic violence that

may eventually prove useful for identifying persons at risk. However, existent literature

does not provide empirical support for these factors. Identifying factors that might assist

clinicians in the recognition of clients who are at risk for domestic violence would help

clinicians to attend appropriately to this potentially dangerous problem.

Ongoing assessment in the context of knowledge regarding correlates of domestic

violence can provide important information for evaluating risk of a particular violent

incident. In addition, strategies for assessing violence and violence risk in both

perpetrators and victims can assist clinicians in approaching this difficult topic in a

clinical setting. A careful assessment of the potential for violence within clients' ongoing

relationships is necessary for clinicians to provide appropriate care (Riggs et al., 2000).








Prior Relationship Aggression

One of the primary risk factors for perpetrating violence against a spouse or

significant other is having committed such violence against that same person previously.

Rarely, if ever, does an incident of spouse abuse occur in isolation; typically, violence

occurs throughout the relationship. In a longitudinal study of violence within marriages,

O'Leary et al. (1989) examined couples engaged to be married and followed them for 3

years. These researchers found that violence was relatively stable over the first few years

of marriage. For example, of men who perpetrated aggression prior to marriage, 51%

also had engaged in aggression during the first 18 months after marriage. In comparison,

the probability of a man who was not aggressive prior to marriage being aggressive 18

months into the marriage was 15%.

In another study of engaged and newly married couples, men who perpetrated

violence against their partners prior to the wedding were significantly more likely than

were men with no history of relationship violence to perpetrate violence within the

following year (Leonard & Senchak, 1996). Examining more established relationships,

Feld and Straus (1989) found that frequency of self-reported aggression was predictive of

aggression in the ensuing year: Almost 50% of men who had been violent prior to the

initial assessment in the study engaged in violence in the following year. In comparison,

only 10% of those who had not been violent prior to the initial assessment engaged in

violence in the following year.

Demographic Characteristics

Rates of domestic violence tend to decrease slightly as the age of couples increases

(Straus et al., 1980). Other demographic risk factors for marital violence appear to be

related to increased stress in the family. For example, men of lower SES are at an








increased risk for perpetrating domestic violence and tend to perpetrate more severe

violence than higher-SES counterparts. Similarly, men who are unemployed appear to be

at an increased risk for perpetration of spouse abuse.

National surveys suggest that men of color are at an increased risk, compared to

Caucasian men, for perpetrating violence against their wives and partners. However, this

difference appears to be at least partially the result of different SES levels in White and

minority samples (Holtzworth-Munroe et al., 1996).

Psychological Characteristics

Men who have perpetrated violence against a spouse tend to differ from those who

have not done so on a number of psychological constructs. Many of these factors,

according to Riggs et al. (2000), are of theoretical interest for understanding why spouse

abuse occurs and may be helpful in identifying men at risk for such violence. For

example, spouse abusers tended to be more angry and hostile in general than their

nonabusive counterparts (Eckhardt, Barbour, & Stuart, 1997). Further, when confronted

with marital conflict situations, abusive men tended to be less assertive and respond more

with anger and hostility than nonviolent men. Particular topics of conflict such as

jealousy and threats of abandonment may have exacerbated this pattern of response

(Holtzworth-Munroe & Anglin, 1991). This latter finding fits with research that

identifies fear of abandonment as an important aspect of abusive men's behavior (Dutton,

Saunders, Starzomski, & Bartholomew, 1994) and may serve to identify specific points in

time when the risk of spouse abuse is particularly high. Unfortunately, few of these

variables have been investigated in ways that translate the findings into specific

indicators of risk.








Specific Psychological Syndromes

Post Traumatic Stress Disorder (PTSD). Researchers have linked men's

symptoms of depression, PTSD, borderline personality disorder, and substance abuse to

the perpetration of violence against their wives and partners. Important with regard to the

present discussion, some of these syndromes have been examined as risk factors rather

than simply as correlates of marital violence. Thus, they may serve as useful markers or

risk factors of future perpetration.

Researchers have examined the link between symptoms of PTSD and the

perpetration of marital violence. Studies in this area indicate that men with PTSD are at

considerable risk for perpetrating marital violence. Jordan et al. (1992) found that wives

of Vietnam veterans with PTSD were about twice as likely (30%) as wives of veterans

without PTSD (15%) to report that their husbands had engaged in marital violence.

Depression. Depressive symptoms have been related to the perpetration of marital

violence in a number of studies (Maiuro, Cahn, Vitaliano, Wagner, & Zegree, 1988; Pan,

Neidig, & O'Leary, 1994). Generally speaking, men who are aggressive toward their

wives exhibit more depressive symptoms than do nonaggressive men. For example, in

one study, significantly more abusive men, as compared to a nonabusive group, scored

within the depressed range of the Beck Depression Inventory (Maiuro et al.). Higher

scores on this inventory have also been related to violence in couples seeking marital

therapy (Vivian & Malone, 1997) and in general population samples (Pan et al.).

Substance abuse. The rates of spouse abuse among men diagnosed with substance

abuse problems indicate that men with diagnosable alcohol problems are at substantially

increased risk for spouse abuse (Gondolf & Foster, 1991; Leonard, Bromet, Parkinson,

Day, & Ryan, 1985; Murphy & O'Farrell, 1994; Stith, Crossman, & Bischof, 1991). For








example, Leonard et al. found rates of marital aggression in men diagnosed with a current

alcohol problem (44%) to be about 3 times greater than in men without an alcohol use

disorder (15%) or a past alcohol problem (14%). Murphy and O'Farrell reported that

about two thirds of a sample of married male treatment-seeking alcoholics had engaged

in marital violence.

Borderline personality disorder. Another disorder that has been linked to the

perpetration of spouse abuse is borderline personality disorder. This disorder is

characterized by identity issues that become salient in intimate relationships and that vary

on three defining features: identity diffusion, primitive defenses, and reality testing.

