Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An ecological assessment of the efficacy of individual and couples treatment formats of Experiential… Grigg, Darryl Norman 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-893788.pdf [ 8.3MB ]
JSON: 831-1.0053992.json
JSON-LD: 831-1.0053992-ld.json
RDF/XML (Pretty): 831-1.0053992-rdf.xml
RDF/JSON: 831-1.0053992-rdf.json
Turtle: 831-1.0053992-turtle.txt
N-Triples: 831-1.0053992-rdf-ntriples.txt
Original Record: 831-1.0053992-source.json
Full Text

Full Text

AN ECOLOGICAL ASSESSMENT OF THE EFFICACY OF INDIWDUALAND COUPLES TREATMENT FORMATS OF EXPERIENTIAL SYSTEMICTHERAPY FOR ALCOHOL DEPENDENCYbyDARRYL NORMAN GRIGGB.A., The University of Waterloo, 1980M.A., The University of British Columbia, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF EDUCATIONinTHE FACULTY OF GRADUATE STUDIESDepartment of Counselling PsychologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAFebruary 1994© Darryl Norman GriggIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of_____________________The University of British Colum IaVancouver, CanadaDate___DE-6 (2/88)IIABSTRACTThis study investigates the differential treatment efficacy of Experiential SystemicTherapy (ExST) with a comparison treatment called Supported Feedback Therapy (SFT) asit is applied to the problem of alcohol dependency. The inquiry also compares the treatmenteffectiveness of ExST when offered to the individual alcoholic (ExST-I) and when providedin couples therapy conjoint treatment (ExST-C).An ecological approach to assessment was developed for the investigation. Self-report questionnaires tapping an array of areas including indices of alcohol use,intrapersonal functioning, couples adjustment, and family characteristics were employed tomeasure treatment effects from the perspectives of father, mother, and eldest child.Participating families met inclusion criteria including an alcoholic dependent father and anon-alcohol abusing mother in a state of marital distress residing in an intact family situationwith at least one child living at home.One hundred and fourteen families were randomly assigned to participatingtherapists and one of three treatment conditions including ExST-I, ExST-C,or SFT. Therapywas conducted at two out-patient clinics, one located in an urban setting and the otheroperating in a rural context. Data were collected from all participating families before andafter treatment. Data were also gathered at a three month follow-up from participants inthe ExST-I and ExST-C treatment conditions.The results of the mixed model multivariate analyses indicated that there were nosignificant differences between ExST and SFT evident at post-treatment; however, bothtreatments were found to have promoted highly significant improvements on measures ofdrinking behavior, intrapersonal symptomology, marital adjustment and family satisfaction.When ExST-I and ExST-C were compared, the results revealed no significant differencesbetween the treatment formats although both parents reported highly significant posttreatment changes on all instruments. Additionally, the significant changes associated with111ExST-I and ExST-C which were reported by both parents at post-treatment were found to beequally durable at the end of a three month follow-up. The results of the analyses based onthe eldest child’s perspective showed that the assessments of family satisfaction wereunaffected by the treatment conditions and remained consistent across all measurementoccasions.Within system analyses which provided detailed examination of the magnitude ofchanges reported by both parents at post-treatment were performed. The within systemresults based on measures probing the assessment domains of alcohol, intrapersonal, coupleand family from the father and mother perspectives, revealed that the improvementsachieved by the treatments were far reaching and touched a wide array of areas instatistically significant and clinically relevant fashions.ivTABLE OF CONTENTSPageAbstract iiTable of Contents ivList of Appendices ixList of Tables xList of Figures xiiAcknowledgement xivCHAPTER 1: INTRODUCTION 1Context of the Problem 1The Problem 6Purpose of the Study 8Definitions 9CHAPTER 2: LITERATURE REVIEW 13Alcoholism 13Biopsychosocial Model of Alcoholism 15Family Systems Theory Overview 17Systemic Theory and Individual Treatment 21Integrative Problem-Centred Therapy 22Internal Family Systems 24Experiential Systemic Therapy 27Couples and Family Therapy of Adult Alcoholism:Treatment Outcome Research 291956-1973 291974-1980 301980-1986 36V1986-1993.38Summary of Family Therapy Alcoholism:Treatment Outcome Research 46Couples and Individual Therapeutic Format Research 47Experiential Systemic Therapy Synopsis 54Rationale for ExST Development 56Summary of Literature Review 59Research Hypotheses 60Hypothesis 1 61Hypothesis 2 61CHAPTER 3: METHODOLOGY 62Research Design 62Supported Feedback Therapy 63Family Inclusion Characteristics 65Family Description 66Family Profile 67Alcohol Consumption Profile 69Previous Treatment Profile 72Family Involvement in the Study 73Drop Out Families 74Complete Treatment Families 74Complete Treatment Families with Missing Data 75Family Participant Involvement 75Data Collection 76SFT Treatment Group 77ExST Individual Treatment Group 78ExST Couples Treatment Group 78Clinical Context 80Therapists 80Treatment Implementation 81viSFT Therapist Training and Selection 82SFT Supervision 82ExST Therapist Training and Selection 83ExST Supervision 83Treatment Implementation Check 84Ecological Assessment Model 86Implementation of Ecological Assessment Model 89Instrumentation 90Alcohol Measures Description 91Michigan Alcohol Screening Test 93Alcohol Dependence Data 95Inventory of Drinking Situations 96Situational Confidence Questionnaire 98Intrapersonal Measures Description 100Shipley Institute of Living Scale 101Symptom Checklist 90 Revised 102Beck Depression Inventory 104Couples Measures Description 105Edmonds Martial Conventionality Scale 105Dyadic Adjustment Scale 107Areas of Change 109Family Measures Description 111Family Information Form 112Family Satisfaction 112Family Environment Scale 114Family Adaptability and Cohesion Evaluation III 116Phases of Data Analysis and Operationalization of Research Hypotheses 118Phase 1. Preliminary Analysis 118Phase 2. Eco-System Analysis 119Marker Variables 119Statistical Procedures 120Phase 3. Within System Analysis 120Phase 4. Therapeutic Process Validation 121Operationalization of Research Hypothesis 1 121Operationalization of Research Hypothesis 2 124CHAPTER 4: RESULTS 129Preliminary Analyses 129Instrument Overview and Pre-treatment ParticipantDescription 129viiAlcohol.129Intrapersonal.131Couple 133Family 135Group Equivalence 137Summary of Equivalence Tests 141Eco-system Analyses of Hypotheses 142First Research Hypothesis: Differential Efficacy of ExST and SFT 142Sub-hypothesis 1-a 142Sub-hypothesis 1-b 144Sub-hypothesis 1-c 146Second Research Hypothesis: Differential Efficacy of ExST-Iand ExST-C 148Sub-hypothesis 2-a 148Sub-hypothesis 2-b 150Sub-hypothesis 2-c 152Sub-hypothesis 2-d . 154Sub-hypothesis 2-e 156Sub-hypothesis 2-f 158Within Systems Analyses 159Within Alcohol System 159Within Intrapersonal System 161Within Couple System 165Within Family System 170Therapeutic Process Validation 173Clinic Site 173Therapist Gender 174Therapist 174CHAPTER 5: DISCUSSION 175Eco-system Analyses 175First Research Hypothesis: Differential Efficacyof ExST-I and SFT 175Second Research Hypothesis: Differential Efficacyof ExST-I and ExST-C 184Within Systems Comparisons 188Within Alcohol System 189Within Intrapersonal System 190Within Couple System 191Within Family System 194Within Systems Summary.195Therapeutic Validation 196Limitations of the Study 196Limitations of Field-based Research 198Generalizability 199Conclusions 201Future Research 203REFERENCES 205viiiixLIST OF APPENDICESPageAPPENDIX A:Schedule of ExST Training Events and Alcohol and DrugProgram Therapist Appraisals of Model Training Experiences 228APPENDIX B:Experiential Systemic Therapy Overview 233APPENDIX C:Father and Mother Weekly Situations Diaries and ReducedVersions of Supported Feedback Therapy Wall Charts 265APPENDIX D:Project Consent FormAPPENDIX E:Project Payment Schedule A and B 284APPENDIX F:The Alcohol Recovery Project Adherence Rating Scale Rating Form 287APPENDIX G:List of Instruments Employed in Ecological Assessmentand Measurement Occasions 290APPENDIX H:Summary of Analyses of Instrument Characteristics 292xLIST OF TABLESPage1. Schematic Design of the Study 622. Perspective of Respondent and Systemic Assessment Level 903. Means, Standard Deviations, and Reliability Estimates of AlcoholMeasures for Fathers and Mothers at Pre-test 1304. Means, Standard Deviations, and Reliability Estimates ofIntrapersonal Measures for Fathers and Mothers at Pre-test 1325. Means, Standard Deviations, and Reliability Estimates ofCouples Measures for Fathers, Mothers and Couples at Pre-test 1346. Means, Standard Deviations, and Reliability Estimates ofFamily Measures for Fathers and Mothers at Pre-test 1357. Means, Standard Deviations, and Reliability Estimates ofFamily Measures for Eldest Children at Pre-test 1368. Summary of Univeriate Tests of Fathers’ Marker Variablesfor Pre-treatment and Post-treatment Differences for ExST and SFT 1439. Fathers’ Pre-treatment and Post-treatment Means and StandardDeviations of Marker Variables for ExST and SFT 14310. Summary of Univeriate Tests of Mothers’ Marker Variables forPre-treatment and Post-treatment Differences for ExST and SFT 14511. Mothers’ Pre-treatment and Post-treatment Means and StandardDeviations of Marker Variables for ExST and SFT 14512. Eldest Children’s Pre-treatment and Post-treatment Means andStandard Deviations of Marker Variable for ExST and SFT 14713. Summary of Univariate Tests of Fathers’ Marker Variables forPre-treatment and Post-treatment Differences for ExST-I and ExST-C 14914. Fathers’ Pre-treatment and Post-treatment Means and StandardDeviations of Marker Variables for ExST-I and ExST-C 149xi15. Summary of Univariate Tests of Mothers’ Marker Variables forPre-treatment and Post-treatment Differences for ExST-I and ExST-C 15116. Mothers’ Pre-treatment and Post-treatment Means and StandardDeviations of Marker Variables for ExSt-I and ExST-C 15117. Eldest Children’s Pre-treatment and Post-treatment Meansand Standard Deviations of Marker Variables for ExST-I and ExST-C 15318. Fathers’ Post-treatment and Follow-up Means and StandardDeviations of Marker Variables for ExST-I and ExST-C 15519. Mothers’ Post-Treatment and Follow-up Means and StandardDeviations of Marker Variables for ExST-I and ExST-C 15720. Eldest Children’s Post-Treatment and Follow-up Means andStandard Deviations of Marker Variable for ExST-I and ExST-C 15821. Summary of Univariate Tests Within Alcohol Level ofAssessment of Pre-treatment and Post-treatment Differences 16022. Summary of Univariate Tests of Fathers’ Within Intrapersonal Levelof Assessment for Pre-Treatment and Post-treatment Differences 16223. Summary of Uriivariate Tests of Mothers’ Within IntrapersonalLevel of Assessment for Pre-treatment and Post-treatment Differences 16424. Summary of Univariate Tests of Fathers’ and Mothers’ Within CoupleLevel of Assessment for Pre-treatment and Post-treatment Differences 16625. Summary of Univariate Tests of Couples Variables for Within CouplesLevel of Assessment for Pre-treatment and Post-treatment Differences 16926. Summary of Univariate Tests of Fathers’ and Mothers’ Within FamilyLevel of Assessment for Pre-treatment and Post-treatment Differences 171xiiLIST OF FIGURESPage1. Family involvement sub-groups 732. Therapy participant sub-groups 763. Overview of TARP research protocol 764. Ecological assessment model 885. Fathers’ and mothers’ mean pre-treatment SCL-90-R sub-scale scorescompared to normal non-patients and in-patient norms 1336. Participating family mean scores on FES at pre-treatment and distressat norm comparison 1367. Fathers’ pre-treatment and post-treatment means of markervariables for ExST and SFT 1448. Mothers’ pre-treatment and post-treatment means of markervariables for ExST and SET 1469. Eldest children’s pre-treatment and post-treatment means ofmarker variable for ExST and SET 14710. Fathers’ pre-treatment and post-treatment means of markervariable for ExST-I and ExST-C 15011. Mothers’ pre-treatment and post-treatment means of markervariables for ExST-I and ExST-C 15212. Eldest children’s pre-treatment and post-treatment means ofmarker variable for ExST-I and ExST-C 15313. Fathers’ post-treatment and follow-up means of markervariable for ExST-I and ExST-C 15514. Mothers’ post-treatment and follow-up means of marker variablesfor ExST-I and ExST-C 15715. Eldest children’s post-treatment and follow-up means of markerfor ExST-I and ExST-C 159xlii16. Fathers’ total sample pre-treatment means and treatment grouppost-treatment means for within alcohol system variables 16117. Fathers’ total sample pre-treatment means and treatment grouppost-treatment means for within intrapersonal system variables 16318. Mothers’ total sample pre-treatment means, and treatment grouppost-treatment means for within intrapersonal system variables 16419. Fathers’ and mothers’ total sample pre-treatment means, and treatmentgroup post-treatment means for within couple system variable DAS 16720. Fathers’ and mothers’ total sample pre-treatment means, and treatmentgroup post-treatment means for within couple system variable AC 16821. Couples’ total sample pre-treatment means, and treatment grouppost-treatment means for within couple system variables 17022. Fathers’ total sample pre-treatment means and treatment grouppost-treatment means for within family system variables 17223. Mothers’ total sample pre-treatment means, and treatment grouppost-treatment means for within family system variables 173xivACKNOWLEDGEMENTThis study has required the involvement, co-operation, and support of manyindividuals and agencies. Over the six years it has taken to complete the research, manypeople have contributed to the work and enriched the investigation in the process. I wouldlike to thank all those whose hands helped shape the study and to share my gratitude fortheir assistance. I would like to recognize a number of people who have played substantiveroles in the realization of this dissertation.I am grateful for the assistance, recommendations, and editorial suggestions offeredto me by my dissertation committee John D. Friesen, Robert F. Conry, and BethHaverkamp. In addition, I would like to thank Cheryl Bate and Warren Weir for theircompanionship and insights over the years it has taken to complete this work.I would like to convey my heartfelt appreciation for the years of support from theDirectors of the Surrey and Summit clinics Jean Mathews and David Todtman. In addition,I would like to offer my thanks to the administrative and support staff at the clinics and theuniversity including Merrilee Shea, Suzzanne Jacques, Gerry Tothill, Bay Gumboc, RobynMoore, and Lori Walker whose cheerful assistance made the journey just that little biteasier.Special appreciation goes to the participating therapists who gave so much ofthemselves to make this study a reality including Natacha Villasenor, Gillian Neuman, GeoffLyon, Gordon Venn, Norma Elbe, Warren Weir, Dan Mitchell, David Todtman, GillianReeves, Dan McGee, and Dale McMullen.Finally, and most importantly, I would like to share my appreciation for my familywho have provided sustained encouragement through the ups and downs of this dissertation.In particular, I would like to honour the loving collaboration of my wife and best friendJennifer Newman who more than anyone has been there when most needed. Lastly, I mustthank our son Robin whose small hands and belly laugh have helped me find the courage tocomplete this task and remember what is really important.1CHAPTER 1: INTRODUCTIONContext of the ProblemThe images of burnt out souls living on skid-row are haunting reminders of the resultsof long term heavy alcohol dependency. With these people relegated to trash bins and darkcorners, it is all too easy to dismiss them as fringe members of the population andmarginalize the problem of alcohol dependency at the same time. Yet the territory ofalcohol problems extends well beyond skid-row. In fact, estimates suggest that only 5% ofthe people struggling with serious alcohol dependency sink to the isolation of the streets(Steinglass, Bennett, Wolin, & Reiss, 1987). The remainder continue to live in homes insome form of family unit. Consequently, for every individual with an alcohol problem, thereare many others whose lives are directly affected by the drinking behaviour and who wouldbenefit from the amelioration of the drinking problem.The relationship between alcohol dependency and the interpersonal contexts in whichthe drinking behaviour is embedded is of considerable interest to theorists, researchers andclinicians alike. Indeed, many see the connection between the two to be so interwoven thatterms such as “Alcoholic Marriage” and “Alcoholic Family” have become popular (e.g.,Bradshaw, 1988; Steinglass et al., 1987; Wegscheider, 1981). These terms are meant todenote networks of relationship in which patterns of abusive drinking are so integrated intothe structures of the interpersonal contexts, that they become the dominant centralorganizing features of the relational nets. (Lawson, Peterson, Lawson, 1983; Steinglass,1980; Steinglass et al., 1987).Early systemic models of alcoholism view the association between problem drinkingand relational fields to be interactive, reciprocal, and homeostatic (Finney, Moos, Cronkite,& Gamble, 1983). Steinglass (1981) articulated this position with regards to the stabilizing2influence or “adaptive consequences” of continued alcohol use to a family system’s balancewhen he wrote some families, equilibrium is restored by increasing interactional distance(the drinker goes off to drink in the basement), or diminished physical contact(the non alcoholic spouse refuses to have sex with someone who is drunk), orreducing tension in the family (family members’ usual patterns of behavior areless tension provoking than unique patterns); whereas in other families,alcohol might be associated with closer interactional distance (the nonalcoholic makes contact by fighting after the alcoholic spouse has beendrinking), disinhibition (the use of alcohol permits ritualized sexual behavior),or maintaining distance from the social environment (the alcoholic fights withneighbors when drunk). (p. 301)From this perspective, every part of a system is so related to its fellow parts that achange in one part implies a change in all and in the entire system. The family is seen as anetwork of relationships that does not behave as a simple composite of independentelements but rather coherently, as an inseparable whole (Watzlawick, Beavin, & Jackson,1961). Consequently, first-order cybernetic models postulate that it is unwise and perhapsdangerous to attempt to change aspects of a system without due consideration being given tothe meaning such changes might have for both the sub-system members and the system as awhole (Auerswald, 1985; Keeney, 1983).Applied to the treatment of alcoholism, the first-order systemic models challengedthe conventional notion that the “problem” resided in the individual alcoholic. Thesetreatment approaches asserted that the social contexts in which behaviors are embeddedwere intrinsically connected to problem drinking. Consequently, a change in an alcoholic’sdrinking pattern necessarily implied a change in the dynamics and the quality of therelational context. In addition, a corollary of this perspective hypothesized that contextualsystems will become unsettled when an individual member attempts to change someimportant behavior. Importantly, this position maintained that marital or family systemsmay struggle against improvements in an alcoholic’s consumption pattern in an effort topreserve the family system’s stability and identity (Jacobson, Munroe, & Schmaling, 1989;3Steinglass et al., 1987; Usher, Jay, & Glass, 1982). As a result, family system models oftreatment for alcoholism which were grounded in first-order cybernetic thought, focusedattention on the families’ interactive process and addressed changes in the family context asa way to assist the recovery process (Davis, 1987; Kaufman & Kaufman, 1979; Kaufman &Pattison, 1981; Lawson, Peterson, & Lawson, 1983).The alcoholism literature has had a long history of seemingly unbridled speculationabout the role of the non-drinking spouse and the family in both the development andpersistence of alcohol dependency. Much of this early work portrayed the social contexts ofmarriage and family as negative influences in an alcoholic’s life and as problematic obstacleswhich must be overcome (Bailey, 1961; Edwards, Harvey, & Whitehead, 1973; Paolino &McCrady, 1977; Steiner, 1969). A troubling example in this regard was provided by Whalen(1953) in which the author concerns himself with the unconscious motivations which cause awoman to marry an alcoholic man. In this article, four pejorative categories were offeredincluding “Suffering Susan”; “Controlling Catherine”; “Wavering Winifred” and “PunitivePolly,” all of whom were thought to play a causal role in the alcoholics’ suffering.The negative view of the families and, in particular, the wives of alcoholics slowlygave way to a more empathic and complex understanding of the connection between contextand problem. The introduction of family systems theories to the field of alcoholism occurredin the late 1960’s and early 1970’s. Spurred on by the “adaptive consequences” model (Davis,Berenson, Steinglass, & Davis, 1974), considerable effort was made to explore the ways inwhich alcoholism came to be gradually incorporated into family life over time. Findingsfrom family interaction research (Billings, Kessler, & Gomberg, 1979; Frankenstein, Hay, &Nathan, 1985; Hersen, Miller, & Eisler, 1973) supported the notion that families with analcoholic member are, as Steinglass et a!. (1987) noted:4highly complex behavioral systems with remarkable tolerance for stress as wellas occasional bursts of adaptive behavioral inventiveness that provoke wonderand admiration in observers. (p. 8)The development of treatment approaches to alcoholism based on a familyperspective lag behind theorizing about the problem of alcohol dependency (Davis, 1987;Kaufman & Pattison 1981). While elegant first-order systemic hypothesizing (e.g., Bateson,1979) about alcoholics proceeded and gained empirical support from some research efforts(e.g., Jacob, Dunn, & Leonard, 1983), family treatment models designed specifically foralcoholism were slower to emerge. However, over time there was an impressive generationof family treatment models focused on alcohol dependency problems (Davis, 1987; Gaeic,1986; Kaufman & Kaufman, 1979; Lawson et al., 1983; Steinglass et al., 1987; Treadway,1989).The relational formulations of first-order cybernetic theory held a measure of appealto practitioners in the field and consequently gained wide acceptance. Indeed, it is nowcommonly assumed that therapeutic efforts with alcoholics should include treatingsignificant people in the alcoholic’s life (Whittingham, 1987). Nonetheless, a lack ofwidespread availability of marital and family therapy in treatment facilities has beendocumented (Camacho-Salinas, O’Farrell, Jones, & Cutter, 1984; Regan, Connors,O’Farrell, & Jones, 1985), and a dearth of efficacy studies testing the clinical implementationof these first-order family systems models has been highlighted (McCrady, 1989; O’Farrell,1992).Research probing the efficacy of the first-order systemic treatment approaches toalcoholism that were articulated has been sadly lacking. None of the systemic therapyapproaches detailed by Davis (1987), Kaufman and Kaufman (1979), Lawson et al. (1983),Steinglass et al. (1987), Treadway (1989), Thomas and Santa (1982), or Gaeic (1986) havebeen formally evaluated in well designed studies, or subjected to widespread testing inclinical settings. The empirical studies which have been conducted on the effectiveness of5couple or family treatments of alcoholism (reviewed in a later section) have predominantlyfocused on behavioural approaches. This disturbing fact was spelled out by Jacobson,Munroe, and Schmaling (1989) who noted:The discrepancy between the predominance of systemic notions leading us tobelieve in the promise of marital treatments and the dearth of research on theefficacy of systemic approaches is striking. (p. 9)Systemic theory continued to develop even as the first-order cybernetic treatmentmodels of alcoholism noted above were articulated. The second-order cyberneticperspective (Hoffman, 1986; Sluzki, 1985) which no longer viewed systems as objectivehomeostatic units outside the observer, began to challenge earlier family formulations (Dell,1985; Keeney, 1983). Since 1985, the field of marital and family therapy has moved toembrace second-order systemic thought and rejected an objectivist epistemology (Simon,1992). Perspectives such as post-modernism (Anderson & Goolishan, 1988; Gergen, 1991)have taken family therapy well beyond its first-order cybernetic beginnings (Cecchin, Lane,& Ray, 1993). New treatment approaches such as the popular narrative approach (White &Epston, 1990) concern themselves with the role that observers play in constructing the realitybeing observed and seek new ways to understand and speak about problems that allow forthe difficulties to be resolved. The generation of these new viewpoints in the field of familytherapy reflects an important shift away from how families and their problems wereunderstood (Hoffman, 1990). In particular, constructs such as homeostasis, resistance,boundaries and family rules have been for the most part abandoned (Cecchin et al., 1993;Goolishan & Anderson, 1992).The important developments in systemic treatments noted above have not beenwidely reflected in the area of alcoholism treatment. Indeed, the marital and familyapproaches which at present dominate the field of alcoholism treatment continue to berooted in first-order cybernetic thought (e.g., Steinglass et a!., 1987). Nonetheless, second-6order systemic models applied to alcoholic problems have recently emerged (e.g., Friesen,Grigg, Peel, & Newman, 1989) and are being applied to alcoholism treatment.The need for innovation in treatment approaches to alcoholism continues to bepressing with many of the currently practiced models proving to result in efficacy rates whichare not as impressive as would have been hoped (Miller & Hester, 1986; Nathan & Skinstad,1987). Jacobson, Munroe, and Schmaling (1989) recently issued a call for the developmentof therapeutic approaches to alcoholism noting that:the need for clinical innovation in alcoholism treatment is most acutebecause no treatment has emerged as consistently effective to a clinicallysignificant degree. (p. 8)Marital and family approaches to treatment and interventions which include broad-spectrumstrategies and relapse prevention procedures have been identified as promising directions(Institute of Medicine, 1992). However, well designed empirical evaluations ofdevelopments in treatment are necessary to promote the careful advancement of thetreatment field.The ProblemExperiential Systemic Therapy (ExST) (Friesen et al., 1989) is an integrativeapproach to the treatment of adult alcoholism. Recently generated at the University ofBritish Columbia, the model has reached the point where empirical examination in a fieldbased study is required prior to further development and wider implementation. Theapproach is introduced below.Experiential Systemic Therapy (ExST) is a treatment method that was designedspecifically for alcohol dependency problems (Friesen, Grigg, Peel, & Newman, 1989).Generated in a rich clinical environment, ExST is a response to the call for furtherdevelopment of treatment approaches for alcoholism noted earlier. ExST is a second-order7,cybernetic systemic therapy that focuses on problems within the interactions betweenmultiple layers of human relations. The ExST approach emphasizes the role of the observerin developing solutions to problems and enables the systemic potential to address andtransform disturbed patterns of relationship between parts of self as well as betweensignificant others, objects, symbols and other contexts in the outer world of the individual.The model exemplifies a broad spectrum approach to alcoholism and includes relapseprevention within the treatment protocol. The ExST model integrates individual and familytherapy concepts and techniques in such a way that a unified set of assumptions, conceptsand techniques can equally be applied to individual, couple and family treatment formats.Accordingly, the model’s versatility is an asset in meeting the changing needs of clientsengaged in the recovery process.ExST evolved out of an effort to train alcohol and drug treatment clinicians employedby the Government of British Columbia in the practice of couples and family therapy. Thedialogue between trainers and trainees gave rise to the generation of an approach totreatment that emphasized the physiological, intrapersonal, interpersonal and spiritualconcerns of those struggling with the multifaceted problems connected to alcoholdependency and a model of supervision which focused on theoretical development, technicalrefinement, and personal growth (Newman, Friesen, & Grigg, 1991).The ExST model gained wide support from counsellors who were trained in thetreatment over the years of 1987 - 1989. A copy of the ExST training events schedule and alisting of therapist comments drawn from workshop evaluations is presented in Appendix A.The therapists’ comments aptly reflect the enthusiasm with which the therapy was receivedby clinicians in the field. Informal reports from therapists employing the model subsequentto the training further supported the development of ExST.8The popularity of ExST led to continued requests for training and it became clearthat the treatment held much promise for the field of alcohol treatment. Yet despite theappreciation and encouraging comments made by previously trained therapists, the efficacyof the model in terms of its measurable effectiveness had not been evaluated. In addition,the issue regarding the comparability and/or the utility of the individual and couplestreatment formats of ExST remained an empirical question. Consequently, widespreadtraining activities were suspended pending the results of a treatment outcome study thatwould probe the efficacy of the model in both its individual and couples form.Purpose of the StudyThis study is one of a series of studies connected to a large-scale research projectentitled The Alcohol Recovery Project (TARP). Carried out over a period of five years,TARP has received funding from the British Columbia Alcohol and Drug Program (nowpart of the provincial Ministry of Health and formerly in the Ministry of Labour andconsumer services) and from the British Columbia Health Research Foundation (HealthServices Research Programme). Other assistance has been extended to TARP by theUniversity of British Columbia and the Humanities and Social Sciences Research Services.These funds and other forms of assistance have enabled the completion of this study, as wellas others resulting from TARP activities. This body of research has been conducted underthe general direction of the Principal Investigator, John D. Friesen, Ph.D., co-investigatorRobert F. Conry, Ph.D., and project coordinator and clinical supervisor, Darryl N. Grigg.While preliminary results of TARP have recently been presented (Grigg, Friesen, & Conry,1993), additional information regarding TARP and other specific studies related to it may beobtained from Professor John Friesen, Department of Counselling Psychology, University ofBritish Columbia, 5780 Toronto Road, Vancouver, B.C., V6T 1L2, Canada.This study was designed to investigate the efficacy of ExST in the treatment ofalcohol dependency. In addition, the inquiry also evaluates the effectiveness of both9individual and couples treatment formats of ExST as they are implemented with alcoholdependent clients.The intentions of the inquiry are reflected in the following research questions.