EFFECTS OF COUPLES' PERCEPTIONS OF GENOGRAM

CONSTRUCTION ON THERAPEUTIC ALLIANCE AND

SESSION IMPACT: A GROWTH CURVE ANALYSIS

by

SCOTT COUPLAND, B.S., M.S., M.A.

A DISSERTATION

IN

MARRIAGE AND FAMILY THERAPY

Submitted to the Graduate Faculty of Texas Tech University in

Partial Fulfillment of the Requirements for

the Degree of

DOCTOR OF PHILOSOPHY

Approved

Accepted

December, 1998

Copyright 1998, Scott Keith Coupland

ACKNOWLEDGEMENTS

When I was considering whether to pursue doctoral studies, a professor in my

master's degree program told me, 'The doctorate degree does not necessarily go to the

most intelligent, but to the one who can endure." There is much truth in that statement.

Now that I am at the end of my doctoral studies journey, I am indebted to many people

who have endured with me.

First, I am grateful to the faculty members of my committee and the student

participants who made this project possible. To Dr. David Ivey, your leadership,

feedback, and emotional support were invaluable. To Dr. Julie Serovich, your

involvement in my training from start to finish was a gift. Thank you for going the "extra

mile." To Dr. Karen Wampler, your kindness and wisdom will continue to direct my

academic life. To Dr. Jeff Elias, the design of this study reflects your methodological

expertise. To the students in the marriage and family therapy training programs at Texas

Tech University and Abilene Christian University who were willing to add one more

thing to their busy lives—this project could not have been completed without your active

participation.

Second, I am beholden to my friends. To Dr. Martin Dawson, Dr. Tim Dinger, and

David Riethmeyer, you have been my lifeline. I thank you for fighting for and along side

me. You are noble men and the truest of friends.

Third, I am thankful to my family. To my parents, Jim and Carolyn Coupland, all I

can say is, "genetics show." To my two daughters, Brynne and Cami, who value me more

as a daddy than a doctor—you are a joy to my heart. Last and most important, to my

wife, Joan—the labor pains that gave birth to this project and doctoral degree are, in

many ways, more yours than mine. The extent of your sacrifice on my behalf brings tears

to my eyes and floods my being with gratitude, for I have known no greater picture of

Jesus Christ's sacrifice than the ways in which you have laid down your life for me. I

love you.

II

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ii

ABSTRACT vi

LIST OF TABLES ix

LIST OF FIGURES xi

CHAPTER

L INTRODUCTION 1

Questions 3

n. REVIEW OF LFFERATURE 5

The Genogram 5

The Development of the Genogram 5

Diversity in Standardization Methods of Genogram Construction... 6

The Significance of the Diversity in Standardization Methods of Genogram Construction 7

The Genogram Construction Process 8

Use of the Genogram 9

Therapeutic-Alliance 12

Development of the Therapeutic Alliance Concept in Individual Psychotherapy 12 12

T4ie Relation Between the Therapeutic Alliance and Outcome in Individual Psychotherapy 14

Therapeutic Alliance in Marriage and Family Therapy 15

A Systemic Therapeutic Alliance Theory 17

Therapeutic Alliance Defined 18

Empirical Findings of the Therapeutic Alliance in Couple Therapy 20

The Relation Between Therapeutic Alliance and Outcome 20

CTAS Content Subscale Research 22

Genogram Construction and Therapeutic Alliance 24

Hypotheses 26

111

ffl. METHODS

Samples and Sampling Procedures 28

Participants 28

Solicitation of Participants 35

Therapists 36

Solicitation of Therapists 37

Procedure 43

Assignment of Therapists to Couple Therapy Cases 43

Assignment of Couples to Treatment Groups 43

Administration of Couple Therapy 43

Administration of Treatment Interventions 45

The TAGE Treatment Intervention 45

The SAGE Treatment Intervention 46

The DAI Treatment Intervention (Control) 47

Administration of the Dependent Measures 47

Monitoring Treatment Interventions and Data Collection 50

Instruments 50

Treatment Intervention Instruments 50

The TAGE 50

The SAGE 51

The DAI 52

Dependent Measures 52

CTAS 52

SEQ 54

Demographic Questionnaires 55

Session Follow-up Questionnaire 55

Design 56

Analysis Plan 56

Preliminary Analyses 58

IV

Stage One Analyses 58

Stage Two Analysis 59

Other Analyses 59

rv. RESULTS

Preliminary Analyses 61

The Reliability of the Dependent Measures 61

Group Differences Analysis 61

Participants 61

Therapists 67

Time for Second Session Treatment Intervention 67

Stage One Analyses 72

Stage Two Analyses 91

Other Analyses 93

Therapy Completers versus Noncompleters 93

Split Alliance 95

V. DISCUSSION

Review of Results 98

Within-Couple Member Differences 98

Between-Couple-Member Differences 99

CTAS Subscale and SEQ Index Intercorrelations 102

Construct Validity of the CTAS Content Subscale 103

Split Alliance 104

Limitations 104

Future Research 105

Clinical Implications 106

Summary 107

REFERENCES 109

APPENDICES

A. CONSENT FORM FOR PARTICIPANTS 127

B. DEBRIEFING FORM FOR PARTICIPANTS 130

C. SAMPLE NOTICE FOR THE SOLICFTATION OF THERAPISTS 132

D. CONSENT FORM FOR THERAPISTS 135

E. SAGE 138

F. DAI 151

G. DEPENDENT MEASURES COVER PAGE 157

H. CTAS 159

L SEQ 164

J. DEMOGRAPHIC QUESTIONNAIRE FOR PARTICIPANTS 166

K. DEMOGRAPHIC QUESTIONNARIE FOR THERAPISTS 168

L. SESSION FOLLOW-UP QUESTIONNAIRE FOR THERAPISTS 171

M. SAMPLE FOLLOW-UP LETTER TO THERAPISTS 173

VI

ABSTRACT

Preliminary evidence of the predictive validity of therapeutic alliance to outcome m

couple therapy has recently been empirically established. Clinicians have asserted that

genogram construction is an effective method for building the therapist-client relationship

and impacting specific aspects of therapy sessions. However, this claim had not been

empirically tested in a therapeutic context.

The purpose of this study was to examine how the construction of a therapist-

administered genogram (TAGE) or a self-administered genogram (SAGE) would affect

the therapeutic alliance and the session character in couple therapy across time. Eight

student marriage and family therapists conducting couple therapy with 17 couples

administered a TAGE, SAGE, or control treatment intervention in the second therapy

session. Clients' perceptions of therapeutic alliance and session impact were measured

with the Couple Therapy Alliance Scale (CTAS, Pinsof, 1986) and the Session

Evaluation Questionnaire (SEQ, Stiles, 1984; Stiles & Snow, 1984a), respectively,

following the first five sessions.

A growth curve analysis was used to investigate within-couple member growth

trajectories for individual patterns of change on the CTAS and SEQ and their

subdimensions, and between-couple growth trajectories for treatment group-based

patterns of change on the CTAS and SEQ and their subdimensions. Also examined was

whether the growth curves of couples who completed five sessions were different from

couples who terminated therapy prior to completing five sessions, and if Pinsof and

Catherall's (1986) concept of split alliance was identifiable.

A series of simple regressions were conducted on mean CTAS and SEQ ratings.

Inspection of individual growth curves revealed minimally positive slopes on both

dependent variables for the sample as a whole. Results of one-way MANOVAs showed

no significant differences in couple members' growth curves based either on their

participation in one of the three treatment groups, or when they were divided into therapy

completer and noncompleter groups. In other words, the efficacy of genogram

VII

construction to affect the therapeutic alliance or the depth and smoothness of sessions in

early therapy was called into question.

When Spearman's rho and Kendall's tau correlation coefficients were used to

examine the associations between domains of CTAS and SEQ by treatment group, only

the CTAS tasks subscale was positively correlated with the SEQ depth index in the

SAGE plus couple therapy (-HCT) group. This finding indicated that couples who

completed a SAGE related the therapist's understanding of, and methods for addressing

their difficulties, with the "deep" work of therapy.

Finally, a two-standard deviation rule was applied to differences in mean slope

scores between couple members to evaluate the presence of a split alliance. When this

mle was applied to differences in mean slope scores for the whole sample, six percent of

the couples experienced a split alliance. When applied within treatment groups, 14% of

couples in the TAGE-fCT group and 13% of couples in the SAGE-i-CT group experienced

a split alliance. Relatively few couples, therefore, had disparate attitudes regarding their

alliance with their therapist. This result may have reflected a strength of genogram

construction.

When considered in light of the studies limitations, particularly a small sample size

and data collection weaknesses, these findings, in general, do not support genogram

construction as a method for building a therapeutic alliance or enhancing the depth and

smoothness of early couple therapy. Recommendations were made for future genogram

constmction research.

V I I I

LIST OF TABLES

3.1 Participant Demographics 29

3.2 Participant Demographics by Group 32

3.3 Therapist Demographics 38

3.4 Crosstabulation of the Student Therapists Self-rated Ability to Constmct Genograms and Student Therapists 39

3.5 Crosstabulation of the Student Therapists Clinical Skill Level and Student Therapists 40

3.6 Crosstabulation of the Percentage of Time The Student Therapists Use the Genogram and Student Therapists 41

3.7 Crosstabulation the Theory Used by the Student Therapists and Student Therapists 42

4.1 Cronbach Alpha Reliability Coefficients for the CTAS and the SEQ 62

4.2 Means, Standard Deviations, and One-way ANOVA Summary Table for Participant's Initial CTAS Mean Scores by Participants' Genders, Ethnic Backgrounds, and Relationship Statuses 63

4.3 Means, Standard Deviations, and One-way ANOVA Summary Table for Participant's Initial SEQ Mean Scores by Participants' Genders, Ethnic Backgrounds, and Relationship Statuses 64

4.4 Simple Regression Analyses of Participant's Initial CTAS Mean Scores on Participants' Ages, Lengths of Relationships, and Levels of Education 65

4.5 Simple Regression Analyses of Participant's Initial SEQ Mean Scores on Participants' Ages, Lengths of Relationships, and Levels of Education 66

4.6 Means, Standard Deviations, and One-way ANOVA Summary Table for Participant's Initial CTAS Mean Scores by Therapist's Gender, Ethnic Background, Degree Program, and Theory Used 68

4.7 Means, Standard Deviations, and One-way ANOVA Summary Table for Participant's Initial SEQ Mean Scores by Therapist's Gender, Ethnic Background, Degree Program, and Theory Used 69

IX

4.8 Simple Regression Analyses of Participant's Initial CTAS Mean Scores on Therapist's Ability to Construct Genogram, Age, Clinical Experience, Clinical Skill Level, Percent of Time Genogram Used 70

4.9 Simple Regression Analyses of Participant's Initial SEQ Mean Scores on Therapist's Ability to Constmct Genogram, Age, Clinical Experience, Clinical Skill Level, Percent of Time Genogram Used 71

4.10 One-way MANOVA Summary Table for Planned Comparisons of Participants' CTAS and SEQ Mean Scores as a Function of Group 94

4.11 Intercorrelations of the CTAS Content Subscales and SEQ Indexes by Treatment Intervention Group 96

LIST OF FIGURES

3.1 3 x 5 Multiple-Group-Multiple-I Times Series Experimental Design 31

3.2 Examples of the Coding Systems for Envelopes and Form Packets 49

4.1 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple One 73

4.2 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Two 74

4.3 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Three 75

4.4 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Four 76

4.5 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Five 77

4.6 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Six 78

4.7 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Seven 79

4.8 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Eight 80

4.9 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Nine 81

XI

4.10 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Ten 82

4.11 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Eleven 83

4.12 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Twelve 84

4.13 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Thirteen 85

4.14 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Fourteen 86

4.15 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Fifteen 87

4.16 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Sixteen 88

4.17 Regression Lines Displaying the Relationship between Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for Couple Seventeen 89

4.18 Box-and-Whiskers Plot of Mean CTAS and SEQ Slope Scores 90

4.19 Regression Lines Displaying the Relationship between Collected Mean Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number for the Entire Participant Sample by Gender 92

x i i

CHAPTER I

INTRODUCTION

"The beginning phase of family therapy is in many ways the most important, since

it sets the stage for all later work. At the same time, it is often the most difficult phase"

(Feldman, 1976, p. 103). From the perspective of the therapist, the complicated nature of

initial marriage and family sessions is in a large part due to the multitude of activities that

must be juggled. For example, common issues that are usually addressed early in therapy

include: engaging all family members, building both a rapport and a working relationship

with all family members, gaining a temporary position of influence within the family,

establishing credibility, collecting information, sifting through content information to

identify important process information, setting goals, directing interventions, assigning

tasks, assessing family functioning, establishing boundaries, avoiding coalitions, and

stmcturing and pacing the sessions. Successful management of such activities can set the

stage for further engagement of the family and for therapy outcome.

One technique widely used in early therapy sessions by marriage and family

therapists to accomplish these management tasks is genogram constmction. Because of

its flexibility and practicality, the genogram has gained tremendous popularity as a

clinical tool. Its use in research, however, has been limited. To date, the genogram has

been used in research for collecting data (de Chesnay, Marshall, & Clements, 1988; Jolly,

Froom, & Rosen, 1980; Rogers & Cohn, 1987; Rogers & Durkin, 1984; Rogers, Durkin,

& Kelly, 1985; Shellenberger, Couch, & Drake, 1989; Talan, 1983; Vukov & Eljdupovic,

1991) evaluating interventions (IngersoU-Dayton, & Arndt, 1990), predicting future

health risks (Haas-Cunningham, 1993; Rogers, Rohrbaugh, & McGoldrick, 1992;

Rohrbaugh, Rogers, & McGoldrick, 1992 ), exploring the meaning to an individual of the

process of constmcting a genogram (Erianger, 1987), examining ways it sensitizes

physicians to clinical care issues (Rogers & Rohrbaugh, 1991), and assessing the

reliability of those who constmct them (Coupland, Serovich, & Glenn, 1995; Jolly et al.,

1980; Rogers et al., 1992; Rogers & Holloway, 1990; Rogers & Rohrbaugh, 1991;

1

Rohrbaugh et al., 1992). Of these studies, only two examined the genogram constmction

process from the nonclinician's perspective (Erianger, 1987; Rogers & Durkin, 1984),

and neither of these studies was conducted in a therapeutic setting. Thus, nothing is

known about how genogram constmction affects either specific processes of therapy, or

client attitudes toward therapy sessions.

Several marriage and family therapists (e.g.. Beck, 1987; Braverman, 1997;

Erianger, 1991; Ingersoll-Dayton & Arndt, 1990; Ivey, Ivey, & Simek-Morgan, 1996;

Kuehl, 1995; Miehls, 1992; Wachtel, 1982) have asserted that one process of therapy that

genogram constmction will impact is the establishment of a therapeutic alliance. This is

significant because several recent reviews of therapeutic alliance outcome literature in

individual psychotherapy have concluded that early therapeutic alliance is an important

mediator of psychotherapy outcome (Gaston, 1990; Henry, Stmpp, Schacht, & Gaston,

1994; Horvath, 1994; Horvath, Gaston, & Luborsky, 1993; Horvath & Luborsky, 1993;

Horvath & Symonds, 1991; Luborsky, 1994). Furthermore, four couple therapy studies

have provided preliminary evidence of the predictive validity of therapeutic alliance

(Bourgeios, Sabourin, & Wright, 1990; Holtzworth-Munroe, Jacobson, DeKlyen, &

Whisman, 1989; Johnson & Talitman, 1997; Quinn, Dotson, & Jordan, 1997).

Other marriage and family therapists have suggested that the process of

constmcting a genogram will influence clients' attitudes about their satisfaction with

(Kramer, 1985), the power (Kramer, 1985) and meaningfulness of (Brock & Barnard,

1988) therapy, and "enhance their overall experience" (Beck, 1987, p. 350). Although

researchers have reported that nonclinical individuals (Erianger, 1987), and family

medicine patients (Rogers & Durkin, 1984) expressed positive benefits from constmcting

their own genograms, these investigations were not conducted in a therapeutic context.

Thus, research has yet to address whether the constmction of a genogram, either by a

therapist or a client, will affect eariy therapeutic alliance or session impact in a

therapeutic setting. Because of the empirical link between therapeutic alliance and

therapy outcome, understanding what activities enhance therapeutic alliance is important.

In addition, knowledge that genogram constmction influences specific client attitudes

about their therapy could provide valuable information for clinicians that employ the tool.

One limitation to previous therapeutic alliance research in individual psychotherapy

and couple therapy has been its assessment with either single time point measurements,

measurements averaged across sessions, or pre/post measurements. While information

gained from these designs can be useful in describing therapeutic alliance ratings at one

or more time points or the amount of change taking place in therapeutic alliance ratings

between arbitrary time points, it can not provide a description of the intricacies of change

in therapeutic alliance ratings across time. Such a description could provide a superior

basis for answering questions regarding systematic interindividual differences in change

over time on therapeutic alliance ratings and session impact ratings.

Some researchers have conjectured about how therapeutic alliance and session

impact might be related. Heatherington and Friedlander (1990) hypothesized that because

tasks and goals represent the "work" of therapy, client and therapist collaboration in these

areas will be reflected in sessions that are perceived as deeper. They also asserted that a

smooth session will be more likely when the emotional bonds between the client and the

therapist are perceived to be relatively stronger. In their study to test these assertions,

these researchers took measurements of therapeutic alliance and session impact at a single

time point (between the third and sixth session) in couple therapy, and found only ratings

on session tasks and depth to be significantly correlated. It remains yet unknown,

however, what the pattern of these relations look like across sessions. Therefore, a

secondary interest of this study is the associations between the tasks, goals, and bonds

ratings on the therapeutic alliance instmment and clients' perceptions of depth and

smoothness on the session impact instmment, across time.

Questions

Based on the brief summaries presented above, two types of questions will be

examined in this study. The first concerns within-couple member growth: Do couple

members' perceptions of therapist-couple alliance and session impact change over time in

couple members who have (a) constmcted a therapist-administered genogram (TAGE)

and participated in couple therapy, (b) constructed a self-administered genogram (SAGE)

and participated in couple therapy, or (c) participated in directed assessment interview

(DAI) plus couple therapy? Such questions naturally precede investigation of how

within-couple member changes differ from one couple member to the next. The second

question, therefore, is about systematic between-couple member differences in growth:

What are the effects of the three experimental groups on the stmcture of couple member

growth curves, including effects on growth rates, on the shape of the curvature of growth

curves, and on the variability of growth trajectories. In other words, are changes in

therapist-couple alliance and session impact systematically related to having either (a)

constmcted a TAGE and participated in couple therapy, or (b) constmcted a SAGE and

participated in couple therapy, or (c) participated in a DAI plus couple therapy?

An secondary question is whether the subscales of the therapeutic alliance

instmment and indexes of the session impact instmment may be changing in the same

way across treatment groups. That is, to what degree will the tasks and/or goals subscales

of the therapeutic alliance instmment be correlated with the depth index of the session

impact instmment over time? Also, to what degree will the bonds subscales of the

therapeutic alliance instmment be correlated with the smoothness index of the session

impact instrument over time?

CHAPTER n

REVIEW OF LFFERATURE

This chapter first will address three major topics related to the genogram: (a; the

development of the genogram, (b) the genogram constmction process, and (c) the use of

the genogram. Next, five major topics concerning the concept of therapeutic alliance will

be examined: (a) the development of the therapeutic alliance in individual

psychotherapy, (b) the relation between therapeutic alliance and outcome in individual

psychotherapy, (c) therapeutic alliance in marriage and family therapy, (d) a systemic

theory of therapeutic alliance, and (e) the relation between therapeutic alliance and

outcome in couple therapy. Finally, conclusions and hypotheses will be presented.

The Genogram

In the following section several areas of the genogram literature will be reviewed.

The review begins by examining how the proliferative use of the genogram across

professional disciplines has led to the development of numerous genogram constmction

methods, and the significance of this for clinical practice and research. This will be

followed by a discussion of the genogram constmction process with a distinction made

between SAGEs and TAGEs. Lastly, examples of common ways the genogram is used in

data collection, assessment, intervention, and intervention evaluation will be provided.

The Development of the Genogram

In the late 1960's, Murray Bowen and Jack Bradt, two family-oriented psychiatrists,

independently began to employ family charts drawn on blackboards or poster boards in

their clinical and theoretical work with families (Bowen, 1966, 1978; Bradt, 1980; Bradt

& Moynihan, 1971). This activity went without a formal name until 1972 when the name

"genogram" as a label for these family schematic diagrams was introduced into the

marriage and family therapy literature by Guerin (1972). The remainder of the decade

saw a growing recognition of the utility and practicality of the genogram in the marriage

and family therapy movement. Writings in marriage and family therapy literature during

these years documented the expanding use and development of the genogram (e.g.,

Barnard & Corrales, 1979; Carter & Orfanidis, 1976; Guerin & Pendagast, 1976;

Lieberman, 1979a, 1979b; Ori'anidis, 1979; Pendagast & Sherman, 1977). In fact,

Pendagast and Sherman (1977) noted that "the use of a genogram in the study of a family

is now as basic a piece of information as the family surname" (p. 3).

The seminal figure in the early group of genogram users was Murray Bowen with

the conceptual framework for genogram constmction and analysis originally an outgrowth

of his family systems theory (Bowen, 1978). Despite its primary association with

Bowen's theoretical perspective, the popularity of the genogram has broadened not only

to other postulated individual, family, and group therapy orientations, but also to

professional disciplines concerned with family dynamics such as career counseling (e.g..

Fobs, 1991; Okiishi, 1987; Splete & Freeman-George, 1985), family medicine (e.g..

Crouch & Davis, 1987; Mullins & Christie-Seely, 1984; Shore, Wilkie, & Croughan-

Minihane, 1994), nursing (e.g., Dobson, 1989; Herth, 1989; Smoyak, 1982), and social

work (e.g., Hartman & Laird, 1983; Ingersoll-Dayton & Arndt, 1990; Schames, 1990). In

addition, the genogram has been tailored to address specific issues in therapy (e.g., client-

generated solutions [Kuehl, 1995], cultural awareness [Estrada & Haney, 1998; Hardy &

Laszloffy, 1992, 1995;McGill, 1992], gender [Sherman, 1991], intimacy patterns

[Sherman, 1993], illness patterns [McDaniel, Campbell, & Seaburn, 1990], childbirth

preparation [Condon, 1985], child adoption placement [McMillen & Groze, 1994], sexual

history and beliefs [Hof & Berman, 1986]). Because of the genogram's utility and

flexibility, its application will likely continue to increase.

Diversity in Standardized Methods of Genogram Constmction

The multiplicity of therapeutic perspectives and professional fields utilizing the

genogram has led to considerable variation in how genograms are constmcted.

According to McGoldrick and Gerson (1985), "Even among clinicians with similar

theoretical orientations, there is only a loose consensus about what specific information to

seek, how to record it, and what it all means" (p. I). In response to this diversity.

McGoldrick and Gerson (1985) formulated the first comprehensive standardization for

constmcting and interpreting genograms based on concepts articulated by Bowen and

other systems theorists. The constmctive impact of McGoldrick and Gerson's (1985)

standardization effort is noted by frequent references to their work in recent marriage and

family therapy literature; nevertheless, unique symbols and formats used in genogram

constmction still speckle this literature (Serovich, Glenn, & Coupland, 1993).

An example of a unique genogram format was created by Friedman, Rohrbaugh,

and Krakauer (1988). They developed the 'time-line genogram " that "is plotted on the

vertical axis so that events can be shown when they actually occurred'" (Friedman et al.,

1988. pp. 293-294). Unlike the static nature of the standard genogram, the time-line is

designed to capture the evolution of relationships and events over time. The authors

acknowledged that one disadvantage of this design is that it can be difficult to read.

Another creative addition to the genogram format suggested for use with either the

standard or time-line genogram is assigning colors to accent characteristics of family

members or issues associated with particular presenting problems (Lewis, 1989).

Although such innovation was encouraged by McGoldrick and Gerson (1985), these two

ideas have received comparatively little recognition in the family therapy literature next

to the standard genogram design.

The Significance of the Diversity in Standardized Methods of Genogram Constmction

A standardized genogram format has appeal because it provides a "default" method

for genogram constmction that promotes consistency of use and understanding for those

who employ it. This is important clinically for case presentations, supervision, and

consultation, and the transference of case records. In practice, however, a standardized

genogram format often serves more as a general framework than a regimented protocol,

thus permitting creativity in the constmction process. Consequently, it is possible that the

specific information recorded on genograms, as well as the method of genogram

constmction, will vary within and across clinicians. Thus, whether clinicians are reliable

in constmcting genograms becomes questionable. Several empirical investigations have

been conducted addressing this concern. In one study, Coupland and colleagues (1995)

found that the accuracy by which marriage and family therapy doctoral students recorded

content data on genograms varied highly across categories of genogram information. In

fact, in their review of genogram research conducted in medical settings, Coupland and

colleagues (1995) reported mixed reliability findings across studies.

Such variability in the reliability of genogram constmctors can be problematic in a

research context. For example, de Chesnay and colleagues (1988) recmited marriage and

family therapists to submit genograms of families where father-daughter incest had

occurred so that the genograms could be analyzed for specific family variables and

stmctures. The authors noted the inconsistencies across the marriage and family

therapist's assessment techniques, which included methods of genogram constmction,

were a major limitation to the reliability and validity of their analysis. It may be

important, therefore, that future genogram research employ a chosen standardized format

that is consistent within and across all therapists involved in the research.

The Genogram Constmction Process

Regardless of the format, a genogram is a diagrammatic representation of a family

covering at least three generations. It displays a picture of family membership and

stmcture as well as demographic and relational data about family members. Constmction

of a basic genogram usually begins in the initial therapy session (Guerin & Pendagast,

1976), but sometimes as late as the third therapy session (Wachtel, 1982). A single-

session genogram interview may take from 30 to 90 minutes. Some clinicians, however,

prefer to spread this process out over several sessions since creating a genogram is a

process, and not a "sterile shorthand of history and relatedness" (Lierberman, 1979b, p.

74). Variance in constmction time may also be due to the number of persons

participating in the interview, and the theoretical perspective and intention of each

clinician.

8

In most cases, therapists constmct genograms as they interview clients during

therapy sessions. Such TAGEs generally are drawn on paper so they can be kept as part

of the case record. Some therapists, by contrast, either send clients genogram forms to

complete prior to attending the initial interview (Paul & Paul. 1982). or have clients

constmct their genograms outside the therapy session (Woolf, 1983). A 'self-

administered genogram" (Rogers & Cohn, 1987) was adapted by Erianger (1987) to

systematically instmct clients how to constmct their genograms apart from the assistance

of a therapist.

Researchers have conducted empirical studies with two dissimilar samples to

examine the constmction process of the SAGE. With the first sample, patients in a

family medicine clinic, the constmction of a SAGE was demonstrated to be a fairly

efficient method for obtaining specific family information (Rogers & Cohn, 1987; Rogers

et al., 1985), with overall mean completion rates ranging between 70% (Rogers &

Rohrbaugh, 1991) and 74% (Rogers & Holloway, 1990). In addition, Rogers et al. (1985)

reported that two-thirds of the patients who completed a SAGE thought the information

would improve their medical care. With the second sample, nonclinical individuals in

their homes, Erianger (1987) used the SAGE to investigate the meaning the genogram

constmction process had for individuals. She found that the quantity and type of

information produced by the SAGE was consistent across age groups and gender, and that

completing a SAGE was a highly positive experience for men and women. These studies

suggest that the SAGE is a good method for collecting family data apart from the

assistance of a clinician, and that individuals who complete them perceive benefit from

the experience. To date, research using a SAGE has not been conducted in a therapy

context. In addition, it is not known if similar results could be obtained with a therapist-

mediated genogram.

Use of the Genogram

Despite when or how genograms are constmcted, the instmment is useful clinically

in four different ways. First, a genogram is an efficient method for collecting, organizing.

processing, and reviewing an extensive amount of family information. Three general

categories of data that can be recorded on a genogram are: (a) family structure (biological

and legal relationships between family members): (b) descriptive information about

individual family members (demographic data, factual or imagined emotional, behavioral,

personality, and medical data); and (c) relationship descriptors (triangles, patterns of

conflict, distance, enmeshment, and estrangement (McGoldrick & Gerson. 1985).