Studies have found that abusive men score higher on measures of borderline personality

disorder than do nonabusive men and that, among men who assault their wives, more

severe violence is associated with higher levels of borderline personality disorder

(Dutton, Starzomski, & Ryan, 1996).

Other Risk Factors

Marital dissatisfaction. As a group, men who perpetrate violence against their

partners are less satisfied with their relationships than are nonviolent men (Alderondo &

Sugarman, 1996; Hotaling & Sugarman, 1986; Sugarman & Hotaling, 1989).

Witness to spouse abuse or victim of child abuse. Many studies have found that

men who perpetrated violence against their wives were more likely than men in

nonviolent comparison groups to report that they had experienced violence in the family

of origin, either as a witness to spouse abuse or as the victim of child abuse (Alderondo &

Sugarman, 1996; Dutton & Hart, 1992; Hotaling & Sugarman, 1986; Kalmuss, 1984;

Sugarman & Hotaling, 1989).








Previous head injury. Rosenbaum et al. (1994) documented a link between a

history of head injuries and the perpetration of spouse abuse. In one study, these

researchers found that men in treatment for abusing their wives reported a significantly

higher rate of head injuries than did a group of nonabusive men.

Ability to Assess Violence Risk

Otto (2000) spoke to the importance of the mental health practitioner having the

ability to assess violence risk in clients. The perception that persons with mental illness

are at increased risk for violence, as compared to their non-mentally ill counterparts, can

be dated at least to the time of Plato (Monahan, 1992). Indeed, among the rationales

offered for establishing some of the first public psychiatric hospitals in this country was

the need to protect the public by confining persons with mental illness who posed a risk

of violence to the community. Flowing logically from the belief that there was a

connection between violence and mental disorders was the assumption that mental health

professionals, as a function of their expertise, were uniquely able to identify and treat

persons whose emotional functioning increased their risk for violence, and could thereby

reduce such risk (Otto).

Violence Risk Assessment Study

In response to the above, Monahan (1984, 1988) identified limitations of research

examining the relationship between mental disorders and violence, as well as mental

health professionals' abilities to assess violence risk. Due to these limitations, Monahan

called for a "second generation" of investigations to better address these issues. This call

resulted in a series of studies (Monahan & Steadman, 1994, for summary) and review

articles (Mossman, 1994; Otto, 1992, 1994) and formed the basis for the Violence Risk

Assessment Study organized by John Monahan under the auspices of the MacArthur








Research Network on Mental Health and the Law. Findings from this "second

generation" of research, which incorporated many of Monahan's (1984, 1988)

recommendations, suggest the following: (a) violent behavior is not necessarily a low

base rate behavior and occurs with some degree of frequency among persons with mental

disorder (Otto, 1992; Steadman et al., 1998; Wessely & Taylor, 1991); (b) persons with

certain mental disorders and symptom clusters are more likely to engage in violent

behavior than are persons without such disorders or symptom clusters (Swanson, 1994;

Swanson, Holzer, Ganzu, & Jono, 1990); and (c) mental health professionals have some

ability to assess violence risk among persons with mental disorders (Mossman; Mulvey &

Lidz, 1998; Otto, 1992, 1994).

Assessing and Managing Risk

It is this body of developing research, along with research examining violence risk

factors among criminal and nonclinical populations, that provides direction for clinicians

faced with the task of assessing and managing risk with their clients. Although one might

question whether findings from one population are applicable to other populations, a

meta-analysis by Bonta, Law, and Hanson (1998) provides some support for the claim

that risk factors for violent behavior may be similar across populations.

It should be no surprise to even beginning clinicians that more remains unknown

than known about risk factors for violence among persons with mental disorders. Good

practice requires that clinicians familiarize themselves with relevant literature and use

informed clinical judgment in cases for which research literature provides no direction

(Otto, 2000).








HCR-20 Item Risk Assessment Instrument

Use of a structured, guided clinical assessment developed in light of the extant

research, such as HCR-20 (Webster, Douglas, Eaves, & Hart, 1997) can help to form the

basis of a comprehensive evaluation that assesses factors relevant to violence risk. The

HCR-20 directs clinicians to cover a total of 20 areas considered to be relevant to

violence risk: 10 historical items, 5 clinical items, and 5 risk management items.

Preliminary data indicate that the HCR-20 can be reliably scored (Belfrage, 1998;

Douglas, & Webster, 1999; Ross, Hart, & Webster, 1998) and has some predictive power

when compared to other risk assessment instruments.

Static and Dynamic Risk

Otto (2000) stated that, broadly speaking, risk factors for violence among persons

with mental disorders fall into one of two categories. Static risk factors are those that

either cannot be changed (e.g., age, gender) or are not particularly amenable to change

(e.g., psychopathic personality structure). Identification of these factors is important in

terms of identifying the client's absolute or relative level of risk; however, these factors

typically have few implications for treatment or management of risk, since the factors, by

definition, cannot be changed. In contrast, dynamic risk factors are those that are

amenable to change (e.g., substance abuse, psychotic symptomatology). Identification of

these factors is important, both in terms of estimating the client's absolute or relative

level of risk and for purposes of treatment planning.

Hanson (1998) made a similar distinction: between stable dynamic factors and

acute dynamic factors. Stable dynamic factors can change but have some enduring

quality over time and across situations (e.g., deviant sexual preferences or alcoholism),

whereas acute dynamic factors (e.g., sexual arousal or alcohol intoxication) are "states"








which can change much more rapidly. Assessing the former category may be more

important for treatment planning and intervention planning when dealing with persons for

whom there are concerns for violence in the future, while the latter category may be more

important in terms of assessing imminent risk and making decisions about immediate

interventions.

Conclusion

Feminist writings have focused attention on domestic violence and challenged the

assumptions of approaches that blame women for their victimization. They support both

the growing emphasis on interdependence and mutuality, rather than on autonomy and

differentiation, in couple relationships and they focus on a more collaborative respectful

relationship with clients (Luepnitz, 1988).