(1) Does the delivery of the ExST result in significant change when compared to acontrast treatment group on indices of alcohol dependency, individualfunctioning, and measures of marital dynamics and family qualities in thetreatment of alcohol dependency problems?(2) Does the delivery of individual or marital treatment formats of ExSTdifferentially affect alcohol dependency, individual functioning, maritaldynamics and family qualities to a significant degree when applied to thetreatment of alcohol dependency problems?DefinitionsOperational definitions of terms used in this study are as follows:Alcoholism: For the purposes of this study the terms alcoholism and alcoholdependency will be used synonymously. All alcoholics in the investigation satisfied theDSM-III-R (1987) diagnostic criteria for severe Psychoactive Substance Dependence asdetailed below:Diagnostic Criteria for Psychoactive Substance DependenceA. At least three of the following:1. substance often taken in larger amounts or over a longer period than the personintended2. persistent desire or one or more unsuccessful efforts to cut down or control substanceabuse3. a great deal of time spent in activities necessary to get the substance (e.g., theft),taking the substance (e.g., chain smoking), or recovering from its effects4. frequent intoxication or withdrawal symptoms when expected to fulfill major role(obligations at work, “high,” intoxicated while taking care of his or her children), orwhen substance use is physically hazardous (e.g., drives when intoxicated)105. important social, occupational, or recreational activities given up or reduced becauseof substance use6. continued substance use despite knowledge of having a persistent or recurrent social,psychological, or physical problem that is caused or exacerbated by the use of thesubstance (e.g., keeps using heroin despite family arguments about it, cocaine-induced depression, or having an ulcer made worse by drinking)7. marked tolerance: need for markedly increased amounts of the substance (i.e., atleast a 50% increase) in order to achieve intoxication or desired effect, or markedlydiminished effect with continued use of the same amountNote: The following items may not apply to cannabis, hallucinogens, or phencyclidine(PCP):8. characteristic withdrawal symptoms (see specific withdrawal syndromes underPsychoactive Substance-induced Organic Mental Disorders)9. substance symptoms of the disturbance have persisted for at least one month, or haveoccurred repeatedly over a longer period of timeB: Some symptoms of the disturbance have persisted for at least one month, or haveoccurred repeatedly over a longer period of time.Criteria for Severity of Psychoactive Substance Dependence:Mild: Few, if any, symptoms in excess of those required to make the diagnosis, andthe symptoms result in no more than mild impairment in occupationalfunctioning or in usual social activities or relationships with others.Moderate: Symptoms or functional impairment between “mild” and “severe”.Severe: Many symptoms in excess of those required to make the diagnosis, and thesymptoms markedly interfere with occupational functioning or with usualsocial activities or relationships with others.(p. 167)In this investigation, all alcoholics are male. Consequently, the pronoun “he” will be used inreference to the alcoholics participating in the study.System: A system has been defined as a complex of elements (subsystems) standingin interaction (Von Bertalanffy, 1968). While distinctions have been made between openand closed systems, recent systemic thought recognizes these distinctions as punctuationsarising out of a participant observer’s point of view as “true” qualities of the system underobservation (Dell, 1985; Keeney, 1983). Consequently, a system and the subsystems whichconstitute it are determined by the level of observation established by the observer. Forexample, just as the respiratory system in an individual can be studied as a system unto itselfbeing comprised of several interacting elements or subsystems (i.e., diaphragm, lungs,11windpipe), so too can the respiratory system be viewed as one subsystem along with otherssuch as the digestive, circulating and neural elements that work in concert to enable thesystem of the individual to maintain his/her physical existence. In this study, systems will beidentified at different levels of abstraction (e.g., intrapersonal, marital, and family). It isunderstood that some systems may be viewed as subsystems of another system at a higherlevel of observation.A systemic perspective is a point of view that focuses on relationships amongelements while maintaining an appreciation for the whole system in which the relationshipsare manifest. This effort seeks to derive a synthesis, and thereby, as Banathy (1987) notes;“capture and define whatever is emerging from our synthesis at a higher level ofunderstanding” (p. 127). A systemic lens is one which particularly attends to the deepinterconnected and interactive essence of things and understands phenomena as inseparablefrom the ecological contexts in which they exist.Experiential Systemic Therapy: In this study, ExST is defined as the systemicallybased experiential and symbolic treatment model specified in manual form which isavailable upon request. An overview of the model (Friesen, Grigg, & Newman, 1991) ispresented in Appendix B. The therapy is designed to address problematic relationshipsthrough a process of intensification which generates relationally novel patterns of interactionand a deeper awareness of interconnection (Friesen et al., 1989).Supported Feedback Therapy: This refers to the comparative treatment conditiongenerated for and employed in this investigation. The Supported Feedback Therapy model(SFT) (Grigg, Friesen, Weir, & Bate, 1991) is a procedure that accents client’s responsibilityand capacity to change. SFT has been articulated in the manual form and is available uponrequest. SFT is a therapeutic procedure that employs charts as visual aids to providefeedback to clients about a variety of areas of life that are of concern to alcoholics and are12monitored on a weekly basis. The SFT weekly monitoring forms and photo-reduced versionsof the feedback charts are provided in Appendix C.Treatment Format: In this study the term treatment format is used to describe theconstituents of the therapeutic system. Format is defined as: the style or manner of anarrangement or procedure (Allen, 1990) and consequently treatment format refers to thecomposition of a given therapeutic system (e.g., individual, couple, family).13CHAPTER 2: LITERATURE REVIEWThis section reviews the literature salient to the investigation. To begin the review,issues regarding the definition of alcoholism are addressed before the biopsychosocial viewof alcoholism is presented. A brief overview of the family systems perspective is thenoffered and followed by a consideration of systemic models which integrate individualtreatment into their formulations. A careful examination of studies focused on the systemictreatment of adult alcoholism is made and subsequently, a look at the research ontherapeutic formats is undertaken. A brief overview of the ExST treatment and therationale for its development are presented, followed by a summary of the literaturereviewed. A statement of the research hypothesis being tested in the investigation concludesthe second chapter.AlcoholismA difficulty in the conceptualization of alcoholism involves defining what constitutesthe problem. There is considerable controversy in the literature regarding the make-up ofthe essential features of the addictive process (Donovan, 1986; Shaffer & Milkman, 1985).Alcoholism has been broadly defined by Keller, McCormick, and Efron (1982) as:the repetitive intake of alcoholic beverages to a degree that harms the drinkerin health or socially, or economically, with indication of inability consistentlyto control the occasion or amount of drinking. (p. 20)The difficulty with this definition is that it is not as precise as needed. Terms such as“harm” or “inability to control” lack the specificity that would allow investigators to agree onindividual cases and, consequently, confusion has often characterized the field.The definition of alcoholism has long been a troublesome issue. Christie and Bruun(1969) characterized the result of efforts to define alcoholism as “a conceptual mess” in asmuch as different things were being discussed under the same label and the same things14were being talked about under different labels. Many definitions and sets of clinical criteriaof alcoholism have been published over the years including a core syndrome of alcoholdependency (Edwards & Gross, 1976), and a culturally sensitive view proposed by the WorldHealth Organization (Edwards, Gross, Keller, & Moser, 1976). Other definitions based onassessing the amount of harm caused by the drinking (American Psychiatric Association,1980; Feigher, Robins, Guze, Woodruff, Winokur, & Munoz, 1972) and perspectives basedon the quantity and quality of alcohol problems (Calahan, 1970) and the loss of the ability tocontrol alcohol drinking (Jellinek, 1960) have also been proposed.Each of these definitions has placed a different emphasis on the aspects ofphysiological, behavioural and social functioning in their descriptions. Consequently,concerns have been expressed regarding the reliability of diagnostic classifications based onthese models (Jolly, Fleece, Galanos, Milby, & Ritter, 1983; Pattison, 1981). Noting that theincidence of alcoholism diagnosed depends on the definition employed, Boyd, Weissman,Thompson, and Myers (1983) demonstrated that by using the seven prominent clinical andresearch criteria, as little as 47% of subjects diagnosed by one set of criteria as alcoholicwould be diagnosed in a similar way by another diagnostic system.The most current diagnostic system has been presented in the Diagnostic andStatistical Manual of Mental Disorders revised edition (DSM-III-R) by the AmericanPsychiatric Association (1987). The chief advantage of this system is that it is basedprincipally on observable behaviour. Helzer (1987) praises the DSM-III-R system describingit as “a robust definition, having both high inter-rater reliability and considerable predictiveutility” (p. 284). In this sense, the present DSM-III-R classification is a substantialimprovement and consequently has been employed in this study. Accordingly, participantsin this investigation all conform to the criteria of severe psychoactive substance dependenceas defined by the American Psychiatric Association (1987).15Biopsychosocial Model of AlcoholismThe field of alcoholism has been complicated by a wide array of models andconflicting theories often linked to different academic disciplines. With proponents of thesevarying perspectives convinced of the validity of their positions, there has been relativelylittle effort to integrate them into a single unified approach (Kissin, 1983). In the last 10years, there has been a growing appreciation for the need to synthesize the variousperspectives and levels of analysis in order to develop a comprehensive view of this problem(Chiuzzi, 1991; Donovan, 1988; Galizio & Maisto, 1985). Such an integrative perspective onalcoholism fits aptly with an integrative treatment model like ExST.The prevailing integrative model of alcoholism has been that adapted from a genericperspective articulated by Schwartz (1982) and has been called the biopsychosocial model(Donovan, 1988). This model has become popular in a wide variety of medical andpsychological domains and it is particularly useful in the addictions field since it recognizesthat such a problem has multiple determinants and that a combination of biological (moreaccurately physiological), psychological and social factors are required for their development(Zucher & Gomberg, 1986). Accordingly, alcoholism is seen as a manifestation of aninteractive process of physiological, psychological and social-environmental factors thatprogress in such a way over time that at some point the dependent condition of alcoholismbecomes evident. Schwartz (1982) has described the biopsychosocial model as organisticand systemic:The essence of systems thinking is that the functioning of a system as a wholeemerges out of the dynamic interactions of its parts (subsystems) and thesystem’s interaction with its environment (the supra system of which thesystem is a part). In terms of medicine, examples of organistic thinkinginclude the belief that specific diseases (constellations of symptoms) representthe complex interaction of specific environmental stresses (including germs)and the organism in question (including its genetic and experiential history)and that biological and behavioral stresses always interact with each other toproduce particular constellations of signs and symptoms in particularindividuals. (p. 1042)16A biopsychosocial model of alcohol dependency bridged the conflicts and differencesof previously mutually exclusive models and a more complex and holistic view of theproblem has emerged (Galizio & Maisto, 1985). This shift in perspective has had a profoundeffect upon aspects of the clinical endeavor, including assessment and treatment.Assessment and treatment activities are intrinsically connected to a perspective of theproblem. In the past, assessment and treatment have tended to be relatively specific andnarrow in definition. The result (from a biopsychosocial perspective) has led assessment andtreatment efforts to ignore entire aspects of the problem and to focus on only a limitedportion of the entire problem.The biopsychosocial model offers a vehicle for a reproachment of previously dividedviews. Offered as a comprehensive view of the etiology of alcoholism (Kissin & Hanan,1982; Zucher & Gumberg, 1986), it has been employed to predict the onset of problemdrinking (Wallace, 1985) and to assess the multiple layers of human existence involved inalcohol dependency (Donovan, 1988; Wanberg & Horn, 1983). The application of abiopsychosocial model in treatment is a recent phenomenon and is reflected in thedevelopment and endorsement of broad spectrum treatment approaches (Institute ofMedicine, 1992; Miller & Hester, 1986) and relapse prevention programs (Chiauzzi, 1991).However, treatment models that assume a biopsychosocial perspective of alcoholdependency remain in an early state of development. Nonetheless, it is now clear that eachalcohol dependent person struggles with a blend of biological, psychological and socialfactors and that treatment efforts must be individually tailored to meet the specifictherapeutic demands of each recovering person. As Kissin and Hanson (1983) noted:As the alcohol dynamic emerges, mechanisms at all three levels appear tooperate sequentially and simultaneously so as to influence the developmentand course of the syndrome. (p. 2)Clearly, therapeutic efforts must address all three levels in order to be maximally helpful ininterrupting the perpetuation of the dependent condition and the eventual degeneration of17the alcoholic individual. The ExST treatment used in this study is informed by thebiopsychosocial perspective on alcohol and assumes this point of view in approaching theneeds of clients.Family Systems Theory OverviewThe family therapy movement with its systemic theoretical perspective had its originsin numerous independent places (Broderick & Schrader, 1981; Guerin, 1979; Olson, 1970),yet by the end of the 1950’s the diverse origins had coalesced into a unified field. Growingout of a disenchantment with the conventional psychotherapeutic treatment of individuals,founders of family therapy laid claim to a treatment procedure which was promoted as bothmore efficient and potent (Bowen, 1976; Jackson, 1959). By 1962, family therapy as apsychotherapy had differentiated itself from the field of psychiatry and psychology and hadformed its own professional associations and published its own journal (Nichols, 1984).The systemic theory which connected the diverse elements of the family therapy fieldwas closely associated with general systems theory (Von Bertalanffy, 1968) which borrowedfrom the biological field and assumed that the system being observed could be objectivelyconsidered as separate from the observer, and cybernetics (Weiner, 1948) which assertedthat all systems share certain attributes such as homeostasis, rules of conduct, andorganization. Consequently, a view of the family evolved that included viewing it as a livingbiological system and at the same time governed by the rules of a machine. This first-ordercybernetic perspective breathed life and vitality into the treatment of the family. The family,when seen as a single system composed of interacting subsystems (individuals), combined insuch a way that the whole was greater than the sum of its parts and created problems toserve some important function for the entire family unit.In this theoretical frame, the family was construed as being governed by rules whichregulate deviation from established patterns or codes of conduct and thus preserve the18integrity of the family or maintain the family homeostasis. Family member problems, oncethe focus of intense individual psychological probing were recast as symptomatic problemswhich communicated something about the entire family. The individual with the problem or“identified patient” was understood as the symptom bearer and it was assumed that thesymptom was generated and maintained by the family as a necessary behaviour to itsequilibrium. Brown (1974) notes that:Systems theory assumes that all important people in the family unit play a partin the way family members function in relation to each other and in the waythe symptom finally erupts ... In systems theory the focus is on the functionalfacts of relationship - on what happened, how it happened, and where andwhen it happened. (p. 30)In this way, the burden of responsibility for symptoms was shifted from the individualinto the relational domain of the family with particular attention being paid to thecommunicative rules, boundaries, organization, roles and control or regulating mechanismswhich operate to maintain the family as it satisfied its needs. This is to say that although anindividual family member’s behaviour might indeed have chaotic and destructive qualities,the family was seen to nevertheless rely on the undesirable behaviour as a functional aspectin the patterns and processes in the family system (Hoffman, 1981).The therapeutic implication of this early systems model was to render the family as athing that was broken and in need of fixing (Keeney, 1983). Consequently, models weredeveloped that out-maneuvered or out-strategized the resistant character of families in orderto help them change (Cecchin et al., 1993).The field of family therapy has grown to expand upon its systemic theory. Whilecurrent innovations have moved the family therapy field away from its organismic andmechanistic formulations of the family in order to explore such notions as the familynarrative (White & Epston, 1990), the family epistemology (Dell, 1985; Keeney, 1983), andfamily ecology (Auerswald, 1985; Bogdan, 1984), the original appreciation for theimportance of the dynamic relationships among family members has continued. For19example, Bodgan (1984) refuted the homeostatic view which reified the family into a thing,but asserted that:the behaviour of family members shows order, pattern, organization orredundancy because the behaviour of each individual is in a sense cognitivelyconsistent with the behaviour of every other individual in the system. Moreexactly, the ideas of each family member lead him to behave in ways thatconfirm or support the ideas of every other family member. (p. 376)The movement away from first-order systemic thinking was nonetheless an importantdevelopment (Anderson & Goolishian, 1993). Disenchanted with many of the practices thatevolved out of a first-order view of the family, theorists branched into two directions:second-order cybernetics and narrative epistemology (Cecchin et al., 1993).The second-order systems view of the family is based upon a deep appreciation forthe cognitive subjectivity of the observer (Sluzki, 1985). From this perspective, the familywas not a thing, but rather a process. The family was an expression of an ecology ofindividuals with various perspectives sharing living experiences which were mediatedthrough language. Through shared experience, family members come to have a profoundinfluence upon one another. Maturana (1978) noted:If the medium is also a structurally plastic system, then the two plastic systemsmay become reciprocally structurally coupled throuh their reciprocalselection of plastic structural changes during their history of interactions. Insuch a case, the structurally plastic changes of state of one system becomeperturbations for the other and vice-versa, in a manner that establishes aninterlocked, mutually selecting, mutually triggering domain of statetrajectories. (p. 36)This kind of thinking shifted the position of the therapist. No longer was the therapistseen as an expert, outside of the family, with a job to fix the dysfunctional machine, butrather a participant observer who was co-constructing relational realities in a linguisticdomain (Keeney, 1983; Maturana & Varela, 1980). In addition, the feminist work ofGoidner (1988) and others, challenged the limitations of first-order cybernetics in explaininggender-related issues and argued that “reality” was the result of a socially constructedprocess. As a result, therapists were urged to consider the unequal social differences20between men and women and to strive to amplify the marginalized voices of women(Taggart, 1985).The post-modern epistemological branch of family therapy moved even farther awayfrom the cybernetic metaphor (Cecchin et al., 1993). This social constructionist perspectiveemerged out of the hermeneutic tradition and asserts that understanding and meaning arecultural, public and inter-subjective (Gergen, 1991). The family is not viewed as a thing fromthis perspective but rather the family is thought of as a linguistic process or story. Humanbeings are seen as immersed in a narrative in which everyone participates and createproblems as well as dissolve them (Goolishian & Anderson, 1992). Thus, the therapist’s taskis to participate as a conversation artist in therapeutic discussions with clients in an effort toco-develop new meanings, new realities and new narratives (Anderson & Goolishian, 1988).The lively debates regarding epistemology and practice continue in the field of familytherapy (e.g., Cecchin et al., 1993; Goolishian & Anderson, 1992; Simon, 1992), however, itseems clear that for the most part, first-order systems view of family and therapy is a thing ofthe past. Presently, even strategic therapy models of treatment traditionally built on thefirst-order cybernetic metaphor and fashioned after the Mental Research Instituteformulations have distanced them from the first-order systems view. As Eron and Lund(1993) state:In line with constructivist thinking, problems are. seen as caused bydefective family systems, and problems arej seen as serving some sort offamily function, that is, preserving stability, regulating boundaries orhierarchies. (p. 294)Despite its development, the field of family therapy continues its tradition ofunderstanding individual problems with a contextual appreciation. All behaviours are stillunderstood as occurring within a relational field and consequently the meaning of a givenbehaviour (as communication) can only be fully appreciated within the relational setting thatit occurs. In this way, an individual’s behaviour (i.e., symptom) is seen as part of an21interactive dynamic process or discourse involving others and not seen as a decontextualizedisolated individual statement.Systemic Theory and Individual TreatmentThe history of family therapy included a break from the mainstream individualtreatments of psychiatry and psychology. Family systems therapy initially worked exclusivelywith family and marital units (i.e., two or more people) and this helped build a separateidentity for systemic therapists that was easily distinguished from individual treatmentmodels (Becvar & Becvar, 1988). In this way, theory (i.e., systemic) and format (i.e., conjointfamily) were initially fused and spoken of as synonyms. Consequently, from a historicalperspective individual and family psychotherapeutic formats have been conceptualized ascompeting orientations and mutually exclusive forms of treatment (Pinsof 1983). Whilemany, such as Whitaker (personal communication, 1990), maintain firm convictions to thebelief that in order to work systemically one must work with the entire family system, recentpositions have been asserted regarding the possibility of working systemically withindividuals (Friesen, et al., 1989; Pinsof, 1992; Steinglass, 1991; Schwartz, 1987a, 1989).With the advent of second-cybernetic models (Sluzki, 1985) and the recognition ofthe importance of the observer (Varela, 1979), therapists were no longer seen as objectiveparties making true observations of reality. Rather, therapists were viewed as participants tothe observation. Since the therapist’s perspective and propensity to make distinctions