Numerous studies (de Chesnay et al., 1988; Shellenberger et al., 1989; Talan, 1983:

Vukov & Eljdupovic, 1991) have utilized genogram constmction interviews to collect

and organize family information, and build databases in efforts to understand particular

family processes. Collectively, these studies lend support to the potential research and

clinical merits of genograms as data collection instmments.

Once family information is collected on a genogram, a second way it can be used is

in assessment. This may involve making tentative hypotheses, interpretations or

predictions about how various stressors (e.g., family pattems, life cycle transitions) have

contributed, are contributing, or will contribute to symptomatology, presenting problems,

or possible therapeutic solutions and preventive care decisions. The ability to use the

genogram to predict future health problems has been addressed in two studies. In the first

investigation (Rogers et al., 1992), six family physicians each evaluated 20 patient

genograms to assess for future health risk. Twenty-six categories of genogram

information were used to make these predictions. When the predictions were compared

to actual patient outcomes three months later, the genogram expert's predictions were no

more accurate than chance levels. In the second study (Haas-Cunningham, 1993), 63

patient genograms were assessed at intake and two to three years later to evaluate the

efficacy of the genograms for predicting patients at risk for physical illness and poor

health status. Predictions were based on 44 genogram items. The results indicated a

significant relation between the genogram and a majority of the health outcome variables.

From these limited findings it would appear that the usefulness of the genogram as a

predictive measure of health outcomes in family medicine settings has potential value for

longer-term predictions (i.e., two to three years). These studies, however, have not been

10

replicated, and genogram assessment studies have yet to be pursued in marriage and

family therapy settings.

A third use of genograms is to plan and administer strategic interventions. Wachtel

(1982) noted that the genogram provides "a context which makes sense" for intervening

with clients (p. 338). Interventions can be catalyzed by previously formulated

hypotheses, interpretations, or predictions devised from genogram information. Some

examples of how genogram constmction has been used clinically as a method of

intervention are: (a) assisting client's recall of negative and positive memories about

family relationships (Crosby, 1989), (b) helping children from fragmented homes

communicate about changing family roles and relationships in a group therapy context

(Davis, Geikie, & Schamess, 1988), (c) reframing and detoxifying family legacies

(Gewirtzman, 1988), and (d) building a therapeutic alliance (Beck, 1987). The ways

genograms can be used to intervene with clients are limited mainly by clinician's

creativity.

Finally, genograms can be employed to assess the extent to which an intervention

has been effective. The single documentation of the genogram having been used as an

intervention evaluation tool was Ingersoll-Dayton and Arndt's (1990) study examining

the quality of 41 elderly person's family relationships following short-term family

therapy. Pre-treatment and post-treatment genograms were constmcted and compared

with two measures of treatment effectiveness to assess change in dimensions of closeness

and positive/negative interactions among family members. Several significant

correlations were found between changes in family dynamics as portrayed by the

genogram and changes on the therapy effectiveness measures. These findings suggest

promise for the use of genograms as treatment evaluation instmments.

In summary, the clinical utility of genograms has been well established during the

past 30 years. The use of this instmment in research, however, has been limited. This

may be due, in part, to diversity in genogram constmction methods, and to mixed

reliability results of genogram constmctors (see Coupland et al., 1995). To date, in a

therapeutic setting, researchers have not employed the SAGE, and have used the TAGEs

11

only for data collection purposes, with one exception (Ingersoll-Dayton & Arndt, 1990).

To say that much future genogram research is necessary is to state the obvious.

Therapeutic Alliance

In this section, literature concerning the concept of therapeutic alliance will be

reviewed, first in individual psychotherapy, and second, in marriage and family therapy.

To begin, the context surrounding the development of a theoretical and research base of

the therapeutic alliance in individual psychotherapy will be traced. Then, a summary of

empirical findings regarding the relation between the therapeutic alliance and therapy

outcome will be provided. The remainder of the section will focus on the therapeutic

alliance in marriage and family therapy. The shift from a pragmatic concern with

relationship factors to a more recent theoretical interest in the therapeutic alliance will be

tracked. A systems-oriented therapeutic alliance theory then will be explicated and

followed by an examination of limited, but supportive research.

Development of the Therapeutic Alliance Concept in Individual Psychotherapy

Therapeutic alliance, as a clinical concept, has a rich and evolving history in

individual psychotherapy dating back to the eariy days of psychoanalysis. It was not,

however, until 1976 that the role and function of the therapeutic alliance began to be

investigated empirically. Since this time, the field of individual psychotherapy has

experienced a surge in research on the therapeutic alliance (also called "helping alliance"

or "working alliance").

The initial impetus behind the development of the body of research on the

therapeutic alliance can be largely attributed to three factors. First, there was an emerging

interest in relationship variables common to all forms of therapy (Frank, 1961; Rogers,

1951, 1957). Second, eariy reviews using either traditional narrative methods of

summarizing comparative outcome research (Bergin & Lambert, 1978; Bergin & Suinn,

1975; Beutler, 1979; Goldstein & Stein, 1976; Kellner, 1975; Lambert & Bergin, 1973;

Luborsky, Singer, & Luborsky, 1975; Meltzoff & Kornreich, 1970; Roback, 1971), or

12

meta-analyses (Smith & Glass, 1977; Smith, Glass, & Miller, 1980). suggested that,

overall, diverse therapies produced comparable therapeutic outcomes. Despite

methodological and conceptual problems, these findings were a harbinger of Lambert and

Bergin's (1994) recent conclusion from their review of psychotherapy outcome literature

that "common factors loom large as mediators of treatment outcome" (p. 167).

Significantly, the therapist-client relationship stands as one of the most likely common

denominators among different approaches (Horvath & Symonds, 1991; Lambert, 1986,

1992; Luborsky, 1984; Oriinsky, Grawe, & Parks, 1994; Oriinsky & Howard, 1986).

Third, the original psychodynamic conceptualization of the therapeutic alliance concept

was broadened and formulated in pantheoretical terms by Bordin (1979). This

formulation permitted the therapeutic alliance to be conceptualized in light of the

possibility that it was a common factor to all forms of treatment.

In this context, operationalizing alliance became a cardinal tasks for researchers.

Several research groups developed instmments independently and more or less

simultaneously to measure therapeutic alliance (e.g., the California Therapeutic Alliance

Rating System [Marmar, Horowitz, Weiss, & Marziali, 1986; Marziali, Marmar, &

Kmpnik, 1981]; the California Psychotherapy Alliance Scales [Marmar, Gaston,

Thompson, & Gallagher, 1989]; the Penn Helping Alliance Scales [Luborsky, 1976;

Luborsky, 1984; Luborsky, Cris-Christoph, Alexander, Margolis, & Cohen, 1983;

Luborsky, McLellan, Woody, O'Brien, & Auerbach, 1985; Morgan, Luborsky, Cris-

Christoph, Curtis, & Solomon, 1982]; the Therapeutic Alliance Scale [Marziali, 1984];

the Vanderbilt Psychotherapy Process Scale [Gomes-Schwartz, 1978; Gomes-Schwartz &

Schwartz, 1978; O'Malley, Suh, & Stmpp, 1983]; the Vanderbilt Therapeutic Alliance

Scale [Hartley & Stmpp, 1983]; the Working Alliance Inventory [Horvath, 1981]). Most •

therapeutic alliance research conducted to date has utilized these instmments and their

derivatives.

13

The Relation Between Therapeutic Alliance and Outcome in Individual Psychotherapy

Several recent reviews of therapeutic alliance research (Gaston, 1990; Henry et al..

1994; Horvath, 1994; Horvath et al., 1993; Horvath & Luborsky, 1993; Horvath &

Symonds, 1991; Luborsky, 1994) have chronicled empirical evidence concerning the

relation between therapist-client alliance and outcome. One of the consistent findings

across a variety of treatment populations, treatment lengths, types of therapies, and

different therapeutic alliance and outcome measures was that a good therapeutic alliance

is predictive of positive therapy outcome.

To add to these findings, Horvath and Symonds (1991) conducted a meta-analysis

of 24 therapeutic alliance studies published between 1978 and 1991. Across these

studies, the mean sample size was 49, and 70% of the participants were female.

Treatment lengths averaged 20.6 sessions, and the therapists averaged 8.1 years of

experience. They found that quality of the therapeutic alliance accounted for a significant

proportion outcome variance, with an overall effect size of .26. This estimate was

conservative, however, because the calculation was based on the cautious assumption that

all correlations observed but not reported by investigators or reported as not significant

equaled zero. It is possible that the tme magnitude of this relation is as high as .32

(Horvath, 1994).

In addition, Horvath and Symonds (1991) observed that eariy therapeutic alliance

(first to fifth session) was slightly more predictive of outcome (effect size = .30) than

therapeutic alliance measures averaged across sessions or taken toward the middle of

treatment (effect size = .21). Even though these values are neariy identical, other studies

(DeRubeis & Freely, 1991; Piper, Azim, Joyce, & McCallum, 1991a; Piper et al., 1991b;

Plotnicov, 1990) conducted since then have confirmed this trend. Based on these

findings, Horvath and Luborsky (1993) suggested that the therapeutic alliance may have

an initial phase of development that occurs within the first five psychotherapy sessions.

Results from several other investigations (Horvath, 1981; O'Malley et al., 1983;

Saltzman, Leutgert, Roth, Greaser, & Howard, 1976) suggested that the therapeutic

alliance probably peaks during session three. Other researchers (Frank & Gunderson,

14

1990: Kokotovic & Tracey, 1990; Plotnicov, 1990; Saunders, Howard. & Oriinsky, 1989)

have demonstrated that if a good therapeutic alliance is not established during the initial

phase of psychotherapy, it will be reflected by poor outcome or dismpted therapy.

Controversy has arisen concerning the relation between therapeutic alliance and

early benefits of therapy. What is at question is whether the therapeutic alliance is a

mediator of early session-level benefits, or whether it is simply an epiphenomenon of

treatment gains. In responding to this question, Horvath and Luborsky (1993) concluded

that the evidence accumulated to date indicates the therapeutic alliance can be

distinguished from early therapeutic progress and is predictive of outcome beyond these

incremental improvements. Horvath and Greenberg (1994) added, "It is noteworthy that

the alliance/outcome relationship is not significantly related to length of treatment. This

finding seems to offer further assurance that the variable is not just the first transient

blush of success, but more likely a significant and important process factor" (p. 7).

In sum, nearly 20 years of therapeutic alliance research has established that

therapeutic alliance is an important constmct in the field of individual psychotherapy.

"The evidence on the alliance strengthens the view that outcome can be predicted from

early ratings of the therapeutic relationship" (Lambert & Bergin, 1994, p. 165). The

implications of this conclusion for future research, theory development, and clinical

practice are myriad.

Therapeutic Alliance in Marriage and Family Therapy

Although the therapeutic alliance theory and research base are well established in

the individual psychotherapy literature, this is not the case in the marriage and family

therapy literature. During the last decade, however, there has been increased concern in

the marriage and family therapy literature with the therapeutic relationship in general,

and, more specifically, the therapeutic alliance (Coady, 1992; Miller, Hubble, & Duncan,

1995; Morrissette, 1990; Pinsof & Catherall, 1986; Schulman, 1993; Wynne, 1988).

Marriage and family therapists have long addressed the pragmatic importance and

influence of relationship factors in the beginning stage of therapy. For example, they

15

have discussed the "useful rapport" (Ackerman, 1966), the "social stage" of therapy

(Haley, 1976), "joining" and "accommodating" (Minuchin, 1974), and the "coalitionary

process" (Sluzki, 1975). Common factors in therapy such as empathy, warmth,

genuineness, respect, and tmst remain prevalent relationship building themes in the

clinical literature (Carpenter & Treacher, 1989; Cimmamsti & Lappin, 1985; Coady,

1992; Friedman, 1991; Giblin, 1996; Ivey et al., 1996; Morrissette, 1990; Nichols, 1988;

Street, 1994; Woody, 1990; Worden, 1994). Even strategic and Milan therapists who

have traditionally discounted relationship variables (i.e., therapist warmth, empathy,

genuiness) now are addressing these relationship factors (Brock & Barnard, 1992;

Duncan, 1992, 1993; Green & Herget, 1991; Ham, 1989; Kleckner, Frank, Bland,

Amendt, & Bryant, 1992; Mashal, Feldman, & Sigal, 1989; Perry, 1993; Wilkinson,

1992). Many therapists have asserted that these therapeutic alliance related conditions are

critical to the course and/or outcome of therapy (Alexander, 1988; Beach, Sandeen,

O'Leary, 1990; Bornstein & Bornstein, 1986; Cimmamsti & Lappin, 1985; Danser, 1998;

Giblin, 1996; Goolishan & Anderson, 1990; Johnson & Talitman; 1997; Mackey, 1985;

Schilling & Gross, 1979; Segraves, 1982; Solomon, 1977; Swift, 1993; Weeks & Treat,

1992; Wells, 1990; Woody, 1990).

Several earlier investigations lend support to the significance of these therapeutic

conditions in therapy with families. Two studies (Alexander, Barton, Schiavo, &

Parsons, 1976; Waxenberg, 1973) suggested that both continuation and outcome in

family therapy are influenced by the therapist's expressed warmth and empathy. In the

Alexander et al. (1976) study these relationship skills accounted for 45% of the variance

in outcome. Similariy, Beck (1972) reported a lack of therapist warmth was positively

related with treatment failure. In addition, therapist self-reports of positive attitudes

toward families (Shapiro, 1974), or the belief that they had a good relationship with the

family (Firestone & O'Connell, 1984) differentiated between family continuance or

termination. One investigation (Shapiro & Budman, 1973), however, contradicted the

trend of these findings. These researchers reported that premature termination or

16

continuance in family therapy was related more to clients' perceptions of the therapist's

activity level than the therapist's expressed empathy or concern.

More recently, results from three studies of Milan-informed therapy (Coleman,

1987; Green & Herget, 1991; Mashal et al., 1989) "converge on the opinion that a

therapeutic stance lacking in warmth and active engagement contributes to poor

outcomes" (Green & Herget, 1991, p. 326). By contrast, in an exploratory study of

family therapy cases. Shields, Sprenkle, and Constantine (1991) found that initial session

stmcture related variables accounted for 25% of the outcome. Although joining related

variables (praise and empathy) were moderately positively correlated with outcome, they

did not enter into the regression equation.

In a literature review, Gurman and Kniskern (1978) concluded that

The ability of the therapist to establish a positive relationship with his or her clients, long a central issue of individual therapy, receives the most consistent support as an important outcome-related therapy factor in marital-family therapy. Therapist empathy, warmth, and genuineness, the 'client-centered triad,' appear to be very important in keeping families in treatment beyond the first interview....Apparently it is important for the marital-family therapist to be active and to provide some stmcture to eariy interviews, but not to assault family defenses too quickly....[A] reasonable mastery of technical skills may be sufficient to prevent worsening or maintain pretreatment functioning, but more refined therapist relationship skills seem necessary to yield tmly positive outcomes in marital-family therapy, (p. 875)

In a subsequent literature review, Gurman, Kniskern, and Pinsof (1986) reconfirmed

these findings. Overall, the evidence indicates that in family therapy, therapist

relationship and stmcturing skills make a difference in outcome.

A Systemic Therapeutic Alliance Theory

Despite these findings, an explicit "theoretical and research perspective on the

therapeutic alliance in marital and family therapy has been slow to develop" (Johnson &

Greenberg, 1989, p. 98). During the past several years, however, one researcher, William

Pinsof, has pioneered a conceptualization of therapeutic alliance theory from a systems

perspective (Pinsof, 1988a, 1988b, 1989, 1994a, 1994b; Pinsof & Catherall, 1986).

17

Pinsof extended Bordin's (1979) conceptualization of the therapeutic alliance from

an individual level to include couple and family levels. This was necessary because

marriage and family therapists often deal with more than one client at a time. Pinsof and

Catherall (1986) described the therapist-client alliance in marriage and family therapy as

existing on three hierarchical levels with each level constituting one component of the

overall alliance. The lowest level, the "individual alliance," consists of the therapeutic

alliance between the therapist and each individual family member. The intermediate

level, the "subsystem alliance," refers to the therapeutic alliance between the therapist and

each of the multi-person subsystems (e.g., marital, sibling) within the family. The highest

level, the "whole system alliance," pertains to the therapeutic alliance between the

therapist and all of the family members as a group. A circular, reciprocal dynamic

characterizes interactions among therapeutic alliance levels, and each therapeutic alliance

relationship influences and is influenced by every other therapeutic alliance relationship

in an ongoing, mutual manner. Thus, it would be theoretically incorrect to consider any

single level of the therapeutic alliance in isolation.

When the three levels are attended to simultaneously, it is possible for differences

of opinion regarding aspects of therapy to occur between family members and the

therapist. For example, one family member may feel helped by the therapist but may

believe (accurately or not) that the therapist is not helping another family member. Pinsof

and Catherall (1986), therefore, distinguished between an "intact alliance" and a "split

alliance" (p. 139). A split alliance occurs when family members' feelings toward the

therapy or therapist differ notably. In an intact alliance, family members' attitudes toward

the therapy or the therapist are uniform. Empirical support for the concept of the split

alliance in couple and family therapy was demonstrated by Heatherington and Friedlander

(1990).

Therapeutic Alliance Defined

Based on this systemic conceptualization, Pinsof and Catherall (1986) defined the

therapeutic alliance as "that aspect of the relationship between the therapist system and

18

the [client] system that pertains to their capacity to mutually invest in, and collaborate on.

the therapy" (p. 139). In this definition the client system "consists of all the human

systems...that are or may be involved in the maintenance or resolution of the presenting

problem" (Pinsof, 1983, p. 20). The therapist system refers to all persons involved in

providing treatment to the client system. Components of the relationship between the

therapist and the client systems not relevant to the therapeutic alliance "would include

sexual feelings, feeling and thoughts about religious, political and socioeconomic

differences or similarities, etc. Of course, these aspects may affect the alliance, but they

are usually not defined as essential ingredients or impediments to the alliance" (Pinsof &

Catherall, 1986, p. 139).

Pinsof and Catherall's (1986) concept of the therapeutic alliance consists of two

major dimensions—content and interpersonal system. The content dimension was derived

from Bordin's (1979) proposed tripartite definition of the working alliance in individual

psychotherapy, and consists of three subdimensions-tasks, goals, and bonds (Pinsof &

Catherall, 1986). The tasks subdimension deals with the extent to which the client feels

understood by the therapist, and believes that the therapist's power, method, and

technique are helpful and apposite to his or her difficulties. The goals component

concerns the extent to which the client and therapist have a common, shared agreement

about the goals for therapy, especially as they are linked to the presenting problems.

Finally, the bonds subdimension has to do with the nature of the human relationship

between the therapist and the client. It includes feelings such as caring, warmth, and tmst

directed toward and perceived to be offered by the therapist.

The second dimension is the interpersonal system. This dimension moves the

conceptualization of the therapeutic alliance beyond just the relationship between the

therapist and the client, and includes the relationships between the therapist and other

relevant members of the client's interpersonal system. The interpersonal system consists

of three subdimensions-self-therapist, other-therapist, and group therapist (Pinsof &

Catherall, 1986), that tap the client's perception of the relationships between the therapist

19

and (a) the individual client, (b) the various client subsystems, and (c) the system or

family as a whole, in terms of tasks, goals, and bonds.

Together, the content and interpersonal system dimensions form a 3 X 3 matrix

that illustrates the theoretical stmcture of the therapeutic alliance. This stmcture led to

the creation of three self-report scales to measure the therapeutic alliance in individual

(Individual Therapy Alliance Scale), couple (Couple Therapy Alliance Scale), and family

(Family Therapy Alliance Scale) therapy (Pinsof & Catherall, 1986).

Empirical Findings of the Therapeutic Alliance in Couple Therapy

Five published empirical investigations have addressed the therapeutic alliance in

couple therapy. Four examined the relation between therapeutic alliance and outcome,

and three focused on the Couple Therapy Alliance Scale (CTAS, Pinsof & Catherall,

1986) content subscales.

The Relation Between the Therapeutic Alliance and Outcome

Many couple therapists have emphasized the importance of the therapeutic alliance

to the outcome of treatment (Broderick, 1983; Dryden & Hunt, 1985a, 1985b; Gurman,

1982; Hester, 1992; Jacobson & Margolin, 1979; Johnson & Greenberg, 1989: Johnson &

Talitman, 1997; Margolin, 1986; Rutan & Smith, 1985; Schroder, 1992). This alleged

relation has been addressed in four empirical investigations. In the first study

(Holtzworth-Munroe et al., 1989), 32 couples were treated by 13 therapists with social

leaming based marital therapy. Clients completed pre- and post-treatment measures of

the Dyadic Adjustment Scale (Spanier, 1976), and both the clients and the therapists

completed parallel rating forms at the end of each session. Individual item scores were

averaged across therapy sessions and then combined into composite scales that were

averaged across therapy sessions. Significant positive correlations were found between

successful outcome and the therapists' perceptions of having been effective in inducing a

collaborative atmosphere and their perceptions of the couples active involvement and

20

compliance in therapy. Similar results were reported from the couples' perspective.

Couples who believed they were actively participating in treatment and complying with

homework assignments showed the greatest post-treatment marital satisfaction. This

active collaboration between the therapist and couple is consistent with Pinsof and

Catherall's (1986) definition of the goals subdimension of the therapeutic alliance.

The remainder of investigations examining the therapeutic alliance employed

Pinsof and Catherall's (1986) CTAS. Bourgeios and colleagues (1990) examined the

relation between therapeutic alliance measured after the third session by the CTAS and

pre- and post-treatment measures of marital distress. Sixty-three couples participated in a

co-therapist led, group marital skills training program for a period of nine consecutive

weeks. Their findings were three fold. First, pre-therapy levels of marital distress were

not significantly associated with therapeutic alliance scores. This suggested that marital

distress had no impact on therapeutic alliance formation. Second, clients' ratings of

therapeutic alliance were positively correlated with some outcome measures with the

proportion of variance of outcome explained by the quality of the therapeutic alliance

ranging from 3% to 10%. When the part of the variance of the therapeutic alliance that

was not independent of the pretherapy scores was excluded, the actual proportion of

variance due to therapeutic alliance was greater than 8%. Third, therapeutic outcome was

more powerfully predicted by the strength of the therapeutic alliance for men than

women.

In Johnson and Talitman's (1997) study, 34 couples received 12 weekly sessions of

Emotionally Focused Marital Therapy (Johnson, 1996). The couples were assessed with

the CTAS following the third session, and several outcome measures at the end of the 12

weeks and at a three month follow-up. The authors found that therapeutic alliance scores

accounted for 22% of the variance in post-treatment measures of satisfaction, and 29% of

the variance at follow-up. Couple alliance scores were also found to be positively

correlated to couples' gains in marital satisfaction at termination and follow-up.

Furthermore, alliance and intimacy scores were significantly related at termination and

follow-up, accounting for 10% of the variance in the couples' post-treatment intimacy

21

level scores and 16% of the variance in the couples' follow-up intimacy level scores. The

couples' initial tmst levels were found to be predictive of a positive therapeutic alliance,

accounting for 13%? of the variance. Pretreatment levels of marital satisfaction were not

significantly related to the couples' alliance scores. No differences between the males

and females in these findings were reported.

CTAS Content Subscale Research

Three of five studies being reviewed have focused on specific subdimensions of the

therapeutic alliance as defined by Pinsof and Catherall (1986). One impetus for this

research has been that some clinicians have hypothesized that the bonds domain of the

therapeutic alliance takes precedence during the initial phase of couple therapy, and that

the tasks domain is more prominent in the middle phase (Dryden & Hunt, 1985b: Johnson

& Greenberg, 1989). In a post-hoc analysis in Johnson and Talitman's (1997) study

discussed above, the investigators sought to determine if any of the content dimensions

(bond, task, and goal) of the CTAS were more powerful in predicting outcome. Only the

task dimension was found to be significantly related to the couples' post-treatment and

follow-up marital satisfaction levels.

Another study investigating the association of CTAS subscales and outcome in

couple therapy was conducted by Quinn et al. (1997). In this study, 17 couples were

administered the CTAS following the third session and a post-termination two question

outcome measure addressing perceived goal attainment and exp)ected continuance of

change. For women participants, significant positive associations were found between all

three CTAS content subscales and treatment outcome, and for male participants,

significant positive correlations were found for the CTAS tasks and bonds subscales and

treatment outcome. It was noted that these associations were stronger for women than

men. In addition, results showed that the therapy outcome was more positive when the

wives' alliance scores were higher than the husbands' alliance scores. The researchers

stated that these findings were not evidence of a causal relationship between CTAS

22

subdimensions and success of therapy, but only partial evidence of how alliance

contributes to positive outcome.

In the final investigation, Heatherington and Friedlander (1990) hypothesized that

couple therapy clients would perceive sessions characterized by client and therapist

collaboration in the areas of tasks and goals as having more depth because they represent

the "work" of therapy. They also hypothesized that sessions marked by strong emotional

concordance (bonds) between the client and therapist would be perceived as having

greater smoothness. To address these hypotheses, 28 couples completed the CTAS and

the Session Evaluation Questionnaire (SEQ, a measure of session impact on the domains

of depth and smoothness [Stiles, 1980; Stiles & Snow, 1984a]) once between the third

and sixth therapy sessions. Only the tasks subscale was found positively correlated with

perceived depth. In addition, no statistical differences in mean ratings were found

between male and female clients.

The following are some conclusions about this research on the therapeutic alliance

in couple therapy. First, the studies that examined gender differences associated with the

therapeutic alliance (i.e., Bourgeios et al., 1990; Heatherington & Friedlander, 1990;

Johnson & Talitman, 1997; Quinn et al., 1997) showed inconsistent results. While client

gender may play a part in the alliance process, it is unclear how. Second, when taken

together, the results of these studies are suggestive of a positive therapeutic alliance-

outcome relation in couple therapy. Third, the results of the three studies specifically

addressing relations between CTAS subdimensions and couple therapy outcome (i.e.,

Heatherington & Friedlander, 1990; Johnson & Talitman, 1997; Quinn et al., 1997) did

not present a clear picture of which therapy sessions specific domains of the therapeutic

alliance take precedence since they involved the measurement of alliance at only one time

point in eariy therapy. Finally, in 1994, Friedlander, Wildman, Heatherington, and

Skowron noted in their review of the therapeutic alliance literature in couple therapy that

no "investigators have assessed how the [alliance] relationship develops over time" (p.

408), and "what strategies are effective in creating a strong alliance" (p. 412). These

observations remain tme today. Further research is needed to address these issues. In

23

paritcular, identifying treatment delivery methods that improve the alliance (Whisman &

Synder, 1997), assessing what the pattern and rates of change in the therapeutic alliance

would look like across sessions, and examining the relation between therapeutic alliance

and session impact across time. This research would also enable a further testing of the

constmct validity of the CTAS.

Genogram Constmction and Therapeutic Alliance

The genogram literature is replete with statements praising the merits of genogram

constmction for therapists as well as for clients. Unfortunately, reports of reputed

benefits derived by clients from the construction of genograms have relied almost

exclusively on therapist self-report. Only two empirical investigations have garnered the

client's experience of constmcting a genogram. Results from an unpublished study

conducted with a nonclinical sample by a marriage and family therapist (Erianger, 1987),

and a study conducted with individual patients in a family medicine setting (Rogers &

Durkin, 1984), suggested that personal benefit can be realized from the constmction of a

SAGE. Surprisingly, client reported benefits from constmcting either a SAGE or a

TAGE have not been examined empirically in a therapeutic setting.

In order to obtain a more complete picture of the clinical relevance of genogram

constmction, more needs to be learned about this process from the client's perspective.

For example, questions remain as to how clients see genogram constmction effecting not

only the therapist-client relationship, but also their therapy experience. McGoldrick and

Gerson (1985) stated the issue as follows: "In the end, the proof of the value of the

genogram will be a documentation of its impact on clinical practice. First of all, research

on the genogram is necessary to establish whether families that participate in genogram

interviews benefit in some way from their involvement" (p. 148). They then proposed

that "In family therapy, one could study the effects of a genogram interview on the

joining process or attitudes toward therapy" (McGoldrick & Gerson, 1985, p. 148).