Graduate counselor education programs would ethically be fulfilling their ethical

code and accreditation standards by including in their curriculum a course in family

violence. Students would benefit by being informed regarding methods of family

violence assessment and could be given vignettes to test their level of mastery in this skill

area. Such training preparation could take place before students are made eligible for

practicums or internships. Requirements such as these would better prepare students to

treat family violence cases effectively.

Various intervention models are used in treating relationship violence. According

to Harrell (1991), the short-term, court-ordered, batterer-only psychoeducational model

had a high rate of recidivism, along with increased amounts of psychological abuse.

Edleson and Grusznski (1988) and Pence and Paymar (1993) found better results from

the short-term psychoeducational model used in the Domestic Abuse Project in Duluth,

Minnesota. On the other hand, Dutton (1995) and Hamberger and Ambuel (1997) found








that interventions having the best results in stopping men from using violence were those

that required attendance for a minimum of 2 years. Dutton's research suggests that many

batterers have serious mental illnesses in addition to problems with power and control

that underlie their use of violence (Walker, 1999). O'Leary (1993) and Geffner (1995)

found that special techniques in family psychotherapy can be effective in helping to stop

violence in the family.

Corsi (1999) developed a model that appears to incorporate the best features from

available programs and is easy to apply across cultures due to its ecological approach

(Walker, 1999). The recommended treatment approach begins with individual

counseling for all parties and then marital and dyad sessions, in addition to family and

group sessions (Giarretto, 1976).

Violence in close relationships is now acknowledged as a highly significant issue

across diverse cultural groups (Walker, 1999). The underreporting of couple violence,

even in couple therapy, is a major finding that has emerged over the past decade. One

line of research has differentiated different patters of violent behavior, distinguishing

those patterns that are more or less likely to be amenable to treatment (Jacobson &

Gottman, 1998). Such research may assist the couple therapist in determining when and

how to intervene.

In general, assessment procedures, risk factors, and treatment feasibility issues in

violent relationships are now beginning to be addressed (Bograd & Mederos, 1999;

Holtzworth-Munroe, Beatty, & Anglin, 1995). Well-defined assessment procedures

(Bograd & Mederos) and differentiated treatment strategies are essential for client safety.

All couple therapists must know how to identify and address patterns of violence, ranging





65


from verbal intimidation and threats to coercion and battering, and they must be able to

make informed decisions about the best interventions to use in particular cases (Johnson

& Lebow, 2000).

This literature review would be appropriate for an audience including mental health

practitioners, faculty, health care workers, trauma and crisis center employees, domestic

violence workers, parents, educators, and law enforcement personnel.













CHAPTER 3
METHODOLOGY

Relationship violence is a problem of extensive proportions in America. There are

several theoretical models for assessing and treating violence survivors in order to

prevent continuing abuse. However, it is unknown which treatment methods, assessment

questions, and prevention models are actually put into practice with this population.

Therefore, the primary purpose of this study was (a) to determine the factors that are most

frequently reported by MFTs to be related to the identification, assessment, and treatment

of relationship violence; (b) to measure how therapists, supervisors, and faculty members

in marriage and family therapy rate themselves on their competence in the identification,

assessment and treatment of relationship violence; and (c) to measure their rating of the

importance of certain competencies and skills in the identification, assessment, and

treatment of relationship violence.

These purposes were accomplished by examining two variables via the survey.

The first variable, Quality of Graduate Training, was assigned two subscales, each with

five items in the survey. The first subscale, Knowledge, was measured by five items

eliciting respondents' self-rating of their knowledge in relationship violence. The second

subscale, Graduate Training Received, was measured by five items asking respondents to

rate the graduate training that they received in relationship violence. The second

variable, Competencies in the Identification, Assessment, and Treatment of Relationship

Violence, was measured by 10 items asking respondents to rate the importance of certain

competencies identified by the researcher as directly related to this variable.








In an attempt to clarify some of the unresolved problems in the existing literature,

this study was designed to answer the following questions:

1. How do MFTs rate themselves on their knowledge in the identification,
assessment, and treatment of relationship violence?

2. How do MFTs rate their graduate training in the identification, assessment, and
treatment of relationship violence?

3. How do MFTs rate the importance of specified competencies and skills in the
identification, assessment, and treatment of relationship violence?

The significance of this study was discussed in chapters 1 and 2, based on (a) the

increase in incidence of relationship violence as indicated by statistics; (b) professional

responsibility: legal, ethical, and therapeutic issues; (c) multiple forms of treatment

(metatheoretical, postmodernism, feminism, sociocultural, and social learning); and

(d) identification of training needs.

The research methodology is described in this chapter. Included are descriptions of

the research method, survey design, rating scales in the social sciences, research

procedures, sample, sources of error in Web-based designs, survey research, research

barriers, and limitations in relationship violence.

Research Method

Approximately 1,000 active members of AAMFT from various geographic regions

of the United States were asked to participate in the research (appendix D). The RVTS

was used to collect demographic data, including the professional counselor's age, gender,

ethnicity, marital status, and highest academic degree.

This study was conducted using a Web-based survey sent to therapists, approved

supervisors, and faculty members in marriage and family therapy. An exploratory factor








analysis was used to identify the related factors of training in the identification,

assessment, and treatment of relationship violence as reported by MFTs.

A preliminary self-designed survey instrument (RVTS) was developed based on

clinical expertise and research reviews. The model's elements were refined, based on

field testing suggestions offered by 10 colleagues with practice expertise and research

backgrounds. These colleagues were recruited based on their expertise in the areas of

measurement, relationship violence, and family therapy, which included experience in

teaching assessment and prevention of relationship violence. Certain variables were

selected over others for the instrument, based on research and clinical knowledge. A

panel of experts in the field reviewed the variables, and a revised selection was made

based on their feedback on the items. This testing was important in establishing validity

of the instrument and served to make improvements in the formatting and structure of the

instrument.