wasunderstood as having very much to do with what was eventually observed (Dell, 1985), it waslogical to question what constituted a legitimate system for a systemic therapist to work within treatment.Some recent texts on family therapy (e.g., Becvar & Becvar, 1988) have suggested thatsystemic treatment can be done with individuals so long as the therapist continues to22consider the individual to be a part of a larger system. Despite such assertions, very fewmodels are currently in existence which strive to bridge systemic theory with an individualtreatment format and attempt to negotiate the integrative task required in so doing (Friesen,Grigg, Peel, Newman, 1989; Pinsof, 1983; Schwartz, 1988). The following section reviews theintegrative family therapy models of treatment noted above.Integrative Problem-Centered TherapyPinsof (1983) articulated Integrative Problem-Centered Therapy (IPCT) as anintegrative and comprehensive framework therapeutic practice. IPCT was described as amodel of psychotherapy that not only combined individual and family treatment modalities,but also linked behavioural, communication and psychodynamic theories of therapy.Described by Pinsof (1983) as a “systemically eclectic and comprehensive model” (p. 20),IPCT construes problems to be the result of unsuccessful problem solving that stem from“blocks” in healthy problem solving processes within the patient system. IPCT is built uponthe assumption that each of the various theories of therapy and their corresponding modesof implementation (individual, couples, family), have their own domains of expertise as wellas limitations. Since it is asserted that no single treatment format or orientation iscomprehensive enough to meet the demands of problems brought to treatment, the IPCTmodel strives to provide a theoretical framework through which a therapist can determinewhich model of treatment and which format of therapy would be the most appropriate tomodify the patient system’s block to successful problem solving. This is to say that IPCTprovides a set of principles for applying different treatment formats, theoretical models, andtechniques in order to maximize both efficiency and effectiveness.Pinsof (1983) introduced the concept of “the patient system” in order to resolve theproblem of defining the patient. In individual therapy, the individual is the patient.Similarly, in marital therapy the patient is the couple, and in family therapy the patient is23seen as the entire family system. Pinsof defined the “patient system” as consisting of “all thehuman systems (biological, individual-psychological, familial-interpersonal, sociooccupational etc.) that are or may be involved in the maintenance or resolution of thepresenting problem” (p. 20). The task of the therapist in treating the patient system is not tochoose between these levels of system but rather to identify and address the relativecontribution of each to both problem maintenance and resolution.The IPTC model is ambitious in its undertaking. By suggesting a problem solvingsupra-theory, it strives to link a variety of otherwise self contained theories and therapyformats to treatment. The model seems to recognize that there is a time and a place toemploy various treatment formats and procedures and strives to provide a blueprint to aid indetermining when to employ a particular approach.A difficulty in the implementation of the model may be the demands it places on thetherapist. An IPCT therapist would need to be competent and comfortable in moving fromone therapeutic role and style to another, including behavioural, structural, interpersonal,and psychodynamic. An IPTC therapist would have to be further versed in workingeffectively in individual, couple and family treatment formats. Such a therapist wouldobviously require considerable training; however this is not an impossible task. A secondconcern revolves around the process of moving from one guiding theory to another astreatment progressed. This altering of therapeutic position could be bewildering to clients.Since clients learn what is required of them in therapy, it seems possible that an IPTC clientmight be unsure of what is expected of them from one moment to another, as a new set ofexpectations and demands were presented to them as the therapist shifted models.Although IPTC strives to integrate the various treatment models, it can be questionedas to the degree to which it bridges the theoretical distances between them. While it clearly24provides a more systematic way of being theoretically and technically eclectic, the modelfails to explain the relationships between the various levels of system except to note that theymay (or may not) be involved in problem maintenance and resolution. Embedded in themodel is the notion that the more resilient the problem the more likely one will end upemploying individual treatment formats. This is to say that the model proposes workingwithin a family and/or marital treatment contexts first before finally resorting to “the type ofwork associated with long-term individual psychotherapy” (p. 31).The IPCT model is in its early stages of development and without empirical supportor widespread clinical demonstration. Until such work is conducted, the model is perhapsbest viewed as one step towards the reconciliation of individual and family treatmenttheories.Internal Family SystemsThe Internal Family Systems model (IFS) developed by Schwartz (1987a, 1987b,1988, 1989) provides a theoretical framework which more specifically establishes integrativelinkages between individual formats of treatment and family systems therapeuticapproaches. Schwartz (1989) asserted that:it is possible to intervene at either level using the same systemic paradigm andtechniques, rather than having to shift from a systemic at the family level to,for example, a psychodynamic or cognitive/behavioural paradigm at theinternal level. (p. 91)In contrast to Pinsof’s (1983) IPTC approach, the IFS model provides a theoreticalframework which enables therapists to move from individual to marital and family treatmentformats while maintaining the same theoretical model.Schwartz (1988) recognized the contribution family therapy has made in viewingbehaviour as intricately related to the social context in which it is embedded. It is clear thatit has been helpful to relate symptoms to patterns of interaction and to evoke concepts liketriangulation, boundaries and hierarchies. However, Schwartz (198Th) also noted that family25systems paradigms can become restrictive and narrow themselves when they imply thattherapists should not consider the individual’s intrapersonal process in their assessment andtreatment because the internal territory has been portrayed as distracting, unimportant orintrinsically non-systemic.The integrative synthesis of the IFS model is made possible by the recognition thatthe individual can be seen as being composed of by a variety of “parts” which exist ininteraction with one another. Drawing on the neuro-psychological works of Gazzaniga(1985) and Ornstein (1986), Schwartz (1987a) established the notion of the multiplicity ofmind. That is, the intrapersonal domain of existence is viewed as a community, family ofparts, or a tribe of mental systems that reside in each individual. The sub-personalities, likemembers of an “external” family struggle for influence, interact sequentially and form avariety of alliances and organizations. The parts may exist in peaceful co-existence or theymay have conflictual or even contemptuous relationships. Each part is viewed as beingorganized around a particular premise or set of premises about the world and how to existwithin it.Through his work with bulimics, Schwartz (1989) came to the conclusion that onecould work with the external family and get it working well and still have the symptomcontinue to be exhibited. He also became aware that the internal relations of his individualclients closely resemble the relationships in the external family prior to treatment. With thisinsight, Schwartz (1987a) began to explore the “inner” family network of parts and torecapitulate the therapeutic work of the external family within the intrapersonal domain.The view of the individual as being composed by many parts or sub-personalities is not anentirely novel idea. Nonetheless, the IFS model clearly establishes a bridge between thesystemic thinking of family therapy and established individual therapy approaches. Schwartz(1987a) recognized that there are some clear similarities between the IFS model and object26relations theory. He also noted theoretical overlap with Gestalt Therapy (Pens, 1951),Psychosynthesis (Assagioli, 1973), Voice Dialogue (Stone & Winkelman, 1985), and NeuroLinguistic Programming (Grindler & Bandler, 1982).The relationship between the intrapersonal and interpersonal processes has beenidentified as the most interesting and underdeveloped aspect of the IFS model (Schwartz,1987a). There are clear parallels between these internal and external families, howevertheir linkage is yet to be articulated. Interpersonal patterns observable in the family havealso been observed in the relationships between parts. For example, in abusive systems,denial at the family level is also replicated at the internal level. Interventions at one levelare hypothesized to effect parallel relations at another level, however this assumption wouldseem to contradict the very history of the model in which changes in the external familysystem did not necessarily transfer into changes in the internal family level.Clearly more work is required to understand the connection between the levels ofsystem in order to assist in guiding therapeutic intervention. By providing a much neededtheoretical bridge between family therapy practice and individual treatment modalities, theIFS model has made a valuable contribution to the rapprochement of individual and familysystems models. The model is the theoretical product of clinical work which has provided itwith a sense or practical credibility. Nonetheless, no outcome studies have been conductedwhich empirically test the model’s application. This would seem the obvious next step in thedevelopment of the approach. Finally, it should be noted that Schwartz (1987b) employed astructural-strategic model of clinical intervention in implementing the IFS model. One mustassume that other models of family therapy can be introduced into the intrapersonal domainthrough the door that the IFS model provides and in fact such examples are beginning toappear in the literature (e.g., Bryant, Kossler, & Shirar, 1992). Consequently, the IFS model27may best be viewed as providing an integrative theoretical bridge which allows for theapplication of a variety of modes of family therapy in the treatment of individuals.Experiential Systemic TherapyThe Experiential Systemic Therapy (ExST) model (Friesen et al., 1989) wasdeveloped as an integrative treatment approach that synthesized individual and familytherapy concepts and techniques. Meant to provide a unified set of assumptions andconcepts that could be equally applied to individuals, couples, and families, the ExSTapproach attends to the multiple layers of human experience and affirms theirinterconnectedness.Built upon the notion that existence is a relational phenomenon (Friesen, Grigg, &Newman, 1991), ExST portrays human existence as an intricate web of systemic connection.Like the IFS model, ExST recognizes the intrapersonal system as being a unity comprised ofa variety of parts that are in a dynamic state of relationship just as the interpersonal systemof the family is made up of a variety of people interacting with one another. However, theExST model extends this kind of thinking further than does the IFS model by including thephysical and biophysiological levels of system at one end of its systemic perspective andinternational and global systems at the other (Newman, 1990).Noting that the experiences that arise from the interactive process between a childand his/her caregivers are influential in a child’s way of being in the world, ExST employsthe work of Maturana (1978) to explain how two or more interactive systems come toestablish patterns over time. While Maturana (1978) is chiefly concerned with thedevelopment of interpersonal relations, Newman (1990) argued that his formulations couldbe applied to any number of interacting systems including those within the intrapersonaldomain.28Friesen et a!. (1989) introduced the descriptive term “substantive relational theme” asa notion to represent the underlying essence of clients relationally based stories such asunlovableness, abandonment, rejection, and unworthiness. Similar to IFS, ExST theoristsrecognized relational parallels shared between the patterns of relationship manifest in thevarious levels of systemic existence. The substantive relational theme concept is meant toaddress the central relational current which runs consistently through the intrapersonal andinterpersonal domains. As a result, the substantive relational themes of clients are the fociof treatment efforts regardless of treatment format in ExST.Similar to Pinsof (1983), ExST theory asserts that it is important to assess which levelof the client system is most amenable to change and suggests that therapists begin at thelevel of system with the most available resources for transformation. The ExST modelassumes that various formats of treatment will be employed as they are required to addressthe various contributions made by the interactive processes at each level of the system to thecontinuation of the repetitive, restrictive and rigid relational theme. In this way, thesubstantive relational theme construct links the physiological, intrapersonal andinterpersonal systems of marriage and family asserting that the systems will tend to beisomorphic to one another as they share the same relational story.ExST is an experiential form of treatment; however like the IFS model, thetheoretical formulations which enable ExST to be applied in individual, marital and familytreatment formats can also enable other treatment models to be implemented in a similartrans-systemic fashion. ExST is the product of therapeutic experience, and as such it isgrounded in clinical practice. It is currently being empirically tested in a large outcome andprocess study.29Couples and Family Therapy of Adult Alcoholism: Treatment Outcome ResearchThe following section is restricted to research probing the treatment of adultalcoholics using couples and family therapy. In the interest of brevity and precision,treatment research on substance abusing adolescents using family therapy approaches (e.g.,Joanning, Quinn, Thomas, & Mullen, 1992; Lewis, Piercy, Sprenide, & Trepper, 1991;Liddle, Dakof, Parker, Garcia, Diamond, & Barrett, 1993; Stanton & Todd, 1982) has beenexcluded from the literature review.1956 - 1973The application of therapeutic procedures that expanded the individual treatmentscope and recognized the need to treat both the alcoholic and his spouse occurred in the mid1950’s. The approach at this time was to offer concurrent but separate group psychotherapymeetings to alcoholics and their wives (Ewing, Long, & Wenzel, 1961 ;Gliedman, 1957;Gliedman, Rosenthal, Frank, & Nash, 1956; Macdonald, 1958; Vogel, 1957). These papers,which were focused on the conduct of what was identified as the traditional manner ofpsychoanalytically oriented psychotherapy, were not research studies. Rather, thesepublications reported on experimental efforts with a new tact in alcoholism treatment.While none of these early studies employed outcome measures, the impressions of outcomeby the authors were favourable and supported the further development of “family treatment.”Mention of working with alcoholics and their wives together in couples groupsappeared at the end of the 1950’s (Gliedman, 1957). However, it is not until the early 1960’sand later that couples group treatment gained momentum (Burton, 1962; Burton & Kaplan,1968a, 1968b; Ewing & Long, 1961; Gallant, Rick, Bay, & Terranova, 1970; Sands &Hanson, 1971).30Reports of family therapy methods applied to the problem of alcoholism appear inthe literature in the late 1960’s (Esser, 1968; Ewing & Fox, 1968). The conjoint therapymodel of Satir (1964) was employed in the early 1970’s (Esser, 1970, 1971; Meeks & Kelly,1970) with some favourable trends in outcome being noted.By today’s standards, these early studies suffered from methodological shortcomingsincluding reports on unspecified treatment procedures, small unrepresentative sample sizesand poorly defined impressionistic outcome indices. However, these initial reports weredescriptive of new approaches and not meant as definitive research efforts. The couples andfamily theoretical frames implicit in many of these works were in formative stages ofdevelopment. Nonetheless, taken together the enthusiasm of many testimonials and casereports offered a measure of credibility which was later supported by the preliminaryempirical efforts of Cadogan (1973) and Corder, Corder, and Laidlaw (1972).These early articles have been reviewed by others (e.g., Miller & Hester, 1980;Steinglass, 1976) and will not be examined further in this document. This cursory review isoffered to provide a historical account of the pioneer work which substantiated the positiveevaluation of systemic approaches to alcoholism that was noted in the day. In particular, thepivotal review of alcohol treatments made by Keller (1974) praised marital and familytherapy approaches as “the most notable current advance in the area of psychotherapy” (p.116). Consequently, a call for controlled outcome studies to evaluate this promisingdirection in treatment was issued at this time.1974 - 1980Between the years of 1974 to 1980 several studies probing the utility of a variety offamily systems oriented treatment approaches were reported (Janzen, 1977). In keeping31with earlier trends, the studies had limitations due to methodological difficulties, howevertheir results were nonetheless favourable and enthusiastic.Davis, Berenson, Steinglass, and Davis (1974) articulated the “adaptive consequences”model of alcoholism. Davis et al. (1974) clearly detailed a link between alcoholconsumption and the social conditions which surround the behaviour by use of four caseexamples. The central thesis of this position was that alcohol consumption was associatedwith behaviours which were clearly adaptive in social situations, but which were somehowexcluded from alcoholics repertoire of behaviour when not drinking (e.g., assertiveness,expressiveness, problem resolution, etc.). While the argument put forward by the authorswas compelling and the treatment implications were clear, there were no results of treatmentpresented which employed the adaptive consequence formulation reported.Some empirical support for family models of treatment was offered by Hedberg andCampbell (1974). In a study comparing behavioural family therapy with other individuallyoriented behavioural treatments including systematic desensitization, covert sensitizationand shock presentation, the family approach was found to be the most effective method oftherapy. The design included assigning 49 subjects to treatment groups and employed a goalattainment measure of success in which improvement was apparently ascertained through aninterview with the patient, the patient’s spouse and the patient’s therapist. No statisticalanalyses were used to test for significant differences between treatment groups and it is notclear from the report how conflicting estimates of improvement between parties washandled. Results were based on a percentage of subjects either attaining their goals orimproving substantially in relation to their goals. Trends in both outcome categoriesrevealed that the family counselling treatment group was the most effective of the fourtreatments offered at the 6 month follow up assessment point.32Building on an earlier pilot work by Hunt and Azrin (1973), Azrin (1976) conducted astudy comparing a community reinforcement procedure with an intensive inpatient hospitaltreatment program. The community reinforcement program was based on a social learningtheory and included marital and family counselling as principal aspects of treatment.Described in some detail by Azrin, Naster, and Jones (1973), the marital therapy portion oftreatment addressed all areas of the marriage. Built on the work of Stuart (1969), theprocedure included maximizing things which were satisfying to each member of thepartnership. This maximizing of satisfactions was done in the context of reciprocalsatisfactions being received from a partner in a kind of “you scratch my back and I’ll scratchyourst’fashion.Twenty men admitted to a hospital for treatment for alcoholism participated in thestudy. The men were matched into 10 pairs and each pair was randomly assigned to eitherthe experimental community reinforcement or the individually oriented in patient hospitaltreatment control group.Standardized outcome measures were not used and procedures were difficult tofollow. Apparently, interviews were employed at post-treatment and at a 2 year follow-up toascertain drinking related criteria including how much drinking participants were engagedin, how much work they had engaged in, how much work they had missed and how muchthey were absent from home. However, there is no indication of pre-treatment assessmenton these variables reported. Results were provided in terms of percentage comparisons andno statistical procedures were employed. Nonetheless, the results strongly supported the useof this procedure even at the 2 year follow-up when compared to the regular individualhospital treatment which was, unfortunately, only minimally described in the study.The provision of marital treatment was an aspect of the treatment proceduresemployed in a study by Edwards, Orford, Egert, Guthrie, Hawker, Hensman, Mitcheson,Oppenheimer, and Taylor (1977). In this study 100 alcoholic married men were randomly33assigned to either a “minimal” or “maximal” treatment condition. While the minimaltreatment group received a single session of advice, the maximal treatment group receivedan array of treatments which included conjoint sessions with husband and wife whenappropriate. Unfortunately, the clinical aspects of the treatments were not detailed andtherefore a clear understanding of what was done in the treatment is impossible. It seemsclear that the maximal treatment was not particularly effective since there were nostatistically significant differences between the groups observed at 12 month follow-up datacollection occasion. An interesting feature which emerged from this research effort thatseems to have relevancy to treatment was reported by Oppenheimer, Egert, Hensman, andGuthrie (1976) who concluded that regardless of treatment condition, marital cohesivenesswas found to be the central variable which predicted positive treatment outcome.Marital and family treatment for alcoholism was also reported in a clinical reportconducted in Israel (Amir & Elder, 1978). Regretfully, as in Edwards et a!. (1977) it isimpossible to discern from the report what the therapeutic procedures entailed. However,unlike Edwards et a!. (1977), the outcome results of this treatment, which focused on thefamily and the community, appeared positive, though limited by a small sample size andpoorly defined outcome criteria.Steinglass, Davis, and Berenson (1977) and Steinglass (1979) reported on aninnovative marital treatment study that built on the adaptive consequence model which waspreviously presented by Davis et al. (1974). The approach taken in this experiment wasparticularly innovative and controversial to many in the field because it included theprovision of alcohol in an inpatient treatment setting. Recognizing that families ofalcoholics were entirely organized around alcohol consumption, Steinglass et al. (1977)coined the terms “alcoholic systems” to describe the participants in the study.34While the original design of the study called for 40 to 50 couples and included a wait-list control group, clients proved to be difficult to recruit. Only 10 couples were eventuallyrecruited for the study and all were desperate for a solution to their problems having longhistories of previous treatment failure. Thus, the study was recast from a treatment outcomestudy into a clinical pilot project. The authors cited political difficulties with the AlcoholicAnonymous lobby, who took exception to the experimentally induced intoxication aspect ofthe treatment program, as the chief reason for the departure from the original study design.The treatment procedures included daily 90 minute multiple-couples groups led by twopsychiatrists trained in family therapy. It is noteworthy that the two therapists varied inunspecified ways in their treatment approaches. The main consistent element of the therapyoffered by these therapists was a common systemic theoretical approach that viewed alcoholintoxication as a functional aspect of couple interaction. Both therapists assumed that thedrinking problem could best be understood by contrasting the drinking behaviour with “dry”phases of the couples’ relationship. The ten subjects were divided into four groups becausethe treatment facility maximum was three couples. The research protocol was divided intothree phases: an initial two week out-patient phase in which participants met for threesessions per week; a 10 week in-patient phase; and finally, a post-hospitalization three weekout-patient phase consisting of two group meetings per week. Following these proceduresgroups continued to meet at six week intervals for a six month period.Therapists observed the couples’ interaction throughout the day and assessed theirinteractional patterns. Couples were encouraged to go about their normal routines beingunrestricted by hospital demands including shopping and cooking meals, arrangingrecreational activities and determining sleeping arrangements.A laudable design feature of the study was to employ standardized state of the artresearch instruments as outcome indices. This aspect of the research broadened the scopeof outcome assessment to include individual symptomology and couple functioning as well asdrinking behaviour.35The treatment effects were rather muted. Results indicated that one of the therapistsseemed to be correlated with more improved functioning. Statistical analyses were notreported, but Steinglass (1979) noted that treatment effects were slight and nonsignificant.The nonsignificant treatment result was softened by the authors who pointed to thechronicity of the subjects’ drinking problems. Nonetheless, the study offered compellingobservations which lent support to an interactional view of alcoholism with clear treatmentimplications.Paolino and McCrady (1976) provided a case report of an innovation in treatmentthat included the joint admission into hospital of both alcoholics and their spouses. Thistreatment approach was the topic of a pilot study reported by McCrady, Paolino,Longabough, and Rossi (1979). The study, which represented the first controlled study ofjoint admission, reportedly assigned 33 patients to one of three experimental groups. Thereader is left confused as to how the random assigmnent resulted in a substantially unevensubject assignment to treatment groups. The 3 treatment groups were: (1) joint admission,in which the alcoholic’s spouse lived in the hospital with the alcoholic. The couple attendedcouples group therapy conjointly and both attended individual therapy groups separately, (2)couple involvement, in which the therapeutic procedures were the same as the jointadmission group except the spouse did not live in the hospital, and (3) individualinvolvement in which the alcoholic attended individual therapy groups but the spouse wasnot included in the treatment protocol. Like Steinglass (1979), this study included standardinstruments measuring alcohol involvement, psychological disturbance and maritaladjustment.Despite a relatively small sample size and a corresponding lack of statistical power,the results of parametric and nonparametric statistics on the pretest and follow-up showedsuperior and comparable outcomes for the two groups which included the alcoholic and hisspouse. Unfortunately, the treatment procedures of the groups employed in this study were36not made very clear however care was taken to provide regular supervision of the therapists.These results clearly supported the notion that marital and family approaches to theproblem seem to be more effective than treatments that do not attend to these contexts.As the 1970’s drew to a close, family oriented treatment approaches to alcoholismseemed to be building a measure of empirical support for their application. However, thereremained a