These are intriguing proposals, because several marriage and family therapists (e.g.. Beck,

1987; Braverman, 1997; Erianger, 1991; Ingersoll-Dayton & Arndt, 1990; Ivey et al..

24

1996; Kuehl, 1995; Miehls, 1993; Wachtel, 1982) have contended that genogram

constmction can enhance therapeutic alliance (which includes joining), and others (e.g..

Beck, 1987; Brock & Barnard, 1988; Kramer, 1985; Kuehl, 1995) have claimed that

genogram constmction can influence clients' attitudes about their therapy experience

(which will be referred to as "session impact").

The importance of eariy therapeutic alliance as a mediator of outcome in individual

psychotherapy has been empirically established, and there is preliminary evidence of the

predictive validity of eariy therapeutic alliance with couples. Because of the popularity of

genogram constmction in marriage and family therapy and its common use in the opening

phases of therapy, the effects of genogram constmction on therapeutic alliance will be the

primary focus in this investigation. A secondary focus will be examining the effects of

genogram constmction on session impact as it could provide valuable information to

clinicians who employ genograms. Finally, Friedlander and Hetherington's (1990)

hypothesized relation between the subscales of the CTAS (tasks, goals, and bonds) and

the indexes of the SEQ (depth and smoothness) will be examined, because of the need to

replicate their constmct validity test for the CTAS.

In previous therapeutic alliance studies in individual psychotherapy or couple

therapy, therapeutic alliance was assessed with either single time point measurements,

measurements averaged across sessions, or pre/post measurements. While information

gained from these designs can be useful in describing therapeutic alliance ratings at one

or more time points or the amount of change taking place in therapeutic alliance ratings

between arbitrary time points, it can not provide a description of the intricacies of change

in therapeutic alliance ratings across time. To address this problem, a growth curve

analysis will be employed in this study. This approach will permit a superior basis for

answering questions regarding systematic interindividual differences in change on

therapeutic alliance slope scores and session impact slope scores.

This study will take new directions in the study of the genogram. Data will be

collected from clients' perspectives in a therapeutic setting, and the effects of genogram

constmction on therapeutic alliance and session impact will be assessed. In addition, for

25

the first time, a growth curve approach will be used to examine therapeutic alliance and

session impact and the relations between their subdimensions.

Hypotheses

Five groups of hypotheses will be tested. The first and second groups of

hypotheses will formally test planned comparisons between the following treatment

groups: the TAGE plus couple therapy(TAGE-hCT); the SAGE plus couple therapy

(SAGE-HCT); and a DAI plus couple therapy (control, DAI+CT). for both dependent

measures.

For therapeutic alliance, the planned comparisons are:

la. The mean rate of change (calculated from individual linear growth curves which

represent a constant rate of change) for the TAGE+CT group will be greater than mean

rate of change for the SAGE+CT group following the administration of the treatments.

lb. The mean rate of change (calculated from individual linear growth curves which

represent a constant rate of change) for the TAGE+CT group and the SAGE+CT group

will be greater the mean rate of change for the DAI+CT group (control).

For session impact, the planned comparisons are:

Ha. The mean rate of change (calculated from individual linear growth curves which

represent a constant rate of change) for the TAGE+CT group will be greater the mean rate

of change for the SAGE+CT group following the administration of the treatments.

nb. The mean rate of change (calculated from individual linear growth curves which

represent a constant rate of change) for the TAGE+CT group and the SAGE+CT group

will be greater than the mean rate of change for the DAI+CT group (control).

The third through fifth groups of hypotheses will investigate the relations between

the tasks, goals, and bonds subscales of the therapeutic alliance instmment and the depth

and smoothness indexes of the session impact instmment over time. Three hypotheses

are proposed for each of the three treatment groups, where hypotheses Ula through IIIc

are for the TAGE+CT group, hypotheses IVa through FVc are for the SAGE+CT group,

and hypotheses Va through Vc are for the DAI+CT group (control).

26

nia, rVa, Va. A change in the tasks subscale will be positively correlated with a

change in the depth index.

mb, rVb, Vb. A change in the goals subscale will be positively correlated with a

change in the depth index.

EQc, rVc, Vc. A change in the bonds subscale will be positively correlated with a

change in the smoothness index.

27

CHAPTER III

METHODS

Samples and Sampling Procedures

Participants

Participants were solicited from the client populations seeking couple therapy

services at two outpatient marriage and family therapy centers located on the campuses of

Texas Tech University (TTU) and Abilene Christian University (ACU). These centers

were staffed by doctorate (TTU) or master's (ACU) level students who were enrolled at

programs accredited by the Commission on Accreditation for Marriage and Family

Therapy Education.

Twenty-one couples participated in the study; however, four of these couples (eight

individuals) were excluded because they completed only one couple therapy session and

one assessment package. Data collected from a single time point was insufficient for

analyzing change in their responses. The resulting sample (see Table 3.1) consisted of 17

couples (17 men and 17 women) who ranged in age from 17 to 47 years old (n = 33 [age

data was not obtained for one male]); M = 28.4; SD = 7.6). Ages for the men (n = 16)

ranged between 21 and 47 years old (M = 29.4; SD = 7.4). Ages for the women (n = 17)

ranged between 17 and 47 years old (M = 27.4; SD = 7.8). Education levels of the

individual participants were either some high school (n = 5), a high school diploma or

equivalent (n = 4), vocational or technical training (n = 2), some college (n = 16), an

associate's degree (n = 2), a bachelor's degree (n = 2), or a master's degree (n = 2). Most

of the participants were Caucasian (n = 22); however, there were also two African-

Americans, one Asian, seven Hispanics, and one Native-American. Education level and

ethnic origin information was not obtained from one participant. The couples (n = 17)

were either dating (n = 1), living together (n = 3), or married (n = 13), and the amount of

time the couples had been in their relationship ranged from five to 238 months (M = 69.1;

SD = 77.9). This purposive sample was obtained by contacting the clinic director at TTU

and the director of research at ACU to solicit their cooperation.

28

Variable

Demographics by Individual

Age (years)

Male

Female

Education Levels'

Some High School

High School Diploma

Partici

Vocational/Technical Training

Some College

Associate's Degree

Bachelor's Degree

Master's Degree

Ethnic Background'*

African-American

Asian

Caucasian

Hispanic

Native-American

Demographics by Couple

Relationship Status

Dating

Living Together

Married

Length of Relationship (months)

Table 3.1 ipant Demographics

n

33a

16=*

17

5

4

2

16

2

2

2

2

1

22

7

1

1

3

13

17

Range

17-47

21 -47

17-47

5-328

M

28.4

29.4

27.4

69.1

SD

7.6

7.4

7.8

77.9

Note. ''Data missing for one participant.

29

This study had a 3 (treatment intervention) X 5 (session) "multiple-group-muliiple-

I" time series experimental design (Glass, Willson, & Gottman, 1975; see Figure 3.1).

The couples were randomly assigned to one of three groups. Seven couples (14

individuals were randomly assigned to group one, eight couples (16 individuals) were

randomly assigned to group two, and two couples (4 individuals) were randomly assigned

to group three. Unequal numbers of participants in the three groups resulted when data

collection was ended after 15 months before acquiring the anticipated sample size. Had

the anticipated sample size been acquired, the random assignment of couples to groups

would have led to equal couple numbers in each group. Data collection was ended after

15 months due to time restrictions in completing this dissertation. The following groups

resulted.

Group one (TAGE+CT) (see Table 3.2) consisted of seven opposite-sexed, paired

couples (14 individuals), aged 20 to 41 (M = 26.6; SD = 6.6). Ages for the men (n = 7)

ranged between 21 and 41 years old (M = 27.9; SD = 6.8). Ages for the women (n = 7)

ranged between 20 and 36 years old (M = 25.3; SD = 6.5). Education levels of these

individuals were two with some high school, one with vocational or technical training, six

with some college, two with an associate's degree, two with a bachelor's degree, and one

with a master's degree. Three ethnic groups were represented by these individuals:

African-American (n = 1), Caucasian (n = 11), and Hispanic (n = 2). One of these

couples was dating, one was living together, and five were married, and the amount of

time the couples had been in their relationship ranged from 19 to 168 months (M = 64.9;

SD = 50.0).

Group two (SAGE+CT) (see Table 3.2) consisted of eight opposite-sexed, paired

couples (16 individuals), aged 17 to 47 (M = 29.5; SD = 8.9). Ages for the men (n = 7

[age data was not obtained for one male]) ranged between 23 and 47 years old (M = 30.4;

SD = 8.8). Ages for the women (n = 8) ranged between 17 and 47 years old (M = 28.6;

SD = 9.6). Education levels of these individuals were two with some high school, four

with a high diploma or the equivalent, one with vocational or technical training, seven

with some college, and one with a master's degree. Five ethnic groups were represented

30

Treatment Group Session 1 Session 2 Session 3 Session 4 Session 5

Group 1

Group 2

Group 3

R

R

R

O

O

O

OI,

OI2

OI3

0

0

0

0

0

0

0

0

0

Figure 3.1 3 X 5 Multiple-Group-Multiple-I Times

Series Experimental Design

Legend: Group 1 = Therapist-administered genogram plus couple therapy Group 2 = Self-administered genogram plus couple therapy Group 3 = Directed assessment interview plus couple therapy (control) R = Random assignment O = Couple Therapy Alliance Scale and Session Evaluation Questionnaire Ii = Therapist-administered genogram 12 = Self-administered genogram 13 = Directed assessment interview

31

a. 3 O O X I

CJ

2 o H D

Q c

"G

cu

cn H

2 < O Q

CN (J D, +

2 a o <

00

H D, +

2 O O <

Q on

c

=1

col

c

Q 00

SI

00 c <n

.2

>

CJ CT3 3

<^, — r-; U-] r^ cr]

ir, o o

r<~, r<~, r<-,

sC sC Tj-r*", r*", r* ,

\ 0 \ 0 sC CN CM rJ

CN r j

i n 00

VTi

OS CN

1 1

3C 00

' ^

d

p -

1 1

CN

(3

sC

cr

sC 00

1 1

00

vO

vd

vO so (N

1 1

o CN

' ^

DC

\ C

O)

<N

1 1

CN

r

^/~,

sd

r^,

( N

1 1

o CN

r

ex 7i

Ogr

E

;a

div

c

JO

C/3 «1

(ye;

OJ 00

<

c

cs — r-

CN — vO CN CN

c

Wom

e

v.

>

Ji c

luca

tio

o o CJ

0 0

0 0

£

Som

e 1

c<3

E o a. Q

hool

CJ

Hig

hS

^^

CJ

"c JS CJ

^

nal

o

Voc

ati

00 c

c«

00

"o u E o cn

00

Q

."2 *CJ o

<

00

Q

o •.J

CQ

gu

32

CN

CO

OJ . ~

T3

3 C

X) C3 C

o H u

2<

CN CJ

5 W P o

00 O

- u D. + 2 O O <

H

Q 00

00

c

DC

c/5|

00

c

ooj

00 c

m — <N

— — OO Tt — 00

— CN

aria

ble

>

C/J

(U >

duca

tion

Le

w

ree

0 0 (L>

Mas

ter'

s D

pun

o k>

thni

c B

ackg

u

rica

n

<u

Afr

ican

-Am

Asi

an

Cau

casi

an

His

pani

c

ican

u^ (U

Nat

ive-

Am

c/3

emog

raph

ic

y C

oupl

e

Q CQ

oupl

es

U

33

CN

r^ ZJ

X Ti

H

T3

3 C

^^ C O

U

E <

CN CJ

a, + 3 Ll 2 O a <

00

H - U o- +

2 O a <

00

to

Q 00

00

c

Q 00

00

c C3

X

r3

(N

CN sO

rJ

CN

oc

00 m CN I m

00

in

sO

00

— — m,

y; X

3

"^ ^^ C/5 diqsu

o

clat

i

cc:

0 0

c

Dat

k a

u X •tairf

0 0

gTo

c

Liv

ied

'r^ r3

^ r*

" ^

J

C! /^ < - •

^-

nshi

p el

ati

c

+ T T 1 batart

r ^ ^ < C/

>^ C-73 ^m

O X

• ^ ^

_o c 3 ^

d •y; m.*

"E. ,_

.>

50

c <u 0 0

-a '_» L -

Sic

• —

d . • •

^ M

-5 . . i

1

y ;

rapi

O - C

II r"* u + LU

o < H , ! c •n Q .

. —• •J

t^ r3 Cu U c o u_

,o 00 c SSI

r-

i2 Q

C3

4J «^ O

z

> . a. r3

o ~^ _^ C-3

'•J

y. _3

a. ^ CJ

> u_ o C

^

:J

V? •y.

O y : -y; r3

T3 IJ CJ

'•5 II

H ^-z

_+ <

>^ C -r3 L .

CJ X

• —

J J C -3

CJ y ; 3

a.

E :« b . 0 0 O

0 0

-a

'E • ~

— -5 r3

34

by these individuals: African-American (n = I), Asian (n = 1), Caucasian (n = 8).

Hispanic (n = 4), and Native-American (n = 1). The education level and the ethnic origin

of one individual were not obtained. Two of these couples were living together, and six

were married, and the amount of time the couples had been in their relationship ranged

from five to 238 months (M = 83.5; SD = 100.3).

Group three (DAI+CT [control]) (see Table 3.2) consisted of two opposite-sexed,

paired couples (4 individuals), aged 26 to 36 (M = 30.5; SD = 5.3). Ages for the men (n

= 2) ranged between 26 and 36 years old (M = 31.0; SD = 7.1). Ages for the women (n =

2) ranged between 26 and 34 years old (M = 30.0; SD = 5.7). Education levels of these

individuals were one with some high school, and three with some college. Two ethnic

groups were represented by these individuals: Caucasian (n = 3), and Hispanic (n = 1).

Both couples were married, and the amount of time the couples had been in their

relationship ranged from 20 to 33 months (M = 26.5; SD = 7.5).

Solicitation of Participants

To be eligible for participation, both couple members had to agree to participate.

This study focused on the therapeutic alliance dynamic in couple therapy, which differs

theoretically from the therapeutic alliance dynamic in individual psychotherapy (Pinsof &

Catherall, 1986). Additionally, a minimum age of 17 was required for inclusion in the

study.

All eligible couples were solicited for participation at the end of the first session by

their therapist. The end of the first session was chosen because this allowed the therapist

to assess whether couple therapy was the most appropriate treatment mode. Therapists

had the freedom not to solicit a couple's participation if in their judgment (a) such

participation would be risky (e.g., violent couples may require individual rather than

conjoint sessions), (b) individual therapy would be more appropriate (e.g., one member

requiring a referral; one or both members needing specialized individual therapy; the

couple decides to separate or divorce and no longer desires conjoint therapy), or (c)

family therapy would be more appropriate.

35

The therapist explained to the couple the general purpose of the study and its

voluntary, anonymous, and confidential nature. Couples were informed that participation

would involve the completion of a brief questionnaire immediately following the first five

sessions. Reading and signing a statement that outlined the study's purpose,

questionnaire requirement, and participant rights (see Appendix A) was accepted as

indication of the participant's informed consent to allow the questionnaire data to be used

in this research project. Care was taken in the obtaining of informed consent not to allow

one couple member to coerce the other to participate. If one couple member refused to

sign the consent form, the couple was eliminated from participation. If (a) at least one

couple member terminated their participation at any point after the first therapy session,

(b) the therapy was not conducted conjointly during at least part of a therapy session, (c)

any other person (e.g., family member, friend, therapist's supervisor, co-therapist)

attended any part of a therapy session, or (d) the case was transferred to another therapist,

the couple was eliminated from further participation in the study. These conditions

helped to control confounds to the couple therapy dynamic. Persons attending individual

therapy that then became couple therapy were not eligible for participation. Participants

were given no reward or compensation for participating in the project, but were given a

debriefing form (see Appendix B) at the end of their participation in the study.

Therapists

At TTU, five students active in the clinic and associated with the marriage and

family therapy doctoral degree program were solicited to participate in the study as

therapists. Regarding these students, two were male and four were female. One of the

male students was completing an internship at this site, and during the course of the study

the other four students were in their first two years of the program. At ACU, two students

active in the clinic and associated with the marriage and family therapy master's degree

program were solicited to participate in the study as therapists. Regarding these students,

both were males in their first year of the program. At both sites students were trained in

and practiced various models of marriage and family therapy.

36

A questionnaire was given to the seven student therapists who agreed to participate

in the investigation to gain information about (a) their demographic characteristics (see

Table 3.3), (b) their self-rated ability to constmct genograms. (c) the percentage of time

they constmcted genograms in therapy, (d) the theory they used in therapy, and (e) their

self-rated clinical skills level. One student from ACU, however, did not complete the

questionnaire. Regarding the six remaining students, their ages ranged from 25 to 47

years (M = 33.6; SD = 9.6), and five were Caucasian and two were Asian. Their mean

clinical experience ranged from three to 90 months (M = 38.8; SD = 27.9). These

students were asked to rate their ability to constmct genograms and their clinical skill

level on a 5-point Likert scale with excellent coded as "5" and poor coded as T' (see

Tables 3.4, 3.5). They were also asked to list the percentage of time they used the

genogram on a 5-point Likert scale with 100% coded as "5" and 0% coded as "1 , " as well

as list the theory they used in therapy (see Tables 3.6, 3.7). Because some students

participated as therapists more than once during the course of the study, their responses to

these questions may have varied.

Solicitation of Therapists

Individual announcements were distributed to students asking for their participation

on a volunteer basis (see Appendix C). The investigator followed up students by

telephone a few days later to answer any questions they had concerning the study and to

solicit their participation as therapists. Students who agreed to participate as therapists

attended a one and one-half hour orientation conducted by the researcher. At this

orientation, each student read and signed a consent form that outlined the study's purpose

and therapist requirements and rights (see Appendix D).

The content of this orientation covered (a) a general presentation of the study, (b)

methods for soliciting client participation in the study, (c) procedures for administering

the three treatment interventions and the dependent measures, and (d) a training session

on genogram constmction. Therapists were kept unaware of the specific purposes and

37

Table 3.3

Therapist Demographics

Therap

TTU

#1

#2

#3

#4

#5

#6

ACU

#7

#8

Note.

ist

' missing

Gender

male

male

female

female

female

female

male

male

data. TTU:

Age

47

45

26

29

38

25

25 a

Ethnicity

Caucasian

Asian

Caucasian

Asian

Caucasian

Caucasian

Caucasian a

= Texas Tech University.

Clinical Experience in Months

4 8 - 5 3

72 -90

36

10-16

3 - 8

22

9 a

ACU = Abilene Christian

Unviersity.

38

Table 3.4 Crosstabulation of the Student Therapists Self-rated Ability

to Constmct Genograms and Student Therapists

Student Therapists'"

Ability to Constmct Genogram

Average

Good

Excellent

Total # of Scores

Mean^

c^#l TTU

3

4

7

4.14

cr#2 TTU

1

3

4

3.75

?#3 TTU

1

1

4.00

9 #4 TTU

2

1

3

3.67

$#5 TTU

3

3

4.00

9 #6 TTU

1

1

5.00

o"#7

ACU

1

1

4.00

Total #of

Scores

6

8

6

20

Note. ?=male. c3" = female. TTU = Texas Tech University. ACU = Abilene Christian University. The numbers in this table indicate the numbers of times student therapists scored their ability to constmct a genogram with a certain rating on a 5-point Likert scale (from "excellent" to "poor"). Students who participated in the study more than once had multiple scores.

Data missing for therapist #8. ^"5" = "excellent" to " 1 " = "poor."

39

Table 3.5 Crosstabulation of the Student Therapists Clinical

Skill Level and Student Therapists

Student Therapists'"

Clinical Skill Level

Average

Good

Excellent

Total # of Scores

Mean''

Note. ? = male.

o"#l

TTU

7

7

5.00

o" = female

o«#2 TTU

I

3

4

3.75

:. TTU

?#3 TTU

1

1

3.00

?#4 TTU

2

1

3

3.33

?#5 TTU

3

3

4.00

= Texas Tech University.

?#6 TTU

1

1

4.00

a'#7 ACU

I

1

4.00

Total #of

Scores

4

9

7

20

ACU = Abilene Christian University. The numbers in this table indicate the numbers of times student therapists scored their clinical skill level with a certain rating on a 5-point Likert scale (from "excellent" to "poor"). Students who participated in the study more than once had multiple scores.

Data missing for therapist #8. ^"5" = "excellent" to " 1 " = "poor."

40

Table 3.6 Crosstabulation of the Percentage of Time the

Student Therapists Used the Genogram and Student Therapists

% Time Genogram Used

25

50

75

100

Total # of Scores

Mean^

Note. ? = male.

o"#l TTU

7

7

2.00

c = female

c/'#2 TTU

3

1

4

4.25

. TTU:

Student Therapi

?#3 TTU

1

I

3.00

?#4 TTU

3

3

2.00

ists"

?#5 TTU

1

2

3

3.33

= Texas Tech University.

?#6 TTU

1

1

4.00

o"#7 ACU

I

1

3.00

Total #of

Scores

11

2

6

1

20

ACU = Abilene Christian University. The numbers in this table indicate the numbers of times student therapists scored the percentage of time they used the genogram with a certain rating on a 5-point Likert scale (from "100%" to "0%"). Students who participated in the study more than once had multiple scores.

b,u 'Data missing for therapist #8. ""5" = "100%" to "l" = "0%."

41

Table 3.7 Crosstabulation the Theory Used by the Student

Therapists and Student Therapists.

Student Therapists'"

Total Theory c #l cr#2 ?#3 ?#4 9 #5 ?#6 o"#7 # of Used TTU TTU TTU TTU TTU TTU ACU Scores

Behavioral/

Social Learning 2 2

Brief 1 I

Eclectic 3 1 4

Emotionally Focused 1 1 Feminist I 2 3 Stmctural 4 1 3 1 9

Total # of Scores 7 4 1 3 3 1 1 20

Note. ? = male, d" = female. TTU = Texas Tech University. ACU = Abilene Christian University. The numbers in this table indicate the numbers of times student therapists used a particular theory in this study. Students who participated in the study more than once had multiple scores.

' Data missing for therapist #8.

42

hypotheses. Therapists were given no reward or compensation for participating in the

project, but they were provided with a summary of the study.

Procedure

Assignment of Therapists to Couple Therapy Cases

Couple therapy cases were assigned to therapists on the basis of availability, a

common method of assignment (see Stanton & Todd, 1982). Shields, Sprenkle, and

Constantine (1991) suggested that in using this assignment method "There is no reason to

believe there [are] any systematic bias in how cases [are] distributed, other than

therapist's schedules" (pp. 5-6).

Assignment of Couples to Treatment Conditions

Couples were assigned at random to the treatment intervention groups through the

following procedure. The three treatment interventions groups were assigned a number

from one to three. The TAGE+CT treatment intervention was group one. The

SAGE+CT treatment intervention was group two. The DAI+CT treatment intervention

(control) was group three. A table of random numbers was used to generate a random list

of the numbers one, two, and three. This list of numbers was the sequential order that

couples were assigned to the three treatment intervention groups after they consented to

participate in the research project. Therapists were able to identify the treatment group

the couple was assigned to by a code marked on prepared dependent measures packets.

The coding system is detailed in a later section of this chapter.

Administration of Couple Therapy

Couple therapy in all three treatment groups (the TAGE+CT group, the SAGE+CT

group, and DAI+CT group [control]) was conducted using a theoretical orientation

chosen by the therapist. Because a consistent finding from individual psychotherapy

research indicates that a good alliance is predictive of positive therapy outcome regardless

of treatment modality (Gaston, 1990; Henry et al., 1994; Horvath, 1994; Horvath &

43

Luborsky, 1993; Horvath & Symonds, 1991; Luborsky, 1994), it was assumed that the

differences across therapists in theoretical orientations would not be a significant source

of systematic variance. Therapists conducted all therapy sessions by themselves:

however, their therapy sessions were under supervision by American Association for

Marriage and Family Therapy Approved Supervisors. The first, third, fourth, and fifth

therapy sessions were scheduled for 50 minutes each. Both couple members were

required to be present during at least part of the session. No other therapist or supervisor

was permitted to enter the therapy room and participate as a co-therapist at any point

during the session. In addition, no family member, friend, or other person(s) were

permitted to participate in any part of the therapy session with the couple. To maintain

consistency in therapy session lengths, if a couple was more than 10 minutes late for a

therapy session, the therapist cancelled the therapy session and rescheduled for another

date. Sessions were generally scheduled on a weekly basis, although the exact number of

days between sessions varied within and across cases. For the 17 couples in the sample,

the mean lapse between sessions was 13.43 days (SD = 6.97).

With the exception of the second therapy session in the TAGE+CT group and the

SAGE+CT group, the therapy sessions did not include the constmction of a genogram or

any other diagrammatic data collecting instmment (e.g., "The Time Line" [Stanton,

1992], an eco-map [Hartman, 1995], a family sociogram [Sherman & Fredman, 1986]).

Because the design of this study was to introduce three different stimuli (i.e., the three

different treatment interventions) at one point in the therapeutic process (i.e., the second

session), and then to measure the carry over responses of couples to these stimuli across

time, the genograms completed during the second therapy session were not displayed

during subsequent therapy sessions. Because it was likely that the genogram constmction

process would raise issues relative to couple's ongoing therapy, therapists were free to

address these issues during the therapeutic dialogue as they deemed necessary or at the

couple's request.

44

Administration of Treatment Interventions

The three treatment interventions; the TAGE, the SAGE, and DAI (control), were

administered during the second therapy session. Waiting until the second therapy session

to administer the treatment interventions allowed therapists to assess if couple therapy

was the appropriate treatment mode, and it gave the couple the opportunity to 'tell their

story" without the therapist being overly directive. One and one half-hour appointments

were scheduled for all three treatment intervention groups for the second therapy session

to allow sufficient time for completion of the genograms and processing of the genogram

constmction experience. The administration of the dependent measures was conducted in

the same manner at both of the study sites.

The TAGE Treatment Intervention

Administration of the TAGE treatment intervention proceeded as follows. The

therapist escorted the couple to a therapy room and following normal social exchanges,

informed them of his or her intent to constmct a genogram during the therapy session.

The administration of the TAGE was justified to the couple as a method for helping the

therapist and themselves to understand the context of their presenting problems. The

therapist conducted a genogram interview with each couple member while drawing their

genograms together on an 18 by 24-inch sheet of poster paper with a pencil. Following

the completion of their genograms, the therapist spent a few minutes having the couple

share insights gained from the genogram data. This involved asking each couple member

to briefly discuss how they perceived the information or pattems of information recorded

on the genogram to be related to their presenting problems and goals for therapy. This

process helped to give meaning to the experience beyond simple data collection.

Therapists did not, in general, probe issues at a level deeper than shared by the couple

members, because doing so may have detracted from completing the genograms in the

allotted time. Therapists, however, did assess whether any emotional harm occurred as a

result of the genogram constmction process. If either couple member reported emotional

harm, this was processed immediately. No out-of-session tasks were assigned to the

45

couple following the second session in order to control for task differences between

couples. Completed genograms became a permanent part of the couple's case record, but

copies of the genograms were provided to the couple members if requested after the study

was completed.

The SAGE Treatment Intervention

Administration of the self-administered treatment intervention proceeded as

follows. The therapist escorted the couple to a therapy room and following normal social

exchanges, provided the couple with a pencil and one SAGE packet which included

complete instmctions and an 18 by 24 inch genogram skeleton chart (see Appendix E).

The administration of the SAGE was justified to the couple as a method for helping the

therapist and themselves to understand the context of their presenting problems. After

briefly reviewing the instmctions with the couple and, when necessary, answering

questions regarding the instmctions, the therapist stayed in the room to clarify the

instmctions if requested, but remained silent. If the couple engaged the therapist more

than three times, the therapist left the room. The therapist's silence allowed the couple to

complete their SAGEs apart from the assistance of the therapist, and it minimized the

level of couple-therapist interaction. Controlling the level of therapist-couple interaction

was significant because this interaction and the therapist's mediator role were what

discriminated the TAGE treatment intervention from the SAGE treatment intervention.