This model was further refined based on a field study given to experts in the above

areas. Results from the field studies were used to improve the model. Approximately

20% of the original items were deleted and another 20% were refined after feedback

about the interpretation of items was given.

The exploratory factor analysis program was conducted on the basis of the

researcher's specifications of the following items: (a) the variables to be factor analyzed;

(b) whether the data were in raw form or in the form of a correlation or covariance

matrix; (c) the number of factors to be extracted or the criteria by which to determine

such a number; (d) whether the diagonal elements of the correlation matrix were to be

replaced by communality estimates and, if so, what types of estimates were to be used;








(e) whether to employ orthogonal or oblique rotation; (f) the particular type of rotation to

be used (Kim & Mueller, 1978); and (g) giving the extracted factors names that would be

inclusive of all the items within that factor loading

Survey Design

In this section the prevalence of using surveys and rating scales in the social

sciences is discussed. Rating scales and their relevance to the RVTS survey development

are highlighted.

A survey design provides a quantitative or numeric description of some fraction of

the population-the sample-through the data collection process of asking questions of

people (Fowler, 1988). One goal of this data collection will be to generalize the findings

on the current status of graduate training in relationship violence from a sample of

responses to a population.

The data in this study were collected and compared via a Web-based survey system

that scored responses electronically. Data from the individual marriage and family

therapists, approved supervisors, and faculty members were examined to identify factors

in the assessment/training provided to marriage and family therapists at respective

schools regarding relationship violence. Factors related to treatment of relationship

violence were correlated. Since this is an exploratory factor analysis, identifying factors

that are important in the training of marriage and family therapists in working with

relationship violence was highlighted.

The purpose of survey research is to generalize from a sample to a population so

that inferences can be made about some characteristic, attitude, or behavior of this

population (Babbie, 1990). Broadly, a sample is a part selected to represent a larger

whole (Warwick & Lininger, 1975). The variables are the concepts or information in








which the researcher is interested. A questionnaire is a series of questions presented to

the sample in person by an interviewer, over the telephone, via Internet, or via computer,

through a self-administered Internet or Web-based instrument or in some other way. The

data analyses and reports are then used to describe the group or to draw inferences about

the variables, their relationships to each other, and their relationships to the population of

interest (Nelson, 1996).

Surveys usually focus on people-facts about them or their opinions, attitudes,

motivations, behaviors, and so on-and the relationship between variables under study

related to these people. For example, survey research might be used to compare

demographic characteristics of a sample of people in a particular location, their access to

mental health services, and their perceptions about the efficacy of those services. This

family therapy research could be used to make recommendations about improving the

curriculum/training provided to marriage and family therapist in relationship violence.

These generalizations have been made in previous research. For example, in their

1989 survey research on family therapy Wetchler, Piercy, and Sprinkle surveyed both

supervisors and supervisees about their impressions of their supervision experiences and

made some suggestions about marriage and family therapy training based on responses to

their survey.

Rating Scales in the Social Sciences

Various types of questionnaires are by far the most-used method of data collection

in psychology and other social sciences, and almost all of them use rating scales as their

primary response mode. Countless articles have followed the seminal work of authors

such as Freyd (1923), Thurstone, (1928), and Likert (1932). A response scale should








fulfill psychometric standards of measurement quality as well as practicality criteria, such

as comprehensibility for respondents and ease of use.

It has been recognized for many years that answers to self-administered

questionnaires are influenced by the way in which the questions and answers are

displayed on questionnaire pages (e.g., Rothwell, 1985; Smith, 1993; Wright & Barnard,

1975, 1978). However, scientific understanding of the natures of those effects is not well

developed. Although it has been argued on theoretical grounds that visual layout and

design make a difference in how people answer questionnaires (Jenkins & Dillman, 1997;

Sless, 1994), little experimental evidence exists that changing the visual presentation of

individual survey questions influences people's answers.

On occasion, rating scales are used in which verbal labels are compressed to

saturate one end of the response continuum. In these scales, differences arise between the

normal meaning of the label and its scalar position. In some instances, however, equally

spaced options across the entire response continuum may not provide the desired

properties in the measurements. For instance, Symonds (1931) recommended the use of

evaluative rating scales of a set of labels that were packed with positive descriptions to

overcome individuals' tendencies to be lenient in their description of others. For a 5-

point rating scale, he recommended using the labels Poor, Fair, Good, Very Good, and

Excellent. Guilford, in his classic Psychometric Methods (1936), reiterated Symonds's

solution for the problem of errors of leniency, hi addition, Guilford recommended that

"in a similar manner in the numerical type of scale, the strength of the descriptive

adjectives may be adjusted so as to counteract the error of central tendency" (p. 272).








French-Lazovik and Gibson (1984) also demonstrated that the labels used

influenced the distribution parameters of rating scale data. By using more positive labels,

they were able to systematically change the ratings in the predicted directions.

The meaning of the verbal label of a scale may depend upon the contexts of the

label. One important context to be considered is the position of the label. Chase (1969)

suggested that the meaning of the scale adjectives be determined by the relative position

of the adjective in a group of response categories rather than by the "standard" definition

of the scale labels. This suggestion was consistent with findings by Wildt and Mazis

(1978) that both label and location had an impact on subjects' responses.

Klockars and Yamagishi (1988) found that the meaning of the labeled position was

defined as a compromise between the label itself and the relative position. They showed

results that suggested the use of rating scales containing verbal anchors predominantly

from one end of a continuum to provide increased discrimination in the portion of the

scale. This provides the respondent with response options that are more discriminating in

one portion of the underlying continuum. Consequently, as stated by Worcester and

Bums (1975), "The problem is not just that different words mean different things but that

the same word can be made to mean different things as the context changes (p. 182)."