clear need for further substantiation. Having documented and reviewed themany treatments available to alcoholics, Miller and Hester (1980) somewhat cautiously statethat “systemic family treatment approaches seem to be a method deserving of furtherexploration in the treatment of problem drinkers” (p. 61).1980 - 1986The assessment of marital and family therapy as a treatment for alcohol dependencyimproved by the time Miller and Hester (1986) wrote their second major review of alcoholtreatment studies. At the time of the second review, these authors had refined theirassessment to include only marital therapy which was viewed as having been found topromote the maintenance of sobriety. Miller and Hester (1986) conclude that the“consistency of positive findings at short follow-up certainly warrants further investigation,and indicates that marital therapy is a worthwhile modality to consider for inclusion inalcoholism treatment” (p. 139). This improvement in the reviewers’ assessment would seemto have been based upon the addition of two more well designed studies that addedadditional weight to the literature that had been accumulating steadily since the earlieststudies first appeared.O’Farrell, Cutter, and Floyd (1985) reported on a well designed study that comparedindividual outpatient counselling with two treatment groups that in addition to receivingindividual treatment also participated in either behavioural marital therapy or an37interactional couples group therapy. Thirty-four subjects were randomly assigned to threetreatment groups. Established self report indices of marital functioning as well asvideotaped interactional tasks were employed as outcome measures in this study. Post-treatment data were collected 12 to 14 weeks after treatment was initiated. No follow-updata were reported at this time. The results supported the hypothesis that the behaviouralmarital therapy was more effective on some of the marital measures than the interactionalcouples therapy and the individual counselling treatment alone. No significant differencesbetween the interactional couples treatment and the individual counselling were reported.However, the interactional treatment couples did improve in terms of their desiredrelationship and positive communication while the individual clients in the control group didnot.The study results must be considered with some degree of caution. The sample sizewas quite small (i.e., n = 10 to 12 per group) which limits the statistical power to identifydifference (Kasdin, 1986). In addition, the delivery of the interactional couples grouptreatment was interrupted by a change in therapist two-thirds of the way through the study.While the treatment procedures for the behavioural treatment were clearly spelled out(O’Farrell & Cutter, 1984), the interactional marital treatment and the individualcounselling procedures of the comparison group were not specified in the report makinginterpretation difficult.The second important study in this time period was presented by McCrady, Moreau,Paolino, and Longabough (1982) who reported the results of a four year follow-up of theMcCrady et al. (1979) study reviewed earlier. After reviewing the data collected four yearsafter the therapy had been completed, the authors concluded that the superior treatmenteffects of the two versions of behavioural marital therapy (joint admission or outpatient)over the individual counselling treatment control group in the earlier study had not been38sustained. This result was interpreted as indicating that marital therapy has an importantshort-term impact but not necessarily an enduring advantage over individual treatment.The remaining studies reported prior to Miller and Hesters (1986) review did notcontribute substantially to an understanding of the efficacy of treatment. A single case studyof Relationship Enhancement therapy as specified by Guerney (1977) was reported byWaldo and Guerney (1983). The results were anecdotal with the authors reportingfavourable outcomes of continued abstinence and improved marital relationship at a sixmonth follow-up.The remaining two studies published prior to 1986 reported on studies in progress.Zweben and Perlman (1983) described their research project design but had no results toreport at the time of the article. Bennun (1985) provided case examples which supportedboth family problem solving and systems therapy, however these data were anecdotal innature.1986 - 1993Several studies have been published after the 1986 review by Miller and Hester thatcontribute to the body of literature regarding family and couples treatment of alcoholdependency.Extending the pilot work and case study published earlier (McCrady, 1982; McCradyet a!., 1979), McCrady, Noel, Abrams, Stout, Nelson, and Hay (1986) conducted a study inwhich 45 alcoholics and their spouses were randomly assigned to one of three outpatientbehavioural treatments: (1) minimal spouse involvement (MSI) that required spouses toattend and observe the alcoholic’s individual therapy, (2) alcohol-focused spouseinvolvement (AFSI) which included the teaching of skills to the spouse aimed at enablingbetter handling of drinking situations in addition to the treatment offered in the MSI39condition, and (3) alcohol behavioural marital therapy (ABMT) that incorporated all theskills taught in the MSI and AFSI conditions with specific behavioural marital therapyinterventions. Unfortunately, while brief synopses of the treatments are provided in thereport, there is no mention of them being specified in manual form. Recognized indices ofalcohol consumption and marital functioning were employed as outcome measures. Whileall three groups improved to a statistically significant level, results showed few statisticallysignificant differences between the groups. The differences that did exist provided somesupport for the view that the marital benefits of ABMT were more stable than the othertreatment groups in that they decayed at a slower rate and to a lesser degree. Again, arelatively small sample size limited the statistical power of the study. Nonetheless, theseresults suggest that a treatment built upon the premise that there is a reciprocal relationshipbetween drinking and family functioning results in a more lasting positive change thentreatments which do not recognize the salience of the familial context to recovery.The results of the research presented by Bennun (1986) completed the study firstreported in Bennun (1985). While the study had sought to test the efficacy of the Milan andsystemic problem solving approaches in treating alcoholism, the focus on alcohol treatmenthad been lost by the time of the 1986 report. Over the course of the study, a number ofsubjects were included that were not experiencing alcohol problems. No reason for thischange in research procedures was provided. Methodological problems concerningtreatment implementation and standardization and the use of only a single index of alcoholdependency and no measures of marital or familial functioning limit the utility of thisresearch. Consequently, while the study offered some support for the two systemicapproaches that were being investigated, lamentably their application in the alcohol fieldwas not advanced in any substantive way.40Zweben, Peariman, and Li (1988) reported on the outcomes of a large comparativetreatment project that contrasted a systemic marital treatment approach with a single sessionof advice counselling. Similar to the Edwards et al. (1977) study of “minimal” and “maximal”treatment reviewed earlier, alcoholics and their spouses were randomly assigned to one ofthe treatment groups. While a sample size of 218 couples began the study, 116 couplescompleted the ambitious project which included a pre, mid, and post treatment design, aswell as 6, 12, and 18 month follow-up data collecting occasions.The marital therapy in this study was not operationalized in manual form and onlyscant descriptions of the treatment is provided. The authors do note that the marital therapywas fashioned upon the adaptive consequence model of alcoholism (Davis et al., 1974;Steinglass et al., 1989), and that the advice counselling was patterned after the work ofEdwards et al. (1977). An array of recognized alcohol measures and indices of maritalfunctioning were employed to test whether eight sessions of marital treatment were moreeffective than a single meeting of advice counselling. Both groups improved to a statisticallysignificant degree on the main marital measure over the course of the study, however, theclinical significance of the improvement is questionable and no substantive between groupdifferences emerged.Again methodological limitations of the study influence the interpretation. The chiefconcern regarding the treatment implementation expressed by the authors centered on therelatively short duration of the marital treatment. In this regard, Zweben et al. (1988)hypothesized that the treatment period may have been simply too short and consequently thetreatment was too weak to actually make a difference. A second difficulty with the study wasthe fact that couples that entered the study scored in the satisfied range on the maritalmeasures at pretest and consequently a ceiling effect on the measure likely precluded anypossibility of positive treatment outcomes on the independent measure of maritalfunctioning. In addition, the authors recognized the possibility of a placebo effect associatedwith the extensive researcher contact required by the data collection procedures.41Notwithstanding these issues, the results were essentially consistent with the study byEdwards et al. (1977) suggesting that a single one hour meeting of advice counselling wasvery effective and that marital treatment benefits could not be differentiated from theminimal contact group.Chapman and Juggens (1988) reported on an experimental study of three treatmentprograms conducted in New Zealand. In this research, 113 alcoholics were randomlyassigned to a six week inpatient program, a six week outpatient program or a singleconfrontational interview. Although the treatments were clearly oriented to the individual,spouses apparently participated in all the treatments in an unspecified manner. Thetreatment approaches were not put in manual form or standardized and were described as“eclectic.” Outcome measures administered at pre-treatment and 6 month and 18 monthfollow-ups included alcohol indices as well as some psycho-social measures. While nostatistical results of the psycho-social tests were included, it appeared that once again nosignificant difference emerged between the three treatment programs. The description ofthe study is somewhat scant. It is impossible to ascertain what role marital or family therapyplayed in the treatments by reading the report and it is unclear what the actual treatmentprocedures involved. In terms of treatment description, the authors noted that alltreatments relied on suggestion, verbal persuasion, and “careful coercion” (p. 76).The anecdotal results of a psychodynamically oriented married couples group werepresented by Davenport and Mathiason (1988). The study did not include a comparisongroup nor recognized measurement indices. Consequently, while the research contributedby identifying the typical psychodynamic issues which emerged through the group process,the report did not extend the evaluation of the efficacy of this treatment approach beyondprevious efforts (e.g., Burton, 1961; Cadogan, 1973; Corder et al., 1972; Ewing et al., 1961;Gallant et al., 1970).42A study that provided support for the viewpoint that important changes ininterpersonal relationships are possible when working with individuals was reported bySisson and Azrin (1986). In this investigation, two different treatments were offered totwelve women who were suffering negative consequences as a result of the alcohol abuse ofa loved one. Nine of these women were concerned about their husbands, two were worriedabout male siblings and one was focused on her father. The procedures included therandom assignment of the women to either a traditional program that included the provisionof information regarding the disease concept of alcoholism and supportive counselling, or toa reinforcement training program that was aimed at teaching skills that would enable thesubject to help get the individual they were concerned about into treatment. While thetraditional treatment program stressed the view that alcoholism was the alcoholics’ problemand counselled the women to distance themselves and emotionally withdraw from therelationship, the reinforcement training emphasized ways for the participants to maintaintheir emotional connection. The women were instructed in ways to motivate their loved oneto treatment. The reinforcement training program employed experiential activities aimed atpreparing these women to enact new behaviors with their husbands, brothers or fathers thathelp instill an interest in treatment. Unfortunately, the treatment procedures employed inthis study were not detailed in manuals and only brief descriptions of the treatments involvedare provided. Results were based on the women’s assessment of the drinking behaviours ofthe alcoholics and a simple accounting of how many of the alcoholics ended up incounselling.The results of this study supported the reinforcement procedures. While none of thealcoholics whose partners attended the traditional program ever attended treatment, all butone of the alcoholics associated with the reinforcement procedures came for therapy.Additionally, the reinforcement group seemed to affect the alcoholics drinking behaviour in43a positive way, while the traditional group had no meaningful effect on the alcoholics’behaviours.The study is tempered by methodological issues that include no articulation oftreatment manuals, a small sample size, questionable random assignment procedures, andno recognized outcome indices particularly those focused on the effects of two treatments onthe women who participated. The therapists were not described and the treatmentprocedure used in the traditional treatment group was not standardized or described.Finally, there was no mention of supervision or checks on therapist competency andadherence to treatment protocols. Despite these substantial methodological problems thestudy offers empirical evidence that the partners of alcoholics have some impact over theirpartners’ drinking behaviour. As a result of this study, it appears that spouses can be moreinfluential in promoting treatment for alcohol dependent family members than traditionalapproaches have assumed.Support for the use of marital and family treatments for adult alcoholism wasprovided by Bowers and Al-Redha (1990). In this recent study, 16 couples, in which onemember was alcoholic, were randomly assigned to either couples group therapy or tostandard individual treatment condition. Couples were assessed prior to treatment and on 1month, 6 month and 1 year follow-up occasions. The standard treatment was offered on anopen-ended basis and included the formation of a therapeutic bond and an examination ofareas in the client’s life in need of change. The mean number of sessions was 7.43 and theaverage amount of time an individual spent in the treatment was 11.15 hours. The couplestreatment began with an eight hour day long session which was followed by 8 sessions lasting90 minutes each. The average amount of time for a couple in treatment was 19 hours. Thecouples treatment included role playing of life situations, communication training andexperiential activities aimed at enabling the expression of feeling and assertiveness. Bothconditions of treatment were administered by the same 2 therapists. Although the44treatments do not appear to have been put into manual form, a description of the couplestreatment is apparently available from the authors.Bowers and Al-Redha (1990) used indices of alcohol consumption, the maritalrelationship and social and employment functioning to measure the effectiveness of thetreatment procedures. Regretfully only one of the marital measures employed was arecognized research tool and the remaining elements of the measurement battery werequestionnaires developed for the study. The results showed that both treatment conditionshad a significant positive impact on the drinking behaviour, but only the couples groupmaintained the changes on the later follow-up occasions. The same pattern of result wasfound with the marital measures. Significant treatment effects for both groups were noted atpost-treatment for both treatments but these changes were retained by the couples grouponly at later data collection points. No significant differences were found with regards to thesocial and employment measures. The methodological limits to the study include a smallsample size as well as the omission of treatment manuals and treatment implementationchecks. Thus, it is difficult to be certain as to what occurred in either treatment. It appearsthat the treatment was not a behavioural marital treatment but beyond this, it is difficult toascertain what theoretical model of therapy informed the treatment. However, the resultsdo appear to support the use of couples treatment indicating strong treatment effectsresulting from 8 sessions of couples group therapy.The final piece of research reviewed in this section was provided by O’Farrell, Cutter,Choquette, Floyd, and Bayog (1992). In this report, additional 2 year follow-up results areprovided for the study originally conducted by O’Farrell et al. (1985). At the time of theearlier investigation, behavioural marital therapy (BMT) had been found to be superior toeither individual counselling or interactive marital therapy (IMT). The present study ofO’Farrell et al. (1992) tempers the original findings with less impressive results. For themost part, the differences between BMT and the individual therapy group had not been45sustained at follow-up measurement points. This is to say that there was no significantdifference found between BMT and the no marital therapy group on measures of drinkingadjustment, negative consequences of drinking or husbands’ marital adjustment. However,the wives who attended BMT reported significantly higher levels of marital adjustment thanthe individual comparison treatment group. Moreover, differences between IMT treatmentgroup and BMT highlighted at post-treatment were not evident at any of the follow-upmeasurement points on any of the measures employed in the research. The main consistentsignificant finding of the follow-up investigation showed that both the BMT and IMT werecomparable and resulted in fewer number of days that the couples were separated over thecourse of the study when compared to the individual treatment group. While the authors doreport on non-significant trends in the results which favor BMT, the substance of thesetrends is questionable considering that the study suffers from important methodologicallimitations, including small numbers of subjects (n = 10-12 per treatment group), issuesregarding pre-treatment group comparability (e.g., the husbands’ mean pre-treatment scoreson the marital adjustment measure were more than one standard deviation greater than themean score for the BMT husbands on the same measure), and finally, treatment differenceconfounds (the BMT alcoholics were all required to take Antabuse as part of the treatmentprocedures while others were not) reports of trends in the data appear unwarranted.Nevertheless, the present report is important because it suggests that the two maritaltreatment groups were more similar in their performance than previously thought and,furthermore, the results indicate that without additional treatment, the individualcounselling approach resulted in follow-up outcomes equivalent to the marital treatments forhusbands but less effective outcomes for wives. The study also found that regardless oftreatment, the alcoholics with the most severe marital and drinking problems prior totreatment had the worst outcomes in the two years after treatment.46Summary of Family Therapy Alcoholism Treatment Outcome ResearchThe literature regarding the application of family and couples therapy to alcoholismis for the most part positive and encouraging. Several comparative studies have beenconducted probing couples formats of treatment. However, with the notable exception ofHedberg and Campbell (1974) there are unfortunately few well designed studies that haveexamined treatment formats that include the entire family with the adult alcoholicpopulation. Nonetheless, the literature which does exist is predominantly favourable,however it is by no means definitive. This fact is reflected in the recent review of researchon the prevention and treatment of alcohol-related problems by the Institute of Medicine(1992). In this review, family therapeutic approaches to alcohol treatment were againrecognized as promising directions for further research, and it is concluded that“interventions to improve the functioning of couples and families may enhance favourableoutcomes” (p. 12). The caution reflected in this conclusion is also evident in other reviewsthat have recently been published (Collins, 1990; McCrady, 1989; O’Farrell, 1992; O’Farrell& Cowles, 1989; Orford, 1990). While these reviews agree that over the past 15 yearsresearch has generated controlled outcome studies that have become increasingly rigorous, ameasure of tentativeness is exercised in assessing the strength of evidence supporting theeffectiveness of the family treatments for alcohol problems.A careful examination of the literature reveals that it is accurate to note that themajority of the studies in the area of couples and family treatments for alcoholism supportthe implementation of behavioural marital and family therapy. Several authors haverecently expressed concern that the other forms of marital and family therapy which havenever been formally tested are benefitting from the overgeneralization of the researchaccomplishments of the behaviourally oriented models (Jacobson et al., 1989; McCrady,1989; O’Farrell, 1992). As Jacobson et a!. (1989) lamented, the family concepts mostcommon in clinical practices are not derived from the behavioural models which have47empirically legitimized their application. Consequently, it is not surprising that thebehaviourally oriented researchers have called for evidence from their family therapycolleagues and challenged them to generate their own empirical support. McCrady (1989)made this abundantly clear when she wrote:clinicians working with these ... perspectives should be challenged to evaluatetheir work or it will be difficult for them to justify the continued use ofuntested practices when carefully evaluated and articulated procedures areavailable. (p. 180)Studies have been conducted that have compared other forms of couples treatment tobehavioural forms of treatment most notably Johnson and Greenberg (1985) and Snyder andWills (1989). In the Johnson and Greenberg project, a marital treatment approach known asEmotionally Focused Therapy was found to be more effective than behavioural maritaltherapy in bringing about the marital improvement in conflicted couples. In the Snyder andWills study an insight oriented marital therapy treatment was found to be as effective as thebehavioural marital therapy in bringing about individual and marital change in conflictedcouples. While studies such as these may soften the complaints like those expressed byMcCrady (1989), the call for other forms of couples and family treatments to be empiricallytested with the alcoholic population remains a pressing matter.Couples and Individual Therapeutic Format ResearchThis study was designed to examine the efficacy of ExST and to probe the relativeeffectiveness of the treatment provided in two different formats (individual and couples).While the ExST treatment conditions share the same underlying treatment philosophy andtechnical range, they differ with respect to format (who is attending the therapy sessions).Clearly, a variety of therapeutic options are both constrained and enabled by the differenttreatment formats. However, in the ExST treatment conditions, differences arise as artifactsof the format of the treatment delivery as opposed to distinctions arising from contrastingtherapy approaches. In previous research comparing individual treatments and maritaltreatments, researchers have compared one kind of individual therapy with another kind of48marital or family therapy (e.g., Barlow, Mavissakalin, & Hay, 1981). Results from this kindof research strategy, although helpful, are ambiguous with respect to the understanding ofdifferences in treatment outcomes that are attributable to different treatment formats.Comparative therapy studies are established on the premise that by measuring thetreatment outcomes of contrasting therapy models, the scientific community can understandwhich of the treatment options under study result in what kinds of changes in the clients.This research approach is helpful in establishing which of the treatments being comparedworks the best for the population being investigated. Consequently, such a researchapproach can be used to compare one individual form of therapy with a couples form oftherapy and can result in information pertaining to the two treatments of concern, howeverthis strategy of research is limited with respect to what it can say about differences betweenindividual and marital therapy formats. This limitation arises because the individual andcouples treatments differ substantially in matters of treatment philosophy, theory, andtechnique, as well as format.Recently, efforts have been made to tease out the differences between individual andcouples treatment formats (Foley, Rounsaville, Weissman, & Chevron, 1989; Jacobson,Dobson, Fruzzetti, Schmaling, & Salusby, 1991; O’Leary & Beach, 1990). This has beenmade possible by comparing individual and couples treatment that share the sametheoretical orientations and differ only with respect to the format in which they areimplemented. This research strategy allows for comparisons of treatment formats which arenot confounded by theoretical and practical differences, as would be the case if one type ofindividual treatment was compared to another type of marital treatment (e.g., gestalt therapyvs. structural-strategic). The following section reviews the research on therapeutic formatthat has been conducted.49Foley et al. (1989) conducted a pilot study in which 18 depressed patients whoconcomitantly complained of marital disputes were randomly assigned to either individualinterpersonal therapy or conjoint marital interpersonal therapy. Patients were given 16weekly sessions of treatment and outcomes were based on pre-treatment and post-treatmentassessments. The measurement procedures employed included recognized research indicesof depression, intrapersonal functioning, marital relationship and social adjustment.Therapists were well trained, and the treatments were conducted in accordance withtreatment manuals which specified the therapy protocols. Treatment adherence andtherapist competency was verified by expert raters.The results of this study indicated significant improvement for both treatmentconditions on the depressive condition and social adjustment, however there was no betweengroup difference on these variables. While there was no significant change on theintrapersonal measure for either treatment, the results of the marital measures indicatedthat the marital treatment condition had in fact resulted in significantly better treatmentoutcomes on both the Locke-Wallace Marital Adjustment Test and the Affectional sub-scaleof The Dyadic Adjustment Scale. In particular, couples in the marital treatment conditionreported significantly higher levels of affectional expression and also greater levels of overallmarital adjustment than did their individual treatment couples counterparts. A measure ofcaution is warranted with respect to the study. The sample size was small and consequentlystatistical power to identify differences was an issue, and a comparative control treatmentcondition was not included. In addition, there is no mention of the kind of statisticalprocedures which were employed and pre-treatment differences on the marital measureswhich may have existed may well have played a statistically significant role in the outcomesreported. For example, in the pre-treatment means on the Dyadic Adjustment Scale whichare provided, only the depressed clients in the couples treatment condition fall in thedistressed range of the instrument. Finally, a follow-up testing occasion would have beenhelpful to establish whether the between group differences were sustained over time.50Nonetheless, the study is exemplary with respect to the treatment implementationprocedures that were used in that they are representative of the state of the art in therapyoutcome studies. The results are best viewed with tentativeness but nonetheless support theview that the marital format performed better on the important dimension of maritalfunctioning than did the individual treatment format.Two other recent investigations have been reported that approximate this researchapproach (Jacobson, Dobson, Fruzzetti, Schmaling, & Salusley, 1991; O’Leary & Beach,1990). Both comparative studies contrasted cognitive behavioural therapy (individualformat) with behavioural marital therapy (couples format) with a third experimentalcondition in the treatment of depressed women. O’Leary and Beach (1990) noted that thetherapies differ in terms of the technical emphasis they place on bringing about change. Forexample, while behavioural marital therapy works to decrease negative interactions in orderto increase feelings of closeness, open sharing of thoughts and concerns and increase positiveinterchanges and problem-solving, cognitive behavioural therapy works on decreasingnegative cognitive distortions and selective memories for negative events and increasingpositive beliefs about self relationships and the future. Nonetheless, in many ways thesetechnical elements are a result of the opportunities available to the therapist that areattributable to the different treatment format. The two treatments share in common thesame philosophical underpinnings and cognitive-behavioural theoretical orientation. Inaddition, both treatments view the therapist role and positioning the same way and bothstrive to generate therapeutic alliances with clients in a manner similar to one another.Accordingly, while cognitive-behavioural and behavioural marital therapies can be construedas different treatments, it can be argued that these studies extend the research on treatmentformat.O’Leary and Beach (1990) presented the results of a research project that furthered apilot study reported earlier (Beach & O’Leary, 1986). In this project, 36 maritally discordantcouples with depressed wives were randomly assigned to either cognitive therapy,behavioural marital therapy or a wait-list control group. The same therapists provided boththe individual and the marital treatments and all therapists provided both the individual andthe marital treatments and all therapists received regular supervision and were judged byexperts to be competent in both treatment formats. The treatments were 16 weeks induration and were punctuated by pre-treatment and a post-treatment data collectionoccasions. A one year follow-up was also conducted. A limited number of standardmeasures were employed in the study, including the Beck Depression Inventory and theDyadic Adjustment Scale. The results of the study showed that both treatment formats weresignificantly effective in reducing level of depression at both post-treatment and follow-up.The marital index suggested that the individual treatment did not affect the marriage,however, the marital treatment significantly raised the level of marital satisfaction at post-test. Follow-up data affirmed that this important treatment difference was durable.The study was limited by a small sample size. It also suffered from othermethodological problems including no clear articulation of the treatments provided inmanual form and a failure to implement a means of verifying levels of treatment adherenceand therapist competency over the course of the project. Nonetheless, the results appearpersuasive and closely resemble the kind of treatment format effects reported by Foley et al.