As the couple constmcted their genograms, therapist-couple interaction was permitted to

clarify instmctions but not to assist either member in any part of the genogram

constmction process.

Following the completion of their genograms, the therapist spent a few minutes

having the couple share insights gained from the genogram data. This involved asking

each couple member to briefly discuss how they perceived the information or patterns of

information recorded on the genogram to be related to their presenting problems and

goals for therapy. This process helped to give meaning to the experience beyond simple

data collection. Therapists did not, in general, probe issues at a level deeper than shared

46

by the couple members, because doing so may have detracted from completing the

session in the allotted time. Therapists, however, did assess whether any emotional harm

occurred as a result of the genogram constmction process. If either couple member

reported emotional harm, this was processed immediately. No out-of-session tasks were

assigned to the couple following the second session in order to control for task

differences between couples. Completed genograms became a permanent part of the

couple's case record, but copies of the genograms were provided to the couple members

if requested after the study was completed.

DAI Treatment Intervention (Control)

In the DAI treatment intervention (control), the therapist escorted the couple to a

therapy room and conducted a therapy session according to the procedural guidelines

described previously for the administration of couple therapy. The only exception was

that the therapist used a DAI outline (see Appendix F) to guide the session. The

questions in this outline approximate questions used in a genogram interview.

Administration of the Dependent Measures

At the end of the first five sessions, both couple members were administered a

dependent measures packet including a cover page (see Appendix G), the CTAS (Pinsof

& Catherall, 1986; see Appendix H), and the SEQ (Form 3; Stiles, 1980; Stiles & Snow.

1984a; see Appendix I). For the first session only, the packet also included a

demographic questionnaire (see Appendix J). Couple members remained in the therapy

room to fill out the forms. Because of the nature of the questions, the therapist left the

therapy room after reviewing the instmctions with the couple. At this time, the therapist

completed a demographic questionnaire following the first session (see Appendix K). and

a session follow-up questionnaire at the end of each session (see Appendix L). When the

couple members and the therapist completed their forms they placed them in a designated

locked box in the waiting room.

47

The random assignment of couples to treatment groups and the distribution of the

appropriate dependent measure packets was managed in the following way. Envelopes

were prepared prior to the onset of the study with each containing four consent forms

(two for the couple members and two duplicates for the researcher's records), 10

dependent measures packets (one packet per couple member per session), one therapist

demographic questionnaire (one per session), five therapist follow-up questionnaires (one

per session), and one participant debriefing form (one for each couple). The packet also

contained one SAGE set of instmctions and skeleton chart for couples assigned to the

SAGE+CT treatment group. Each of the three treatment groups required a specific

combination of the above forms, depending on the session. A two-part code was marked

on each envelope (see Figure 3.2). The first part identified the case number (from "01" to

"21"), which represented the order couples were randomly assigned to treatment groups.

The second part identified the treatment group ("TA" = therapist-administered genogram

plus couple therapy, "SA" = self-administered genogram plus couple therapy, or "DA" =

directed assessment interview plus couple therapy [control]) to which the couple was

assigned. By looking at the letter code on the envelope, the therapist was able to identify

which treatment to administer to the couple during the second session. A three-part code

was marked on upper right-hand corner of each page of the forms in each envelope (see

Figure 3.2). The first two parts of the code matched the code on the envelope, and the

third part identified the session number ("1" to "5"). This enabled the therapist to

administer the appropriate form packet at each session. This coding system also enabled

the primary researcher to keep track of the data and identify which couple and therapist

forms went together.

The coded forms were then placed in matching coded envelopes in the sequential

order (session one to session five) they were administered. The envelopes were placed in

a box in the clinic office and stacked in the sequential order based on case number, with

the envelope marked case number "01" on top of the stack and the envelope marked case

number "21" on the bottom of the stack. Before entering the first session with a couple.

48

(a) Envelope codes.

OOITA 080SA 114DA

(b) Form codes.

OOlTAl 080SA3 114DA5

Legend. (a)

(b)

Figure 3.2 Examples of the Coding Systems for

Envelopes and Form Packets

The first three numbered columns refer to the case number and ranged from "001" for case number one to "114" for case number 114. The fourth and fifth lettered columns refer to the three treatment groups where "TA" = therapist-administered genogram plus couple therapy, "SA" = self-administered genogram plus couple therapy, or "DA" = directed assessment interview plus couple therapy (control).

The first three numbered columns and the fourth and fifth lettered columns are identical to the envelope codes. The sixth numbered columns refers to the session number and will range from "1" for the first session to "5" for the fifth session.

49

the therapist took the top envelope in the box for use with this couple. If the couple was

deemed ineligible for the study, the envelope was returned to the top of the stack.

In the event an envelope or specific forms were misplaced or destroyed,

replacement forms were made available along with a chart indicating which forms should

be used for each session depending on the treatment group. The therapist was responsible

for writing the appropriate codes on the forms before administering them. These

procedures were conducted in the same manner at both of the study sites.

Monitoring Treatment Interventions and Data Collection

The primary investigator was physically present at the centers for the therapist

orientations and then worked closely with the clinic director and the therapists via the

telephone to answer questions or address any complication that arose. Periodic letters

(see Appendix M) were sent to the therapists expressing gratitude for their participation

and appreciation for their attention to detail in the administration of the treatment

interventions and the dependent measures packets. Arrangements were made to have the

completed dependent measures packets and therapist demographic questionnaires mailed

to the primary investigator.

Instmments

Treatment Intervention Instmments

Three treatment interventions were employed in this study: the TAGE, the SAGE

(Erianger, 1987; Rogers & Cohn, 1987; see Appendix E), and the DAI (see Appendix F).

The TAGE

The TAGE forms and instmctions were adapted from those designed by Rogers and

Cohn (1987) and Erianger (1987). The format and symbols for the TAGE were based on

a standardized genogram description developed by McGoldrick and Gerson (1985).

Categories of information collected on the TAGE for three or four generations of a family

included: names, ages, dates of births and deaths, occupations, level of education.

50

ethnicity, medical problems, emotional stresses, legal problems, personal characteristics

(i.e., descriptive words or nicknames that characterize individuals), locations, and

relationship descriptors (e.g., overly close, conflicted). For the constmction process, the

therapist was present in the room with the couple to interview and constmct both

participants' genograms on a poster board.

Because genograms are not standardized measurement tools like psychological

tests, their psychometric properties are not relevant (Coupland et al., 1995). Genograms

are instead graphical representations of family information and dynamics. They serve as

databases that are used heuristically in therapeutic settings. For example, data collected

on genograms may assist clinicians and family members to gain a contextual

understanding of presenting problems. It may also help clinicians formulate hypotheses

and interventions as part of the change process. For the purposes of this study, the

primary concern was the participants experience of constmcting their genogram rather

than the accuracy or completeness of the information they recorded on the genogram, or

the manner by which the collected information was used in future sessions.

The SAGE

Like the TAGE, the SAGE was based on the genogram standardization design of

McGoldrick and Gerson (1985). The format, symbols, and categories of information

recorded on the SAGE were identical to those of the TAGE. The only difference was that

the SAGE was presented to participants as an eight-page packet containing an

explanation of the genogram, step-by-step constmction instmctions, and a genogram

skeleton chart upon which the individual participants filled in the categories of

information listed under the explanation of the TAGE. Therapists had no part in the

constmction process except to initially clarify the instmctions, although they were present

in room with the participants as they constmcted their genograms. As with the SAGE,

the psychometric properties of the TAGE are not relevant.

51

The DAI

The DAI was a list of questions that the therapist used to guide the second session.

These quesUons were used to gain information similar to those that would be obtained in

a genogram constmction interview.

Dependent Measures

The questionnaire for participants included three sections: (a) the CTAS (Pinsof &

Catherall, 1986), (b) the SEQ (Form 3; Stiles, 1980; Stiles & Snow, 1984a), and (c) a

demographic questionnaire. Therapists also completed a demographic questionnaire.

CTAS

The CTAS (Pinsof & Catherall, 1986) is a 29-item self-report measure of the

client's view of the therapeutic alliance on three content and three interpersonal system

subscales (see Appendix H). The content subscales were derived from Bordin's (1979)

definition of the working alliance. The "tasks" items (n = 13; e.g., "The therapist has the

skills and ability to help me") refer to the client's belief in the therapist's power and

method, a perception of feeling understood by the therapist, and the relevancy of tasks to

presenting problems. The "goals" items (n = 6; e.g., "The therapist understands my goals

in therapy") address the extent the client and therapist agree on and are working toward

the same goals to seek resolution of the presenting problems. The "bonds" items (n = 10;

e.g., "I tmst the therapist") concerns the affective and relational qualities of the

therapeutic relationship. The interpersonal system subscales are "self-therapist" (11

items; e.g., " The therapist cares about me as a person"); "other-therapist (11 items; e.g.,

"My partner feels accepted by the therapist"); and "group-therapist (7 items; "The

therapist is helping my partner and me with our relationship").

All items are rated on 7-point Likert scales, from "completely agree" (1) to

"completely disagree" (7). With reverse scoring on selected items, ratings are summed.

Higher ratings reflect perceptions that are in greater agreement with the statement. The

results can be scored as an index of global alliance (total score) or broken down into six

52

subscales: the tasks alliance, the goals alliance, the bonds alliance, the therapeutic

alliance of the individual, the therapeutic alliance of the partner, and the overall

therapeutic alliance of the couple.

In two pilot studies with small sample sizes, Pinsof and Catherall (1986) reported

overall rate-rerate reliabilities of .79 (n = 17) and .84 (n = 35). High internal consistency

coefficients (alpha) for the overall couple score were obtained in several studies: .96

(Johnson & Greenberg, 1985), .95 (n = 126) by Bourgeois et al. (1990), .93 (n = 66) by

Heatherington and Friedlander (1990), and .95 (n = 119) by Thomas (1994). Cronbach's

alpha reliability coefficients have also been obtained for the six subscales: from .85

(Johnson & Greenberg, 1985) to .88 (Heatherington & Friedlander, 1990) to .93 (Thomas,

1994) for the tasks subscale, from .70 (Heatherington & Friedlander, 1990) to .89

(Thomas, 1994) to .92 (Johnson & Greenberg, 1985) for the goals subscale, from .78

(Thomas, 1994) to .85 (Heatherington & Friedlander, 1990) to .88 (Johnson & Greenberg,

1985) for the bonds subscale. In addition, Heatherington and Friedlander (1990) found

no systematic statistical differences in client reports based on gender.

Evidence of constmct validity, in particular the extent to which the subscales are

measuring distinct underlying dimensions, has been demonstrated. Intercorrelations of

the six subscales were reported to range from .63 to .95 (Bourgeois et al., 1990) and from

.52 to .84 (Heatherington & Friedlander, 1990). A substantial proportion of unaccounted

for variance in these correlations supports their utility as separate subscales. In one study

(Heatherington & Friedlander, 1990), however, the authors reported that only one of three

CTAS subscales (tasks) was significantly correlated with one of two SEQ indexes

(depth). Initial indications of predictive validity have been shown by Bourgeois et al.

The concept of the split alliance was determined to be reliably identifiable by

Heatherington and Friedlander (1990). By comparing individual couple members' self-

therapist subscale ratings, they found that 43% of couples could be classified as

experiencing a split alliance based on a one standard deviation difference in ratings.

When a two standard deviation mle was applied, 14% of couples were said to be

experiencing a split alliance.

53

SEQ

The SEQ (Form 3; Stiles, 1980; Stiles & Snow; 1984a; see Appendix H) is a 10-

item self-report instmment that measures session impact. Stiles and colleagues (1994)

defined impact as "a session's immediate subjective effects, including clients' evaluations

of the session, their assessments of the session's specific character, and their postsession

affective states" (p. 175). This instmment measures session impact on two factor-

analytically-derived indexes. The "depth" index (n = 5; items I - 5) is comprised of the

following bipolar scales: deep-shallow, full-empty, powerful-weak, valuable-

worthless, and special—ordinary. The "smoothness" index (n = 5; items 6 10) is

comprised of the following bipolar scales: comfortable—uncomfortable, smooth—rough,

easy—difficult, pleasant—unpleasant, and relaxed—tense. Mean ratings on a 7-point

semantic differentials result in scales that each range from "1" to "7," with higher scores

indicating more positive perceptions of the session's impact.

Internal consistency alphas on the depth and smoothness indexes of .87 and .93

have been reported with an American sample (Stiles & Snow, 1984a), and of .90 and .92

with a large British sample (Stiles et al., 1994). Constmct validity of the depth and

smoothness indexes has been established in several factor analyses with American

samples (Stiles, 1980; Stiles & Snow, 1984a, 1984b) and one British sample (Stiles et al.,

1994). In each of these studies the author's concluded that depth and smoothness are

internally consistent, independent dimensions of participants' evaluations of sessions. In

addition. Stiles et al. examined the theoretical relation between the SEQ and the Session

Impact Scale (SIS) and found high intercorrelations among the depth and smoothness

indexes with all but one of the SIS indexes.

The concurrent validity for client reports on the depth and smoothness indexes has

strong empirical support. Stiles et al. (1994) reported the indexes to be highly correlated

with the SEQ's postsession "positivity" and "good therapist" indexes and the SIS's

"understanding," "problem-solving," and "relationship" indexes, as well as with single-

time global evaluations scales, "good-bad," and "helpful-hindering." In addition, the

indexes have been significantly correlated with specific covert counseling processes

54

(Fuller & Hill, 1985; Hill, Helms, Spiegel, & Tichenor, 1988a; Regan & Hill, 1992) and

overt counseling processes (Hill et al., 1988b). By contrast, research findings show

minimal or no support for the predictive validity of the depth and smoothness indexes.

Ghent reports on the depth index were significantly predictive of working alliance and

specific client-identified target concerns, but these correlations were not significant on the

smoothness index (Mallinckrodt, 1993). Neither index was found significantly related to

therapy outcomes on the Beck Depression Inventory (Mallinckrodt, 1993; Stiles, Shapiro,

& Firth-Cozen, 1990), the Symptom Checklist 90 (Stiles et al., 1990), the Rosenberg

Self-esteem Scale (Stiles et al., 1990), or the Counseling Center Follow-up Questionnaire

(Mallinckrodt, 1993).

Demographic Ouestionnaires

Questionnaires for participants (see Appendix J) and therapists (see Appendix K)

were used to obtain demographic data. For participants, age, gender and race,

relationship status, length of relationship, and educational level were assessed. For

therapists, age, gender, and race, year in degree program, extent of clinical experience,

and primary theoretical orientation to therapy were assessed. In addition, therapists were

asked to identify the percentage of time in therapy they use a genogram, and to rate their

clinical skills and ability to constmct a genogram on 5-point Likert scales from

"excellent" to "poor." The demographic data for participants and therapists was collected

to provide general sample descriptions and for group comparisons.

Session Follow-up Questionnaire

A session follow-up questionnaire (see Appendix L) was used to determine if the

conditions for the administration of couple therapy were conducted according to protocol.

Therapists were asked to note whether (a) both couple members were present during at

least part of the session, (b) another therapist or supervisor entered the therapy room and

participated as a co-therapist at any point during the session, and (c), any family member

friend, or other person(s) participated in any part of the therapy session besides the

55

couple. In addition, therapists were asked to provide necessary explanations. Following

the second therapy session, the therapists were asked to record how many minutes it took

the couple to complete either the TAGE, SAGE, or DAI. This information was used to

assess for significant group differences on the time variable by treatment group

intervention.

Design

This research had a 3 (treatment intervention) x 5 (session) "multiple-group-

multiple-I design" (Glass et al., 1975; see Figure 3.1). which is a variation on the basic

time-series experimental design. The purpose of this design is to provide a description of

level and direction changes due to treatment intervention effects. It also permits

comparison of treatment intervention effects between groups.

Analysis Plan

A growth curve analysis, referred to as hierarchical linear modeling (Bryk &

Raudenbush, 1987, 1992), was employed to investigate hypotheses la through lib.

According to Bryk and Raudenbush (1987), this growth modeling strategy enables

an integrated approach for (a) studying the stmcture of individual growth and estimating important statistical and psychometric properties of collections of growth trajectories; (b) discovering correlates of change, that is, factors that influence the rate at which individuals develop; and (c) testing hypotheses about the effects of one or more experimental or quasi-experimental treatments on growth curves, (p. 155)

This approach has a two stage, hierarchical character. The first stage is a within-

subjects model expressing individual growth over time. In the equation for this model,

the status on some outcome is represented as a function of an individual growth trajectory

plus random error. The second stage is a between-subjects model describing systematic

interindividual differences in growth. In the equation for this model, variability in

individual growth parameters is represented as a function of differences between subjects

in background and experience.

56

Two advantages of hierarchical linear modeling made it the appropriate analysis for

this study. First, hierarchical linear modeling has an advantage over traditional analytic

methods for longitudinal data (e.g., trend analysis in analysis of variance or multiple

analysis of variance, profile analysis in multiple analysis of variance) in that the number

and spacing of observations can vary across subjects (Bryk & Raudenbush, 1987. 1992).

Thus, "each individual in the data set can possess an empirical growth record containing

different numbers of waves of data with randomly assigned temporal spacing" (Willett &

Sayer, 1994, p. 379). In this investigation, therapy dropouts led to missing data, and

therapy sessions were usually not spaced equally. Second, hierarchical linear modeling

can handle any number of waves of longitudinal data. The advantage here is that this

approach allows change to be viewed as a continuous process of growth over time. In

doing so, it avoids inadequate descriptions of change inherent in pre/post ("two-wave")

designs (Francis, Fletcher, Stuebing, Davidson, & Thompson, 1991; Rogosa, Brandt, &

Zimowski, 1982; Willett, 1988; 1989; Willett, Ayoub, & Robinson, 1991). The

intricacies of the ongoing process of growth cannot be captured by taking a "snapshot" of

status before and after as in pre/post measurement; however, hierarchical linear modeling

is able to capitalize on the richness of growth trajectories (Willett et al., 1991; Willett &

Sayer, 1994).

In this study, measurement was based on two to five waves of data depending on

the number of couple therapy sessions and dependent measures packets the couples

completed. The two to five waves of data corresponded to the first two to five successive

couple therapy sessions. This decision was based on findings from Horvath and

Symonds' (1991) meta-analysis of 24 individual psychotherapy therapeutic alliance

studies. They found that eariy therapeutic alliance (first to fifth session) was slightly

more predictive of outcome than therapeutic alliance measures averaged across sessions

or taken toward the middle of treatment. To add, in their review of the therapeutic

alliance-outcome literature in individual psychotherapy, Horvath and Luborsky's (1993)

concluded that the therapeutic alliance may have an initial phase of development that

occurs within the first five therapy sessions.

57

Preliminary Analyses

Three preliminary analyses of the data were conducted. In the first analysis,

Cronbach's alpha was employed to test the reliability of CTAS and the SEQ. In the

second analysis, a series of one-way analysis of variances (ANOVA) and simple

regressions were conducted on mean intercept therapeutic alliance and session impact

scores by dependent variable to assess for significant group differences and to check the

integrity of the random assignment of couples to treatment groups. Discrete variables of

the participant's ethnic background, gender, and relationship status, and the therapist's

gender, ethnic background, degree program, and theory used in therapy were analyzed by

one-way ANOVAs. Continuous variables of the participant's age, education, length of

relationship, and the therapist"s age, clinical experience, clinical skill level, frequency of

use a genogram in therapy, and ability to constmct a genogram were analyzed with simple

regression analyses. In the final analysis, a simple regression analysis was conducted on

means for the time it took to conduct the second session treatment intervention to assess

for significant group differences.

Stage One Analyses

Modeling individual change occurred as follows. The selection of a suitable

mathematical function to represent tme individual growth was the first step in this

analysis. Willett (1989) and Willett and Sayer (1994) suggested that the choice of a

mathematical function should be guided by theory so that the individual growth

parameters will have interpretable meaning in light of the psychological process being

modeled. When no underiying theory of growth is available, however, a well-fitting,

lower-order polynomial should be used to approximate the "tme" growth function

(Willett, 1989; Willett & Sayer, 1994).

In this study, visual inspection of the individual time series by dependent variable

(i.e., therapeutic alliance and session impact) was used to examine each individual's

pattern of change. The visual tests indicated that for both dependent variables, a least

squares regression line representing constant change or straight line growth should be

58

fitted through each individual's data points. Slope scores were then calculated to

approximate each individual's tme change. Thus, each individual had two regression

lines with two intercept scores and two slope scores. A line was assumed an adequate

description for individual growth because the growth process in this study was not theory

driven, and there were a relatively small number of data collection waves (a total of five

waves).

Stage Two Analyses

Modeling systematic interindividual differences in growth occurred as follows. In

order to assess the effects of the treatments groups on the growth curves, mean slope

scores were calculated for the treatment groups by the dependent variables. Thus,

there were three mean slope scores for each dependent variable. Next, one-way multiple

analysis of variances (MANOVA) were used to test planned comparisons la through lib.

MANOVA was selected because there are two continuous dependent variables, and it

would give an idea of the relative effect difference associated with each dependent

variable.

Other Analyses

Three additional sets of analyses were conducted on the data. The first was drawn

from a suggestion for future research in Quinn's et al. (1997) study of the relation

between the therapeutic alliance and outcome in couple therapy: "Do couples...that

terminate prior to reaching goals or completing therapy differ in their perceptions of

alliance?" (p. 436). Data from couples who did not complete five measurements (those

who completed two to four measurements), "noncompleters," were examined in the

following manner. For each of the dependent variables, intercept and slope scores from

noncompleters were collapsed into one group and intercept and slope scores from couples

completing all five measurements, "completers," were collapsed into a second group, t

tests were conducted by dependent variable on mean intercept scores to assess whether

59

significant initial differences existed between the groups, and on mean slope scores to

determine if the groups' growth patterns were significantly different.

Second, to test hypotheses Ilia through Vc. regarding the relations between the

tasks, goals, and bonds subscales of the CTAS and the depth and smoothness indexes of

the SEQ over time, mean slope scores for the specific subscales and indexes were

correlated using Spearman's rho or Kendall's tau correlation coefficients. Spearman's

rho was used for calculating correlations within the TAGE+CT and SAGE+CT groups

because the number of cases in these groups was less than 30 (Issac & Michael, 1990).

Kendall's rho was used to calculate correlations within the DAI+CT group because the

number of cases in this group was less than 10 (Issac & Michael, 1990). These

correlations served to test the constmct validity of the CTAS content subscales.

Third, the concept of the split alliance was determined to be reliably identifiable by

Heatherington and Friedlander (1990) when assessed at one point in couple therapy.

Whether the concept of the split alliance is empirically viable in couple therapy across

time was assessed by comparing couple member's self-with-therapist mean slope scores

and applying a two standard deviation decision mle. This decision mle was selected

because Heatherington and Friedlander (1990) suggested that a one standard deviation

decision mle "reflects a small measurement difference on the instmment and could easily

be accounted for by subjects' response styles" (p. 305). The result was the percentage of

couples in the sample said to be experiencing a split alliance with the therapist.

60

CHAPTER IV

RESULTS

Preliminary Analysis

The Reliability of the Dependent Measures

Cronbach alpha (internal consistency) reliability coefficients were calculated for the

CTAS and the SEQ and their subscales and indexes, respectively (see Table 4.1). The

overall alpha for the CTAS was .97, and the alphas for the three subscales—bonds, tasks,

and goals—were .94, .95, .84, respectively. For the SEQ, the alpha for the total test was

.88, and .93 for the smoothness index and .89 for the depth index. The reliability

coefficients were strong and demonstrated that the overall scales and subscales or indexes

of both dependent measures had internal consistency.

Group Differences Analyses

Participants

Demographic data from the participants was analyzed by a series of one-way

ANOVAs and simple regressions on the initial mean therapeutic alliance and session

impact scores to assess for any significant group differences. Discrete variables of

participant's ethnic background, gender, and relationship status were analyzed by one­

way ANOVAs. Continuous variables of the participant's age, education, length of

relationship and were analyzed with simple regression analyses. No significant group

differences were found (see Tables 4.2 - 4.5). These results suggested the random

assignment of couples to treatment groups, but they must be interpreted in light of small

participant numbers per treatment group (group one, n = 14; group two, n = 16; group

three, n = 4) which lowered the statistical power for detecting significant group

differences.

61

Table 4.1 Cronbach Alpha Reliability Coefficients

for the CTAS and the SEQ

Dependent Measure Cronbach Alphas

CTAS

Overall (29 items) .97

Bonds Subscale (10 items) .94

Tasks Subscale (13 items) .95

Goals Subscale (6 items) 84

SEQ

Overall (10 items) 88

Smoothness Index (5 items) .93

Depth Index (5 items) 89

62

Table 4.2 Means, Standard Deviations, and One-way ANOVA

Summary Table for Participant's Initial CTAS Mean Scores by Participants' Genders,

Ethnic Backgrounds, and Relationship Statuses

Variable n Mean Scores

SD df

Gender

Male

Female

Ethnic Background

African-American

Asian

Caucasian

Hispanic

Native-American

Relationship Status

Dating

Living Together

Married

Note. The range of scores

17

17

2

1

22

7

1

2

6

26

4.79

5.10

5.14

4.14

5.03

4.80

5.14

5.05

5.13

4.89

for the CTAS is 1 - 7.

.54

.63

.68

.59

.63

.02

.80

.58

1,32

4,28

2.41 13

.73 .58

2,31 .42 .66

63

Table 4.3 Means, Standard Deviations, and One-way ANOVA

Summary Table for Participant's Initial SEQ Mean Scores by Participants' Genders,

Ethnic Backgrounds, and Relationship Statuses

Variable n Mean Scores

SD df

Gender

Male

Female

Ethnic Background

African-American

Asian

Caucasian

Hispanic

Native-American

Relationship Status

Dating

Living Together

Married

Note. The range of scores

17

17

2

1

22

7

1

2

6

26

4.34

4.54

3.70

3.90

4.45

4.80

3.30

4.55

4.38

4.45

for the SEQ is 1 - 7 .

.82

.70

1.56

.74

.49

.64

.74

.80

1,32

4,28

.58 .45

1.65 19

2,31 .04 .96

64

^

Tab

le 4

.

00 < H U 15

Init

ant'

s

a. 131

C

lyse

s of

]

c3 C

< c

_o

ress

00 <u DC

JJ o. E

«^ o c/3

ngth

s,

Le

<u 00 < "c/s

artic

ipan

CL c o c/3 (U u. O CJ

c/3

ean

S

tion

!d

uca

UJ

o c/3

13

and

Lev

c/T

hip

c/3 C o

00

3 00

13 -a o

c/3

c «J

o U

UHI

<

c^l

Oi

cd

CQ

4—1

u CQ

ui 00

r^ CN — CN

c*-) m

o o

ON O N ON VO

m

O

t ^ 00 O O

VO

O ^ Tt

r^ oo

o o

VO

o

O

JJ X

.2 >

X

o 2 -2 DC

CJ 3

.- 0 0 (U

< .2 ^ 65

a CO

4—>

'E c«

^ O cd 4>

C cd .£-"o 'E cd

»o cu ^ o

id ^

c < c

_o ' c ^

c/3

2 00 OJ

DC OJ

£.2

c/3

x: c

= td CJ

si < o

c/3 «-> QJ

^ C > CJ

CU c o

2 o CJ

00 c cd u

• a c cd

c/T O ,

X c/3 c o

a:

.E c^

cd

E E 3

00

13 -o o

c/3

C CJ

o

u

tui

<

QC\

a

cd 4—•

CQ

cd 4—1

u CQ

ui

00

X I

00

o

o

o

o

VO CN

CN O

OO — r - NO

oo —' CN >n

CN —'

o o

o

- ^ CN — CO m m

CN O

ON —

o —

NO

ON —

o —

00 O

o

JJ JO .2 cd

>

X

o o

.2 u

Cd CJ 3

T3 W

< ^ ^

66

Therapists

Demographic data from the therapists was analyzed by a series of one-way

ANOVAs and simple regressions on the initial mean therapeutic alliance and session

impact scores to assess for any significant group differences. Discrete variables of the

therapist's gender, ethnic background, degree program, and theory used in therapy were

analyzed by one-way ANOVAs. Continuous variables of the therapist's age, clinical

experience, clinical skill level, frequency of use a genogram in therapy, and ability to

constmct a genogram were analyzed with simple regression analyses. No significant

group differences were found (see Tables 4.6 - 4.9). It was concluded that the

assignment of therapists to the treatment groups was random. These results, however,

must be interpreted in light of low statistical power.