Usually, rating scales (category scales in psychometric terms) offer between 4 and

11 response alternatives (ordinal scale points which are supposed to be equidistant).

Numbers or words or graphic symbols ( or a combination thereof) can be used to denote

the categories, but verbal labeling has become the dominant approach to facilitate

communication (Rohrman, 2002). Instead of labeling every point on the scale, end points

may be verbalized. In the Relationship Violence Training Survey (RVTS) instrument








constructed for this study a 6-point rating scale was used, with only the labels Very Poor

and Excellent to define the two end positions for items 1-10 and Not Important and Very

Important for items 11-20; intermediate positions were unlabeled. The respondent was

allowed to discriminate between the two end points to define the meaning of the response

without the judgment of labels used in other positions.

Research Procedures

The researcher requested permission and obtained approval from the University of

Florida Institutional Review Board to proceed with the study as proposed. Following this

approval, the researcher sent research packets to professional counselors across the

United States who were active therapist members of AAMFT, approved supervisors, and

COAMFTE graduate training program faculty members.

Informed consent was obtained from participants. The participants were given the

option to refuse participation in the study. All participation was voluntary and

confidential. The results from the study are summarized as group findings and will be

available to participants who express an interest in the results. The individual responses

of counselors were not reported to their institutions/agencies or anyone else. See

appendix D for the letter of invitation to participate in the survey and the accompanying

consent form.

Sample

The first in a series of three Internet Web-based RVTS was sent electronically to a

random sample of licensed marriage and family therapists. The list of Internet mail

addresses was obtained from the AAMFT. The second questionnaire in the RVTS series

was sent to a random sample of marriage and family therapy faculty members who teach

in training programs in major university settings. This list was generated from approved








marriage and family therapy training programs from the AAMFT organization branch

COAMFTE, which monitors and credentials marriage and family therapy training

programs. The third questionnaire in the RVTS series was sent to a random sample of

AAMFT approved supervisors from a list obtained from the AAMFT organization,

retrieved from their national list of approved supervisors.

The questionnaire requested information about demographic variables (age, gender,

etc.) and information about education and training. Therapists, supervisors, and faculty

members were asked to rate themselves, using a 6-point scale, on their knowledge of

factors related to the identification, assessment, and treatment of cases relating to

relationship violence. They were also asked to rate their marriage and family therapy

graduate training in the identification, assessment, and treatment of violence. They were

also asked to rate the importance of certain competencies identified by the researcher to

be related to the identification, assessment, and treatment of relationship violence.

Due to the initial response rate on the electronic response (error rate or nonresponse

rate), an electronic follow-up mailing (appendix E) was sent out 2 weeks after the first

electronic mailing to those who had not responded, with a reminder announcement listing

the Web site link for participants to contact to be involved in the survey.

Sources of Error in Sample Web Surveys

The remarkable power of a sample survey is its ability to estimate, with precision,

the distribution of a characteristic in a defined population. In addition, that estimate can

usually be made by surveying only a small portion of the population under study. Sample

surveys are subject to four major sources of error, and each must be attended to in order

to have confidence in the precision of the sample survey estimates (Groves, 1989). These

errors are (a) coverage error, the result of all units in a defined population not having a








known nonzero probability of being included in the sample drawn to represent the

population; (b) sampling error, the result of surveying a sample of the population rather

than an entire population; (c) measurement error, the result of inaccurate responses that

stem from poor question wording, poor interviewing, survey mode effects, and/or some

aspect of the respondent's behavior; and (d) nonresponse error, the result of nonresponse

from people in the sample, who, had they responded, might have provided different

answers to the survey questions from those given by persons who responded to the survey

(Dillman & Bowker, 2001).

All four of these sources of error are as applicable to the design, and implementation

of Web surveys as they are to mail interview surveys. However, the early

implementation of Web surveys suggests that some aspects of error, and in particular

coverage and nonresponse, have been mostly ignored. Sampling error, although not

being neglected, is instead often inferred when it is not appropriate to do so. For

example, many Web surveys are conducted using samples of convenience or availability,

and thus depend heavily on the solicitation of volunteer respondents, as described by

Bandilla (2001).

One of the basic assumptions in surveying is the recognition that, for simple

random samples of a defined population, the precision of results is closely related to

completed sample size, or the number of respondents. For example, simple random

samples of 100 have a precision of 10 percentage points, and those of 1,100-1,200 (the

size commonly used for election survey prediction) have a precision of 3 percentage

points. Sampling error is decreased by about half when sample size is quadrupled. Thus,

samples of several thousand are expected to have precision measure in tenths of a








percentage point, assuming no other sources of error. Large number of volunteers

respondents, by themselves, have no meaning. Ignoring the need to define survey

populations, select probability samples, and obtain high response rates together provide a

major threat to the validity of web surveys (Dillman & Bowker, 2001).

Regarding coverage error, using the Web to survey the general public remains quite

limited. Moreover, Black and Hispanic households are about two fifths as likely to have

Internet access as are White households, and rural Americans about half as likely to have

access as urban Americans with comparable incomes. Nua Internet Surveys (1997)

estimated that 179 million people, or about 3% of the world's populations, has been

online at least once, and over 40% of U.S. households now own computers but only one

quarter of all households have Internet access (National Telecommunication and

Information Administration [NTIA], 1999).

However, this does not mean that researchers are unable to conduct scientifically

valid Web surveys. Some populations-employees of certain organizations, members of

professional organizations, certain types of businesses, students at many universities and

colleges, and groups with high levels of education-do not exhibit large coverage

problems. When nearly all members of a population have computers and Internet access,

as is the case for many such groups, coverage is less of a problem.

Another serious source of potential error in Web surveys is the nonresponse

problem associated with positing a Web questionnaire and inviting people to respond.

Number of contacts (or call-backs) has always been a major influence on response rates

to other survey methods, and the tool most depended upon to reduce nonresponse error.