(1989) and support the use of marital treatments when symptoms are associated with maritaldiscord.The last study conducted to date probing the treatment formats was conducted byJacobson et al. (1991). In this investigation, 72 depressed women and their spouses served assubjects and were randomly assigned to either an individual treatment format (cognitivebehavioural therapy), a couples treatment format (behavioural marital therapy) or a52treatment condition that offered both individual and couples treatment sessions (combinedtherapy). Marital discord was not a prerequisite for subject participation in this investigationas it was in the projects of Foley et al. (1989) and O’Leary and Beach (1990).Jacobson et al. (1991) exemplified a well-executed psychotherapeutic outcome study.Therapists were well trained and carefully selected by raters employing standard criteria.Therapy manuals were employed for the individual and marital treatments and the therapywas monitored throughout the inquiry to ensure both treatment manual adherence andtherapist competency. The same therapists provided all three treatments which were 20sessions in duration. Recognized research indices were employed to measure depressionand marital satisfaction at both the pre-treatment and post-treatment occasions.In Jacobson et al. (1991), the 3 treatment formats were contrasted in terms of howwell they performed with maritally distressed and maritally non-distressed clients. As aconsequence the results are somewhat complex. While all treatments led to significantreduction in depression severity regardless of pre-treatment levels of marital satisfaction, thecouples treatment did not perform as well as the individual treatment format in cases wheremarital distress was not reported and this difference was statistically significant. In cases ofmarital discord the individual and couples treatment formats were comparable in theireffectiveness in terms of the depression index and there was a trend for the combinedtreatment to fare more poorly than either the individuals or couples therapy.The results derived from the marital measures were somewhat mixed. Two kinds ofdata were generated, one kind was based on a self report questionnaire, and the other used amarital interaction coding. The self report measure results indicated that all treatmentgroups evidenced improvement in marital satisfaction, however among the maritallydistressed sub-sample only the marital treatment format led to statistically significantimprovement. It is interesting to note that among the non-distressed sub-group thecombined treatment format was the only treatment format offered that posted positive53statistically significant change. The quality of marital interaction, a direct observationalsource of data, failed to reflect any significant differences between treatment groups,however it is noteworthy that the combined treatment was the only treatment format thathad any significant impact on the couples behaviour both reducing husband and wifeaversive behaviour and significantly increasing wife’s facilitative behaviour.The authors of the study spent considerable time speculating about why the combinedtreatment format which was a mix of individual and couples sessions did not produce thelargest treatment affects as they had anticipated. Several explanations were offeredregarding dosage and therapist competency behaviour, however one is left wondering aboutthe degree of continuity offered by this format. It appears that there was not a manualprepared for this treatment condition and is not clear as to how the individual and couplestreatments were wed. Judging by the methods used in measuring treatment adherence, itappears that in fact the therapists were expected to keep the individual sessions distinctlydifferent from the couples meeting. This procedure casts doubt regarding the level oftechnical synthesis and theoretical integration present in the combined treatment that isrequired to bridge the individual and couples treatment traditions.The results of this study are, in many ways, similar to the previous two studies ontherapy format and support the notion that a couples therapeutic context is the treatmentformat of choice in situations where marital distress is present. This, of course, makesintuitive sense since the couples problems can no doubt be seen as contributing to themaintenance of the problem, compounding the individual’s difficulty and perhaps impedingtherapeutic movement.More research in this area is required. Interpersonal and cognitive-behaviouraltherapies have been investigated in this way, however the level of theoretical integration inboth cases appears to leave room for improvement by other treatment approaches. All ofthe three studies conducted to date on treatment format have focused on the symptom of54depression. Additional work is necessary to explore whether or not these results areparticular to this type of client or generalize to other problems. The present researchinitiative contributes to this literature by investigating an integrative experiential andsystemic therapy in the treatment of alcoholics implementing the therapy in an individualand a couples treatment condition.Experiential Systemic Therapy SynopsisThis section summarizes the ExST model employed in this research. A thoroughoverview of the treatment is beyond the scope of this report because this level ofspecification would cause the document to be prohibitively long. As mentioned earlier, amanual which articulates and operationalizes the therapy has been written and is availableupon request. The following section is a brief review of the treatment which will enable acursory understanding of the treatment.In the domain of psychotherapeutic approaches, the ExST model of therapy is locatedat the confluence of the experiential and systemic streams of practice (Friesen et al., 1989).It is the infusion of the philosophical, theoretical and technical elements of these two richtraditions of treatment which gives rise to the uniqueness of the model.From the time of its inception, the ExST model has sought to integrate as wide abreadth of methods of practice as possible. This has been done in order to provide as greata range of technical options to therapists while remaining consistent with the ExST theory.ExST assumes that the key to the therapeutic enterprise rests in the creativity, spontaneityand ingenuity of the therapist and consequently, the model empowers clinicians to trustthemselves as the journey with clients unfolds.ExST understands therapy as a culturally sanctioned change ritual or rite of passage(Koback & Walters, 1984). As such, “therapy” is seen as a symbol of healing and change and55the therapist is viewed as the symbolic facilitator of the ritual itself. Consequently, all actionrelated to the treatment and all behaviours connected to it are views being imbued withsymbolic significance related to constructs of healing and change.En keeping with the underpinnings of other experiential forms of treatment, ExSTassumes that clients require therapeutic experiences as opposed to a therapeuticexplanations. Seeking to make the therapeutic experience as profound and meaningful aspossible, ExST therapists attempt to generate with their clients a therapeutic environmentthat is warm, caring and respectful and also unique from normal everyday experiences.Action-oriented therapeutic activities that engage the entire being of the client areevoked whenever possible in this treatment. As a result, the therapy takes on a quality whichis much more than simply a verbal discourse about ideas, feelings and facts. Therapists seekto heighten client awareness of their ExST model using the condition in as intense a way asappropriate in order to perturb changes which spontaneously arise out of the client’s wealthof previously untapped potential. In this way, moments of change and transformation aregenerated in this here and now therapy. Symbolic representations of troubled relationshipsare often evoked in ExST (Friesen et al., 1991). This procedure has been recognized as apotent way of approaching entrenched problems. Indeed, Selekman (1993) noted thatsymbolic evocation was:useful with families that have been oppressed by a symptom for a long time, havenot responded well to the basic Solution-Oriented approach, and are in the same orworse categories. (p. 148)In particular, the symbolic evocation of alcohol in the therapeutic context has been found tobe an effective way of helping alcoholic and other family members transform theirrelationship to the addictive substance (Todtman, Friesen, Newman, & Grigg, 1993).56The ExST model views relationships as the bedrock of human existence and perceivesindividuals as inseparable from the webs of connection which contextualize behaviour.Problems are understood as relational difficulties which are characterized as rigid, repetitive,and restrictive patterns of interaction that dominate people’s lives. The process of change isa process of emancipation from such dominating relationships, be they with “the bottle” orwork, a partner or family member.The therapeutic venture is conceived as a collaborative process shared mutually bytherapists and clients. Therapists act as benevolent companions with knowledge of theterritory being traversed. However, it is the client’s experiences and needs which ultimatelyguide the therapeutic journey on the path which is taken and therapists put their trust in thedevelopmental impetus which motivates client’s participation in therapy.As mentioned in an earlier section, ExST is an integrative model that can be appliedin a variety of treatment contexts including individual, marital and family therapy. For thepurposes of this study, the individual and marital formats of the therapy were employed andconstitute different treatment conditions. The essential difference between the two formatsof ExST is the constitution of the therapeutic system (i.e., the spouse either attends meetingsin the marital format or does not attend meeting in the individual format). Process andcontent differences between the two formats arise as a consequence of the therapeuticsystem composition since the philosophy/theory and technical elements which drives theExST treatment remains constant across both treatment condition.Rationale for ExST DevelopmentThe need for an integrative model of clinical practice was the central reason for thedevelopment of the ExST model. In a survey of therapists conducted prior to theestablishment of ExST, Geiss and O’Leary (1981) found that over 80% of the clinicians that57responded to their questionnaires reported that they were eclectic in their approaches toproblems. It seemed that the theoretical models which guided practitioners wererestraining clinicians and were not addressing the clinical requirements of most therapists.Accordingly, clinicians were going outside of their theoretical orientations in order to meetthe needs of clients. ExST aimed to provide an integrative theoretical framework whichwould embrace as wide a spectrum of technique as possible and thereby provide a theory ofpractice which would not restrict therapist efforts.At the time that ExST was being generated, the alcoholism treatment field seemedfragmented. Many models of therapy had been introduced which viewed the problem ofalcohol dependency from a particular point of view (see Biopsychosocial section pg. 15).The developers of ExST sought to provide a model of therapy which would bridge thevarious perspectives by employing a biopsychosocial model of alcoholism as a foundationalaspect of its formulations. Consequently, a systemic perspective which included anappreciation of the physiological and bio-chemical issues of alcoholism as well as anunderstanding of intrapersonal, marital, family and social factors came to be a crucial aspectof ExST.Another area of fragmentation addressed by the ExST model pertains to therapeuticmodality specialization. Traditionally, those who did individual therapy did not do grouptherapy or marital and family therapy. Indeed, the training contracts which provided theclinical environment for the generation of ExST were geared towards this kind ofspecialization. However, it soon became evident that for practical reasons, the specialistmodel of treatment delivery was not appropriate for the treatment mileau in which many ofthe therapists in training performed their duties. A model which was versatile in terms of itsapplication in a number of therapeutic situations including individual, couples and familieswas required to meet the clinical needs of the therapists. Consequently, the developers ofExST sought to integrate individual practice with a family systems theoretical orientation.58A final area of integration intrinsic to ExST is the model’s flexible means of problemformulation and treatment planning. The spectrum of difficulties which confront recoveringalcoholics include behavioral, affective, and cognitive problems. While many models oftherapy address one or two of these problem areas directly, ExST concerns itself with allthree domains of experience within its theoretical formulation. Accordingly, the model canbe aptly applied to the wide array of changing problems which clients present as they movethrough the recovery process. Additionally, the readiness of alcoholics to address the manyproblems in their lives varies dramatically client to client. For some, breaking the behavioralpattern of alcohol consumption must be the main focus of the treatment. For others, movingswiftly to resolve issues of profound childhood trauma is the central concern. Accordingly, itappeared to the developers of ExST that a model of treatment was required that assessedthe needs of each client and, imbued with considerable flexibility, addressed the uniqueproblems of the client with an individualized treatment plan.The integrative quality of ExST is the sine qua non of the therapeutic model. Theimpetus to synthesize fragmented aspects of clinical practice provides the model with aunique position in the alcohol treatment field. With its commitment to build bridges ratherthan fences and to regard matters of health in a holistic fashion, ExST has been developedto offer therapists a way of working effectively to help clients establish more harmoniousrelationships in the many domains of their lives.It is important to acknowledge that the author of the present dissertation is one of thedevelopers of the ExST approach. As a clinician, and the author is committed to workingwith clients using the ExST model as his guiding theoretical orientation. While every efforthas been made to ensure that no researcher bias has resulted from the clinical allegiance ofthe author, it must be recognized the ExST treatment is dear to the author’s heart.59Summary of Literature ReviewThis review has examined the outcome research pertaining to couples and familytherapy approaches to alcohol. The majority of the well designed studies on marital andfamily treatments for adult alcoholics have investigated variations of behavioural couplestherapy and the body of results has shown that effecting change at the level of the maritalrelationship is an important goal of treatment. Systemic treatments have for the most partnot been placed under empirical scrutiny despite a proliferation of these models and apopularization of their perspectives. In addition, new couples and family treatments whichhave emerged since second-ordered cybernetic theory became the dominant perspective inthe couples and family field are in need of empirical validation.The literature review has also considered recent research focused on exploring thetreatment efficacy of individual and couples formats of therapy. These studies seem tosuggest that while the individual and couples formats perform comparably in terms ofsymptom reduction, the couples treatment format has the added benefit in bringing aboutpositive relational change in distressed couples that is most pronounced at follow-upmeasurement occasions.There remains a great deal of work to develop more effective means of helpingalcoholics and their families negotiate the many obstructions on the road to recovery. In thisconnection, Jacobson et al. (1989) called for therapeutic innovation in the area of alcoholismtreatment sighting the lack of consistent results demonstrating the superiority of onetreatment over another. For some, the search for a single approach to the problem ofalcohol dependency seems a misguided venture. Miller and Hester (1986, 1989) essentiallyagree with Jacobson et al. (1989) and put forward an argument for the need to match clientswith appropriate treatments based on careful assessments.60The matching argument asserts that particular treatments may be more effective thenother approaches given the particular client needs at any moment in time and that treatmentpotency can be optimized by carefully matching client needs with the application of theappropriate treatment. This thinking is similar to the notions that underlie the IPTC modelof Pinsof (1983). The main difference in the thinking however, is that unlike Miller andHester (1986, 1989), Pinsof (1983) believes that the same therapist can match the client’sneeds with different treatments when the therapy is based on both an assessment of theproblem and the therapeutic model which is most appropriate at the time. Similarly, theExST model, with its emphasis on therapist creativity, assumes that the therapist canindividualize the treatment procedures to match the unique needs of each client.The ExST model presently under investigation is a response to Jacobson’s (1989) callfor new treatments. Because of the integrative qualities of ExST, treatment can be offeredin a variety of ways including individual, couples, family and group formats. ExST alsoallows for the utility of a broad spectrum of techniques from which therapist may choose tomeet the changing needs of each client. Consequently, ExST is meant to be a flexibleapproach to treatment that can be adapted to fit the requirements of different clients.The need for further research in the area of alcohol dependency that examines theefficacy of a new approach to treatment and explores the effects of individual and maritalformats of this approach has been highlighted. The following section articulates thehypotheses which have been formulated for empirical scrutiny.Research HypothesesThe following research hypotheses have been formulated to reflect the intent of theresearch as articulated in the research questions that appeared earlier in this document. Thehypotheses are focused on the differential treatment efficacy of ExST and SFT, and the61differential treatment format efficacy of ExST-I and ExST-C. The hypotheses are stated indirectional terms that favor ExST and the couples format.Hypothesis 1The first hypothesis of this study is focused on the differential treatment effects of thetherapies under study and it is: When compared to the families in which the alcoholic fathercompleted SF1’, the families in which the alcoholic father completed ExST will reportsignificantly greater improvement at post-test on measures contained in the ecologicalassessment package employed in the study.Hypothesis 2The second hypothesis of the study centers on the post-treatment and follow-updifferential treatment effects of the two formats (individual and couple) of ExST under studyand is: When compared to the families in which the father completed ExST-I, the families inwhich both father and mother completed ExST-C will report significantly greaterimprovement at post-test and/or follow-up as measured by selected instruments in theecological assessment package used in the study.The two hypotheses will be then broken into sub-hypotheses related to the differentfamily members and operationalized in terms of the specific instruments in the ecologicalassessment package used for the analysis at the conclusion of the following chapter.62CHAPTER 3: METHODOLOGYResearch DesignThe investigation used a mixed model experimental design: three groups by threeoccasions. The design is presented below in Table 1 using the notation given by Stanley andCampbell (1963).Table 1.Schematic Design of the StudyRandomized Pre-test Treatment Post-test Follow-upGroupsR1 01 X1 02R2 03 X2 04 05R3 06 X3 07 08R1 = Experimental Group 1 X1 = SFTR2 = Experimental Group 2 X2 = ExST-I (Individual format)R3 = Experimental Group 3 X3 = ExST-C (Couples format)As revealed in Table 1, participants were randomly assigned to one of the three levelsof the independent variable, treatment. Two of the three treatment groups provide forms ofExperiential Systemic Therapy (X2,X3), with the third group implementing the SupportedFeedback Therapy (X1). Participants were also randomly assigned to one of the cadre oftherapists that were involved in the provision of the treatments. The dependentmeasurement variable occurred on three occasions: pre-test, post-test, and follow-up.In the original design of this investigation, an ExST family treatment group and await-list control group were included as treatment conditions. In the early stages of the63project, the family treatment group was dropped from the design in the early stages of theproject when the difficulty of executing the design was realized. However, the wait-listcontrol group was replaced by the SFT comparison treatment group at the last possiblemoment. Just prior to data collection, it became evident that the wait-list control group wasnot a viable option. Clinic discomfort with the wait-list group, and anticipated clientreluctance to accept this treatment option necessitated the implementation of a comparisontreatment. While the wait-list control group would have been preferable from theperspective of determining the efficacy of ExST and its treatment formats by controlling fortesting variables, the pragmatic alternative of a comparison treatment group was the onlyavailable alternative. Although the implementation of a known comparison treatment groupsuch as BMT would have been the preferred option, time constraints precluded this choice.Accordingly, the SFT model was generated, operationalized in manual form and institutedas the comparison treatment condition in this study.Supported Feedback TherapyThis treatment condition was developed after it became clear that a wait-list controlgroup was not a practical alternative for the research study. The SFT treatment conditiongenerated and operationalized in manual form by Grigg et al. (1989), was meant to be aquasi-treatment and control group. In order to be suitable for implementation, thetreatment had to offer enough therapeutic qualities that it would be acceptable to theparticipants, and at the same time control for some of the variables that would have beencontrolled for by a wait-list group. Accordingly, SF1’ coupled the process of weekly selfmonitoring on alcoholics that was to be done by all research participants with regular(weekly or bi-weekly) meetings with a therapist. A number of wall charts were devisedwhich transferred the weekly monitored behaviour contained in the Weekly SituationsDiaries (WSD) onto a series of graphs (a copy of the WSDs completed by fathers andmothers in this study and samples of the charts can be found in Appendix C). In this way,64the behaviour monitoring information recorded each week by the alcoholic was shifted ontocharts which revealed the alcoholic’s life process and provided a focus for the therapists andclients.. The areas of concern covered in the self-monitoring process included alcohol, self,marriage, family, friends and work.The SFT model is built on a caring, warm, non-judgemental and supportivetherapeutic relationship based on the work of Rogers (1951, 1961). The focus of thetherapy was on charting the previous week(s) and examining and learning from the feedbackthat was available in the charts. As a result, the therapy was oriented in time to the presentand the near past and the future was meant to be de-emphasized. Therapy meetings wereexpected to accent the client’s personal responsibility for recovery and whenever possible,were designed to be conducted on a bi-weekly basis.With this treatment protocol, the SFT condition was conceived of as an elaboratequasi-control group at the time of its inception. By offering contact time with therapists,SFT was thought of as controlling for the portion of therapeutic improvement attributable tothe participants’ involvement in therapy. In addition, the SFT treatment condition wasmeant to control for changes in clients associated with the increased attention connected totheir involvement in a research study. Thus, it was hoped that the SFT comparison groupwould control for bias engendered by the measurement of participants with the extensive setof questionnaires included in the ecological assessment battery used in this study.The SFT treatment was designed to systematically contrast ExST in a number ofcritical therapeutic dimensions. While ExST takes an unstructured creative approach toeach session, SFT follows a predicable structured session procedure. Accordingly, ExSTtreatments are highly individualized and tailored to fit each client’s particular