Time for Second Session Treatment Intervention

Seventeen couples participated in the second session treatment intervention. The

mean length of time it took to conduct this intervention for the three treatment

intervention groups was 60.29 minutes (SD = 13.80). To assess for significant group

differences on this variable by treatment intervention group, mean times were calculated

and a simple regression analysis was conducted. The mean time for the seven couples in

the TAGE+CT group was 60 minutes (SD = 4.23). The mean for the eight couples in the

SAGE+CT group was 63.75 minutes (SD = 5.32). The mean for the two couples in the

DAI+CT group was 47.50 minutes (SD = 12.50). The results of a simple regression

analysis performed on these means were not significant (F[l, 15] = .37, p = .55), which

suggested that there were no group differences based on the time it took to conduct the

second session treatment intervention. As discussed above, these results must be

interpreted in light of low statistical power.

67

Table 4.6 Means, Standard Deviations, and One-way ANOVA

Summary Table for Participant's Initial CTAS Mean Scores by Therapist's Gender,

Ethnic Background, Degree Program, and Theory Used

Variable n Mean Scores

SD df

Gender

Male

Female

Ethnic Background

Asian

Caucasian

Degree Program

Doctoral

Masters

Theory Used

Behavioral/ Social Learning

Brief

Stmctural

Other

10

7

5

11

15

2

4.72

4.88

5.01

4.74

4.80

4.71

.53

.58

.66

.48

.54

.81

2

1

5

8

4.86

4.07

5.10

4.75

.24

.65

.47

Note. The range of scores for the CTAS is 1 - 7.

1, 15

1,14

1, 15

3, 12

.37

.84

.04

.56

.38

.84

1.19 .36

68

Table 4.7 Means, Standard Deviations, and One-way ANOVA

Summary Table for Participant's Initial SEQ Mean Scores by Therapist's Gender,

Ethnic Background, Degree Program, and Theory Used

Variable n Mean Scores

SD df

Gender

Male

Female

Ethnic Background

Asian

Caucasian

Degree Program

Doctoral

Masters

Theory Used

Behavioral/ Social Learning

Brief

Stmctural

Other

Note. The range of scores

10

7

5

11

15

2

2

1

5

8

4.49

4.61

5.04

4.28

4.47

5.05

5.25

5.00

4.48

4.30

for the SEQ is 1 - 7.

.79

.61

.27

.74

.72

.21

.35

.62

.79

1, 16

1, 14

1, 15

3, 12

12

4.83

1.21

.73

.06

.29

1.13 .38

69

c/3 < E-U t> cd 2 4-> C/3

X o c/3 U

'^ o O K 4->

CJ

c c CJ

X < c«

4 - ^

C/3

CJ

• • 5

^ 2 2 cd £

c o c/3

2 o CJ

0 0 ^ (U c

OC «« (O

.E CO

ex „ tu cu cd "7^ o (U

-a OJ c/3

E ii > 2

X ^ g

0 0 c/3 <U

<id.i c _o

'(A C/3 (L)

5d .S g)CJ c

o

X o

Cd

E E 3

00 1 3 O

c/3 4 . ^

c

O

u

tui

<

0C\

Oi

cd OJ

PQ

Cd

4—>

CQ

ui 00

X

.2 cd

q

o

ON

r*-*

O

r<~,

CN

O

>n

q —'

o

O

<N

O

o o

— CN

O cn

— o O

VO OO CN —'

Tf r- \o — ON o — m

r- r^ VO m O oo Tt O

— CN NO OO O cn CN —

— o

— CN — Tt

o o o —

»n

o

00

o

ON CN

NO

ON CN

I

CJ

2

Con

st!

o 1 •taj cd

^ 0 0 • - O ~ c X U < o

(U 0 0

<

CJ c

xper

ie

w

ical

c *^ U

(U >

kil

lLe

00

ical

c ;X u

Tim

e

n

ent

<

CJ

Ou

I Use

d

E cd b« 00

o c (U

a 70

ble

4.9

Cd

a u 0 0 * -_ CJ

.2 S 4—1 * - >

. — c/5

^ o ." U ^ • «

C O cd * -CU >%

rtic

bi

li es

of P

a pi

st's

A

Ana

lys

n T

hera

c o . 2 C/5 c/3 JJ C/3 i -(U p ^ CJ

ooco Oi § (U iJ

"CuS E

c U 4->

•r c 9- ^ X 0)

U CU

Clin

ical

11

Lev

el,

ram

. A

ge,

linic

al S

ki

g)U c (U

O

T3 <U c/3

enog

ram

U

Tim

eG

o

00

Cd

E E 3

00

1 3 T3 O

c/3

c ' c j

u

tui

<

ctil

Oi

3 QQ

Cd 4—>

<u CQ

ui 00

JJ X

.2 Cd

CN

r i

O

q

o

q

o

(T)

—

o

r n

O

o

00

o

ON CN

O

CN

O

ON CN CN O

O

o

r^ CN —

— o o

00 CN

O

ON

Ov

ON

CN

O

—

NO

O N

Ov O

CN 0 0

CN

.29

o

i n

0 0 i n

m

00 O

CJ

5

Con

st!

1

o 5 w cd

. ^ 0 0

•- o .— c X u < a

(U 0 0

<

nee

xper

ie

u •

ica

c JX U

vel

kill

Le

0 0 •

ica

c "^ U

Tim

e

o

ent

CJ

cu

lUse

d

E cd ka 0 0

o c OJ

O 71

Stage One Analyses

Hypotheses la - lib

It was hypothesized that for therapeutic alliance and session impact, the mean rate

of change for the TAGE+CT group will be greater than mean rate of change for the

SAGE+CT group. For therapeutic alliance and session impact, the mean rate of change

for the TAGE+CT group and the SAGE+CT group will be greater the mean rate of

change for the DAI+CT group.

For the stage one analyses, simple regression analyses were performed on the

therapeutic alliance and session impact scores for 17 couples (34 individuals) who

completed from two to five couple therapy sessions resulting in 76 simple regressions.

For each individual, these regressions yielded two intercept scores and two mean slope

scores. The results of these regressions are displayed by couple in Figures 4.1 - 4.17.

Inspection of the within-couple member growth trajectories suggested that a straight-line

growth model provided a robust and convenient summary of changes in therapeutic

alliance and session impact over time.

Upon examination of the individual slope scores, participants 13 and 14's (couple

seven) slope scores on the CTAS (1.41, 1.60, respectively) and the SEQ (1.69, 2.90,

respectively) were notably higher than those of the other participants. A box-and-whisker

plot identified the scores from this couple as outliers or extreme values (see Figure 4.18).

In addition, the results of a Shapiro-Wilks test on the CTAS and SEQ slope scores were

significant (S-W[34] = .81, p = .01; S-W[34] = .86, p = .01, respectively), indicating that

the data were not normally distributed. When the slope scores from couple seven were

removed from the sample, the Shapiro-Wilks test lost significance for the CTAS and SEQ

(S-W[32] = .97, p = .63; S-W[32] = .96, p = .37, respectively), supporting the hypothesis

that the data without couple seven's slope scores were normally distributed. This finding

is important to statistical inference. For these reasons, the slope scores from couple seven

were considered outliers and eliminated from the remainder of the analyses except for the

split alliance test.

72

7.0

6.5

6 . 0 -

5.5

."5.0

ka o CJ

on g. 4.0

on q 3.5 on "O a .3.0 -I

C/2

< U 2.5 c

S 2.0

d ^

Session Number

Figure 4.1 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple One

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

''Female's CTAS intercept score = 6.48, female's CTAS slope score = -.21. ^Male's CTAS intercept score = 4.35, male's CTAS slope score = -.14. ''Female's SEQ intercept score = 5.80, female's SEQ slope score = -.90. ' Male's SEQ intercept score = 2.80, male's SEQ slope score = .60.

73

7.0 T

6.5

6.0

5.5 •

5 0 -

D •*••> k.

o on

c/3

O 3.5 C/5 "O c

c

3.0 c/n <

U 2.5

S 2.0

Session Number

Figure 4.2 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Two

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

' Female's CTAS intercept score = 4.99. female's CTAS slope score = .28. ''Male's CTAS intercept score = 5.46, male's CTAS slope score = . 11. ''Female's SEQ intercept score = 4.32, female's SEQ slope score = .32. ' Male's SEQ intercept score = 4.15, male's SEQ slope score = .33.

74

7.0

6.5

6.0

5 5

5.0

S 4.5 O CJ

on ^ 4.0

_o c/> a 3.5 4 u on

i 3.0 H on < H U 2.5 c a <u S 2.0

Session Number

Figure 4.3 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Three

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

^Female's CTAS intercept score = 5.39, female's CTAS slope score = -. 15. ''Male's CTAS intercept score = 3.73, male's CTAS slope score = .14. 'Female's SEQ intercept score = 5.42, female's SEQ slope score = -.24. ''Male's SEQ intercept score = 3.63, male's SEQ slope score = .21.

75

Session Number

Figure 4.4 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Four

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 5.38, female's CTAS slope score = -. 14. ''Male's CTAS intercept score = 5.49, male's CTAS slope score = .28. ''Female's SEQ intercept score = 4.27, female's SEQ slope score = .15. ''Male's SEQ intercept score = 4.47, male's SEQ slope score = .70.

76

L .

o 'J

C/5 c

y5

a u on T3 C CQ

C/3 < H U c :3

7.0 T

6.5

6.0

5.5 -

5.0

4.5 •

4.0

3.5

3.0

2.5

2.0

Session Number

Figure 4.5 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Five

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 5.14, female's CTAS slope score = -.17. ''Male's CTAS intercept score = 4.90, male's CTAS slope score = .35. 'Female's SEQ intercept score = 5.23, female's SEQ slope score = -.20. ''Male's SEQ intercept score = 6.67, male's SEQ slope score = -1.10.

77

Session Number

Figure 4.6 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Six

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

^Female's CTAS intercept score = 4.81, female's CTAS slope score = . 19. ''Male's CTAS intercept score = 4.30, male's CTAS slope score = .20. "Female's SEQ intercept score = 3.54, female's SEQ slope score = .32. ''Male's SEQ intercept score = 4.49. male's SEQ slope score = .02.

78

y.

'J

g :/5 -3 c

on <

U c n y

70

6.5 •

6 . 0 -

5.5 •

5.0

4.5 -

4.0

3.5

3.0

2.5

2.0

Session Number

Figure 4.7 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Seven

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'^Female's CTAS intercept score = 3.35, female's CTAS slope score = 1.69. ''Male's CTAS intercept score = 4.00, male's CTAS slope score = 1.41. "Female's SEQ intercept score = .50, female's SEQ slope score = 2.90. ''Male's SEQ intercept score = 3.20, male's SEQ slope score = 1.60.

79

c/l lU k-

o CJ

on

7.0

6.5

6.0

5 5

5.0

4 5

4 . 0 -

on

O 3.5 U c/)

i 3.0 CO

<

U 2.5 H c

S 2.0

Session Number

Figure 4.8 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Eight

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 5.21, female's CTAS slope score = -.55. ''Male's CTAS intercept score = 4.31, male's CTAS slope score = .28. "Female's SEQ intercept score = 1.7, female's SEQ slope score = .90. ''Male's SEQ intercept score = 4.90, male's SEQ slope score = -.40.

80

res

o CJ C/1

K o on O U c/ T3 C

on < H U c ea u s

/ . u •

6 . 5 .

6 . 0 -

5.5 •

5 .0 -

4.5 •

4 0 -

3 . 5 -

3 . 0 -

2 . 5 -

2.0 1

/ •^^^^^'^ " " ' " • ' - d

I" k

/ b /

/ /

/ /

/ •

1 1 1 1 1

Session Number

Figure 4.9 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Nine

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.81, female's CTAS slope score = .26. ''Male's CTAS intercept score = 5.03, male's CTAS slope score = -.01. "Female's SEQ intercept score = 4.12, female's SEQ slope score = .38. ''Male's SEQ intercept score = 5.80, male's SEQ slope score = -.04.

81

7.0 T

Session Number

Figure 4.10 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Ten

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.53, female's CTAS slope score = .42. ''Male's CTAS intercept score = 4.80, male's CTAS slope score = .28. "Female's SEQ intercept score = 3.85, female's SEQ slope score = .25. ''Male's SEQ intercept score = 5.56, male's SEQ slope score = -.02.

82

u ka o CJ

6/3

7.0

6 . 5 -

6.0

5.5

5.0

4.0

on O 3.5 U c/) •o

i 3.0-1 c/3 < H U 2.5 c C3

2 2.0

Session Number

Figure 4.11 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Eleven

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 6,38, female's CTAS slope score = -.45. ''Male's CTAS intercept score = 4.62, male's CTAS slope score = .35. "Female's SEQ intercept score = 4.60, female's SEQ slope score = .30. **Male's SEQ intercept score = 2.30, male's SEQ slope score = .70.

83

7.0 T

6.5

6.0

5.5

5.0

( U - » • • k a o CJ on

_o c/3

a 3.5 U c/3

i 3.0 on < H U 2.5 c «a u S 2.0

Session Number

Figure 4.12 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Twelve

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.24, female's CTAS slope score = .03. ''Male's CTAS intercept score = 3.93, male's CTAS slope score = .69. "Female's SEQ intercept score = 4.10, female's SEQ slope score = .50. ''Male's SEQ intercept score = 4.80, male's SEQ slope score = -.80.

84

7.0 T

6.5

6.0

5.5

5.0

</J 1 > ka

o CJ on 2L o on <y u

4.5

4.0

3 5

C/3

"O

i 3.0 on < H U 2.5 c 4J

2 2.0.

Session Number

Figure 4.13 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Thirteen

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.00, female's CTAS slope score = .26. "Male's CTAS intercept score = 5.50, male's CTAS slope score = -.10. "Female's SEQ intercept score = 5.80, female's SEQ slope score = -.28. ''Male's SEQ intercept score = 4.10, male's SEQ slope score = .50.

85

7.0

6.5

6.0

5.5

5.0

<U • • • - ' ;_ O CJ

C/3

^ 4.0 _o c/3 q 3.5 on •o S 3.0 H

on

U 2.5 e «o

S 2.0

Session Number

Figure 4.14 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Fourteen

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.03, female's CTAS slope score = .21. ''Male's CTAS intercept score = 4.17, male's CTAS slope score = -.03. "Female's SEQ intercept score = 3.68, female's SEQ slope score = .08. ''Male's SEQ intercept score = 5.39, male's SEQ slope score = -.29.

86

7.0

6.5

6.0

5.5

5.0

2i o CJ

C/l

on O 3.5 U on

S 3.0 c/3 < H U 2.5 H c ea ii S 2.0

^ ^ ^ > ;

Session Number

Figure 4.15 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Fifteen

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 5.02, female's CTAS slope score = -.08. ''Male's CTAS intercept score = 4.29, male's CTAS slope score = .19. "Female's SEQ intercept score = 4.35, female's SEQ slope score = -.15. ''Male's SEQ intercept score = 3.85, male's SEQ slope score = .24.

87

7.0 T

6.5

6.0

5.5 .

5 . 0 -

C/5

ka

o CJ C/3

^ O on

a uu c/3 -o c ea

c/3 < H U c C9 <U

4 5

4.0

3 5

3.0

2.5

c ' •

- - 7 - ^ .^

^ K

" v

v ^

Session Number

Figure 4.16 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Sixteen

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.87, female's CTAS slope score = .00. ''Male's CTAS intercept score = 4.26, male's CTAS slope score = .02. "Female's SEQ intercept score = 5.65, female's SEQ slope score = -.11. ''Male's SEQ intercept score = 5.96, male's SEQ slope score = -.11.

88

7 0 T -

65

6.0

5 5

50

!/) OJ k .

o CJ on Si o c/3 O UJ on •o c on < H U

4 . 5 -

4 . 0 .

3 . 5 -

3 . 0 -

2 . 5 .

S 2.0

/ c

Session Number

Figure 4.17 Regression Lines Displaying the Relationship between Mean

Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session

Number for Couple Seventeen

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

^Female's CTAS intercept score = 5.52, female's CTAS slope score = .10. "Male's CTAS intercept score = 4.62, male's CTAS slope score = .04. "Female's SEQ intercept score = 5.14, female's SEQ slope score = .08. ' Male's SEQ intercept score = 5.11, male's SEQ slope score = -.09.

89

V)

o (J

(fi 0) a. _o c/) O

UJ

T3

c CO

< O c CO 0) IS

2 -

1 •

-)fl4

* 1 3 0 3

I Mean CTAS

Slop« Scores

I Mean SEQ

Slope Scores 34

Number of Participants

Figure 4.18 Box-and-Whiskers Plot of the Mean

CTAS and SEQ Slope Scores

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire. The box-and-whiskers plot displays summary statistics for the distribution of mean CTAS and SEQ slope scores. The horizontal line in the box represents the median slope score for the sample. Fifty percent of the participants have values within the box. Lines above and below the box represent the largest and smallest observed values that are not outliers. Values more than one-and-one-half box-lengths from the 75th percentile are outliers, and values more than three box-lengths from the 75th percentile are extremes.

* = extreme value.

O = outlier.

90

After removal of the outliers, the collected mean slope score for the sample (n = 32)

on the CTAS was .08 (SD = .25), and the collected mean slope score for the sample (n =

32) on the SEQ was .06 (SD = .45). When broken down by gender, the collected mean

slope scores on the CTAS were .16 (SD = .28) for females (n = 16) and .05 (SD = 15) for

males (n = 16). The collected mean slope scores on the SEQ were .16 (SD = 32) for

females (n= 16) and .10 (SD = .21) for males (n= 16) (see Figure 4.19). These slope

scores indicated the slightest positive within-couple member growth across time on the

therapeutic alliance and session impact variables. For the CTAS, the maximum observed

positive and negative changes on the 7-point Likert scale over five sessions were +1.40

and-1.10. These numbers for the SEQ were +2.10 and-1.8. To illustrate, if the

participant who demonstrated the greatest positive rate of change had an average CTAS

rating of 4.00 at the first session, his or her average CTAS rating at session five would

have been 5.40. If the participant who demonstrated the greatest negative rate of change

had an average CTAS rating of 4.00 at the first session, his or her average CTAS rating at

session five would have only been 2.90.

Stage Two Analysis

Hypotheses la - lib (Continued)

It was hypothesized that for therapeutic alliance and session impact, the mean rate

of change for the TAGE+CT group will be greater than mean rate of change for the

SAGE+CT group. For therapeutic alliance and session impact, the mean rate of change

for the TAGE+CT group and the SAGE+CT group will be greater the mean rate of

change for the DAI+CT group.

To assess the systematic between-couple member differences in growth one-way

MANOVAs were used to test two planned comparisons on two dependent variables:

CTAS and SEQ. The independent variable was group. Because the group sample sizes

were not equal, the Pillai's Trace statistic was used to test for multivariate significance

(Tabachnick & Fidell, 1989). The first comparison was performed between the

TAGE+CT (n = 14) and SAGE+CT (n = 16) groups. The F test was not significant (F(2,

91

y

'.J

CU

C on T3 C ed

00 <

•Yi

O CU

E o U

SessKDn Number

Figure 4.19 Regression Lines Displaying the Relationship between Collected Mean

Slope Scores on the Couple Therapy Alliance Scale and the Session Evaluation Questionnaire and Session Number

For Entire Participant Sample by Gender

Note. CTAS = Couple Therapy Alliance Scale. SEQ = Session Evaluation Questionnaire.

'Female's CTAS intercept score = 4.78, female's CTAS slope score = 1 6 . "Male's CTAS intercept score = 4.80, male's CTAS slope score = .05. "Female's SEQ intercept score = 4.32, female's SEQ slope score = 1 6 . ''Male's SEQ intercept score = 4.59, male's SEQ slope score = . 10.

92

27) = .41, p = .67, see Table 4.10). The second planned comparison was performed

between the combined TAGE+CT and SAGE+CT groups (n = 30) and the DAI+CT

group (n = 2). Again, the F test was not significant (F(2, 29) = 1.35, g = .28, see Table

4.10). These findings did not support hypotheses la - lib, and indicated that the treatment

interventions did not significantly affect the mean rates of change m the combined

dependent variables. It should be noted, however, that the observed power was .11 for

the first comparison and .27 for the second comparison.

Other Analyses

Therapy Completers versus Noncompleters

Data from the previous stage one analysis were used to examine the impact on the

growth curves based on whether a couple completed five therapy sessions (completers) or

completed two to four therapy sessions (noncompleters). Mean intercept and slope scores

from eight completer couples (n = 16) and eight noncompleter couples (n = 16) were used

for this analysis. To assess for initial differences between groups, t tests were performed

by dependent variable on initial mean intercept scores. Results were not significant for

either the CTAS (t [30] = 1.61, p = . 19) or the SEQ (t [30] = .63, g, = .54). This finding

supported the random assignment of couples to treatment groups. Next, t tests were

conducted by dependent variable on mean slope scores to assess the effects of the two

groups on the growth curves. Significant results were not found for either the

CTAS (t [30] = .98, E = .33) or the SEQ (t [30] = .23, p = .82). Growth curves, therefore,

were not impacted by a couple's completion of five sessions.

Hypotheses lUa - Vc

It was hypothesized that for the three treatment groups (TAGE+CT, SAGE-i-CT,

and DAI+CT): (a) a change in the CTAS tasks subscale will be positively correlated with

a change in the SEQ depth index, (b) a change in the CTAS goals subscale will be

positively correlated with a change in the SEQ depth index, and (c) a change in the CTAS

bonds subscale will be positively correlated with a change in the SEQ smoothness index.

93

Table 4.10 One-way MANOVA Summary Table for Planned

Comparisons of Participants' CTAS and SEQ Mean Scores as a Function of Group

Source of Pillai's Hypoth. Error Variance Trace'' df df F p Power

Group

Comp. I" .03 2 27 .41 .67 .11 b

Comp. 2" .09 2 29 1.35 .28 .27

Note. 'Pillai's Trace is recommended to evaluate multivariate significance when cell sample sizes are not equal (Tabachnick & Fidell, 1989). "Comparison one was between the TAGE+CT and SAGE+CT treatment intervention groups. "Comparison two was between the combined TAGE+CT and SAGE+CT treatment intervention groups and the DAI+CT treatment intervention group.

94

Mean slope scores were calculated for 16 couples (n = 32) on three CTAS subscales

(tasks, goals, bonds) and on two SEQ indexes (depth, smoothness). Spearman's rho

correlation coefficients were calculated to obtain intercorrelations within TAGE-hCT and

SAGE+CT treatment groups between (a) the CTAS tasks subscale and the SEQ depth

index, (b) the CTAS goals subscale and the SEQ depth index, and (c) the CTAS bonds

subscale and the SEQ smoothness index (see Table 4.11).

In the TAGE+CT group (n = 14), no significant correlations were found. In the

SAGE+CT group (n = 16), a significant positive correlation was obtained between the

CTAS tasks subscale and the SEQ depth index (r = .73, p < .01). This significant

correlation supported the hypothesis that a positive change over time in the CTAS tasks

subscale was related to a positive change over time in the SEQ depth index. In other

words, couples who participated in a SAGE+CT felt increasingly understood and

confident about the help offered by their therapist, and equated this with greater depth in

their therapy as the sessions progressed.

Kendall's rho correlation coefficients were calculated to obtain intercorrelations

within DAI+CT treatment group between the CTAS tasks subscale and the SEQ depth

index, the CTAS goals subscale and the SEQ depth index, and the CTAS bonds subscale

and the SEQ smoothness index (see Table 4.11). In the DAI+CT group (n = 2), no

significant correlations were found.

Split Alliance

To assess the concept of split alliance, CTAS self-with-therapist subscale ratings

from 17 couples (n = 34) were used to compute mean slope scores. The mean slope score

for the combined treatment groups and the mean slope scores for the three treatment

groups were calculated to serve as points for applying the two standard deviation

decision rule. The two standard deviation difference in mean slope scores was .94 for the

combined groups, .89 for the TAGE+CT group (seven couples), .50 for the SAGE+CT

group (eight couples), and 1.52 for the DAI+CT group (four couples). A comparison of

the differences between the couple members' mean slope scores with the combined

95

Table 4.11

Intercorrelations of the CTAS Content Subscales and SEQ Indexes by Treatment

Intervention Group

Group

SEQ Indexes

CTAS Subscale SEQ Depth SEQ Smoothness

TAGE+CT

SAGE+CT

DAI+CT"

CTAS Tasks

CTAS Goals

CTAS Bonds

CTAS Tasks

CTAS Goals

CTAS Bonds

CTAS Tasks

CTAS Goals

CTAS Bonds

-.12'

-.25'

.73**'

.19''

1.00'

l.OO''

.07'

.07'

1.00'

Note. CTAS = Couple Therapy Alliance Scale; SEQ = Session Evaluation Questionnaire; TAGE+CT = therapist-administered genogram plus couple therapy; SAGE+CT = self-administered genogram plus couple therapy; DAI+CT = directed assessment interview plus couple therapy.

'n=14 . "n=l6 . "n = 2. ''Spearman's rho correlation coefficient. 'Kendall's tau correlation coefficient.

**P< .01 (one-tailed).

96

deviations revealed one couple (6%) in the sample who expenenced a split alliance with

their therapist. Within the treatment groups, one couple m the TAGE+CT group (14%,

the same couple as above), and one couple in the SAGE+CT group (13%) experienced a

split alliance with their therapist. Relatively few couples who participated in a genogram

interview and couple therapy had disparate attitudes about their individual alliances with

their therapist.

97

CHAPTER V

DISCUSSION

This study was undertaken to examine the process of genogram construction in

couple therapy from the client's perspective. More specifically, the researcher explored

how couples' experiences' of this process affected their alliance with the therapist, and

impacted the specific character of the therapy itself during early sessions. The

investigation was designed so that the intricacies of the change over time in the

therapeutic alliance and the session's impact on couples could be examined. Of primary

interest was whether individual couple member's perceptions of the therapeutic alliance

and the session's impact changed over time given their participation in a TAGE, SAGE,

or control treatment intervention and couple therapy, and whether these changes differed

between couples in systematic ways given their participation in a specific treatment

intervention. A secondary concern was the relation between the therapeutic alliance and

the session's impact over time. This chapter reviews the results of the study, discusses its

limitations, makes recommendations for future research, and suggests clinical

implications of the findings.

Review of Results

Within-Couple Member Differences

The first set of hypotheses (la - Hb), was tested in a two stage analysis. In the first

stage, within-couple member growth curves were plotted and examined by dependent

variable. These growth trajectories were best described with straight lines that showed

little individual variation and minimal collective positive rates of change on the

dependent variables.

The initial stage in the statistical modeling of growth over time involved an

exploratory plotting and inspecting, for each participant, of empirical growth trajectories.

The purpose of this step was to investigate whether couple members' perceptions of the

dependent variables, therapeutic alliance and session impact, changed with the same rates

98

and patterns over time after receiving a second session treatment intervention along with

two to five sessions of couple therapy. Results of the inspection of the within-couple

member growth curves led to the adoption of a linear function as the best approximation

to the "true" growth function. In other words, growth in therapeutic alliance and session

impact were summarized by fitting straight line growth models using ordinary least

squares regression, two per participant. Most of the individual participants showed slight

variation in their rates of growth on both dependent variables. As a collective sample, the

rates of change were minimally positive on therapeutic alliance and session impact

variables. This suggested that session-to-session rates of change on the two dependent

variables for the whole sample were essentially stable.