There is little doubt that procedures can be developed for achieving response rates to








Web surveys that are reasonably comparable with those obtained by other methods

(Dillman, 2000).

The advent of Web surveying presents measurement challenges not previously

faced by survey methodologists and for which research has not yet provided solutions.

The enduring problem is that what the designer of a Web questionnaire sees on the screen

may differ significantly from what some, and in other cases most, respondents see on

their screens. Evaluation of the Web surveys located by D. Bowker (1999) revealed

much variability in methods of construction. When tested on various levels and types of

Web browsers, operating systems, screen configurations, and hardware, the visual

stimulus of the survey items (i.e., physical placement and presentation) was often

different from what had originally been intended by the designer.

Although measurement error effects represent one of the most serious threats to the

conduct of quality Web surveys, they are also among the most easily addressed through

various design controls by programmers. In the construction of this survey HyperText

Markup Language (HTML) was used in conjunction with SurveyWiz0, a program that

was originated by Michael Bimbaum in 1998 (Birnbaum, 2000). This program was

recommended by Dr. Roger L. Worthington, Ph.D., from the Department of Educational

and Counseling Psychology at the University of Missouri-Columbia while attending a

Web-based research methods forum at the 2001 Southern Association for Counselor

Education and Supervision (SACES) convention of the American Counseling

Association in Athens, Georgia. Dr. Worthington shared some of his experiences and

explained how he had enhanced the programming to alleviate some of the measurement

errors mentioned in this section. Dr. Worthington encouraged development of the survey








for the present study using the SurveyWiz formatting, as he believed that it would help

greatly in coding and reducing error while working with the data set prior to running

factor analysis.

During the survey development, the researcher worked closely with Mr. Gary Sipe

from Stetson University CIT Media Services to learn how to run SurveyWiz (Bimbaum,

2000) programming and how to set up the HTML documents. These files were edited

each time a revision was made to the items on the survey and saved on a working

clipboard and then resaved into the actual document. The final versions of all three

surveys were then saved onto the Stetson University Secured Server under an HTML file

reserved for the researcher as an adjunct professor. All responses to the survey were

forwarded from the Stetson server automatically to the researcher's email address, where

they were automatically coded by question number. One way of reducing nonrespondent

errors in this study was that each respondent was automatically coded by response time

and email address (appendix F). This assured that no respondent could submit responses

more than once. This process also helped in sending nonresponse notices to the correct

nonrespondents. Finally, the data were placed into a standard computer application

spreadsheet and prepared for run through a leading computer software statistical analysis

(Statistical Package for the Social Sciences; SPSS, 1975) for the factor analysis and

demographics.

Survey Research in Family Therapy

In family therapy research, many surveys have been designed to determine what

clinicians think or do. Survey research has been used to ask clinicians about their use of

assessment instruments (Boughner, Hayes, Bubenzer, & West, 1994), about how they act

when faced with ethical dilemmas (Green & Hansen, 1989), about their preferred models








of therapy (Quinn & Davidson, 1984), about how they use or view their clinical training

(Carter, 1989; Coleman, Myers Avis, & Turin, 1990; Keller, Huber, & Hardy, 1988) and

admission and program requirements (O'Sullivan & Gilbert, 1989), as well as about

issues related to ethnicity and gender in curricula (Coleman et al., 1990; Wilson & Stith,

1993). Students have been surveyed about their ethnic minority status as therapists in

training (Wetchler, 1989; Wetchler et al., 1989). Supervisors have been asked about their

training practices (Lewis & Rohrbaugh, 1989; Nichols, Nichols, & Hardy, 1990), about

what they view as essential basic family therapy skills (Figley & Nelson, 1989; Nelson &

Figley, 1990; Nelson, Heilbrun, & Figley, 1993), and about the essential elements of

marriage and family therapy and its supervision (White & Russell, 95).

On occasion, the general population or a class of clients has been surveyed to

determine their experience with a particular issue. Examples include wives' experiences

of their husbands' post-traumatic stress symptoms or combat stress reactions (Solomon,

Ott, & Roach, 1986), couples' experiences of marriage encounter weekends (Doherty,

Lester, & Leigh, 1986), and the effects of differing wake-sleep patterns on marital

relationships (Larson, Crane, & Smith, 1991). Halik, Rosenthal, and Pattison (1990)

measured personal authority (Bray, Williamson, & Malon, 1984) of daughters of Jewish

Holocaust survivors or immigrants. These examples of survey research pertain to family

therapy by virtue of the factors measured, which are often easily extrapolated into family

therapy interventions.

The research enterprise in family therapy has undergone two major transformations

since its early investigations. In the first transformation the field moved from its

impressionistic beginnings to an emphasis on quantitative and experimental research. It








also challenged family therapy researchers to develop reliable and valid measures, which

in conjunction with respectable research designs, helped to gain credibility for a fledgling

discipline in a skeptical clinical world (Sprenkle & Bischoff, 1995).

The second transformation involved a shift from a strict adherence to quantitative

methods to incorporation and gradual acceptance of alternate methodologies, especially

qualitative methods (Hoshmand, 1989; Moon, Dillon, & Sprenkle, 1990; Sprenkle &

Bischoff, 1995). Critics also argued that family therapy had made its quantitative leap

too soon, before clearly delineating what was meant by systemic constructs (Bednar,

Burlingame, & Masters, 1988). Therefore, concepts were operationalized before they

were truly understood and consequently seemed removed from clinical reality. This

called for more attention to contextual variables (Atkinson, Heath, & Chenail, 1991).

In terms of practice, marriage and family therapists and other mental health

professionals routinely assess violence potential for children and adolescents and make

related management decisions in psychiatric emergency services, civil psychiatric

hospitals, juvenile justice, and outpatient clinics. Each of these settings may have

different policy requirements for the evaluations, the amount and quality of available

information may vary, and the nature and decisional thresholds may differ. Each of these

factors can influence the way in which the risk assessment is conducted. Aware of this

diversity, this paper outlines some broad principles for violence risk assessment that may

be useful for marriage and family therapists in assessing risk of general violent

recidivism in various contexts.