needs and SFTtreatments are relatively uniform and consistent regardless of specific client differences.The two therapies differ with respect to their temporal orientation and intensity level.Unlike ExST, which is a very intense here-and-now focused treatment, SFT maintains a lowlevel of intensity and largely aims at learning from the there-and-then events of the recent65past. The ExST model is an experiential form of therapy which is deeply symbolic in itsimplementation, and built upon professionally intimate therapeutic relationships withclients. In contrast, the SFT model is a cognitive-behavioural form of therapy which is veryliteral and concrete in its procedures, and oriented to more distant therapist and clientrelationships. Consequently, SFT can be seen as a carefully crafted comparison treatmentwhich differs from ExST in a variety of salient therapeutic features which allows for thetesting of important components of the ExST approach.Family Inclusion CharacteristicsOne hundred and fourteen families were screened for participation in this study. Thefamilies were recruited for the project from a number of sources. Many clients wereidentified through the normal client intake procedures at the participating alcohol and drugtreatment agencies. Other clients responded to presentations at in-patient residentialtreatment programs or were sent to the project by a network of referral sources. Still otherparticipants were self-referred to the project, responding to the media attention innewspapers and on radio and television that the project received. In order for families to bescreened for the research, they had to meet the following inclusion criteria:(1) the fathers had to be struggling with alcohol dependency problems and had to haveconsumed alcohol within the last 3 months to be considered eligible,(2) mothers had to have no reported dependency problems within the last 5 years,(3) the couple had to be complaining of marital distress but still living together andstating that they valued the preservation of the relationship,(4) the couple had to have been living together in either a marriage or common-lawsituation for a minimum of 1 year,(5) The couple had to be willing and able to participate in marital therapy should they beassigned to this treatment condition,66(6) Families had to include at least one child either living at home or in regular contactwith the family and all children above 9 years of age were asked to participate in theproject,(7) remarried or blended families were welcomed into the project and the children in thepresent families could be the offspring of either parent.Participating families that met the inclusion criteria above were excluded from thestudy at the point of screening if one of the following exclusion criteria were evident:(1) the father’s alcohol problems was not severe enough and scored below the criticalcut-off score of 5 on the Michigan Alcoholism Screening Test (MAST) developed bySeizer (1971),(2) the mother’s alcohol use was too severe and she scored greater than the critical cutoff score of 4 on the MAST,(3) the couple’s level of marital distress was negligible with both members of therelationship scoring above the value of 99 on the Dyadic Adjustment Scale (Spanier,1976),(4) a severe psychiatric disturbance was evident in the screening interview and the fatheror mother scored exceptionally high on either the psychiatric or depression sub-scalesof the Symptom Checklist-90 Revised (Derogatis, 1983).Family DescriptionConsiderable information about the 114 families that entered the study wasgenerated at the time of screening. The following section describes the families in terms ofprofiles of the family, alcohol consumption, and past treatment.67Family ProfileThe participating families ranged in size from families with one child to families withfive children. Fifteen families or 13.16% of the sample had only one child. The largestnumber of families of the participants had two children (n = 56 or 49.12%); however, asizable portion of the sample had three children (n = 33 or 28.95%). Only a small numberof families had four children (n = 7 or 6.14%) and even a smaller proportion of participantshad five children (n = 3 or 2.6%).The fathers were, on average, in the middle years of their lives ( = 39.69 yrs., SD =8.73 yrs.) however, the ages of the fathers ranged from as high as 70 in some cases to as lowas 26 in others. As a group, the mean age of the mothers was in the late thirties (1 = 37.58yrs., SD = 8.87 yrs.) and like their male partners, ranged considerably in age with the oldestwife being in her 66th year of life and the youngest being 21 years of age.The average ages of the children in these families have been calculated in terms ofbirth order. The average age of the 114 first born children in this study is in the early teenyears (I = 13.86 yrs., SD = 8.22 yrs.), and their ages ranged from 37 years to only 6 monthsof age. The 99 second born children were living in the participating families and were onaverage just shy of the teen years (X 10.82 yrs., SD = 7.59 yrs.), again the range of ageswas considerable (31 yrs. - 6 months). The 43 third born children living in the participatingfamilies were averaged as a group in the 9th year of life ( = 9.44 yrs., SD = 7.78 yrs.), butvaried considerably in years of age with the youngest again only half a year old and the eldestbeing 31 years of age. As a group, the 10 fourth born children were somewhat older than thethird born children owing to the fact that fewer younger families participated with thisnumber of children (? = 11.1 yrs., SD = 8.28 yrs.). Only three families in the study had fivechildren and were somewhat mature in age. The fifth born children group were on average68the oldest ( = 15.0 yrs., SD = 7.81 yrs.) with a somewhat reduced range of ages, with theoldest fifth born child being 20 and the youngest being 6 years old.In 79.8% of the families, the parents were legally married (n = 91). In the remaining23 families or 20.2% of the sample, the parents were living in a common-law situation. For67% of the husbands and 69.4% of the wives, this marriage was their first and only maritalrelationship. Another 24.1% of the men and 23.4% of the women in this study had divorcedand now remarried their present partner. In terms of previous marital relationships, 22.8%of the husbands and 23.7% of the wives had been married only once before. However, 7.0%of both husbands and wives had been married twice before the present relationship and asmall fraction, .9% of the spouses, reported three previous marriages. The average numberof years that the couples had been living together was considerable ( = 11.83 yrs., SD =7.94% yrs.) and ranged from as short a duration of 6 months to as long a duration of 31 yearstogether.The majority of fathers in this study had some form of employment with 65.7% of thefathers reported having full employment and another 5.6% of the participating fathersemployed on a part-time basis. Nonetheless, almost one quarter (24.1%) of the fathers inthis study reported being unemployed at the time of their involvement in the project.The employment status record of the mothers in the study again showed that themajority of the women were working. Almost half of the women (46.9%) reporting havingfull-time jobs and close to another third (29.2%) reported being involved in part-time work.Unfortunately, many of the wives (17.7%) reported that they were unemployed and werepresently seeking work.69With respect to total family income, the majority of the participating families wouldseem to have been earning amounts that would place them in the middle class. Of thefamilies, 64.3% indicated making a joint family income ranging from $20,000.00 to 59,000.00yearly; however, some families in the study were clearly struggling financially with 17.6% ofthe sample earning below $19,000.00 per year. A smaller percentage reported familyincomes were considerable with 15.2% of the participating families earning over $60,000.00per year.Finally, it should be reported that the majority of the participating families werepredominantly white. Ninety-seven percent of the fathers and mothers were Caucasian. Ofthe remaining sample, 2.7% of the husbands and .9% of the wives were from the first nationsand .9% of the fathers and 1.8% of the mothers were of Asian descent.Alcohol Consumption ProfileOf the 114 alcohol dependent men participating in this research, 92.1% believedthemselves to be dependent on alcohol and 83.3% were convinced they could not drinkalcohol in a controlled fashion. Alcohol had been a problem for a considerable length oftime and 78% of the men indicated that drinking had been of difficulty for more than eightyears. All but 9.9% of the men had tried but failed to stop drinking in the past. In fact, 35%of the men in this study reported having tried to stop drinking on more than 10 occasions.Of the father participants, 53% reported marked signs of tolerance to alcohol and 72.3%indicated that they frequently drank until they were intoxicated.Alcohol played a major role in the lives of the fathers in this investigation. In fact,almost half (45.6%) reported that the longest duration of time they had gone withoutconsuming alcohol was less than one week. Only 3.5% of the men in the sample reportedhaving successfully stopped drinking for more than one year. Of the men in the study, all but701.8% had been engaged in alcohol dependent drinking within the last 3 months and in thesefew cases, the period of abstinence had been imposed on the participants as a consequenceof incarceration.In terms of drinking pattern, 59 of the fathers or 5 1.75% consumed about the sameamount of alcohol on steady daily basis. The remaining 55 fathers or 48.25% of the sample,had a pattern of alcohol misuse which was less predictable, occurring on a binge basis. Inthis pattern, periods of relative little or light alcohol use would be broken up by episodicbouts of very heavy alcohol use.The context of the fathers’ drinking varied. Many of the alcoholics in this studypreferred to drink at home (48.5%) and frequently alone (60.0%). Others, however, drankoutside of the house in bars, cars and other places (5 1.5%) with a variety of other people(40.0%). These other people were typically friends (77.2%) and rarely their spouses(45.5%). In fact, 29.7% of the men reported that they never drank with their wives and only6.9% indicated that they frequently consume alcohol with their partners.Despite efforts to keep alcohol out of their places of employment, alcohol use had adeleterious effect on the occupational lives of the participants (62.5%). Over half of thesample (52.5%) reported having missed work due to alcohol use and only 44.6% felt certainthat they had not lost a job due to alcohol problems.With respect to their involvement with medical facilities as a result of their alcoholproblems, over half the sample (5 1.5%) had gone to see a doctor in this connection. Inaddition, 46.5% of the fathers in this study had caused themselves bodily harm as a result ofdrinking. A total of 13.9% of the sample had been hospitalized due to alcohol consumption,and another 12.9% had gone to emergency wards with alcohol related medical concerns.71Ambulance services had been provided to 5.9% of the sample at one time or another due totheir problem with alcohol dependency.Legal problems and contact with the police are often associated with drinkingproblems and the present sample was consistent with this relationship. One third of thefathers in this study (33.7%) had been arrested for drunken and disorderly conduct. Half ofthe sample (50.5%) had been arrested for driving while intoxicated. Only one third of themen in this study had never been arrested, and only 36.6% had never been convicted of acrime related to their drinking. In point of fact, 11.9% of the men had been convicted fortheir drunken and disorderly conduct and 46.5% had been criminally convicted of drivingwhile under the influence of alcohol. Additionally, 22.8% of the sample reported that theyhad been convicted of some offense that was directly connected to their alcohol problems.Interpersonal conflict which escalates to levels of verbal and/or physical abuse isoften a problem in which alcohol abuse is implicated. The vast majority of the men in thisstudy reported heated verbal fighting with their spouses and 35.8% of the alcoholicsreporting having physically fought with their partners while intoxicated. Over half the men(55.7%) said they had been in verbal quarrels with relatives (sons, daughters, partners, etc.)but this had deteriorated to physical abuse for only 15.5% of the sample.Verbal fights with friends were noted by 55.9% of the men in this research and actualphysical conflict with friends occurred for almost one quarter of the men (24.7%). Inaddition to friends, over half (56.8%) reported verbal fighting with other people and theseconflicts deteriorated into physical fights for 37.8% of the men participating in this study.Many of the alcoholic participants in this study came from family lineage’s that hadalcohol problems running through them. For the alcoholic participants, 9.1% reported thattheir maternal grandmothers had alcohol problems and 23.7% indicated the same for their72maternal grandfathers. With respect to paternal grandmothers, 14.9% of the men notedtheirs had a drinking problem and 22.8% reported that their paternal grandfathers alsostruggled with alcoholism. The parents of the alcoholics in the study often seem to have hadtheir own difficulties with alcohol. Close to a third (28.7%) indicated that their mothers hada drinking problem and well over half (62.2%) identified their fathers as having been alcoholdependent. The uncles and aunts of the men in this study were also identified as frequentlyhaving alcohol problems. Alcohol dependency was noted by participants’ mothers’ sister(15.2%) and for fathers’ sisters (15.1%). Even more pronounced, the participatingalcoholics noted many alcohol problems in their mothers’ brothers (44.3%) and their fathers’brothers (33.0%). With respect to their own siblings, over one quarter of the alcoholicsubjects (26.7%) recognized their sisters’ alcohol problems, and almost half (49.1%) felttheir own brothers had significant troubles with alcohol.Previous Treatment ProfileThe final piece of information pertaining to the alcoholics participating in this studypertains to treatment. Fathers were asked to provide an accounting of their past therapeuticefforts and to specify their goals for the therapy they were about to commence. The majorityof men had previous treatment histories, with only 27.4% of the participants having reportedthat they had never received treatment. Over one quarter of the sample had attendedcounselling sessions for emotional and/or personal problems and 61.9% of the men hadbeen to treatment for their alcohol problem prior to their involvement with the presentinvestigation. Regarding the past treatments that the participants had received, 25.4% of themen had used detox services, 34.2% had spent time in residential, in-patient treatmentcenters, and another 35.1% had tried out-patient counselling.The treatment goals that the men maintained for themselves with respect to alcoholuse varied somewhat, however, most (78.4%) were committed to trying to establish lives73characterized by total sobriety. Of the total sample, 15.3% hoped to stop drinkingcompletely for at least six months, 3.6% wanted to be able to enjoy the occasional socialdrink, and a small fraction, .9% aimed at being moderate social drinkers.Family Involvement in the StudyThe families that were screened into the investigation may be broken into four groupswhich include pre-treatment drop-outs who received no treatment, incomplete treatmentdrop-outs who received some treatment, complete treatment participants, and completetreatment participants with missing data. The four groups are presented graphically belowin Figure 1.IncompleteDataPre-treatnientDrop-out17.5%MissingData8.8%CompleteTreatment52.6%Figure 1. Family involvement sub-groups.As Figure 1 reveals, the largest group of participants completed their involvementwith the study; however, an appreciable percent did not complete treatment. The fourgroups are discussed below.74Drop Out FamiliesA number of families who met the recruitment criteria and were accepted into theproject after an extensive 2 to 3 hour screening interview did not complete treatment. Of the114 families who entered the study, a total of 44 families or 38.60% of the total sampledropped out of the study before completing treatment. A total of 20 families or 17.54% ofthe total sample (45.45% of the total drop-out cases) failed to attend a single therapysession. Of these 20 families, 10 had been randomly assigned to SFT, 7 had been randomlyplaced in the individual ExST treatment condition, and 3 had been randomly assigned toreceive the couples version of ExST.The number of families who attended at least one therapy session but did notcomplete the treatment was 24. This category of family represented 2 1.05% of the totalsample (54.56 of the total drop-out sub-sample). The rate of families who dropped outwithout finishing the treatment they started for the three treatment groups wasapproximately equal. For the SET treatment, 8 families that started treatment did notcomplete the course of therapy. Comparably, 7 individual format ExST families and 9couples format ExST families decided to discontinue therapy before the treatment wascompleted. The mean number of meetings that families in this group of drop-outparticipants completed was 3.46 sessions with a standard deviation of 2.35. The reasonsprovided for withdrawing from treatment centered on marital dissolution of one kind oranother. In some instances, husbands’ relapses to drinking triggered marital separations; inothers, failing health and criminal incarceration were the causes of the prematuretermination of treatment.Complete Treatment FamiliesA total of 60 of the 114 families or 52.63% of those originally screened into the studycompleted treatment and were included in the outcome analyses. In terms of treatment75groups, the complete treatment group sub-divided evenly in three with 20 participantsreceiving SET, ExST-I and ExST-C.Complete Treatment Families with Missing DataIn this study, 10 families, or 8.77%, of the initial sample of 114 families completedtreatment but were omitted from the analyses because of missing data. In all of these cases,entire sets of data were either missing (i.e., post-test) or invalid (response sets).Consequently, the family had to be dropped from the data analytic procedures. In four ofthe cases, the data were expected by TARP, however, in these cases the post-test data hadnot been received at the time that the analyses for the present research were performed.A second form of family with incomplete data sets is not included in the estimateabove. This second group represent those who had yet to return follow-up questionnaires.As this study was parcelled out of an ongoing research project, a number of families includedin the present inquiry had not completed the follow-up wait period at the time that datacollection closure was determined. Consequently, data from four of the families in theExST-I and seven of the families in the ExST-C treatment conditions were unfortunately notavailable for the follow-up analyses.Family Participant InvolvementAlthough the percentage of those who completed treatment and were not missingdata does not appear as high as would be preferred, the percentages of the groups aregreatly affected by the inclusion of the pre-treatment drop-out group in the calculations.The 20 families that dropped out without ever presenting themselves to a single session oftherapy are best not viewed as therapy drop-outs. Figure 2 presents a recalculation of theparticipating family groups percentages, excluding the pre-treatment drop-out group fromthe calculations.76CompleteTreatment63.8%Figure 2. Therapy participant sub-groups.• IncompleteData25.5%MissingData10.6%As revealed in Figure 2, 63.8% of the participants that actually started therapy werein fact in the complete treatment group. Only 25.5% of the participants that enteredtreatment failed to complete their course of therapy. Finally, just 10.6% of the sub-samplethat started therapy were not included in the outcome analyses due to missing data sets.Data CollectionAn overview of the TARP research protocol is presented below in Figure 3. Asdisplayed in the figure, prospective participants attended an initial screening interview (Ti)INFIETiC GA1siI0 TIlES:TI • Initial Project Heetin12 Pr.-Treat.ent Qeitfw,eiree13 • Kid-Treat.ent 0u.stiowirec14 • Pest-Treatment 0uestti,airesTI FoLIow-.p 3Xesti.,resTREATMENTASSIGNMENTInitialClinic TIContactTREATMENT FOLLOW-UPTREATMENTFORMATS,portedFee&ack —,TherapyEXITI,ividual —, 12TherapyWeltCouples —‘Therapy—, 7 —,Weeks—, lb —pWeeks—.‘ 10 —.WeeksT3—, S —,Weeks_._._—, 10 —,Weeks—‘ 10 •—‘W.ekl E •—. 15 —X4ek5- I5-’Weak.Figure 3. Overview of TARP research protocol77and received a battery of tests to verify their suitability for inclusion in the study. Thecomposition of the screening test battery is provided in the instruments section of thisdocument. The screening interview, which included both the husband and wife of thepotential participating family, was conducted at the treatment clinics and extendedapproximately 2 hours in duration.The one hundred and fourteen male alcoholic husbands and their partners thatpassed through the screening procedures were randomly assigned to one of the threetreatment groups and to one of the therapists providing the treatment. A pre-establishedrandom walk generated through the use of a random numbers table was used to assignparticipants into a treatment condition in successive order of inclusion into the study.Although it could be assumed that the randomization procedures were adequate to balanceout pre-treatment group differences, a series of preliminary analysis of pre-treatment datawere conducted (see Results section, pg. 137) and verified treatment group comparabilityprior to the provision of therapy. These analyses were based on data collected at screening(T1) and pre-treatment (T2). Families, including fathers, mothers and children over the ageof 9, were required to complete the pre-treatment questionnaire package before they werepermitted to commence treatment. The following is a description of the remaining datacollection procedures involved for each treatment group once screening was completed.SFT Treatment GroupThis treatment group involved fathers attending regular weekly or biweekly sessionsof individual treatment with a therapist over a maximum 16 week period. The mean numberof sessions was 7.6, SD = .87. As part of this treatment, both father and mother carefullymonitored aspects of their lives on a weekly basis using Weekly Situation Diaries (WSD)provided by the study. Father was asked to answer a short post-session review form aftereach meeting with his therapist. In addition to completing the pre-treatment questionnaire78(T2),families in this group were asked to complete a short mid-treatment (T3) (not includedin this study) and post-treatment (T4) questionnaire booklets. The mid-treatmentquestionnaires were completed after seven weeks of therapy while the post-treatmentquestionnaires were completed at the conclusion of therapy after 16 weeks of treatment hadelapsed. Since the SET group was functioning as a quasi-control group, SET families werenot asked to participate in the follow-up portion of the study and were informed at screeningthat further treatment would be available once the post-treatment questionnaires had beencollected.ExST Individual Treatment GroupIn this treatment group fathers attended regular sessions of individual treatment witha therapist over a maximum twenty week period. The mean number of sessions was 12.8,SD = 2.56. In addition to the pre-treatment questionnaire booklets (T2), families in thisgroup were asked to complete a short mid-treatment (T3), post-treatment (T4) and followup (T5) test batteries. Data from the mid-treatment questionnaires were collected after thefifth session but are not employed in the present study. Participants in this group wererequired to wait 15 weeks after the treatment had concluded before undertaking anyadditional therapy. The follow-up questionnaires were collected at the end of this waitperiod. Father was asked to answer a short post-session review after each therapy meetingand both father and mother were asked to monitor aspects of their lives on a weekly basisusing WSD forms at the end of each week.ExST Couples Treatment GroupThis treatment involved fathers and mothers attending regular sessions of maritaltreatment with a therapist for a maximum of 15 sessions over a 20 week treatment period.The mean number of sessions was 13.3, SD = 2.40. Like the ExST individual treatmentgroup, participating families in this treatment condition were asked to complete midtreatment (T3), post-treatment (T4), and follow-up (T5) questionnaire booklets in addition79to the pre-treatment questionnaires (T2) they completed before entering therapy. Brief mid-treatment questionnaires were completed after the fifth therapy session or tenth week oftreatment, however, none of the information from this data collection occasion is included inthis research. Post-treatment questionnaires were collected at the conclusion of treatment.Participants in this group were also asked to wait 15 weeks after the treatment had finishedbefore undertaking any further therapy. Once this 15 week period had passed, these familiescompleted the follow-up questionnaires. Father and mother were both asked to fill out shortpost-session reviews after each therapy meeting and also requested to monitor their lives ona weekly basis using the WSD forms at the end of each week.It should be noted that none of the data collected at mid-treatment has been used inthe present inquiry. The inclusion of the mid-treatment data collection occasion in thissection’s description is for the purpose of procedural clarity with respect to the researchactivities that participants completed.All therapy sessions conducted as part of this study were videotaped with the consentof the participants. The consent form may be found in Appendix D. Clients andparticipating therapists routinely responded to the post-session reviews after each therapymeeting. Participants were not discouraged from attending support groups such asAlcoholics Anonymous over the course of their involvement with the project.The questionnaire booklets completed at pre-treatment, mid-treatment, posttreatment, and follow-up were extensive (see Instrument section, pg. 90) and requiredconsiderable commitment on the part of participating families. In order to recognize thetime and effort involved in completing each test battery and to help minimize attrition,participants received an honorarium of up to $200.00 over the course of the study which wasdistributed on an established incremental pay schedule after each test battery was received80by the investigation. The payment schedules employed by the study are provided inAppendix E.Clinical ContextThe present investigation was conducted at multiple sites. The research was done incooperation with two alcohol and drug treatment centers funded by the British ColumbiaProvincial Government. The Surrey Alcohol and Drug Programs Clinic (Surrey Clinic) andSummit Clinical and Consulting Services in Duncan (Summit) were selected as treatmentsites. While the Surrey Clinic operated in a busy, expanding urban context, Summit offeredservices in a much smaller, more stable rural setting. Consequently, Surrey Clinic had alarger treatment and support staff to meet the greater demand for services than did theDuncan operation and clinic procedures reflected the different demands of an urban andrural clinical context.Despite these kinds of differences, both clinical sites shared important qualities thatmade them suitable clinical sites including administrations that were supportive of research,access to clients appropriate for the study, facilities enabling the project such as, researchoffice spaces, large therapy rooms, one way mirrors and videotaping capacities and teams ofqualified and enthusiastic therapists.TherapistsA total of 12 therapists participated in the delivery of the treatment underinvestigation. All therapists had completed studies at the Master’s level or more inPsychology, Social Work or a related field in order to participate in the study, and all had tohave a minimum of 3 years direct work experience providing therapeutic services to alcoholdependent individuals and their families. Of the 12 participating therapists, 5 were involvedin providing the individual