Between-Couple Member Differences

The purpose of the second analysis phase was a continued test of hypotheses la -

nb. The estimated slopes of the within-couple member fits provided unbiased estimates

of the session-to-session rates of change in therapeutic alliance and session impact that

became the basis of the subsequent between-couple member analyses. Planned

comparisons were performed to examine predicted systematic differences in the session-

by-session growth rates between TAGE+CT and SAGE-i-CT treatment intervention

groups, and the combined TAGE+CT and SAGE+CT groups with the DAI+CT treatment

intervention group. No significant group differences were found in either comparison on

the combined dependent variables. Given the minimal variation in rates of change in

stage one, small sample sizes per group (TAGE+CT, n= 14; SAGE+CT, n = 16;

DAI+CT, n = 2), and low observed power (.11 for comparison one and .27 for

comparison two), it was not surprising that none of the hypothesized between-groups

differences were supported. The findings indicated that the treatment interventions had

no bearing on the growth rates of either therapeutic alliance or session impact. In the first

comparison, having therapists construct genograms (the TAGE+CT group) made no

greater impact on the growth curves than when therapists were not involved in the

genogram construction process (the SAGE+CT group). In the second comparison, the

99

construction of genograms (TAGE+CT and SAGE+CT groups) made no greater impact

on the growth curves than when genogram-like data was collected but genograms were

not constructed (SAGE+CT control group). Nor was it unexpected that when the

participants were divided into two groups of therapy completers (couple-members who

completed five sessions) and noncompleters (couple-members who completed two to four

sessions) that no significant between-groups differences for rates of change were noted on

either dependent variable.

Two plausible explanations exist for the lack of significant differences in these

statistical tests. First, it could be that the genogram construction process does not have

the efficacy to affect the therapeutic alliance or other client attitudes about therapy as

purported by many clinicians (e.g.. Beck. 1987; Braverman, 1997; Brock & Barnard,

1988; Erianger, 1991; Ingersoll-Dayton & Arndt. 1990; Ivey et al., 1996; Kramer, 1985;

Kuehl, 1995; Miehls, 1993; Wachtel. 1982).

Second, various methodological limitations may have confounded these results. As

noted above, one weighty limitation in this study was a small sample size per treatment

intervention group (TAGE+CT. n= 14; SAGE+CT, n = 16; DAI+CT. n = 2). Stevens

(1992) sample size table for a three-group MANOVA for a medium effect size, power at

.80, and alpha at .05, recommended 44 participants (22 couples) per group in order to be

able to detect significant differences in the linear combination of dependent variables

based on group effects. Given the small sample size per group in this investigation, there

may not have been sufficient power to detect significant group differences. It should be

noted again, however, that even if the recommended sample size had been obtained,

significant group differences may still not have been realized due to a potential true

absence of relation between genogram construction and therapeutic alliance and session

mpact, or other methodological limitations.

The design of the interventions may have minimized the ability to detect significant

statistical differences between the treatment groups in two ways. In the first way, the

interventions may not have been sufficiently different from each other. The distinction of

the interventions between the TAGE+CT and SAGE+CT groups was that the therapist

100

1

constructed the genogram in the TAGE+CT group, and the couple constructed the

genogram in the SAGE+CT group in the presence of a silent therapist. In the DAI+CT

group, the therapist collected the same data as in the TAGE+CT and SAGE+CT groups,

but genogram construction was not included. It may be possible that differences of either

the active involvement of the therapist in genogram construction, or the absence of the

graphical component of genogram construction, were not sufficiently distinct in a

therapeutic sense such that a measurable difference could have been detected between the

treatment groups.

A second way the design of the interventions may have affected the detection of

differences between treatment groups was that the interventions might not have been

strong enough to affect an assessable change in the dependent variables. The nature of

collecting genogram data, with or without the pictorial element, may not have stood out

for couples in early sessions as an experience that sufficiendy distinguished it from the

normal processes of therapy.

The minimal variation in the growth trajectories may have been affected by a

limitation of the measurement process. It is plausible that the dependent vanable

measures, the CTAS and SEQ, were not sensitive enough to detect changes over time in

client attitudes about the therapeutic alliance and session impact. Since the CTAS and

SEQ have not been employed in any other published longitudinal studies to date, it is

difficult to assess their sensitivity for measuring change across sessions.

A more likely and serious limitation contributing to the slight variation in the

growth curves is a matter of the data collection procedure. A visual examination of the

coded data revealed that 23 of 32 participants, on at least one. and up to all. of the

assessment packages they completed, either rated all or almost every question with the

same rating. The most extreme example of this was participant number 33, who, on the

CTAS over five sessions, rated 133 of 145 questions with a neutral response, 11

questions with "agree" (one point above neutral), and one question with "disagree" (one

point below neutral). On the SEQ over five sessions, this same participant rated 27 of 50

with a neutral response and the remaining questions with the response of "agree" (one

IOI

point above neutral). A more common example was participant number six. who. on the

CTAS over five sessions, rated questions from two sessions with minimal variability, and

the questions from three sessions with essentially no variability. On the SEQ over five

sessions, this same participant rated questions from four sessions with minimal

variability, and the questions from one session with the same rating.

There are several conceivable explanations why this occurred. First, the research

design dictated that the participants complete the assessments apart from the presence of

the therapist to guard against the Hawthorne effect. This left the couples together and

unmonitored as they filled out the questionnaires. Instances of identical ratings or similar

trends in ratings within some couple members on particular sets of assessments suggested

that couple members may have conferred with one another over questions, and/or copied

one another's responses.

Second, the participants may not have received adequate or repeated

encouragement from the therapist as to the importance of filling out the forms

independently and reflectively. If this were the case, it may have indicated either a lack

of motivation or poor communication on the part of the therapist, or that the therapists

were not properly trained or encouraged by the researcher.

Third, neither the participants nor the therapists received any form of tangible

remuneration for their part in this investigation. Without tangible incentive for their time

and effort, it is unreasonable to have expected most persons to earnestly give themselves

repeatedly and consistently to this project. The motivation of the therapist may have been

further reduced by the requirement that they implement a specified treatment intervention,

particularly if they believed the timing of the intervention or the intervention itself was

inappropriate to their therapeutic perspective or the needs of the couple.

CTAS Subscale and SEQ Index Intercorrelations

With regard to hypotheses Ola -Vc that investigated predicted positive

intercorrelations between the specific CTAS subscales and SEQ indexes within the three

treatment groups over time, only one correlation in one group significantly supported the

102

predicted association of the hypothesis. This occurred in the SAGE+CT group where the

CTAS tasks subscale was strongly as.sociated with the SEQ depth index. Evidently,

having the couple members construct a genogram by themselves and then discuss it with

their therapist created an initial and sustained trust in the therapist's methods and

techniques as appropriate to dealing with their difficulties. It is plausible that couples

likened the process of constructing a SAGE to other formal self-report assessments like

personality tests familiar to many people. Such measurement instruments often elicit

from the assessee a degree of credibility about the assessor. In the same way, the SAGE

construction and interpretation process may have established early in therapy the couple's

confidence in the therapist's as a person capable of understanding and assisting them with

their problems.

In their study examining the relation between the CTAS subscales and the SEQ

indexes from measurements taken after the third session in couple therapy, the only

positive correlation Heatherington and Friedlander (1990) found was between the CTAS

tasks subscale and the SEQ depth index. This is consistent with the findings of this

investigation, and challenges the proposal of others (i.e., Dryden & Hunt. 1985b; Johnson

& Greenberg, 1989) that the tasks domain does not take prominence until the middle

phase of therapy.

Construct Validity of the CTAS Content Subscales

If CTAS successfully tapped the tasks, goals, bonds subdimensions of therapeutic

alliance, then it was expected that they would be related to the depth and smoothness

subdimensions of the SEQ as hypothesized. The findings, as a whole, did not uphold

these expected relationships. Even if the correlations from the DAI+CT group were

dropped due to a small number of cases (n = 2), only one of six correlations in the other

two groups (TAGE+CT and SAGE+CT) was significantly positively associated as

predicted. One conclusion is a lack of construct validity for the CTAS content subscales.

However, given that Heatherington and Friedlander's (1990) test of the same predicted

associations between the CTAS subscales and the SEQ indexes produced only one

103

significant correlation, the theorized relationships between these measures are called into

question.

Split Alliance

A test of Pinsof and Catherall's (1986) concept of split alliance was conducted to

determine whether attitudes of any couples toward their therapist were notably different.

The empirical data provided some evidence of the theoretical description of a split

alliance. Only one couple in 17 (6%) was said to have experienced a split alliance with

their therapist when the differences between the couple members' mean slope scores were

combined into one group. This means one couple had disparate attitudes about their

relationships with their therapist. It is plausible that participation in a particular treatment

group contributed to the experience of a split alliance. If so. differences in the

percentages of couples experiencing a split alliance would have been expected between

groups. This was not, however, generally the case. One couple in each of the TAGE+CT

and SAGE+CT groups (14% and 13%, respectively) experienced a split alliance. No split

alliances were found in the DAI+CT group, but it only contained two couples. By

contrast, Heatherington and Friedlander's (1990) found 43% of the couples in their

sample experienced a split alliance. The high percentages of intact alliances found in this

study may suggest that something about the process of constructing genograms or

collecting genogram-like information helped create a common attitude within couples

about their individual alliance relationships with their therapist. On the other hand, the

high percentage of intact alliances may reflect the general lack of deviation in mean

CTAS slope scores.

Limitations

In addition to the limitations in this study that were previously discussed, there are

several others that must be taken into account as part of the review and interpretation of

the results of this research. The first of these—insufficient inclusion criteria—are ones

that may have influenced that nature of the participant sample. Any persons who

104

presented at the research site clinics requesting couple therapy, were at least 17 years of

age, and then judged by the therapist as appropriate for couple therapy treatment and

capable of participating in the study were eligible for inclusion in the investigation. One

missing criterion was no prerequisite for a length of co-habitation. Couples who have

been together under cenain periods of time may not have had sufficient time together to

establish themselves as a couple. In this study, three couples who were initially included

as sample constituents had been together nine months or less.

Furthermore, the therapist's decision for inclusion was based entirely on subjective

judgment and did not include any form of objective screening for pretreatment levels of

marital distress or satisfaction, recent or present histories of substance abuse or

dependency, or other psychiatric problems. Assessing pre-therapy levels of marital

distress or satisfaction may not be relevant, however. Pre-treatment levels of mantal

distress (Bourgeios et al., 1990) and marital satisfaction (Johnson & Talitman. 1997)

were shown not to be significantly associated with therapeutic alliance. By contrast, pre-

screening persons for particular recent or present psychiatric problems could be important

because couple therapy may not be the preferred treatment modality for them.

A second limitation regards a measurement issue. By measuring therapeutic

alliance and session impact for only the first five sessions, critical data about the nature of

change on these variables may have been missed. For example, the significance of

associations between the CTAS subscales and SEQ indexes may have altered if they had

been measured beyond five sessions. By extending the number of data collection points,

it may have been possible to identify "critical" periods of change between the dependent

variables in response to the treatment effects.

Future Research

The results of this study suggest three projects for future investigation, A

replication of this study is needed to determine if the small sample size in this study led to

an observed power level too low to detect significant differences between treatment

groups on the dependent variables. Conducting a study with the recommended sample

105

size would more adequately address the question of whether and how genogram

construction impacts the therapeutic alliance and specific aspects of the therapy itself. It

would also aid in assessing whether the therapist's involvement in genogram construction

(the TAGE versus the SAGE) makes any difference in the alliance process. Additionally,

it would be helpful to extend the data collection beyond five sessions. This would allow

for the detection of potentially complex pattems of change over time in the dependent

variables.

A question that was not pursued in this study that would be helpful to investigate is

whether differences exist in how men and women view genogram construction as

impacting their relationship with their therapist. Tracking the strength and pattern of the

CTAS subdimensions (bonds, tasks, and goals) by gender across sessions would help to

clarify the distinct ways genogram construction contributes to the development of the

therapist-client alliance for men and women.

A final question of interest to pursue is whether genogram construction decreases

the incidence of a split alliance between couples and the their therapists significandy

more than other methods believed or known to enhance therapeutic alliance. This could

identify a specific way in which genogram construction may be a mediator of therapeutic

alliance.

Clinical Impliactions

In light of the results of this research, couple therapists should be cautious about

employing genogram construction with the sole intent of either trying to build the

therapeutic alliance, or take "shallow" sessions "deeper" or smooth out difficult sessions.

Nevertheless, when the long and ingrained history of genogram construction in marriage

and family therapy is considered, it is likely that genograms will be continued to be used

by clinicians with the perspective that they are valuable in the development of the

therapist-client relationship and for impacting the nature of therapy sessions. It may,

therefore, benefit couple therapists who use genogram construction, to elicit attitudinal

feedback from their clients as to how this intervention effects the alliance and nature of

106

their therapy over time. Such discussions would encourage the processuai use of

genogram construction, which may. in turn, help configure the nature of the therapeutic

alliance and the therapy itself.

One area where genogram construction may have clinical profit as an intervention

tool is with couples who seem to have disparate attitudes about their individual

relationships with their therapist. In such instances, this disparity could lead to a split in

the alliance that could impede or halt further goal-oriented progress. For example, an

extremely negative attitude held by one couple member towards the therapist might be an

effort to avoid particular issues. Having couples complete a TAGE or SAGE could help

prevent such efforts and resultant alliance disruptions by helping to create united attitudes

of couple members toward their therapist.

The SAGE used in this study was modified from earlier SAGE versions developed

by Rogers and his collegues (Rogers & Cohn. 1987; Rogers et al., 1985) and Erianger

(1987) so that the instructions would be appropriate for a therapeutic environment and

easier for participants to understand and follow as they constructed their genograms.

With the improved format of the instructions, this SAGE version is even more suitable

for use in clinical settings and research. This SAGE will also extend the clinical

application of genograms beyond their traditional in-session usage by employing them as

part of a pre-therapy assessment package or an out-of-session homework assignment.

Summary

This study was conducted to examine whether and how constructing either a TAGE

or SAGE effected couple members' attitudes about the therapeutic alliance and specific

aspects of their therapy over the first five sessions. Given the preliminary predictive

validity of the therapeutic alliance in couple therapy, it is important to identify clinical

methods that can enhance this relationship. The overall findings of this study, however,

did not lend support to the conviction held by many couple therapists that genogram

construction is an effective way of building a therapeutic alliance or of influencing the

depth and smoothness of therapy sessions. Completing a SAGE in couple therapy may

107

contribute to couple members confidence in the therapists methods for addressing their

problems, but this finding needs to be replicated. An additional benefit of genogram

construction for couples may be that it helps to promote a shared attitude about their

individual alliance relationships with their therapist.

This investigation was the first of its kind in genogram research in that data was

collected from clients' perspectives in a therapeutic setting, and change over time in

therapeutic alliance and session impact were examined with a growth curve analysis.

Since understanding pattems of change taps the heart of the interactional process of

effective marriage and family therapy (Friedlander et al., 1994), this study was an

important step for future genogram construction research and clinical application.

108

REFERENCES

Ackerman, N. W. (1966). Treating the troubled family. .New York: Basic Books.

Alexander, J. F. (1988). Phases of family therapy process: A framework for clinicians and researchers. In L. C. Wynne (Ed.), The state of the art in family therapy research: Controversies and recommendations (pp. 175-185). New York: Family Process.

Alexander, J.. Barton, C, Schiavo. R. S., & Parsons. B. V. (1976). Systems-behavioral intervention with families of delinquents; therapist characteristics, family behavior and outcome. Joumal of Consulting and Clinical Psychology. 44. 656-664.

Barnard, C. P., & Corrales. R. G. (1979). The theory and technique of family therapy. Springfield, IL: Charies C. Thomas.

Beach, S. R. H., Sandeen. E. E., & O'Leary, K. O. (1990). Depression in marriage: A model for etiology and treatment. New York: Guilford.

Beck, D. F. (1972). Progress on family problems: A nationwide study of clients' and counselors' views on family agency services. New York: Family Service Association.

Beck, R. L. (1987). The genogram as process. American Journal of Family Therapy. 15. 343-351.

Bergin, A. E., & Lambert, M. J. (1978). The evaluation of therapeutic outcomes. In S. L. Garfield & A. E. Bergin (Eds.), Handbook of psychotherapy and behavior change: An empirical analysis (2nd ed.. pp. 139-189). New York: Wiley.

Bergin. A. E., & Suinn, R. M. (1975). Individual psychotherapy and behavior therapy. Annual Review of Psychology. 26. 509-556.

Beutler, L. E. (1979). Toward specific psychological therapies for specific conditions. Joumal of Consulting and Clinical Psychology. 47. 882-892.

Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory. Research and Practice. 16. 252-260.

Bornstein. P. H., & Bronstein, M. T. (1986). Marital therapy: A behavioral communications approach. New York: Pergamon.

109

Bourgeois, L.. Sabourin, S., & Wright, J. (1990). Predictive validity of the therapeutic alliance in group marital therapy. Journal of Consulting and Clinical Psychology. 58^608-613.

Bowen. M. (1966). The use of family theory in clinical practice. Comprehensive Psychiatry. 7. 345-374.

Bowen. M. (1978). Family theory in clinical practice. New York: Jason Aronson.

Bradt, J. O. (1980). The family diagram: Method, technique and use in family therapy. Washington. D. C: Groome Center.

Bradt, J. O.. & Moynihan, C. J. (1971). Opening the safe: A study of child-focused families. In J. O. Bradt & C. J. Moynihan (Eds.). Systems therapy. Selected papers: Theory, technique, research (PP. 1-24). Washington, D. C: J. O. Bradt and Carolyn J. Moynihan.

Braverman, S. (1997). The use of genograms in supervision. In T. C. Todd & C. L. Stanton (Eds.), The complete systemic supervisor: Context, philosophy, and pragmatics (pp. 349-362). Boston: Allyn and Bacon.

Brock, G. W., & Barnard, C. P. (1988). Procedures in family therapy. Boston: Allyn and Bacon.

Brock, G. W., & Barnard, C. P. (1992). Procedures in family therapy (2nd ed.). Boston: Allyn and Bacon.

Broderick, C. B. (1983). The therapeutic triangle: A sourcebook on marital therapy. Beveriey Hills, CA: Sage.

Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to assessing change. Psychological Bulletin. IOI, 147-158.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage.

Carpenter, J., & Treacher. A. (1989). Problems and solutions in marital and family therapy. New York: Basil Blackwell.

Carter, E. A., & Orfanidis, M. M. (1976). Family therapy with one person and the therapist's own family. In P. J. Guerin (Ed.), Family therapy: Theory and practice (pp. 193-219). New York: Gardner.

110

Cimmarusti. R. A., &Lappin. J. (1985). Beginning family therapy. In D. C. Breulin (Ed.). Stages: Pattems of change over time (pp. I6-25V Rockville, MD: Aspen.

Coady, N. F. (1992). Rationale and directions for an increased emphasis on the therapeutic relationship in family therapy. Contemporary Family Therapy, 14. 467-479.

Coleman. S. (1987). Milan in Bucks County. Family Therapy Networker. 26.405-413.

Condon. J. J. (1985). Therapy of the expectant family: The foetus as a force to be reckoned with. Australian and New Zealand Journal of Family Therapy. 6, 77-81.

Coupland, S. K., Serovich, J., & Glenn, J. E. (1995). Reliability in constructing genograms: A study among marriage and family therapy doctoral students. Joumal of Marital and Family Therapy. 21. 251-263.

Crosby, J. F. (1989). Museum tours in genogram construction: A technique for facilitative recall of negative affect. Contemporary Family Therapy. 11, 247-258.

Crouch, M. A., & Davis, T. (1987). Using the genogram (family tree) clinically. In M. A. Crouch & L. Roberts (Eds.), The family in medical practice: A family systems primer (PP. 174-192). New York: Springer-Veriag.

Danser, D. B. (1998). Recasting the presenting problem: An initial intervention in marital therapy. Marriage and Family: A Christian Joumal, I, 138-141.

Davis, L., Geikie, G.. & Schamess, G. (1988). The use of genograms in a group for latency age children. Intemational Journal of Group Psychotherapy, 38, 189-209.

de Chesnay, M., Marshall, E., & Clements. C. (1988). Family stmcture. marital power, maternal distance, and paternal alcohol consumption in father-daughter incest. Family Systems Medicine. 6, 453-462.

DeRubeis, R. J.. & Freely, M. (1991). Determinants of change in cognitive therapy for depression. Cognitive Therapy and Research. 14. 469-482.

Dobson, S. (1989). Genograms and ecomaps. Nursing Times. 85. 54-56.

Dryden, W., & Hunt, P. (1985a). Therapeutic alliances in marital therapy: I. Pre-therapy influences. In W. Dryden (Ed.), Marital therapy in Britain: Vol. I. Context and therapeutic approaches (pp. 121-143). London: Harper and Row.

Ill

Dryden, W., & Hunt, P, (1985b). Therapeutic alliances in marital therapy: II. Process issues. In W. Dryden (Ed.), Marital therapy in Britain: Vol. 1. Context and therapcucic approaches (pp. 144-168). London: Harper and Row.

Duncan, B. L. (1992). Strategic therapy, eclecticism, and the therapeutic relationship. Joumal of Marital and Family Therapy. 18. 17-24.

Duncan, B. L. (1993). Creating opponunities for rapid change in marital therapy. In \l. A. Wells & V. J. Giannciti (Eds.), Ca.<iebook of the brief psychotherapies (pp. 31S-327). New York: Plenum.

Erianger, M. (1987). An exploratory study of the genogram (Doctoral dissertation. University of Georgia, 1987). Dis.sertaiion Abstracts International. 48 (11) 2989

Erianger, M. (1991). Using the genogram with (he older client. Joumal of Mental Health Counseling. 12. 321-331.

Estrada, A. U., & Haney, P. (1998). Genograms in a multicultural perspective. Journal of Family Psvchotherapv. 9. 55-62.

Feldman, L.B. (1976). Goals of family therapy. Journal of Marital and Family Counseling. 2, 103-113.

Firestone, A.. & O'Connell, B- (1984). Does the therapeutic relationship matter? A follow-up study of adherence and improvement in family therapy. Australian Joumal of Family Therapy. 2. 17-24.

Fohs, M. W. (1991). Family systems assessment: Interventions with individuals having a chronic disability. Career Development Ouarteriv. 39. 304-311.

Francis, D, L, Fletcher, J. M., Stuebing, K. K., Davidson, K. C, & Thompson, N. M. (1991). Analysis of change; Modeling individual growth. Joumal of Consulting and Clinical Psychology. 59. 27-37.

Frank, A. F., & Gunderson, J. G. (1990). The role of the therapeutic alliance in the treatment of schizophrenia. Archives of General Psychiatry. 47. 228-236.

Frank, J. D. (1961). Persuasion and healing: A comparative study of psychotherapy. Baltimore, MD: Johns Hopkins.

Friedlander, M. L., Wildman, J., Heatherington, L. & Skowron, E. A. (1994). What we do and don't know about the process of family therapy. Joumal of Family Psychology. 8. 390-416.

112

Friedman. H.. Rohrbaugh, M., & Krakauer, S. (1988). The time-line genogram: Highlighting temporal aspects of family relationships. Family Process. 27. 293-303.

Friedman. R. (1991). Ten commandments for the family therapist. Clinical Supervisor, 9, 181-186.

Fuller. F., & Hill, C. E. (1985). Counselor and helpee perceptions of counselor intentions in relation to outcome in a single counseling session. Journal of Counseling P.svchology. 32. 329-338.

Gaston, L. (1990). The concept of the alliance and its role in psychotherapy: Theoretical and empirical considerations. Psychotherapy. 27. 143-153.

Gewirtzman, R. C. (1988). The genogram as a visual assessment of a family fugue. Australian Journal of Sex. Marriage and Family. 9. 37-46.

Giblin. P. (1996). Empathy: The essence of marriage and family therapy. Family Joumal. 4. 229-235.

Glass. G. v., Willson. V. L.. & Gottman, J. M. (1975). Design and analysis of time-series experiments. Boulder, CO: Colorado Associated University.

Goldstein, A. P., & Stein, N. (1976). Prescriptive psychotherapies. New York: Pergamon.

Gomes-Schwartz, B. (1978). Effective ingredients in psychotherapy: Prediction of outcome from process variables. Joumal of Consulting and Clinical Psychology. 46u 1023-1035.

Gomes-Schwartz, B., & Schwartz. J. M. (1978). Psychotherapy process variables distinguishing the "inherently helpful" person from the professional psychotherapist. Joumal of Consulting and Clinical Psychology, 46. 196-197.

Goolishan, H. A., & Anderson. H. K. (1990). Discussions: Engagement techniques in family therapy. Intemational Joumal of Family Therapy. 2. 95-96.

Green, R., & Herget, M. (1991). Outcomes of systemic/strategic team consultation: UI. The importance of therapist warmth and active stmcturing. Family Process. 30. 321-336.

Guerin, P. (1972). Study your own family: Part I. In A. Ferber. M. Mendelsohn, & A. Napier (Eds.), Book of family therapy (pp. 446-459). Boston: Houghton Miffin.

113

Guerin. P. J., & Pendagast. E. G. (1976). Evaluation of family system and genogram. In P. J. Guerin (Ed.), Family therapy: Theory and practice (pp. 450-464). New York: Gardner.

Gurman, A. S. (1982). Creating a therapeutic alliance m marital therapy. In A. S. Gurman (Ed.). Ouestions and answers in the practice of family therapy (pp. 15-19). New York: Bmnner/Mazel.

Gurman, A. S., & Kniskern, D. P. (1978). Research on marital and family therapy: Progress, perspective, and prospect. In S. L. Garfield & A. E. Bergin (Eds.), Handbook of psychotherapy and behavior change: An empirical analysis (2nd ed., pp. 817-901). New York: Wiley.

Gurman, A. S., Kniskern, D. P., & Pinsof, D. P. (1986). Research on the process and outcome of marital and family therapy. In S. L. Garfield & A. E. Bergin (Eds.), Handbook of psychotherapy and behavior change (3rd ed., pp. 565-624). New York: Wiley.

Haley, J. (1976). Problem-solving therapy. San Francisco, CA: Jossey-Bass.

Ham, M. (1989). Empathic understanding: A skill for "joining" with immigrant families. Joumal of Strategic and Systemic Therapies. 8. 36-40.

Hardy, K. V., & Laszloffy, T. A. (1992). Training racially sensitive family therapists: Context, content, and contact. Families in Society, 73, 364-370.

Hardy, K. v., & Laszloffy, T. A. (1995). The cultural genogram: Key to training culturally competent family therapists. Journal of Marital and Family Therapy. 21, 227-237.

Hartley, D. E., & Strupp, H. H. (1983). The therapeutic alliance: Its relationship to outcome in brief psychotherapy. In J. Masling (Ed.), Empirical studies of analytic concepts (Vol. l;pp. 1-37). Hillsdale, NJ: Lawrence Eribaum Associates.

Hartman, A. (1995). Diagrammatic assessment of family relationships. Families in Society. 76. 465-476.

Hartman, A., & Laird, J. (1983). Family-centered social work practice. New York: Free.

Haas-Cunningham, S. M. (1993). The genogram as a predictor of families at risk for physical illness (Doctoral dissertation, Syracuse University, 1993). Dissertation Abstracts Intemational. 54 (9) 4590.

114

Heathenngton. L.. & Friedlander, M. L. (1990). Couple and Family Therapy Alliance Scales: Empirical considerations. Journal of Marital and Family Therapy. 16. 299-306.

Henry, W. P., Stmpp, H. H., Schacht, T. E.. & Gaston, L. (1994). Psychodynamic approaches. In A. E. Bergin & S. L. Garfield (Eds. ), Handbook of psychotherapy and behavior change (4th ed.. pp. 467-508). New York: John Wiley and Sons.

Herth. K. A. (1989). The root of it all: Genograms as a nursing assessment tool. Joumal of Gerontological Nursing. 12. 32-37.

Hester. B. (1992). Training for couple therapy. In D. Hooper & W. Dryden (Eds.), Couple therapy: A handbook (pp. 294-3 inv Philadelphia. PA: Open University.

Hill. C. E.. Helms. J. E., Spiegel, S. B., & Tichenor, V. (1988a). Development of a system for categorizing client reactions to therapist interventions. Joumal of Counseling Psychology. 35. 27-36

Hill, C. E., Helms, J. E., Tichenor. V., Spiegel, S. B., O'Grady, K. E.. & Perry, E. S. (1988b). Effects of therapist response modes in brief psychotherapy. Joumal of Counseling Psychology. 35. 222-233.