In his article on risk assessment Borum discussed historical, clinical, and contextual

categories as the factors that show the most robust empirical support. He stated that the








history of violence would include being a victim of abuse or marital conflict. In his

clinical factors category he included substance abuse problems, mental or behavioral

problems, lack of empathy/remorse and attitudes that support violence. Included in

contextual factors were negative relationships, lack of social support, stress and losses,

community disorganization, and availability of drugs (Borum, 2000).

Data Analyses

The statistical technique used in this study was an exploratory factor analysis

(EFA). The variables operationalized in this study are summarized in Table 1. Validity

and reliability of the RVTS instrument are discussed in this section.

The Nature of Factors

The purpose of factor analysis is to discern and to quantify the dimensions

supposed to underlie mathematical entities, which can be thought of as a classificatory

axis with respect to which the test in a battery can be "plotted." The greater the value of

a test's co-ordinate, or loading, on a factor, the more important is that factor in

accounting for the correlations between the test and other factors in the battery.

An exploratory factor analysis is mainly used as a means of exploring the

underlying factor structure without prior specification of number of factors and their

loadings. In this study the responses were extracted into factors of what was most

commonly identified by the participants in the prevention, assessment, and treatment of

relationship violence.

The common factor model incorporates several parameters worthy of review in this

study. This term is not to be confused with the "common factor" theory across models

within the marriage and family literature, which implies a common theory of ideas

between different theoretical approaches. In this study, common factors were defined as








Table 1

Variables Used for the Relationship Violence Training Survey (RVTS)


Exploratory factor analysis variables


Quality of graduate training
in the identification, assessment,
and treatment of relationship
violence as reported by
Marriage and Family Therapists


Importance of competencies
and skills in the identification,
assessment, and treatment of
relationship violence as reported
by Marriage and Family Therapists


Variables


Rating my knowledge and skills in RV
Assessment and treatment in RV
Working with gay and lesbian clients
Obtaining restraining orders
RV skills today versus 5 years ago
RV screening in premarital counseling

Rating my graduate training received in RV
MFT graduate training programs overall
MFT graduate training that I attended
Intake/assessment of RV
Treatment approaches
Continuing education units in RV

Importance of competencies and skills
Self-knowledge in assessment skills
Identifying/assessing imminent danger
Performing protection assessments rating
Recognizing batterer typologies
Identifying client resources
Recognizing signs and symptoms/
cycle of violence
Assessing ethical standards
Assessing through a multicultural model
Identifying risk factors in relationship
violence
Recommending risk assessment/
instruments


Note. RV = relationship violence.

an unmeasured (or hypothetical) underlying variable that is the source of variation in at

least two observed variables under consideration( Kim & Mueller, 1978).

Thurstone (1947) originally advocated the simple structure principle as reflecting

truth about the psychology of cognition; this is where the concept originated. Thurstone,








at the time of the introduction of simple structure, explicitly regarded factoring as a

scientific revival of an old, discredited, unscientific notion of the principle of parsimony

to supplement the first, by which all of the correlations are explained by as few factors as

possible and each correlation is explained with as few of those factors as possible.

Thurstone stated that factor analysis is a technique to show the correlation of all

tests of mental ability. Thurstone found that all of the mental ability tests were positively

correlated, indicating a common factor among them. The analysis indicated the

following seven primary mental abilities: verbal, number, spatial, perceptual, memory,

reasoning, and word fluency (Thurstone, 1947).

Spearman (1904) viewed factor analysis as a data reduction procedure whereby a

matrix of obtained measurements of N individuals on n experimental variables is replaced

by a smaller matrix of factor coefficients or loadings, relating every variable to each of r

factors, each an underlying variable assumed to represent an ability or other kind of trait,

which is conceived as a vector in r-dimensional space (N > n > r).

According to McDonald (1985), the factor is "most like" the variables that increase

most rapidly as the factor score increases. It is unlike the variables with zero loadings, as

these do not vary as the large factor varies, and least like those variables that have large

negative regression weights on it (i.e., the variables that decrease most rapidly as the

factor score increases).

Stages in a Factor Analysis

The factor analysis in the present study was conducted in three stages.

1. A matrix of correlation coefficients was generated for all of the variable
combinations.








2. Factors were extracted from the correlation matrix. The most common method
is called principal factors (often wrongly referred to as principal components extraction,
hence, the abbreviation PC).

3. The factors (axes) were rotated to maximize the relationship between the
variables and some of the factors. In this study a Promax rotation method was used.
Promax is an oblique rotation method through which a simple structure is sought; factors
are rotated without imposing the orthogonality condition (i.e., that they be kept at right
angles), and resulting terminal factors are in general correlated with each other.

A fourth stage can be added in which the scores of each subject on each of the

factors emerging from the analysis are calculated. It should be stressed that these factor

scores are not the results of any actual test taken by the subjects; they are the estimates of

the subjects' standings on the supposed latent variables that have emerged as

mathematical axes from the factor analysis of the data set. Factor scores can be very

useful, because they can subsequently be used as input for further statistical analysis.

In this research study it was deemed advisable to carry out only the first stage

initially, in order to be able to inspect the correlation coefficients in the correlation matrix

R. Since the purpose of this analysis is to link variables into factors, those variables must

be related to one another and therefore have correlation coefficients larger than a

Cronbach's alpha of .70 and factor loadings greater than .30. These numbers are

consistent with the minimums needed in the social sciences for statistically significant

correlations. Any variables that showed no substantial correlation with any of the others

were removed from R in subsequent analysis. It is also advisable to check that the

correlation matrix does not possess the highly undesirable properties of multicollinearity

and singularity. The former is the condition in which the variables are very highly

(although imperfectly) correlated; the latter arises when some of the variables are exact

linear functions of others in the battery, as when the variable C is constructed by adding

the subjects' scores on variables A and B. Should either multicollinearity or singularity








be present, it would be necessary to drop some of the variables from the analysis (Gray &

Kinnear, 1998).