and marital forms of ExST and the remaining 7 delivered the SFT81treatment. The average age of the therapists was in the late thirties (1 = 38.75 yrs., SD =6.58 yrs.). The therapists providing ExST made up a somewhat older group ( = 42.40 yrs.,SD = 7.80 yrs.) than the therapists delivering the SFT treatment (.? = 36.14 yrs., SD = 4.45yrs.). The average number of years of therapists practicing psychotherapy was considerable(1 = 8.17 yrs., SD = 4.37 yrs.) and the number of years working with alcoholics indicates thatall therapists were well established in this work (Y = 5.70 yrs, SD = 3.25 yrs.). Thecomparison between the two groups of therapists is in keeping with their differences in age.The ExST therapist group had practiced psychotherapy a little longer (1 = 9.2 yrs., SD =4.55 yrs.) than their SFT counterparts ( = 7.4 yrs., SD = 4.43 yrs.). Similarly, the ExSTtherapist had worked in the alcohol field slightly longer ( = 6.10 yrs., SD = 2.46 yrs.) thanthe SFT practitioners ( = 5.43 yrs., SD = 3.74 yrs.).Treatment ImplementationIn order to ensure treatment fidelity, Kasdin (1986) proposed 3 principal steps. Toestablish treatment integrity (the extent to which treatments have been conducted asintended), the first step is to systematically train the therapists to implement the treatment.The second step is to establish that therapists adhere to the treatment procedures throughthe course of treatment delivery. The third and final step is to assess the extent to whichtreatment procedures were in fact followed.In this study, the treatments have been implemented in a manner that was in keepingwith Kasdin’s principles. This is to say that therapists were: (1) systematically trained toprovide SFT and ExST and, (2) supervised on a regular basis over the course of the study.The formal assessment of the therapists at the conclusion of the investigation in terms of theextensiveness to which they adhered to treatment protocols and the degree to which theyemulated a competent therapist in the therapy they were providing has not been included in82this study. A rationale for this omission is provided in this section. The procedures followedin implementing the SFT and ExST therapies are presented below.SET Therapist Training and SelectionTherapists in this treatment condition volunteered to participate in the study. A totalof nine therapists from the participating clinics expressed an interest in providing thistreatment. All but one of the SFT therapists had previous training in ExST. Each therapistwas provided with a copy of the SET treatment manual which was thoroughly read prior toparticipating in a series of training meetings conducted at the clinics. The SFT training wasconducted over a 6 week period and lasted 20 hours in duration. The training focused ontherapist mastery of the treatment as specified in the manual. Training procedures includeddidactic presentations, discussion groups, role plays, video tape presentations and videotaped therapy rehearsals. Seven therapists completed the training and were able tosatisfactorily demonstrate competent treatment implementation in role play situations to thetrainers and developers of the treatment approach and were consequently selected for thestudy.SET SupervisionSupervision of the SET therapists was conducted in a group setting on a regularbiweekly basis by developers of the treatment approach. Group supervision proceduresincluded case management and planning, as well as videotaped reviews of sessions anddidactic presentations regarding the principles of the therapy and the implementation of themodel. In addition to the group supervision meetings, individual supervision was availableon a case by case basis to the therapists who requested more in-depth supervision and also atthe request of the supervisor in those instances where further supervision appearednecessary in order to maintain a high level of treatment adherence.83ExST Therapist Training and SelectionA total of 14 therapists from the two clinics volunteered and received training inExST. All therapists were provided with the ExST training manual which provided afoundation and a focus for a 12 day training workshop conducted over a 3 month period inthree 4 day training events spaced one month apart. Since the treatment involved complextechniques and considerable clinical skill, each therapist received 10 hours of direct orvideotape supervision sessions subsequent to the training. The supervision focused onvarious aspects of treatment implementation with individuals and couples.In order to be selected for the investigations, each therapist was required to collectand submit 5 videotapes of individual and couples therapy sessions. From the pool of 10tapes provided by prospective therapists, two individual and two couples therapy sessionswere selected at random and reviewed independently by two trained adjudicators whodetermined the therapists capacity to implement the treatment in both individual andmarital therapy conditions with a sufficient level of integrity and competency. Of the 14therapists that received training, 5 were selected through this procedure and served astherapists delivering the individual and marital treatments of ExST.ExST SupervisionThroughout the course of the research, therapists continued to receive supervision ona regular weekly basis. Group and individual supervision formats were employed by 2trained supervisors who provided live and videotape consultation of both individual andmarital therapy treatments to the ExST clinicians. These consultation meetings includedreviews of theory and technique with particular attention focused on ensuring that thetherapists treatment practices emulated the ExST model and minimized therapist drift fromthe model.84The ExST model of supervision (Newman, Friesen, & Grigg, 1991) guided theorientation of the consultation meetings. This model of supervision places emphasis onthree domains of practice including theoretical development, technical refinement andpersonal growth. Supervision sessions often employed experiential activities which wereconsistent with the ExST therapy being implemented by the therapists in order to addressthe three domains of practice (e.g., role play, role reversal, enactment, artwork, sculpting,symbolic evocation, ritual conduct, and process recall). The focus of consultation tended tobe on the here and now. The belief that therapists did not need to talk Habout the problemsthey were having with clients, but rather needed to experience the problems they werehaving with clients and work through them in a safe supervisory setting was shared bysupervisors and therapists. Thus, supervision meeting provided a rich and generative spacefor therapists and supervisors to encounter the ExST model in a direct way in order to morefully appreciate its practical implementation and ensure model adherence and therapistcompetency.Treatment Implementation CheckA therapist adherence and competency rating scale has been devised by TARP staffin consultation with some of the originators of the two treatment approaches. TheAdherence Rating Scale (TARS) (Thompson, Friesen, Grigg, Weir, & Mitchell, 1993) wasmodeled after the Collaborative Study Psychotherapy Rating Scale developed by Hollon,Waskow, Evans, and Lowery (1984). TARS was designed to measure treatment adherencein terms of extensiveness of implementation competency, and emulation of the idealtherapist as described in the treatment manuals.TARS is comprised of items which tap unique aspects of SFT and ExST, and itemsthat measure areas shared by the treatments. There are a total of 15 ExST items, 15 SFTitems and 15 overlap items that have been randomly ordered in the instrument. Each item is85scored in two parts. Both parts are concerned with some element of treatment (e.g.,Therapeutic Rapport). The first part of the item asks raters to determine the extensivenessof the rapport between therapist and client and to rate this element in terms of a 5-pointLikert scale ranging 0 = none at all to 4 = thoroughly. The second part of each item istermed “emulation” and is concerned with the competency of therapists’ activities. Theemulation portion of the item asks raters to determine how close to the ideal (as specified inthe treatment manuals) the therapist came in the segment being assessed. Raters determinetherapist emulation in terms of a 5-point Likert scale that ranges from 0 = not at all to 4 =very close.Each item in TARS is carefully described and accompanied by examples and notes ofclarification. Every effort was made in the generation of TARS to ensure that the qualitiesbeing measured were observable constructs. While a good working knowledge of the twotreatments is important to the employment of TARS to rate videotapes of actual sessions,raters do not need to be experts in the conduct of either of the therapies. A copy of theTARS manual is available upon request; however, a copy of the rating form of theinstrument appears in Appendix F.The work of implementing TARS and quantifying the levels of treatment adherenceand therapist competency is currently ongoing as part of the TARP research protocol.However, since the TARS study is currently in progress, the results pertinent to thisinvestigation are unavailable at this time. Consequently, the measures of treatmentimplementation have not been included in the present investigation.Notwithstanding the omission of the measure of implementation verification, a highlevel of treatment fidelity characterized this study. The salient components whichcontributed to the high level of therapist competency and treatment model adherence were86the operationalization of the treatments in manuals which clearly specified the parametersof the treatments, the thorough training of therapists to implement the treatments based onthe manuals, the careful selection of the participating ExST and SFT therapists from thestable of therapists that had been trained based on their work implementing the models, andthe regular weekly or bi-weekly supervision of all project therapists by qualified supervisorsover the course of the entire study.Ecological Assessment ModelIn order to adequately answer the research questions posed in this investigation, acontextually sensitive assessment perspective was necessary. The approach to measurementhad to reflect the complex systemic flavor of the inquiry, yet it had to be structured in such away as to be clearly organized in a meaningful fashion.It was argued that the adoption of an ecological perspective allows for acomprehensive appreciation of the process and outcome of change in terms of the contextsin which action is manifest. The pioneer work of psychologist Urie Bronfenbrenner (1977,1979) construed human development as occurring in a series of hierarchically embeddedcontexts. From this point of view, individuals were viewed as growing entities who, whileimpacted by the environment, also affected and had impacts on the environment. This is tosuggest that individuals and environment exist in a dynamic relationship with one another.Both individuals and environment are viewed as standing in a reciprocal fashion such thateach has an influence on the other, and through a process of mutual accommodation, twoway interaction results. Articulating the definition of an “ecological experiment”Bronfenbremier (1979) suggested that such an inquiry should be:an effort to investigate the progressive accommodation between a growinghuman organism and its environment through a systematic contrast between 2or more environmental systems or their structural components, with a carefulattempt to control other sources of influence either by random assignment(planned experiment) or by matching (natural experiment). (p. 36)87While Bronfenbrenner (1979) was chiefly concerned with focusing on environmentalimpacts, the ecological approach was adapted by Conger (1981) to include the focus ofindividual development. Conger (1981) argued cogently that individual measurement mustgo hand in hand with ecological assessment in order to enable a comprehensive appreciationfor change and development.This kind of assessment is capable of reflecting change at a variety of levels ofsystemic observation. Thus, the ecological perspective aptly addressed the centralmeasurement concerns of the present investigation. Consequently, an ecological model ofassessment was adopted by the investigation to examine the differential treatment effects ofboth the two treatments (SFT and ExST) and individual and couples ExST treatmentformats. The assessment model which was developed concerned itself with themeasurement domains of alcohol dependency, individual functioning, marital relationship,and family functioning. The ecological assessment model assumed that each of theassessment domains would tap salient aspects of nested systems that interact in some waywith one another.From an ecological assessment perspective, change in the reference problem ofalcohol consumption is contextualized and related to measurable changes in individual,marital and family levels of system. Treatments (ExST and SF!’) and format (individual andcouple) are to be evaluated in a number of systemic domains. The assessment model used inthis study is in Figure 4.88SYSTEM / SUB-SYSTEM ASSESSMENT LEVELAlcohollndMdualCoupleFamilyFigure 4. Ecological assessment model.There is consistency between the systemic questions of the investigation, the researchdesign and the ecological assessment model that was used. It should be noted that theecological approach to measurement addresses an important concern in the alcoholismtreatment literature. It has been noted that treatment efficacy studies in the past havefocussed primarily on alcohol related dimensions as indices of effectiveness (Nathan &Skinstad, 1987). The preoccupation with changes in alcohol consumption has left manyimportant dimensions of recovery uncharted. Concerned by this tendency, Spicer (1980)stressed that:One of the common mistakes in the area of defining treatment success is touse the abstinent/non abstinent criteria. Repeatedly researchers have foundthat although drinking can be an accurate indicator of program outcome, notall clients who are abstinent are doing well in other areas of their lives. (p. 47)89Clearly, the ecological assessment approach used in this study responds to the callsfor broader scopes of evaluation (Billings & Moos, 1983; Emrick & Hansen, 1983; Friesen,1983; Spicer, 1980) and offers a model for the rigorous evaluation of other treatmentprograms.Implementation of Ecological Assessment ModelIn this study, data regarding each assessment domain or level of system were notobtained from every family member. Children were not asked to comment on father’s intrapersonal condition, or their parents’ marriage. Rather, as members of the family system,children reported on the family domain only. Similarly, wives were not asked to assess theirhusbands’ relationship with alcohol. The systemic assessment levels and family memberperspectives that were tapped are presented below in Table 2.As revealed in Table 2, perspective (response on questionnaire) was obtained bythose directly connected and experiencing a particular systemic domain. Consequently,while fathers reported about their own alcohol consumption, their own intrapersonalcondition, the marriage and the family, they were not asked to assess their wife’s intrapsychicfunctioning. Similarly, mothers reported on themselves, their marriage and their family, butwere not asked to focus on their husband’s alcohol consumption or intrapersonal situation.In this study, the eldest child was chosen to report on his or her views of the family. Thisdecision was made in order to avoid violating the assumption of independence and the dataanalytic difficulties related to uneven numbers of children responding about one family. Theecological assessment model represents a multidimensional, multiperspective measurementapproach that enables a comprehensive evaluation of treatment efficacy when it is applied ina pre, post, and follow-up experimental design.90Table 2Perspective of Respondent and Systemic Assessment LevelPerspective Alcohol Intrapersonal Marriage FamilyFather X X X XMother X X XChildren xInstrumentationThe instruments that were selected for the investigation tap important aspects of thesystemic domains specified by the ecological assessment model. The array of instrumentschosen to operationalize the ecological approach had to satisfy all of the following criteria:(1) The tests had to measure specific qualities, characteristics and/or behaviors that werecentral to the treatment of alcohol dependency at each level of ecological assessment, (2)Instruments considered for the study had to display adequate psychometric properties andhad to be sensitive enough to measure change, (3) Instruments had to have been used inprevious outcome studies and have demonstrated utility in them. In addition to theinstruments used in the change assessment battery, several other instruments were selectedfor specific pragmatic purposes of the studies design (e.g., screening).Preceding the assembly of the TARP instrument array, a large test battery waspiloted and some instruments were dropped as a result of the pilot test analyses. Theinstruments included in this study are a sub-set of a larger test battery employed in TARP.91They were selected for the present investigation prior to any statistical analyses of the largertest battery.The instrument package employed in this inquiry presented below has beensubdivided in terms of the ecological assessment level to which the instrument belonged: (1)alcohol measures, (2) intrapersonal system measures, (3) couples systems measures, (4)family system measures. A description of all measures follows in the section below. A listingof the measures employed and the test occasion schedule followed in this investigation isprovided in Appendix G. In the interest of brevity, a copy of the instrument package has notbeen included in this report, however a copy is available upon request. A description of eachindex used in the study follows. The descriptions have been presented in terms of themeasurement domains of the ecological assessment. While the questionnaire batteriesordered the measures in terms of ecological assessment level (e.g., alcohol, intrapersonal,couple, family), the order of instrument presentation within each level was randomlydetermined for each measurement occasion.1. Alcohol Measures DescriptionAll of the instruments used in this study to measure alcohol use were self-reportmeasures. Some researchers have expressed concern over the validity of self-reportmeasures of alcohol consumption asserting that alcoholics may be prone to obscure andunder report their consumption rates (Guze, Tuason, Stewart, & Picken, 1963; Midanik,1982; Miller, Crawford, & Taylor, 1979). Research probing the question of the reliability ofalcoholic self-report accuracy has, in the main, assuaged the controversy. Studies havedemonstrated a high level of consistency of self reported alcohol use on collateral reports(Maisto, Sobell, & Sobell, 1979) and independent primary sources of information such asarrest records and similar accounts (Sobell & Sobell, 1975; Sobell, Sobell, & Samuels, 1974;Sobell, Sobell, Riley, Schuller, Pavan, Cancilla, Klajner, & Leo, 1987). These findings92support the use of self-reported alcohol measures as a reliable and valid method ofascertaining the clinical status of drinking pattern.In this study, the alcohol dependent husbands were asked to report on their ownalcohol use and their spouses were not asked to act as collaterals. The rationale for notasking wives to report on their partners’ alcohol use was two-fold. Firstly, the activity wasjudged to be unnecessary. In a recent outcome study in which spouses were used tocorroborate alcohol use estimates by their partners (Bowers & Al-Redha, 1990), a rate ofover 90% agreement between spouses was observed at post-treatment and one year follow-up. With such a high level of agreement being observed between couples and the meaningof the differences in the remaining 10% of the estimates being ambiguous at best, the effortto collect such information was questionable. Secondly, the activity was in conflict with thetreatment goals. An important treatment goal was to promote the alcohol dependentindividual’s ownership of the problem and to encourage the non-abusing spouse to focus onother important concerns in her life. In addition, by asking the spouse to also report on herhusband’s drinking, the research procedures would have in essence implied that thealcoholics could not be trusted to report this information reliably and thereby would haveplaced the spouses in awkward positions in terms of observing and recording their husband’sdrinking activities. Consequently, the research procedure of asking the non-drinkingpartners to estimate their husband’s alcohol use for the purposes of research was at oddswith the therapeutic efforts of the study and not employed.The instruments used as a means of measuring the identified problem behavior ofalcohol consumption were:(I) Michigan Alcohol Screening Test (MAST) developed by Seizer (1971),(II) Alcohol Dependency Data (ADD) developed by Raistrick, Dunbar, and Davidson(1983),93(III) Inventory of Drinking Situations (IDS-42) developed by Annis (1982),(IV) Situations Confidence Questionnaire (SCQ-39) developed by Annis and Graham(1988).While the MAST was employed for screening purposes, the ADD was selected from thesealcohol indices as the marker variable to be used as the primary indicator of change in thisdomain in the ecosystem analyses (see Data Analysis section, p. 117). Details concerning thereliability and validity of each instrument follow.I. Michigan Alcohol Screening Test (MAST)The MAST was designed by SeIzer (1971) to provide a consistent, quantifiablemeasure for the detection of alcoholism. The MAST, is a 25 item questionnaire that uses aforced choice, yes/no response format. Items are weighted and total scores may range from0 to 53. A score of 5 or greater indicates a diagnosis of alcoholism. In this study, the MASTwas selected to identify alcoholics at screening and to detect any alcoholic drinking problemswith which the spouses might have been struggling.The author of the instrument reports that many of the items included in theinstrument have been used by other surveys of alcoholism which may account for its popularacceptance as a measure of alcoholism among researchers and clinicians. The MAST hasbeen used with many different subject groups including: alcoholics, persons convicted ofdriving while intoxicated, social and problem drinkers, drug abusers, psychiatric patients,general medical patients, pregnant women, college students, hospital personnel, andconvicted felons (Hedlund & Viewag, 1984).The instrument is constructed to provide a stable identification of a drinking problem.In fact, MAST scores are not affected by current drinking status to such a degree that an94individual who was once heavily dependent upon alcohol would score as an alcoholic on theMAST even after years of sobriety (Hedlund & Vieweg, 1984). Reliability evidence issomewhat scant considering the instrument’s widespread use (Gibbs, 1983). Internalconsistency estimates from 6 studies reviewed by Hedlund and Vieweg (1984), reportedalpha coefficients for the MAST ranging from .83 to .95. Test-retest reliability coefficientshas been reported as .97 for a 1 day test-retest interval, .86 for a 2 day interval, and .85 for a3-day interval. Skinner and Sheu (1982) reported a reliability coefficient of .84 for anaverage 4.8 month test-retest interval for 91 acute psychiatric admissions.In addition to the original validation studies, MAST total scores have been shown tobe significantly correlated to a variety of related instruments including the GeneralAlcoholism Factor of Alcohol Inventory (Skinner, 1979), the MacAndrews Alcoholism Scale(Friedrich & Loftsgard, 1978), the Alcohol Volume Index and the Alcohol Pattern Index(Sokal, Miller, & Debanne, 1981). Factor analytic studies probing the factor structure of theinstrument reported by Zung (1980a, 1980b, 1982) found that one major factor accounted for49 to 78 percent of all common variance which has been interpreted as “General AlcoholicImpairment”. Other factors that have been noted as reasonably consistent across otherstudies (Hedlund & Vieweg 1984) include recognition of alcohol problem, help seeking,marital discord, and legal, work and social problems.The validity of the instrument was originally assessed by searching the records oflegal, social and medical agencies and reviewing respondent’s driving and criminal recordsand linking these to MAST classification. Five different groups participated in the validationstudy. The groups included: hospitalized alcoholics, drivers convicted of driving under theinfluence of alcohol, persons convicted of drunk and disorderly conduct, drivers who hadincurred 12 penalty points in 2 years for traffic violations and accidents and a control group.95This study established that the MAST could be used to classify alcoholics andnonalcoholics even when distortion or minimization of the problem was anticipated. In asecond study, the MAST was given to hospitalized alcoholics who were instructed to lieabout their drinking. Despite this instruction, the MAST apparently correctly identified 92%of these individuals using a cut-off score of 5 or greater.II. Alcohol Dependence DataThe ADD was developed by Raistrick, Dunbar, and Davidson (1983) as aninstrument to measure the severity of alcohol dependence as described by Edwards andGross (1976). In designing the questionnaire, the authors sought to ensure the instrumentwas suitable for clients seeking help with drinking problems and measured the clients’alcohol dependence in its present state. In addition, the instrument was constructed toreflect the full range of dependence and to be sensitive to change in dependency level overtime.The ADD is comprised of 39 items which are assessed on a 4-point Likert typefrequency scale ranging from never = 0 to nearly always = 3 and yielding a maximumdependence score of 117. The instrument generates a single dependency index score andranges of dependency levels have been provided by the authors that stratify the index scoresinto no, mild, moderate and severe dependency groups. While a score of 0 indicates nodependency, a score ranging from 1-30 suggests mild dependence, 3 1-60 signals moderatedependence, and 61-117 indicates severe levels of alcohol dependency.A 15 item shortened form of ADD has been generated with the correlation betweenthe full questionnaire and the shortened form reported as highly significant (r = 0.92). TheSplit-half reliability estimates based on the shortened form was high (r = .87). Furtherevidence of internal consistency of the instrument is based on Spearman Rank correlationsbetween items and total score with significance levels ranging from <0.03 to <0.001.96Studies probing the validity of the instrument have been based on the shortenedversion of the questionnaire (Davidson & Raistrick 1986; Davidson, Bunting, & Raistrick,1990). The construct validity of the instrument is closely tied to the validity of the Edwardsand Gross (1976) notion that dependence is a single unidimensional phenomenon. Theresults of 3 separate factor analytic studies confirmed that there is a strong commonalitylinking all items. With the exception of a single question, the items can be best representedin terms of one strong first factor. Attempts to identify a clear and consistent second factorwhich might underlie the instrument have proven to be impossible. Consequently, theauthors assert that the validity of the dependency construct is supported by factor analysis.Concurrent validity of the instrument has been assessed by comparing the test scoreswith a variety of other measures related to aspects of alcohol dependency including liverfunction tests, other recognized tests of alcohol dependency, and semi-structured clinicalinterviews. The results from 3 separate studies reported in Davidson and Raistrick (1986)support claims of concurrent validity.III. Inventory of Drinking Situations (IDS)Developed by Annis (1982), the IDS is a situation specific measure of drinking aimedat identifying an individual client’s high risk situations with regards to heavy alcoholconsumption. The drinking situations assessed by the questionnaire were based upon earlierwork by Marlatt and his associates (Marlatt, 1978,1979; Marlatt & Gordon, 1980) who foundthat high risk drinking situations could be categorized into 2 major classes, either PersonalStates or Situations Involving Other People. In the IDS, the Personal States class wasfurther subdivided into 5 categories: Unpleasant Emotions, Physical