Hof, L., & Berman, E. (1986). The sexual genogram. Journal of Marital and Family Therapy. 12. 39-47.

Holtzworth-Munroe, A., Jacobson, N. S., DeKlyen, M., & Whisman, M. A. (1989). Relationship between behavioral marital therapy outcome and process variables. Joumal of Consulting and Clinical Psychology, 57, 658-662.

Horvath. A. O. (1981). An exploratory study of the working alliance: Its measurement and relationship to outcome (Doctoral Dissertation, University of British Columbia, 1981). Dissenation Abstracts International. 42 (6) 2503.

Horvath, A. O. (1994). Research on the alliance. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research, and practice (pp. 259-286). New York: John Wiley and Sons.

Horvath, A. O., Gaston, L., & Luborsky, L. (1993). The therapeutic alliance and its measure. In N. E. Miller, L. Luborsky, J. P. Barber, & J. P. Docherty (Eds.), Psychodynamic treatment research: A handbook for clinical practice (pp. 247-273). New York: Basic Books.

115

Horvath, A. O., & Greenberg, L. S. (1994). Introduction. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research, and practice (pp. 1 -9). New York: John Wiley and Sons.

Horvath, A. O., & Luborsky, L. (1993). The role of the therapeutic alliance in psychotherapy. Joumal of Consulting and Clinical Psychology. 61. 561 -573.

Horvath, A. O., & Symonds. B. D. (1991). Relation between working alliance and outcome in psychotherapy: A meta-analysis. Journal of Counseling Psychology. 38*138-149.

Ingersoll-Dayton, B., & Arndt. B. (1990). Uses of the genogram with the elderiy and their families. Journal of Gerontological Social Work. 15. 105-120.

Issac, S., & Michael. W. B. (1990). Handbook in research and evaluation (2nd ed.). San Diego, CA: Edits.

Ivey, A. E., Ivey, M. B., & Simek-Morgan, L. (1996). Counseling and psychotherapy: A multicultural perspective (4th ed.). Boston: Allyn & Bacon.

Jacobson, N. S., & Margolin, G. (1979). Marital therapy: Strategies based on social leaming and behavior exchange principles. New York: Bmnner/Mazel.

Johnson, S. M. (1996). The practice of emotionally focused marital therapy: Creating connection. New York: Brunner/Mazel.

Johnson, S. M., & Greenberg, L. S. (1985). The differential effects of experimental and problem-solving interventions in resolving marital conflict. Journal of Consulting and Clinical Psychology. 53. 175-184.

Johnson, S. M., & Greenberg, L. S. (1989). The therapeutic alliance in marital therapy. Joumal of Cognitive Psychotherapy. 3. 97-110.

Johnson. S. M., & Talitman, E. (1997). Predictors of success in emotionally focused marital therapy. Joumal of Marital and Family Therapy. 23. 135-152.

Jolly, W., Froom, J., & Rosen, M. G. (1980). The genogram. Joumal of Family Practice. 10.251-255.

Kellner, R. (1975). Psychotherapy in psychosomatic disorders: A survey of controlled outcome studies. Archives of General Psychiatry. 35, 1021-1028.

Kleckner. T., Frank, L., Bland, C , Amendt, J. H., & Bryant, R. D. (1992). The myth of the unfeeling strategic therapist. Joumal of Marital and Family Therapy, 18,41-51

116

Kokotovic. A. M.. & Tracey, T. J. (1990). Working alliance in the early phase of counseling. Journal of Counseling Psvchologv. 37. 16-21.

Kramer. J. R. (1985). Family interfaces: Transgenerational patterns. New York: Bmnner/Mazel.

Kuehl. B. R. (1995). The solution-oriented genogram: A collaborative approach. Joumal of Marital and Family Therapy. 21, 239-250.

Lambert. M. J. (1986). Implications of psychotherapy outcome research for eclectic psychotherapy. In J. C. Norcross (Ed.), Handbook of eclectic psychotherapy (pp. 436-462). New York: Bmnner/Mazel.

Lambert. M. J., & Bergin. A. E. (1973). Psychotherapeutic outcomes and issues related to behavioral and humanistic approaches. Comell Journal of Social Relations. 8. 47-61.

Lambert. M. J., & Bergin, A. E. (1994). The effectiveness of psychotherapy. In A. E. Bergin & S. L. Garfield (Eds.), Handbook of psychotherapy and behavior change (4th ed., pp. 143-189). New York: John Wiley and Sons.

Lewis, K. G. (1989). The use of color-coded genograms in family therapy. Joumal of Marital and Family Therapy, 15. 169-176.

Lieberman. S. (1979a). Transgenerational analysis: The geneogram as a technique in family therapy. Joumal of Family Therapy. 1. 51 -64.

Lieberman, S. (1979b). Transgenerational family therapy. London: Groom Helm.

Luborsky, L. (1976). Helping alliances in psychotherapy: The groundwork for a study of their relationship to its outcome. In J. L. Claghom (Ed.), Successful psychotherapy (pp. 92-116). New York: Bmnner/Mazel.

Luborsky, L. (1984). Principles of psychoanalytic psychotherapy: A manual for supportive-expressive treatment. New York: Basic Books.

Luborsky, L. (1994). Therapeutic alliances as predictors of psychotherapy outcomes: Factors explaining the predictive success. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research, and practice (pp. 38-50). New York: John Wiley and Sons.

117

Luborsky, L., Cris-Christoph, P., Alexander. L.. Margolis, M.. & Cohen, M. (1983). Two helping alliance methods for predicting outcomes of psychotherapy: A counting signs vs. global rating method. Journal of Nervous and Mental Disease. 171,480-492. " '

Luborsky. L. McLellan. T., Woody, G., O'Brien. C. & Auerbach. A. (1985). Therapist success and its determinants. Archives of General Psychiatry. 42. 602-611.

Luborsky, L., Singer, B., & Luborsky, L. (1975). Comparative studies of psychotherapies: Is it tme that "Everyone has won and all must have prizes"? Archives of General Psychiatry. 32. 995-1008.

Mackey, S.K. (1985). Stages of family therapy with severely disturbed adolescents. In D. C. Breulin (Ed.), Stages: Patterns of change over time (pp. 75-87). Rockville, MD: Aspen.

Mallinckrodt, B. (1993). Session impact, working alliance, and treatment outcome of brief counseling. Joumal of Counseling Psychology. 40. 25-32.

Margolin, G. (1986). Ethical issues in marital therapy. In N. S. Jacobson & A. S. Gurman (Eds.), Clinical handbook of marital therapy (pp. 621-638). New York: Guilford.

Mannar, C. R.. Gaston. L., Thompson, L. W., & Gallagher, D. (1989). Alliance and outcome in late-life depression. Joumal of Nervous and Mental Disease. 177. 464-472.

Marmar, C. R., Horowitz, M. J., Weiss, D. S., & Marziali, E. (1986). The development of the therapeutic alliance rating system. In L. S. Greenberg & W. M. Pinsof (Eds.), The therapeutic process: A research handbook (pp. 367-390). New York: Guilford.

Marziali. E. (1984). Prediction of outcome of brief psychotherapy from therapist interpretive interventions. Archives of General Psychiatry, 41, 301 -305.

Marziali, E., Marmar, C, & Kmpnick, J. (1981). Therapeutic Alliance Scales: Development and relationship to psychotherapy outcome. American Joumal of Psychiatry, 138. 361 -364.

Mashal. M.. Feldman, R. B.. & Sigal, J. J. (1989). The unraveling of a treatment paradigm: A follow-up study of the Milan approach to family therapy. Family Process. 28. 457-470.

118

McDaniel. S.. Campbell. T., & Seaburn, D. (1990). Familv-oriented primary care: A manual for medical provider ; New York: Springer-Veriag.

McGilI, D. W. (1992). The cultural story in multicultural family therapy. Families in Society. 73. 339-349.

McGoldrick. M.. & Gerson, R. (1985). Genograms in family assessment. New York: Norton.

McMillen, J. C, & Groze, V. (1994). Using placement genograms in child welfare practices. Child Welfare. 73. 307-318

Meltzoff, J., & Kornreich, M. (1970). Research in psvchotherapv. New York: Atherton.

Miehls. D. (1992). Conjoint treatment with narcissistic couples: Strategies to increase empathic interaction. Smith College Studies in Social Work. 64. 3-17.

Miller. S.. Hubble, M., & Duncan, B. (1995). No more bells and whistles. Family Therapy Networker. 19. 52-63.

Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University.

Morgan, R., Luborsky, L., Cris-Cristoph, P., Curtis, H.. & Solomon, J. (1982). Predicting the outcomes of psychotherapy by the Penn Helping Alliance Rating Method. Archives of General Psychiatry. 39. 397-402.

Morrissette. P. J. (1990). Drawing the curtain on family therapy. Family Therapy. I. 67-73.

Mullins. H. C. & Christie-Seely, J. (1984). Collecting and recording family data-The genogram. In J. Christie-Seely (Ed.), Working with the family in primary care: A systems approach to health and illness (pp. 179-191). New York: Praeger.

Nichols, W. C. (1988). Marital therapy: An integrated approach. New York: Guilford.

Okiishi, R. W. (1987). The genogram as a tool in career counseling. Joumal of Counseling and Development. 66. 139-143.

O'Malley, S. S., Suh, C. S., & Strupp, H. H. (1983). The Vanderbilt Psychotherapy Process Scale: A report on the scale development and a process-outcome study. Joumal of Consulting and Clinical Psychology. 51. 581 -586.

119

Orfanidis, M. M. (1979). Problems with family genograms. American Journal of Family Therapy. 7. 74-76.

Oriinsky, D. E., Grawe, K., & Parks, B. K. (1994). Process and outcome in psychotherapy. In A. Bergin & S. Garfield (Eds.), Handbook of psychotherapy and behavior change (4th ed., pp. 270-376). New York: John Wiley and Sons.

Oriinsky, D. E., & Howard, K. I. (1986). Process and outcome in psychotherapy. In S. L. Garfield & A. E. Bergin (Eds. ), Handbook of psychotherapy and behavior char|ge.(3rded., pp. 311-381). New York: John Wiley and Sons.

Paul. N. L., & Paul. B. B. (1982). Death and changes in sexual behavior. In F. Walsh (Ed.), Normal family processes (pp. 229-250). New York: Guilford.

Pendagast, E. G., & Sherman. C. O. (1977). A guide to the genogram. The Family. 5. 101-112.

Perry, R. (1993). Empathy—still at the heart of therapy: The interplay of context and empathy. Australian and New Zealand Journal of Family Therapy, 14. 63-74.

Pinsof, W. M. (1983). Integrative problem-centered therapy: Toward the synthesis of family and individual psychotherapies. Journal of Marital and Family Therapy. 9. 19-35.

Pinsof, W. M. (1988a). Strategies for the study of family therapy process. In L. C. Wynne (Ed.), The state of the art in family therapy research: Controversies and recommendations (pp. 159-174). New York: Family Process.

Pinsof. W. M. (1988b). The therapist-client relationship: An integrative systems perspective. Joumal of Integrative and Eclectic Psychology. 7, 303-313.

Pinsof. W. M. (1989). A conceptual framework and methodological criteria for family therapy research. Joumal of Consulting and Clinical Psychology, 57. 53-59.

Pinsof, W. M. (1994a). An integrative systems perspective on the therapeutic alliance: Theoretical, clinical, and research implications. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research, and practice (pp. 173-194). New York: John Wiley and Sons.

Pinsof, W. M. (1994b). An overview of Integrative Problem Centered Therapy: A synthesis of family and individual psychotherapies. Joumal of Family Therapy, 16. 103-120.

120

Pinsof. W. M., & Catherall. D. R. (1986). The integrative psychotherapy alliance: Family, couple, and individual therapy scales. Joumal of Marital and Family Therapy. 12. 137-151.

Piper, W. E.. Azim. H. F. A.. Joyce. A. S.. & McCallum. M. (1991a). Transference interpretations, therapeutic alliance and outcome in short-term individual psychotherapy. Archives of General Psychiatry. 48, 946-953.

Piper. W. E., Azim, H. F. A., Joyce. A. S.. McCallum. M., Nixon, G. W H., & Segal. P. S. (1991b). Quality of object relations vs. interpersonal functioning as predictor of therapeutic alliance and psychotherapy outcome. Journal of Nervous and Mental Disease. 179.432-438.

Plotnicov. K. H. (1990). Early termination from counseling: The client's perspective (Doctoral dissertation. University of Pittsburgh, 1990). Dissertation Abstracts Intemational. 51 (1) 5038.

Quinn. W. H.. Dotson. D., & Jordan. K. (1997). Dimensions of therapeutic alliance and their associations with outcome in family therapy. Psychotherapy Research. 7. 429-438.

Regan, A. M.. & Hill.C. E. (1992). Investigations of what clients and counselors do not say in brief therapy. Journal of Coun.seling Psychology. 39. 168-174.

Roback, H. B. (1971). The comparative influence of insight and noninsight psychotherapies on therapeutic outcome: A review of the experimental literature. Psychotherapy: Theory. Research and Practice. 8, 23-25.

Rogers. C.R. (1951). Client centered therapy. Cambridge, MA: Riverside.

Rogers. C. R. (1957). The necessary and sufficient conditions of therapeutic personality change Joumal of Consulting and Clinical Psychology. 22. 95-103.

Rogers, J., & Durkin, M. (1984). The semi-stmctured genogram interview: I. Protocal, n. Evaluation. Family Systems Medicine. 2. 176-187.

Rogers, J., Durkin, M., & Kelly, K. (1985). The family genogram: An undemtilized clinical tool. Joumal of the Medical Society of New Jersey, 82. 887-892.

Rogers, J.. & Holloway, R. (1990). Completion rate and reliability of the self-administered genogram (SAGE). Family Practice. 7, 149-151.

Rogers, J. C , & Cohn. P. (1987). Impact of a screening family genogram on first encounters in primary care. Family Practice. 4. 291 -301.

121

Rogers, J. C, & Rohrbaugh, M. (1991). The SAGE-PAGE trial: Do family genograms make a difference? Journal of the American Board of Family Practice. 4. 319-326.

Rogers, J. C, Rohrbaugh, M.. & McGoldrick. M. (1992). Can experts predict health risk from family genograms? Family Medicine. 24. 209-215.

Rohrbaugh, M.. Rogers, J. C, & McGoldrick, M. (1992). How do experts read family genograms. Family Systems Medicine. 10. 79-89.

Rogosa. D. R.. Brandt, D., & Zimowski. M. (1982). A growth curve approach to the measurement of change. Psychological Bulletin, 90. 726-748.

Rutan. J. S.. & Smith. J. S. (1985). Building therapeutic relationships with couples. Psychotherapy. 22. 194-200.

Saltzman, C, Leutgert, M. J., Roth. C. H., Greaser. J.. & Howard. L. (1976). Formation of a therapeutic relationship: Experiences during the initial phase of psychotherapy as predictors of treatment duration and outcome. Journal of Consulting and Clinical Psychology. 44. 546-555.

Saunders, S. M., Howard, K. I.. & Oriinsky, D. E. (1989). The therapeutic bond scales: Psychometric characteristics and relationship to treatment effectiveness. Psychological Assessment, 1. 323-330.

Schames, G. (1990). New directions in children's group therapy: Integrating family and group perspectives in the treatment of at risk children and families. Social Work with Groups. 13. 67-92.

Schilling, S. M.. & Gross, E. (1979). Stages of family therapy: A developmental model. Clinical Social Work Journal. 7. 105-1 12.

Schroder. T. (1992). Approaches to couple therapy. In D. Hooper & W. Dryden (Eds.). Couple therapy: A handbook (pp. 59-89). Philadelphia, PA: Open University.

Schulman, G. L. (1993). The therapeutic relationship in family therapy: A comment on Coady. Contemporary Family Therapy. 15. 341-342.

Segraves, R. T. (1982). Marital therapy: A combined psychodvnamic-behavioral approach. New York: Plenum.

Serovich, J. M., Glenn. E. J.. & Coupland, S.K. (1993, January). CAUTION!! Genogram Under Constmction. Paper presented at the meeting of the Texas Association for Marriage and Family Therapy, Houston, TX.

122

Shapiro, R. J. (1974). Therapist attitudes and premature termination in family and individual therapy. Journal of Nervous and Mental Disease, 159. 101-107.

Shapiro, R. J., & Budman, S. H. (1973). Defection, termination, and continuation in family and individual therapy. Family Process, 12, 55-67.

Shellenberger. S.. Couch. K. W., &. Drake. M. A. (1989). Elderiy family members and their caregivers: Characteristics and development of the relationship. Family Systems Medicine. 7. 317-322.

Sherman, R. (1993). Marital issues of intimacy and techniques for change: An Alderian systems perspective. Individual Psychology. 49. 318-329.

Sherman, R.. & Fredman. N. (1986). Handbook of stmctured techniques in marriage and family therapy. New York: Bmnner/Mazel.

Sherman. V. L. K. (1991). Women's experience and perceptions of the salient elements of healthy functioning families (Doctoral dissertation. Texas Woman's University, 1991). Dissertation Abstracts Intemational. 53 (2). 642.

Shields, C. G., Sprenkle. D. H., & Constantine. J. A. (1991). Anatomy of an initial interview: The importance of joining and stmcturing skills. American Joumal of Family Therapy. 19.3-18.

Shore, W. B., Wilkie, H. A., & Croughan-Minihane, M. (1994). Family of origin genograms: Evaluation of a teaching program for medical students. Family Medicine. 26. 238-243.

Sluzki, C. (1975). The coalitionary process in initiating family therapy. Family Process. H , 67-77.

Smith, M. L., & Glass, G. V. (1977). Meta-analysis of psychotherapy outcome studies. American Psychologist. 32. 752-760.

Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The benefits of psychotherapy. Baltimore, MD; Johns Hopkins University.

Smoyak, S. (1982). Family systems: Use of genograms as an assessment tool. In I. W. Clements & D. M. Buchanan (Eds.), Family therapy: A nursing perspective (pp. 245-250). New York: John Wiley and Sons.

Solomon, M. A. (1977). The staging of family treatment: An approach to developing the therapeutic alliance. Joumal of Marriage and Family Counseling. 3. 59-66.

123

Spanier, G. (1976). Measuring dyadic adjustment. Journal of Marriage and the Family. 1 1 113-126.

Splete. H.. & Freeman-George, A. (1985). Family influences on the career development of young adults. Journal of Career Development. 12. 55-64.

Stanton, M.D. (1992). The Time Line and the "Why now?" question: A technique and rationale for therapy, training, organization consultation and research. Journal of Marital and Family Theraov. 18. 331 -343.

Stanton, M. D., & Todd, T. C. (1982). The family therapy of dmg abu.se and addiction. New York: Guilford.

Stevens. J. (1992). Applied multivariate statistics for the social sciences (2nd ed.). Hillsdale, NJ: Lawrence Eribaum.

Stiles. W. B. (1980). Measurement of the impact of psychotherapy sessions. Joumal of Consulting and Clinical Psvchologv. 48, 176-185.

Stiles, W. B.. Shapiro, D. A., & Firth-Cozens. J. A. (1990). Correlations of session evaluations with treatment outcome. British Journal of Clinical Psychology. 29. 13-21.

Stiles, W. B., & Snow, J. S. (1984a). Counseling session impact as viewed by novice counselors and their clients. Journal of Counseling Psychology. 31. 3-12.

Stiles, W. B.. & Snow, J. S. (1984b). Dimensions of psychotherapy session impact across sessions and across clients. British Journal of Clinical Psychology. 23. 59-63.

Stiles. W. B.. Reynolds. S.. Hardy, G. E., Rees. A.. Barkham. M., & Shapiro, D. A. (1994). Evaluation and description of psychotherapy sessions by clients using the Session Evaluation Questionnaire and the Session Impacts Scale. Journal of Counseling Psychology. 41, 175-185.

Street, E. (1994). Counselling for family problems. London: Sage.

Swift, W. J. (1993). Brief psychotherapy with adolescents: Individual and family approaches. American Journal of Psychotherapy. 47, 373-386.

Tabachnick. B. G.. & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: HarperCollins.

124

Talan. R. L. (1983). A case study relating mid-life transition to level of differentiation from mother in re-entry women (Doctoral dissertation. University of Pittsburgh, 1983). Dissertation Ab.stracts International. 45 (6) 1646.

Thomas, E. M. (1994). Values, therapeutic alliance, and marital therapy: A .study of the therapist-client relationship in marital therapy (Doctoral dissertation, Texas Tech University, 1994). Dissertation Abstracts International, 55 (5) 1826.

Vukov. M. G., & Eljdupovic, G. (1991). The Yugoslavian dmg addict's family stmcture. Intemational Journal of the Addictions, 26. 415-422.

Wachtel. E. F. (1982). The family psyche over three generations: The genogram revisited. Journal of Marital and Family Therapy. 8. 335-343.

Waxenberg, B. R. (1973). Therapist's empathy, regard, and genuineness as factors in staying in or dropping out of short-term, time-limited family therapy (Doctoral dissertation. New York University, 1973). Dis.sertation Abstracts International. 34 (3) 1288.

Weeks. G. R.. & Treat. S. (1992). Couples in treatment: Techniques and approaches for effective practice. New York: Bmnner/Mazel.

Wells. R. A. (1990). Engagement techniques in family therapy. International Joumal of Family Therapy. 2. 75-94.

Whisman. M. A.. & Snyder, D. K. (1997). Evaluating and improving the efficacy of conjoint couple therapy. In W. K. Halford & H. J. Markman (Eds.), Clinical handbook of marriage and couples interventions (pp. 679-693). New York: John Wiley and Sons.

Wilkinson, M. (1992). How do we understand empathy systemically? Journal of Family Therapy. 14.45-52.

Willett, J. B. (1988). Questions and answers in the measurement of change. In E. Z. Rothkopf (Ed.), Review of research in education (Vol. 15, pp. 345-422). Washington, DC: American Educational Research Association.

Willett, J. B. (1989). Some results on reliability for the longitudinal measurement of change: Implications for the design of studies of individual growth. Educational and Psychological Measurement, 49, 587-602.

125

Willett, J. B.. Ayoub, C. C, 8L Robinson, D. (1991). Using growth modeling to examine systematic differences in growth: An example of change in the functioning of families at risk of maladaptive parenting, child abuse, or neglect. Journal of Consulting and Clinical Psychology. 59. 36-47.

Willett, J. B., & Sayer, A. G. (1994). Using covariance stmcture analysis to detect correlates and predictors of individual change over time. Psychological Bulletin. 116.363-381.

Woody, J. D. (1990). Clinical strategies to promote compliance. American Journal of Family Therapy. 18. 285-294.

Woolf, V. V. (1983). Family network systems in transgenerational psychotherapy: The theory, advantages and expanded applications of the genogram. Family Therapy. 10,219-237.

Worden, M. (1994). Family therapy basics. Pacific Grove, CA: Brooks/Cole.

Wynne, L. C. (1988). An overview of the state of the art: What should be expected in current family therapy research. In L. C. Wynne (Ed.), The state of the art in family therapy research: Controversies and recommendations (pp. 249-266). New York: Family Process.

126

APPENDIX A

CONSENT FORM FOR PARTICIPANTS

127

Consent Form Code #

I hereby give my consent for my participation in the project entitled: "Effects of couples' perceptions of genogram constmction on therapeutic alliance and .session impact: A growth curve analysis." I understand that the persons responsible for this project are Dr. David Ivey (806) 742-2775 and Scott Coupland (806) 742-3000. One or both have explained that this study is a part of a project that has the following objectives:

I. To determine if and how constructing a genogram will affect couple's perceptions of the therapist-client relationship and their therapy.

The researchers have explained that I may take part in the constmction of a genogram as a normal part of therapy, and that I will be asked to respond to a questionnaire following the first five therapy sessions. I understand that how 1 respond to questions conceming my relationship with my therapist and my therapy are of experimental interest to the researchers.

I further understand that dealing with sensitive matters in therapy or as a part of genogram constmction often involves discomfort. Dealing with sensitive material may produce negative emotions such as anger, fear, or sadness. I also understand that if I constmct a genogram I will have a chance to process the experience with my therapist following completion of the genogram drawing, and that at the conclusion of my participation in the study I will receive a debriefing letter from one of the researchers.

The benefits that can come to me from participating in this project include having the opportunity to address your present problems with a therapist, and contributing to the improvement of couple therapy as a professional activity.

I also understand that I am in no way required to participate in the project and can terminate my participation at any point with no repercussions.

It has further been explained to me that I will be required to give about five (5) to ten (10) minutes of my time at the end of the first five therapy sessions for a total of 25 to 50 minutes; that only the researchers involved with this study will have access to the records and/or data collected; and that all data associated with this study will remain strictly confidential.

Dr. Ivey or Scott Coupland has agreed to answer any inquires I may have conceming the procedures and has informed me that I may contact the Texas Tech University Institutional Review Board for the Protection of Human Subjects by writing them in care of the Office of Research Services, Texas Tech University, Lubbock, Texas 79409, or by calling (316) 742-3884.

128

If this research project causes any physical injury to participants in this project, treatment is not necessarily available at the Texas Tech University or the Student Health Center, nor is there necessarily any insurance carried by the University or its personnel applicable to cover any such injury. Financial compensation for any such injury must be provided through the participant's own insurance program. Further information about these matters may be obtained from Dr. Robert M. Sweazy, Vice Provost for Research. (316) 742-3884, Room 203 Holden Hall, Texas Tech University, Lubbock, Texas 79409-1035.

I understand that I may not derive therapeutic treatment from participation in this study. I understand that I may discontinue this study at any time I choose without penalty.

Signature of Subject Date:

Signature of Project Director or Authorized Representative Date:

Signature of Witness to Oral Presentation Date:

129

APPENDIX B

DEBRIEFING FORM FOR PARTICIPANTS

130

We would like to thank you for participating in this study, "Effects of couples' perceptions of genogram constmction on therapeutic alliance and .session impact: A growth curve analysis." Your participation is valuable to us and to the processes of couple therapy and couple therapy research. We would like to take a moment to tell you a little bit about this study.

Therapists u.se many methods for getting to know their clients and for understanding the stmggles that bring clients to therapy. One popular method is the constmction of genograms or "family trees." As a participant in this project, you may have been involved in constmcting your genogram. The researchers in this study are trying to gain a better understanding of how constmcting a genogram effects what couples think about their relationships with their therapist and their therapy experiences. Your participation in therapy and completion of the questionnaires will help us to examine this.

We predicted that constmcting a genogram would enhance the relationship between your therapist and you and your spouse or partner, and positively impact your therapy experience. Your participation in this study will help us to learn if and how our predictions were on target. We hope that as a result of this study, we will be able to use genograms more effectively with couples in therapy.

Again, thank you for your time. If in the process of your participation in this project you did not constmct a genogram. please contact your therapist about the possibility of doing so. The preliminary results should be available in the spring of 1996. If you would like to know the results, please feel free to contact Scott Coupland by writing him at Texas Tech University, Department of Human Development and Family Studies, Box 1162, Lubbock, Texas 79409-1162. or by calling (316) 742-3000. We appreciate your input.

131

APPENDIX C

SAMPLE NOTICE FOR THE SOLICITATION

OF THERAPISTS

132

Scott Coupland, M.S.

1234 Main Street Lubbock. Texas 01234 (806)555-1212

Month XX. 199X

Dear Doctoral Student in Marriage and Family Therapy,

I am conducting a dissertation research project entitled. "Effects of couples' perceptions of genogram constmction on therapeutic alliance and session impact: A growth curve analysis." My interest is in understanding how constmcting genograms effects what couple's think about their relationships with their therapist and their therapy experiences. This study is particulariy directed to students in AAMFT accredited Marriage and Family Therapy programs. I am. therefore, writing to enlist your help for our project.