Rotation of Factors

Factor analysis is a variable reduction technique that simplification of data by

combining numerous variables into a much smaller set of synthetic variables called

"factors." Factor analysis is "designed to identify factors, or dimensions, that underlie

the relations among a set of observed variables" (Pedhazur & Schmelkin, 1991, p. 66).

As Tinsley and Tinsley (1987) noted:

Factor analysis is an analytic technique that permits the reduction of a large number
of interrelated variables to a smaller number of latent or hidden dimensions. The
goal of factor analysis is to achieve parsimony by using the smallest number of
explanatory concepts to explain the maximum amount of common variance in a
correlation matrix. (p. 414)

The ability of factor analysis to detect underlying factors makes it an extremely useful

tool for researchers who want to demonstrate that their results have construct validity.

Similarly, Gorsuch (1983) stated that "a prime use of factor analysis has been in the

development of both the operational constructs for an area and the operational

representatives for the theoretical constructs" (p. 350).

Thurstone's goal in developing his set of guidelines for rotating factors was that

"the factor pattern of any given variable would be constant when the variable was

included in another factor analysis containing the same common factors" (Gorsuch, 1983,

p. 177). This leads to findings that are more replicable across studies. As Gorsuch noted,

"Thurstone showed that such rotation leads to a position being identified for each factor

that would be independent of the number of variables defining it. Therefore, a simple

structure factor should be relatively invariant across studies" (p. 177).








In the present study an oblique rotation was used with a Promax procedure, since it

was the assumption of the researcher that the primary factors might be related. The

generalizability and replication of this research were better served with an oblique

rotation.

Validity in Factor Analysis

Validity is the strength of conclusions, inferences, or propositions. More formally,

Cook and Campbell (1979) defined it as the "best available approximation to the truth or

falsity of a given inference, proposition or conclusion" (p. 93). In short, "Were we

right?"

Validity can be established in a number of ways. The determination of the most

appropriate way depends on the kind of measure. Face validity is usually not enough. If

the variable is meant to assess mastery of subject matter, a test of content validity is

valuable. Usually, a panel of experts agree that all important content areas have been

covered. Construct validity can be determined by an appraisal of the correlation of the

test with other measures of the same trait or ability. Factorial validity confirms the test

with other measures of the same trait or ability. Factorial validity confirms the construct

by showing the strong presence of expected factors in the tests. Criterion-related validity

demonstrates that the test measure correlates highly with the concurrent validity or

predicts future performance (predictive validity). Study validity refers to the validity of

the measure, experiment, and people for the specific designated purpose. No test has

omnibus validity; that is, no one test does it all (Metzloff, 1998).

Professionals have consistently distinguished between actual validity and face

validity. Anastasi (1988) began a section on face validity as follows:








Content validity should not be confused with face validity. The latter is not validity
in the technical sense; it refers, not to what the test actually measures, but to what it
appears superficially to measure. Face validity pertains to whether the test "looks
valid" to the examinees who take it, the administrative personnel who decide on its
use, and other technically untrained observers. (p. 144)

In the present study, content validity, construct validity, and factorial validity were

considered to be important and tests of these forms of validity were applied. The survey

was given to a panel of experts to review the content and constructs, the wording of the

questions, and the extent to which the constructs accurately reflect the variables to be

measured in the study. Feedback from both groups was incorporated into the final survey

format. Tests of the factorial validity were done, once the data were collected and

correlated for various factors.

Reliability in Factor Analysis

Reliability is the consistency of the measurement, or the degree to which the

instrument measures the same way each time it is used under the same conditions with

the same subjects. In short, it is the consistent repeatability of the measure. A measure is

considered reliable if a person's scores on the same test, given twice, are similar. It is

important to remember that reliability is not measured; it is estimated. Reliability is

usually estimated in one of two ways: test/retest or internal consistency. For this study,

the RVTS will be administered only one time, and internal consistency will be used to

estimate reliability.

Internal Consistency

Internal consistency estimates reliability by grouping questions in a questionnaire

that measure the same concept. In the present study two groups of 10 questions each

measured the same concept (e.g., training in relationship violence). This permitted a








measurement of correlation between responses to those two groups of 10 questions to

determine whether the instrument was reliably measuring that concept.

This study applied the common way of computing correlation values among the

questions on the instrument: calculation of Cronbach's Alpha (Cronbach, 1951).

Cronbach's alpha splits all questions on the instrument in every possible way and

computes correlation values for all such combinations. The computerized statistical

analysis software generates one number for Cronbach's alpha; as with a correlation

coefficient, the closer this alpha to 1, the higher the reliability estimate of the instrument.

Cronbach's alpha is a less conservative estimate of reliability than test/retest. The

primary difference between test/retest and internal consistency estimates of reliability is

that test/retest involves two administrations of the measurement instrument, whereas the

internal consistency method involves only one administration of that instrument.

MFT Training Programs

Due to the nature of this research topic, it may be important to look at the status of

graduate-level training being offered by counselor training programs in prevention and

treatment of family violence. This was done to answer the following questions: Is the

curriculum using the current factors that are highlighted in the literature and

incorporating assessment for risk factors? What are the current assessments tools in

family violence prevention and treatment? Is imminent danger being assessed and are

safety plans being employed for victims at risk? Is administration of these instruments

being taught in counselor preparation programs; if so, are the assessment instruments for

violence used with each case that the marriage and family therapy graduate sees?

Rating of the graduate training program themselves will help in a correlation of

marriage and family therapy graduates, licensed practitioners, faculty, and the training