I seek your participation in this project primarily as a therapist. That is. your major role will be to conduct therapy with couples as you normally do. Below is a brief explanation of the procedures that you would be administrating as a therapist in this study. Your participation will involve initially attending a two hour orientation conducted by me. The content of this orientation will cover (a) a general presentation of the study, (b) methods for soliciting client participation in the study, (c) procedures for administering the self-administered genogram and dependent measures, and (d) a training and review session on TAGE constmction. This orientation will thoroughly familiarize you with the study procedures so that you are prepared to carry out your role and answer any questions the couple might have about the investigation. I will arrange two dates to offer this orientation. For those students unable to attend either scheduled orientation, individual meetings will be arranged. Once the study begins, you will accept couple intakes as your schedule permits and according to your clinics procedures. At the end of the first session, you will solicit couples for participation in the study. This will involve a brief presentation of the study, answering questions, and having them sign consent forms. Following the first five sessions, you will give each couple member a questionnaire which they will complete on their own, and you will fill out a questionnaire which will take you about one minute to complete. If the couple is assigned to one of the two genogram groups, then during the second session you will either give the couple a self-administered genogram to complete on their own, or you will constmct a therapist-administered genogram with the couple. If the couple is assigned to the control group, you will conduct therapy as usual, but without employing a genogram or any other schematic drawing. You will be free to use a theoretical orientation of your choice throughout the therapy.

133

At this time, investigators cannot pay you for your participation. You will, however, be provided with a summary of the results following the data collection. Your participation in this project is very important to the field of Marriage and Family Therapy and is appreciated by the researcher.

I will be contacting you by telephone shortly after you receive this notice to answer any questions you may have conceming the study and hopefully to solicit your participation. Further information will be sent to Dr. X , the clinic director at your university, as the orientation dates draws near and he will keep you up to date on the times and locations. Remember, if these times or dates do not fit into your schedule, an individual orientation can be arranged.

Again, I thank you in advance for your interest and time and look forward to seeing you in June.

Sincerely,

Scott Coupland. M.S.

134

APPENDIX D

CONSENT FORM FOR THERAPISTS

135

Consent Form for Therapists Code #

I hereby give my consent for my participation in the project entitled: "Effects of couples' perceptions of genogram constmction on therapeutic alliance and session impact: A growth curve analysis." I understand that the persons responsible for this project are Dr. David Ivey (806) 742-2775 and Scott Coupland (806) 742-3000. One or both have explained that this study is a part of a project that has the following objectives:

1. To determine if and how constmcting a genogram will affect couple's perceptions of the therapist-client relationship and their therapy.

The researchers have explained that I will conduct couple therapy, may facilitate the constmction of a genogram as a normal part of therapy, or may clarify genogram constmction instmctions for couples. I also understand that I will be asked to respond to a demographic questionnaire following the first five therapy sessions. I understand that my conducting therapy, facilitating the constmction of a genogram, and how I respond to questions about myself is of experimental interest to the researchers.

I further understand that there are no significant risks in this project, but that I will be given a debriefing form and a summary of the study at the end of the data collection.

The benefits that can come to me from participating in this project include gaining a better understanding of the genogram, and contributing to the improvement of couple therapy as a professional activity.

I also understand that I am in no way required to participate in the project and can terminate my participation at any point with no repercussions.

It has further been explained to me that I will be required to give about one (I) minute of my time at the end of the first five therapy sessions for a total of 5 minutes; that only the researchers involved with this study will have access to the records and/or data collected; and that all data associated with this study will remain strictly confidential.

Dr. Ivey or Scott Coupland has agreed to answer any inquires I may have conceming the procedures and has informed me that I may contact the Texas Tech University Institutional Review Board for the Protection of Human Subjects by writing them in care of the Office of Research Services, Texas Tech University, Lubbock, Texas 79409. or by calling (316) 742-3884.

If this research project causes any physical injury to participants in this project, treatment is not necessarily available at the Texas Tech University or the Student Health Center, nor is there necessarily any insurance carried by the University or its personnel applicable to cover any such injury. Financial compensation for any such injury must be provided

136

through the participant's own insurance program. Further information about these matters may be obtained from Dr. Robert M. Sweazy, Vice Provost for Research. (316) 742-3884, Room 203 Holden Hall. Texas Tech University, Lubbock, Texas 79409-1035.

I understand that I may not derive therapeutic treatment from participation in this study. I understand that I may discontinue this study at any time 1 choose without penalty.

Signature of Subject ^ Date:

Signature of Project Director or Authorized Representative Date:

Signature of Witness to Oral Presentation Date:

137

APPENDIX E

SAGE

138

Your Family Genogram* Code #.

Today's Date

A genogram is a family tree, or "map" of your family over several generations, h includes information about births, deaths, divorces, relationships, occupations, events in the family life cycle and much more. By making your family genogram you will look at your own complex family system as it has developed over generations, and the cultural values, "mies" of behavior and general pattems of living that have been transmitted to you, and which you pass on to the generations which follow you.

By drawing your family genogram you will help your therapist and yourself to understand more about you, your family, and the difficulties that brought you to therapy. Many people have reported that creating their family genogram is a highly positive experience. If for any reason drawing your family genogram causes you distress, please stop and discuss this with your therapist.

Please remember that there is no "right" or "wrong" way to make a genogram-each one is unique to the individual and his or her family system. If you have a large family, you may feel cramped for space—don't worry if your genogram is not neat!

You may not be able to give a complete family history "off the top of your head," but please take time to fill out the form as fully and accurately as you can. On the following pages are step-by-step instmctions designed to help you fill in the accompanying chart entitled, "Your Family Genogram." If you have any questions about the instmctions, your therapist will explain them to you. It will take you about 45 to 75 minutes to complete your family genogram.

Most people who complete a family genogram want to keep a copy for their own family use. Your therapist will make of copy of your genogram so it can become part of your permanent family records.

•Adapted from materials developed by John C. Rogers, M.D. and Mary A. Erianger, Ph.D.

139

Family Genogram Instructions

Stepl

Take a moment to look at the chart entitled, "Your Family Genogram," accompanying this set of instmctions. This is the framework of your family genogram and when it is completed, it should contain much of the information you know and remember about your family. It may appear complicated, but the instmctions that follow will help you to make sense of it. The figure below explains the symbols on the genogram.

Squares represent males Circles represent females

This male is the child of the man and woman above

The lines that connect the man and woman above indicate that they are married or living together

Take another look at the family genogram chart and note that it provides symbols or space for four generations: your grandparents, your parents, you and your spouse or partner, and your children (if you have any). In addition, beside each symbol there is a space for you to list the name, date of birth, level of education, and other basic information of each person, but please do not fill in this infoimation until instiucted to do so. Below are explanations for each category of information.

**Naine" - place the first and last name of the person. fiirth Date" - list the month, day, and year of the person's birth. 'Ethnicity" - indicate the person's ethnic background. For example, African-American,

Asian, Canadian, German, Hispanic, Italian, Native-American. 'Educatioif' - list the person's highest level of education completed. For example, grade

school, high school, some college, associate's degree, bachelor's degree, master's degree, doctorate.

*<Occiipatioii" - indicate the person's present job. For example, high school teacher, truck driver, cashier, nurse, homemaker, mechanic. If the person is unemployed, then write "unemployed" in this space. If the person is retired, then write "retired" after the job.

'Death Date" - if the person has died, list the month, day, and year of his or her death.

140

'*Medical Problems" - note any significant medical problems, for example, diabetes, cancer, stroke, heart disease, physical disability, or dmg and/or alcohol problems.

' Emotional Stresses" - note any significant sources of emotional stresses, for example, depression, physical and/or sexual abuse, "nervous breakdown," or schizophrenia.

'*Legal Problems" - indicate any significant problems with the law, for example, committed felony, time in jail or prison, or driving while intoxicated (DWI).

'"Characteristics" - include a nickname, such as "loner," "genius," or "family clown,"' or other words that describe the person. It may help to think of adjectives that you or others have used to describe this person, for example, "hard-discipliner," "quiet and distant," "compassionate," or "sarcastic."

'"Location" - write in the name of the city and state (and country if outside the USA) where the person presently lives, for example, Denver, Colorado; Dallas, Texas; London, England.

"Present Age" - write each person's present age (e.g., "22," "48") in the center of the square or circle symbol. See the example below.

Right below this list, notice there are spaces for you to list dates of marriage ('*M"), separaUon ("S"), and divorce ("D"). If the couple is or was not married, but is or was living together ("LP*), then write the amount of time they have been or were living together in the space.

Here is an example of what the genogram will look like when all the above information is filled in for one couple.

Present Age written inside symbol 0 George Mueth - Name - Anna Mueth

8/5/26-Birth Date-7/16/28 German - Ethnicity - German

High School - Education - B. A. Constmction contractor - Occupation - Teacher

1/3/90-Death Date-Lung cancer - Medical Problems - Alzheimer's disease

Emotionally unavailable - Characteristics - Kind but shallow Physically abused by father - Emotional Problems -

none - Legal Problems - none - Location - Austin, Texas

M. 6/23/48 S. 10/6/77 D. 2/20/79

Using the steps that follow, fill in the facts as clo.sely as you remember them. If names, dates, or other pieces of information do not come easily to mind, simply put in your best guess or move on to those you know.

141

step 2

The instmctions in this step are for your side of the family.

1. To begin, find yourself~"YOU"-on the family genogram. Draw an extra square or circle around the symbol that represents you, as in the examples below.

YOU - male YOU - female

2. Then, fill in the spaces nearest this symbol with information about yourself. For example, list your name, birth date, education, and so forth.

3. Next, fill in the information for your mother and father and your grandparents.

4. Finally, put an "X" in the circle or square if death has occurred (see example below).

142

Step 3

The instmcfions in this step are for your side of the family.

1. Under the line connecting your parents, draw in symbols (squares and circles) for your brothers and si.sters in the oiderof their birth, and write names, ages, and any other significant information about them next to each's symbol. Here are examples of how to add children, along with special symbols for twins, adopted or foster children.

Second bom

6 Third born

Fourth and Fifth born

Sixth and Seventh born

o

Adopted or Foster child

Fraternal twins Identical twins

and pregnancies.

Z\ M

o X

Current pregnancy

(sex unknown)

Male Female Spontaneous Induced .stillborn stillborn abortion abortion

2. Now, add symbols for your father and mother's brothei and sisteis (your aunts and uncles) in the order of their birth, and write names, ages, and any other significant information about them next to each's symbol.

143

3. If you wish, add spouses or partnere of your siblings and your nieces and nephews; you may also wish to add cousins, especially if they have been close to you.

Step 4

The instmctions in this step are for your spouse or partner's side of the family.

I. Now, fill in as much information as you can about your spouse or partner and his or her family.

Steps

The instmctions in this step are for both sides of the family.

I. Draw in any previous marriages for either you or your spouse or partner, or for your parents or your spouse or partner's parents.

Use these symbols to show separation or divorce:

Separation

O v^

Divorce

Here are examples showing previous spouses:

First wife Second wife

V ^ M . 4/2/83 V - > ^ D. 11/9/84

7 ^ ^ ^

. 9/5/85 D. 12/4/89

Current marriage

Husband Wife O

M. 6/11/90

First husband

M. 5/1/82

o 144

Step 6

The instmctions in this step are for both sides of the family.

1. Now move to the bottom of the chart and add your children (if you have any) and their spouses if they are married. If they have children-your grandchildren-add them to the chart. Give names, ages, and any other significant information about them next to each's symbol.

2. Be sure to draw in any children from previous marriages, for you, your spouse, or your parents.

Step?

The instmctions in this step are for both sides of the family.

1. When the information on your chart is complete, use the following symbols to show some of the relationships between persons within your family and your spouse's family and of the two families.

Overly close - i ^ ^ — — ^ - ^ Conflictual ^ ^ V X

Distant ' Estranged I I

Here are definitions of these types of relationships (.see examples below):

Overiy close - reading each other's mind, talking for each other.

Distant - hardly talking or seeing each other anymore.

Conflictual - always fighting or bickering or arguing.

Estranged - being cut off or split apart forever.

2. Now, place a large dashed circle around the members of your household where you now reside or most recently having been living.

Here is an example:

145

H 6 Sister Brother

Steps

The instmctions in this step are for you.

1. Finally, list on the lower right-hand side of the page the major moves or experiences in your life and the approximate dates. Include such things as major geographical relocations, career changes, traumatic losses, and larger events that have had a major impact, such as war, economic depression, or natural disaster.

Step 9

The instmctions in this step are for both sides of the family.

1. Now, look over your family genogram. Is there any other information about your family that you would like to include? Put it on the chart wherever it is appropriate (or where you have room).

Step 10

Your are now finished drawing your family genogram. For many people, drawing a family genogram can be an "eye opening" experience, and may help them see some the influences on their present problems. Your therapist will now want to ask you are few questions about your experience of drawing a family genogram.

146

d <

w X H

O U Q CO

ua X H O

IO c

o

C cd

o

u Q ' c o . Q c o .t: • - rt cn c

^3

J ea c o

-y :

•y, rj •y, n ~r-

i o ' J

Z C Q Q U U J U J O S O : U J C 7 5 - ) C ;

o c u

"O ^ </5

U a:

on

O

o

(U

d <

PL]

X Pu O u Q CO

CO

PQ X H < Pu

IO a

CC

4) Q c o r^ c o -•= • - eg "« E

u ea y u eS D. u _u facial a S ' f - ^

Z C 0 Q U U J U J O 2 Q . U J C > 0 - J C U

1 y )

5? _ ^ •y n r^

y i O

'yi '

'C u <u y — c CJ (U 2 TD ea c/) .c u U Oi

H .J

•

Cd PU

c/)

(4—i

-o c cd

o 147

o

Z

-C

CQ

- 3

•n

.c • • ^

•J y, 3 r3

u

lion

-

u 3

• o UJ

aiio

n cu

p

O

eg

s

1 y;

£ x;

a.

c

/—•

LU

esse

s -

c75

r3 Si)

-J

y.

O X

y • >

Cte

nst

r3

c

1

'_> •J c

;2 "y

Qc:

a:

• Q

e« . S

Z ca

efl

Q T3 C e«

c

eo CJ 3

UJ

1) 3

U

C

w 13

O S

E x> o

c o

W c/3

•y)

E __ JJ eg 3 (U E

-yi t j

U «.* o eg eg

u

CJ

c

'(^ u

Oi

C/)

c t : cd

u t/3 3 O Cu

CO

c

x: U 3 O

148

d <

PU X H PU O PQ

e CO

CO DC PQ X H O

<u o c Cd

o

<u

C cd Urn

o

a:

eg

4, Q eg "O <*-

Q C C <u _ '^ «

C S S

eg

£ g • - - c s eg - > o eg CL u u •-. c CJ 3 : i : - ^ •^

e g . s e g { g - i : - Q o ' i g c ^ ( U L i Z c o Q u u J U J O S c u U J c / n - J a ,

-yv — _ S

•y)

U oi

J

on

u

o

u

d < PU PQ X H PU O PQ

Q CO

CO

c^ PQ X H <

cd

o o

•T3 C cd

o

eg

ta T3 i ^ >~. o

<u _ ' <u eg eg E

eg

CJ eg C o l U - — r^ ^^ o CJ

Z o a Q u u J U J O S c u U J o n J o , C s 2 .c

CJ 3 U 9J O

(/) CJ

C/5 '

c « « o — c CJ (U

2 " eg (/5

.c u U Oi

H J

c/3

x: Cd

149

oi

1)

Q

eg

Q

yi

eg . 3

z oa

c eg <u eg

Q

CJ •yi 3 eg

U

eg u 3

T3

C

eg u "5

c/5

E eg c o

X) o

o "^ .ii ^ u. C „ UJ O 2 o, W on

E eg "T;

U O.

yi kZ u o eg k. eg

u

J

Urn

c Cd

c/3 3 O CU

CO (U CJ c <u

;^ 'c« O

c/3

oo

s 150

APPENDIX F

DAI

151

Directed As.sessment Interview Outline

PARTNER ONE

Medical Problems (e.g., diabetes, cancer, stroke, heart disease, physical disability, dmg and/or alcohol problems) -

Emotional Stresses (e.g., depression, physical and/or sexual abuse, "nervous breakdown," schizophrenia) -

Legal Problems (e.g., committed felony, time in jail or prison, DWI/DUI) -

Birth Order (e.g., first, second, or third born) -

Significant Sibling Information -

Significant Aunt or Uncle Information

Significant Nephew or Niece Information -

Relationship History (i.e., dates of marriage, separation, and divorce, or length of time living together for all relationships) -

Relafional Dynamics with Other Family Members (i.e., significant conflictual, distant, cut-off, enmeshed relationships between this person and other family members) -

Who Lives in Your Household -

152

PARTNER TWO

Medical Problems (e.g., diabetes, cancer, stroke, heart di.sease, physical disability, dmg and/or alcohol problems) -

Emotional Stresses (e.g., depression, physical and/or sexual abuse, "nervous breakdown, schizophrenia) -

Legal Problems (e.g., committed felony, time in jail or prison, DWI/DUI) -

Birth Order (e.g., first, second, or third born) -

Significant Sibling Information -

Significant Aunt or Uncle Information -

Significant Nephew or Niece Information

Relationship History (i.e., dates of marriage, separation, and divorce, or length of time living together for all relationships) -

Relational Dynamics with Other Family Members (i.e., significant conflictual, distant, cut-off, enmeshed relationships between this person and other family members) -

Who Lives in Your Hou.sehold -

153

PARTNER ONE'S PARENTS

Father Mother

First Name -- Birth Date -

- Age-- Date and Cause of Death -

- Ethnicity -- Education Level -

- Occupation -- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -

- Residence -- Other Significant Information -

PARTNER TWO'S PARENTS

Father Mother

- First Name -- Birth Date -

- Age-- Date and Cause of Death -

- Ethnicity -- Education Level -

- Occupation -- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -

- Residence -- Other Significant Information -

154

PARTNER ONE'S FATHER'S PARENTS (PATERNAL GRANDPARENTS)

Father Mother

- First Name -

- Birth Date -- Age-

- Date and Cause of Death -- Ethnicity -

- Education Level -- Occupation -

- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -

- Residence -- Other Significant Information -

PARTNER ONE'S MOTHER'S PARENTS (MATERNAL GRANDPARENTS)

Father Mother

- First Name -

- Birth Date -- Age-

- Date and Cause of Death --Ethnicity -

- Education Level -- Occupation -

- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -

- Residence -- Other Significant Information -

155

PARTNER TWO'S FATHER'S PARENTS (PATERNAL GRANDPARENTS)

Father Mother

- Name -

- Birth Date -- Age-

- Date and Cause of Death -- Ethnicity -

- Education Level -- Occupation -

- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -- Residence -

- Other Significant Information -

PARTNER TWO'S MOTHER'S PARENTS (MATERNAL GRANDPARENTS)

Father Mother

- Name -- Birth Date -

- Age-- Date and Cause of Death -

-Ethnicity -- Education Level -

- Occupation -

- Medical Problems -

- Emotional Stresses -

- Legal Problems -

- Characteristics -

- Residence -- Other Significant Information -

156

APPENDIX G

DEPENDENT MEASURES COVER PAGE

157

Today' s Date Code #

Survey of Opinions About Your Relationship with Your Therapist and Your Therapy Session

Texas Tech University Department of Human Development and Family Studies

A. The statements on the next few pages are designed to learn what you think and feel about your therapist and the therapy session you had today. Please respond to the statements frankly and honestly. We are interested in your opinions and feelings, so there are no "right" or "wrong" responses.

B. It is important that you do not discuss the statements or your responses with your partner or therapist. Also, remember that your responses are CONFIDENTIAL and ANONYMOUS. They WILL NOT be shown to your partner or therapist and will only be used for research purposes.

C. Be sure to follow the instmctions given for responding to sets of statements, and then read each statement carefully. Respond to each statement—do not skip statements. Please respond to each statement in order, and take as much time as you need. If you want to make additional comments, feel free to write them in the margins.

D. When you have finished, please place this booklet through the slot in the locked box marked "RESEARCH" located in waiting room. Do not give this booklet to your therapist. Thank you for your participation in this project, on behalf of the researchers and those who will benefit.

This survey requires 5-10 minutes to complete

Please Turn To The Next Page

158

APPENDDC H

CTAS

159

Instructions

The followmg statements refer to your feelings and thoughts about your therapist right NOW. Please work quickly. We are interested in your FIRST impressions. Although some of the statements appear to be similar or identical, each statement is unique. Please be sure to respond to each statement.

Each statement is followed by seven different responses. Please circle the response that best describes the way you feel AT THIS TIME.

Here is an example of a statement to which you will be asked to respond.

A. The therapist cares about me as a person.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

As you think about your therapist right now, if you sort of believe that the therapist cares about you as a person, then you would circle "AGREE," like in the example above (A). If you do not believe the therapist cares about you as a person at all. then you would circle "COMPLETELY DISAGREE," like in the example below (B).

B. The therapist cares about me as a person.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

1. The therapist cares about me as a person.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

2. The therapist and I are not in agreement about the goals for this therapy.

COMPLETELY STRONGLY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE

COMPLETELY DISAGREE

3. I tmst the therapist.

COMPLETELY STRONGLY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE

COMPLETELY DISAGREE

Please Continue On The Next Page

160

4. The therapist lacks the skills and ability to help my partner and myself with our relationship.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

5. My partner feels accepted by the therapist.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

6. The therapist does not understand the relationship between myself and my partner.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

STRONGLY COMPLETELY DISAGREE DISAGREE

7. The therapist understands my goals in therapy. COMPLETELY STRONGLY

AGREE AGREE AGREE NEUTRAL DISAGREE STRONGLY COMPLETELY

DISAGREE DISAGREE

8. The therapist and my partner are not in agreement about the goals for this therapy.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

STRONGLY COMPLETELY DISAGREE DISAGREE

9. My partner cares about the therapist as a person.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

STRONGLY COMPLETELY DISAGREE DISAGREE

10. The therapist does not understand the goals that my partner and I have for ourselves as a couple in this therapy.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

11. My partner and the therapist are in agreement about the way the therapy is being conducted.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

12. The therapist does not understand me.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

Please Continue On The Next Page

161

13. The therapist is helping my partner and me with our relationship.

COMPLETELY STRONGLY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE

14. I am not satisfied with the therapy.

COMPLETELY STRONGLY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE

15. The therapist understands my partner's goals for this therapy

COMPLETELY STRONGLY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE

16. I do not feel accepted by the therapist.

COMPLETELY DISAGREE

COMPLETELY DISAGREE

COMPLETELY DISAGREE

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

17. The therapist and I are in agreement about the way the therapy is being conducted.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

18. The therapist is not helping me.

STRONGLY DISAGREE

COMPLETELY DISAGREE

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

19. The therapist is in agreement with the goals that my partner and I have for ourselves as a couple in this therapy.)

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

20. The therapist does not care about my partner as a person.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

STRONGLY COMPLETELY DISAGREE DISAGREE

21. The therapist has the skills and ability to help me.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

Please Continue On The Next Page

162

22. The therapist is not helping my partner

COMPLETELY STRONGLY AGREE

STRONGLY COMPLETELY AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

23. My partner is satisfied with the therapy.

COMPLETELY STRONGLY AGREE AGREE AGREE NEUTRAL DISAGREE

24. I do not care about the therapist as a person.

STRONGLY DISAGREE

COMPLETELY DISAGREE

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

25. The therapist has the skills and ability to help my partner.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

26. My partner distmsts the therapist.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

27. The therapist cares about the relationship between my partner and myself.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

28. The therapist does not understand my partner.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

29. The therapist does not appreciate how important the relationship between my partner and myself is to me.

COMPLETELY STRONGLY STRONGLY COMPLETELY AGREE AGREE AGREE NEUTRAL DISAGREE DISAGREE DISAGREE

163

APPENDIX I

SEQ

164

Instructions

The following statements refer to your feelings and thoughts about today's therapy session. Each statement has a pair of words that represent opposite extremes, like "DEEP" and "SHALLOW." In between the pair of words are seven numbers that represent variations between the extremes. Please circle the number that best describes your feelings and thoughts about today's session.

Here is an example of a statement to which you will be asked to respond.

DEEP 1 SHALLOW

As you think about today's therapy .session, if you sort of believe that the .session was deep, then you would circle the number "5," like in the example above (A). If you believe the session was extremely shallow, then you would circle the number "1 . " like in the example below (B).

DEEP 1 SHALLOW

TODAY'S SESSION WAS:

1.

2.

3.

4.

5.

DEEP

FULL

POWERFUL

VALUABLE

SPECIAL

1 SHALLOW

I EMPTY

1 WEAK

1 WORTHLESS

1 ORDINARY

6 COMFORTABLE 1 UNCOMFORTABLE

7. SMOOTH 1 ROUGH

8. EASY 1 DIFFICULT

8. PLEASANT 1 UNPLEASANT

10. RELAXED 1 TENSE

165

APPENDDC J

DEMOGRAPHIC QUESTIONNAIRE FOR PARTICIPANTS

166

Instructions

Please answer the following questions about yourself.

I. My age is:

2. I am:

years, (fill in number)

female _ male (check one)

3. My race is: (check one)

African-American Asian Caucasian Hispanic Native-American Other

The status of the relationship with my therapy partner is: (check one, and fill in numbers for years and months)

Dating Living together Married Separated Divorced

For how long? For how long? For how long? For how long? For how long?

years years years years years

months months months months months

The highest level of education I have completed is: (check one)

Less than high school Some high school High school diploma or equivalent Vocational/technical Some college Associate's degree Bachelor's degree Master's degree Doctorate degree

Thank You Please Place This Booklet In The Locked Box Maiked ' Research"

167

APPENDIX K

DEMOGRAPHIC QUESTIONNAIRE FOR THERAPISTS

168

Demographic Questionnaire for Therapists

Today's date is (fill in month/day/year)

Please answer the following questions about yourself.

1. My age is: years, (fill in number)

2. I am: female male (check one)

3. My race is: (check one)

African-American Asian Caucasian Hispanic Native-American Other

4. l ama 1st, 2nd year student master's degree student, (check one)

5. I have years, months of clinical experience, (fill in number(s))

6. My primary theoretical orientation to therapy is best described as: (check one)

Behavioral/Social learning

Brief Experiential Intergenerational Problem-focused Stmctural Strategic Other (please specify)

7. I use a genogram % of the time in my therapy, (check one)

0% 25% 50% 75% 100%

Please Continue On The Next Page

169

8. I would rate my ability to constmct a genogram as: (check one)

Excellent Good Average Fair Poor

9, I would rate my clinical skills as: (check one)

Excellent Good Average Fair Poor

Thank You

170

APPENDIX L

SESSION FOLLOW-UP QUESTIONNAIRE

FOR THERAPISTS

171

Session Follow-up Questionnaire for Therapists

Please answer the following questions.

1. Today's date is (fin in month/day/year)

2. My clinic number is: (fill in number)

The next three questions (3, 4, and 5) refer to the therapy .session you just completed:

3. Were both couple members present during at least part of the session?

YES NO

4. Did another therapist or supervisor enter the therapy room and participate as a co-therapist at any point during the session?

YES NO

5. Besides the couple, did any family member, friend, or other person(s) participate in any part of the therapy session?

YES NO

6. If you answered "NO" to question 3, OR if you answered "YES" to either question 4 or 5, please explain what happened.

7. If the therapy session you just completed was the second therapy session, how many minutes did it take the couple to complete either a therapist-administered genogram, a self-administered genogram, or a directed assessment interview?

MINUTES (fill in number)

172

Thank You

173

APPENDDC M

SAMPLE FOLLOW-UP LETTER TO THERAPISTS

174

Month XX, 199X

Dear Doctoral Student,

I am writing to you regarding the study, "Effects of couples' perceptions of genogram constmcuon on therapeutic alliance and session impact: A growth curve analysis.' I want to thank you for your participation as a therapist in this study, and for taking the time to attend the orientation. As a graduate student myself I know how busy graduate school schedules can be, so your time and effort are especially appreciated.

During the orientations, two of the most common questions raised about the procedures were...Here are some reminders on how to address these issue should they arise...Your careful attention to the procedures will help the study fiow smoothly.

Should you have any questions about the procedures that I can clarify, or if any problems arise that you need to discuss, please call me at (806) 555-1212, or write me at Texas Tech University, Department of Human Development and Family Studies, Box 1162, Lubbock, Texas 79409-1162. Again, thank you for your time.

Sincerely,

Scott Coupland, M.S.

175