Discomfort, PleasantEmotions, Testing Personal Control and Urges and Temptations. The Situations InvolvingOther People was subdivided into 3 categories which include: Conflicts with Others, SocialPressure to Drink and Pleasant Times with Others.97The IDS generates Problem Index scores for each sub-scale which are derived bydividing the obtained score on the sub-scale by the maximum total possible on that sub-scaleand then multiplying by 100. The manual provides interpretative ranges for problem indexscores which are based on normative data from 202 male and 134 female subjects. Aproblem index score of 0 on any sub-scale indicates the participant has never consumedalcohol in a heavy fashion on the last 3 months and is in the “Low Risk” category andunlikely to develop alcohol problems in this area. The problem index range of 1-33 indicatesModerate Risk and indicates participant rarely drank heavily in these types of situations.When the problem index score is with the range of 34-60 on any sub-scale, the respondent isin the “High Risk” category and has heavily consumed alcohol frequently in these situations.Finally, scores in the range of 67-100 indicates the “Very High Risk” category and indicatesthat participants very frequently drank heavily in these types of situation.The original IDS instrument was 100 items in length, however, a shortened 42 itemform has been developed for research purposes (IDS-42). The present study selected theIDS-42 version because it provided the same kinds of information in a reliable fashion usingless items. Annis, Graham, and Davis (1987) report that the relationship between the IDS42 and the original 100 item form was found to be very strong with sub-scale correlationsranging between .93 and .78. The internal consistency reliability (alpha) estimates for theshortened sub-scales range from .80 to .92 which was only marginally lower from the original100 item sub-scale alpha estimates which ranged from .87 to .96.In terms of the validity of the instrument, Annis et al. (1987) reported that themeasure demonstrates good content validity. Five expert clinicians were consulted to ensureitem clarity and adequacy of item coverage of common relapse situations. In addition, whenthree trained raters were used to validate the classification system and asked to sort the 100items into the 8 categories, a high inter-rate reliability of item placement was observed (92%to 99% agreement rate).98Estimates of the external validity of the IDS were made by correlating the sub-scalescores with measures of alcohol consumption and alcohol dependence. The estimates oftotal quantity of alcohol consumed during the past year were significantly correlated witheach sub-scale (range of 8 r’s = .27 to .43) as was typical daily drinking quantity (range of 8i’s = .12 to .27). This relationship indicates that clients reporting higher levels of drinkingalso received higher scores on the IDS. The IDS was also compared to the AlcoholDependence Score (ADS), an instrument developed by pioneers in the area (Skinner &Horn, 1984) and moderate correlations with the IDS sub-scales (range of 8 r’s = .23 to .52)were found. This relationship established the connection that those who reported morefrequent heavy drinking were also exhibiting more signs of alcohol dependency. In additionto these correlations, Annis et al. (1987) also correlated sub-scale scores with informationregarding clients’ social context of drinking and also numbers of years of problem drinkingand reported significant relationships between these variables. The authors conclude thatthe convergent validity evidence supports the claim that the IDS not only measuresfrequency of drinking but also reflects situation specific patterns of heavy alcoholconsumption.TV. Situational Confidence Questionnaire (SCO)The SCQ was developed as a direct offshoot of the IDS by Annis and Graham (1988).The measure was created as a tool to assess the development of a client’s self efficacy orconfidence in relation to specific drinking situations over the course of treatment and as aindex for the study of treatment outcomes. The SCQ shares the same conceptual tie with thework of Marlatt and his associates (Marlatt, 1978,1979; Marlatt & Gordon, 1980) as does theIDS, and is structured in the same fashion (8 sub-scales and 2 major classes). Accordingly,the SCQ, which is 39 items in length, offers a kind of mirror of the IDS. This is to say thatwhile the IDS provides information regarding how much alcohol was consumed in whichsituations, the SCQ provides measures of the level of confidence that clients’ feel regarding99their ability to avoid drinking heavily in the same personal (Unpleasant Emotions, PhysicalDiscomfort, Pleasant Emotions, Testing Personal Control, and Urges and Temptations) andsocial situations (Conflict With Others, Social Pressure to Drink, and Pleasant Times WithOthers).The internal structure of the SCQ was evaluated using a series of factor analyticprocedures including an exploratory and confirmatory factor analysis. The latter procedureallows for the specification of hypothetical factor structures which are then tested with anobserved data matrix to test the adequacy of fit between model and actual data. Thisprocedure resulted in the authors dropping 3 of the original 42 items from the test andarriving at the 39 item measure.The reliability of the SCQ was based on item-total correlations and internalconsistency (alpha) estimates. While the item-total score correlations with each of the sub-scales were acceptable and ranged from .59 to .91, the internal consistency estimate for eachsub-scale was high with alpha ranging from .81 to .97.The SCQ allows for normative comparisons based on a group of men who wereentering treatment for a variety of alcohol related problems. In addition, scores areinterpreted as indicating the percentage of confidence an individual has regarding thepossibility that he/she will not drink heavily in a particular situation (e.g., score 40 = 40%confident).The SCQ is theoretically linked to the construct of self-efficacy as conceptualized byBandura (1977). The measure is concerned with respondents’ confidence in their abilities toresist the urge to drink and to subjectively identify situations in which they are confident theywill not drink heavily. Estimates of construct validity for the instrument were made by100correlating sub-scale scores with measures of alcohol consumption, the social context ofdrinking and other indices which reflect the construct of self-efficacy. With regards to actualdrinking, the relationship between SCQ sub-scales and consumption were found to begenerally low, but in the main, significant and in predicted directions. In addition,correlational patterns with measures relevant to the self-efficacy construct conformed topatterns predicted by theoretical association.The SCQ has been shown to demonstrate good criterion-related validity. Miller,Ross, Emmerson, and Todt (1987) showed that the instrument could correctly classify 92%of long term abstainers and 65% of new clients entering an alcohol treatment center. Themeasure has also demonstrated predictive validity with Solomon and Annis (1988) reportingthat SCQ scores obtained at intake to treatment predicted average consumption on drinkingdays of clients who drank following treatment discharge. While the instrument failed topredict both the occurrence and the frequency of drinking occasions during follow-up, it wasa strong predictor of the quantity of alcohol consumed when clients relapsed.2. Intrapersonal Measures DescriptionThe second assessment domain, intrapersonal functioning, was of great concern to thestudy. Understanding how the individual system was affected by the treatment procedureswas seen as a critical assessment aim. Consequently, the following indices were selected totap this system. The intrapersonal measures schedule of administration is included inAppendix G. In the study, both father and mother participants completed the intrapersonalquestionnaires despite the fact that 66% of the mothers in the investigation did not directlyparticipate in therapy sessions. Jacob et al. (1983) established an association betweennondrinking spouses intrapersonal functioning and their partners’ alcohol consumptionpatterns. Accordingly, it was important to assess the individual functioning of both parentsthroughout the treatment process.101The following instruments were employed:(I) Shipley Institute of Living Scale (SILS) developed by Zachary (1986),(II) Symptom Checklist 90 Revised (SCL-90-R) developed by Derogatis (1983),(III) Beck Depression Inventory (BDI) developed by Beck (1987).In this investigation, the SILS was used as a descriptive tool and the marker variableselected for the eco-systems analyses was the global measure of the SCL-90-R. The BDI andSCL-90-R sub-scales were employed as more detailed measures and were included in withinsystem analyses.I. Shipley Institute of Living Scale (SILS)The SILS is a measure designed to assess general intellectual functioning in adultsand adolescents and to assist in the detection of cognitive impairments. For the purposes ofthis study, the SILS was used as a descriptive tool and to screen out potential participantswhose English literacy level was too low to answer questionnaire batteries.The instrument consists of two sub-tests including a vocabulary test of 40 items, and a20 item test of abstract thinking. Both sub-tests are timed with 10 minutes allotted for each.The Vocabulary sub-test employs a multiple choice format in which respondents are askedto match a specified target word with one of the four possible words provided. TheAbstraction sub-test uses a completion format. Individuals are provided a logical sequenceand asked to respond with the number or letters that best complete the sequence. While thetest generates 6 summary scores, only 2 concern the present study. These include theVocabulary score and the Abstraction score both of which are obtained directly from the testand represent summary scores from the two sub-tests.102The SILS appears to be a reliable instrument. Split-half reliability correctedcorrelation coefficients were .87 for Vocabulary and .89 for Abstraction (Zachary 1986).The Test-retest reliability estimates across 7 studies reported by Zachary (1986) had anaverage of reliability coefficient of .60 for Vocabulary and .69 for Abstraction with a meantest-retest interval of 9.7 weeks.The SILS has been found to be quite highly correlated with a number of other indicesof intelligence. The various Wechsler intelligence tests have been of particular concern.The correlation between the Shipley and the Wechsler-Bellview ranged from .68 to .79.Similarly, correlations between the SILS and the WAIS and WAIS-R were high (.73 to .90with the WAIS, and .74 with the WAIS-R) (Zachary, 1986). Construct validation work onthe Shipley has also included correlating the instrument with measures of intelligence andacademic achievement including the Army General Classification Test, the SlossonIntelligence Test, The Raven, The Quick Word Test, The Wide Range Vocabulary Test, andthe California Short-Form Test of Mental Maturity. All correlations with these otherinstuments reached statistical levels of significance in the predicted direction.II. Symptom Checklist 90 Revised (SCL-90-R)The SCL-90-R is a 90 item self report symptom inventory designed principally tomeasure the psychological symptom patterns of disturbed clients. The instrument uses a 5-point Likert scale ranging from 0, “not at all” to “4, extremely”. The checklist taps 9 primarysymptom dimensions and generates an additional 3 global indices of distress. The 9 sub-scales include: (1) Somatization, (2) Obsessive/Compulsive, (3) Interpersonal Sensitivity,(4) Depression, (5) Anxiety, (6) Hostility, (7) Phobic Anxiety, (8) Paranoid Ideation, and (9)Psychoticism. The 3 global scores are: (1) Global Severity Index (GSI), (2) PositiveSymptom Distress Index (PSDI), and (3) Positive Symptom Total (PST). The global scoresprovide an overall assessment of a respondent’s psychosymptomatic status. Derogatis (1983)103reported that the measure was well suited for pre-post treatment evaluations since he hasbeen unable to detect any significant “practice” effects that might bias the profile onrepeated administrations.The SCL-90-R is a popular self report symptom inventory and has been widely usedas a measure for clinical assessment and treatment outcomes across a wide number of areasincluding depression, sexual disorders, stress, heart disease, pregnancy, schizophrenia, andsubstance abuse.The instrument has been successfully used as an index of change in medical,psychopharmacological, and psychotherapeutic studies. Efforts to demonstrate the SCL-90-R’s concurrent validity have contrasted instrument scores with various scales of the MMPI,(Derogatis, Rickels, & Rock, 1976), the Hamilton Depression Scale, the Social AdjustmentScale (Weissman, Sholmskas, Pottenger, Prusoff & Locke, 1977), the Maudsley Obsessional-Compulsive Inventory (Sternberger & Leonard, 1990), and the Cancer Inventory of ProblemSituations (Schag, Heinrich, & Ganz, 1983). In all of these studies, the correlation patternsof the SCL-90-R sub-scales and the other instrument(s) were significant and in thetheoretically predicted directions. The reliability measures of the SCL-90-R’s 9 primarysymptom dimensions are acceptable. Internal consistency alpha levels ranged from .77 to.90, and test-retest correlations for the symptom dimensions ranged from .78 to .90.With regards to construct validity, a degree of ambiguity exists concerning theindependence of 9 symptom dimensions of the SCL-90-R. The yield from factor analyticapproaches to this question have been mixed. Cyr, Doxey, and Vigna (1988) reportedfinding only 4 of the 9 dimension reliably derived through an analysis of the data from 295psychiatric inpatients and 177 industrially injured workers. In a second study, Brophy,Novell, and Kiluk (1988) reported identifying 6 of the 9 dimension as relatively stable andhomogeneous factors. In addition, these researchers conducted a principle component104analysis which revealed that the first factor accounted for a large percentage of the variancewhich suggested that in the main, the instrument taps a general dimension ofpsychopathology. In contrast to these 2 studies, Derogatis and Cleary (1977) reportedconvincing factor analytic support for the 9 dimensions based on data from 1,002 psychiatricoutpatients.In this investigation, the global symptom index (GSI) was chosen for the eco-systemanalyses and the 9 dimension sub-scales were used in within system analyses.III. Beck Depression Inventory (BDI)The BDI is a 21 item questionnaire designed to measure the severity of depression.Each inventory item is a group of 4 statements that provide a varying range of responsespertaining to a particular aspect of depression. For example, item 7 reads:0 I don’t feel disappointed in myself1 I am disappointed in myself2 I am disgusted with myself3 I hate myselfBeck and Beamesderfer (1974) have provided cut-off scores to assist ininterpretation. They suggest that scores from 0 to 9 are within the normal symptomaticrange, scores of 10 to 18 signal a mild to moderate level of depression, scores of 19 to 29suggest a moderate to severe depressive condition and scores of 30 to 63 indicate anextremely severe depressive state.The BDI is one of the most widely accepted measures for assessing depression in bothpsychiatric patients (Piotrowski, Sherry, & Keller, 1985) and normal populations (Steer,Beck, & Garrison, 1985). In addition, it has been successfully used as a change index in105treatment outcome studies (Beck, Steer, & Garbin, 1988). The instrument has beentranslated into numerous languages including French, Spanish, Japanese, Chinese, andDutch. Importantly, the BDI has been used to investigate and demonstrate the associationbetween alcoholism and depression (Jacob, Dunn, & Leonard, 1984; McLellan & Thomas,1985; Tamkin, Carson, Nixon, & Hyer, 1985).The alpha coefficient for the BDI has been reported as .86. Accordingly, theinstrument can be viewed as an internally consistent and reliable means of measuring theunderlying dimension of depression (Beck & Steer, 1984).3. Couples Measures DescriptionThe marital system was the next assessment domain tapped in this research. Probingthe ways in which the couple was affected by the different treatment procedures was ofcentral concern to the study. Both members of the spousal dyad responded toquestionnaires pertaining to the marriage. In most cases, husbands’ and wives’ scores werekept separate; however, in some situations the scores were combined to generate agreementestimates. In this study, 3 instruments were employed to quantify important aspects of themarital system according to the schedule of administration included in Appendix G. Themarital relationship indices include:(I) Edmonds Marital Conventionality Scale (EMCS) developed by Edmonds (1967),(II) Dyadic Adjustment Scale (DAS) developed by Spanier (1976),(III) Areas of Change (AC) developed by Weiss and Birchler (1975).I. Edmonds Marital Conventionality Scale (EMCS)This measure was employed in the study to ascertain the degree to which couplesresponded to questionnaires in a conventional fashion. Designed by Edmonds (1967) todetermine an individual’s tendency to respond with a socially desirable bias, this instrumentasked couples to identify whether statements were true or false with regards to their views of106their marital relationship. For example, one item statement asserts: There are times whenmy mate does things that make me unhappy.There are 15 items of the EMCS, however, the author suggests that the items are bestmixed with other items that do not measure socially desirability in order to obfuscate thepurpose of the instrument. Consequently, in this study the 15 items were interspersed with 5items drawn from the Marital Status Inventory (Weiss, & Cerreto, 1980).Developed in the late 1960’s, the instrument was administered to 100 randomlyselected married university students. This sample established a mean score of 34 and astandard deviation of 30. The original instrument was 50 items long and it was subsequentlyreduced to a 15 weighted item questionnaire through the selection of the top discriminatoryitems. The correlation between the short weighted version of the instrument and theoriginal long form was very high (r = .99). The reported internal consistency reliabilityestimates range from 0.80 to 0.93 (Zweben, Peariman, & Li, 1988). The 15 EMCS questionweights range from 4 to 10 and the range of the weighted scores for the instrument is 0-89.The instrument is built on the assumption that couples scoring higher in the EMCSare less likely to reveal accurate assessments of their marital relationship on other maritalmeasures. In this study, this EMCS was used to gain an estimate of bias related to socialdesirability.Marital Conventionality has been demonstrated to play a statistically significant rolein the couples scores on both the Locke-Wallace Scale of Marital Adjustment (Edmonds,Withers, & Dibatista, 1972) and the Relationship Inventory (Schumm, Boliman, & Jurich,1980). Marital Conventionality has also been suggested as an important factor in alcoholics’assessment of their marriages (Rychtarik, Tarnowski, & St. Lawrence, 1989).107It has been suggested that the EMCS could be used as a screening device to disallowthose who evidence high scores on the instrument entrance into research studies. The cutoff value of 20 on the EMCS has been suggested as a possible critical value (Edmonds, 1967;Edmonds et al., 1972; Rychtarik et al., 1989). However, at this time, such a strategy does notseem warranted due to the fact that no work has been done on verifying the appropriatenessof such a cut-off score.There is no clear understanding as to the processes involved in social desirableresponse sets. Paulhus (1984) has proposed a two-component model that makes adistinction between self-deception (the individual actually believes his/her positive reports)and impression management (the individual consciously attempts to distort his/her trueassessment). The need to screen out research participants or control for response sets of thiskind would seem to depend largely on the motivation behind the set. Until a method ofdetermining the underlying processes involved in marital conventionality has been found, itwas judged most appropriate to add this aspect to the descriptive information regarding thesample and to temper interpretations accordingly.II. Dyadic Adjustment Scale (DAS)The DAS is a widely used self-report instrument for assessing marital satisfaction.The measure is a 32 item instrument that taps 4 dimensions of the marital relationship with atotal score ranging from 0 to 151 (Spanier 1976). The sub-scales of the DAS include: (1)Dyadic Consensus (the degree to which couples agree on matters important to therelationship) consisting of 13 items, (2) Dyadic Satisfaction (the degree to which the coupleis satisfied with the present state of the relationship and is committed to it’s continuance)with 10 constituent items, (3) Affectional Expression (the degree to which the couple issatisfied with the expression of affection and sex in the marriage) derived from 4 items, and(4) Dyadic Cohesion (the degree to which the couple experiences a sense of togetherness)108with 7 items. Along with generating 4 sub-scale scores, the DAS produces a total scorewhich represents the overall marital adjustment of the couple (Spanier & Filsinger, 1983).Spanier (1976) reported the internal consistency estimate (Cronbach’s coefficientalpha) to be .96. The reliability of the sub-scales range from a low of .73 for the AffectionalExpression sub-scale to a high of .94 on the Satisfaction sub-scale with Consensus andCohesion having internal consistency estimates of .90 and .86 respectively.The sub-scale structure was originally validated by factor analysis (Spanier 1976) andconfirmatory studies on the instrument have provided additional support (Antill & Cotton,1982; Spanier & Thompson, 1982). In addition to identifying a supporting sub-scalestructure, factor analytic studies have identified a strong single principle factor of“adjustment” that underlies the entire instrument (Antill & Cotton, 1982; Kazak, Jarmas &Snitzer, 1988; Sharpley & Cross, 1982).The concurrent validity of the DAS is based on the correlation of the instrument withan array of other instruments measuring similar qualities of marital relationship includingthe Marital Adjustment Scale, the Georgia Marriage Q-sort, the Intimate RelationshipScale, and the Marital Satisfaction Scale. In all cases, the DAS is significantly correlated inexpected directions.The instrument has been shown to have predictive validity with the measure beingshown to reliability discriminate between married and divorced samples (Spanier, 1976;Spanier & Thompson, 1982). While Spanier (1976) reported norms for married anddivorced couples based on mean total scores as 114.8 (SD = 17.8) and 70.7 (SD = 23.8)respectively, Spanier and Filsinger (1983) warned that the norm for divorced couples may beinaccurate and low.109Following the suggestion of Burger and Jacobson (1979), the DAS total score of 100was used as a cut-off value for entry into the study. Accordingly, one member of the coupleshad to score under the critical value of 100 on the DAS to be included in the study. In thisinvestigation, the total DAS score served as the marker variable at the marital system leveland the sub-scales scores were used in the within system analyses.III. Areas of Change (ACThe AC is a self-report inventory designed by Weiss and Birchier (1975) to assess thedesire for spousal change in their partner’s marital behaviour and to measure each partner’sperception of the changes their spouse desired from them.The AC consists of 34 items that identify specific areas of concern e.g. have mealsready on time. The questionnaire is structured into two parts. In the first part, the 34 itemsfollow the stem statement “I want my partner to...”. In the second part of the instrument, thesame 34 items follow the statement stem that reads “It would please my partner if I...”.Responses are made on a 7-point Likert scale ranging from -3 “Much less” to + 3 “MuchMore”, with the mid point 0 indicating that no change is required.The AC yields a number of score that can be calculated for husbands, wives andcouples (Weiss & Birchler, 1975). The instrument has been scored in two different ways.One scoring procedure focuses on the overall numbers of items regardless of sign and theother focuses on the perceptual accuracy dimension of the AC. The first type of scoringrequires a simple summation procedure that generates separate overall change scores foreach part. Desired Change is the summary score derived from this process for the first partof the measure and Perceived Change is the summary score associated with the second part.A global Perceptual Accuracy measure is generated by comparing the Perceived Changescore from one partner with the Desired Change score from the other.110The second scoring procedure takes into account the perceptual accuracy of the ACby identifying instances in which responses reflect either agreement or disagreement on thedesirability of changing particular behaviours. Agreements are scored when partner A wantsa change on an item and partner B is both aware of the desired change and correctlyindicates the direction in which the change is requested. Disagreements are scored eitherwhen partner A wants a change and partner B does not recognize this, or when partner Bthinks a change is desired when it is not identified by partner A.The validity of the first scoring procedure has been supported by research (Margolinet al., 1983), however, the validity of the perceptually based scores (as indicated through thesecond scoring procedure) is less certain. Despite the fact that this method of scoring is verypopular (Mead, Vatcher, Wyne, & Roberts, 1990), Margolin et al. (1983) were unable todemonstrate that perceptual scores could discriminate between distressed and nondistressedcouples. Indeed, they concluded that there appeared to be virtually no association betweenperceptual and overall marital satisfaction. Nonetheless, the spouses’ agreement anddisagreement scores allow for the computation of a Total Change scores which aregenerated by summing the two perceptual accuracy measures. A ratio of Agreement:Disagreement is the last score generated by AC and this is taken to be a measure of theextent to which spouses are aware of what behaviours to change and of the direction of thechange desired (Margolin, Talovic, & Weinstein, 1983).A number of studies have explored the validity of this instrument. It may be assumedthat couples and individuals with greater numbers of complaints (as indicated through thefirst scoring procedures) would be less well adjusted in their marriage and this assumptionhas empirical support. The AC has been shown to discriminate well between distressed andnon-distressed couples (Birchler & Webb, 1977; Margolin et al., 1983; Margolin &Wampold, 1981). The instrument has also been found to be moderately negatively correlate111with measures of marital adjustment as found by Weiss et al. (1973) (r = -.71) and others(Margolin et al., 1983; Rabin et al., 1986). It has also been shown that the AC is not able todiscriminate between couples in which the husband is an alcoholic and conflicted couples,but does discriminate both from nonconflicted couples (O’Farrell & Birchler, 1987).Limited normative information has been provided by Margolin et al. (1983).Weiss et al. (1973) reported that the instrument has a high level of internalconsistency (r = .89). The AC has been found to be sensitive to change in therapy outcomestudies (Bavcom, 1982; Margolin & Weiss, 1978), however, no stability studies exploring test-retest reliability have been reported to date. Both methods of scoring will be used in thisstudy at the within system level of analyses.4. Family Measures DescriptionA crucial aspect of the research is concerned with the functioning of the family andthe ways in which the family may change as a result of one of three therapeutic efforts. Thisresearch took the perspective that each family member had a valid and important experienceof the family. Consequently, the family was assessed from the points of view of father,mother, and child and no effort was made to aggregate these individual assessments into“family”data. The instruments listed below were employed to operationalize the familysystem in the ecological assessment package. Each instrument is reviewed in the followingsection. The family assessment instruments include:(I) Family Information Form (FIF) developed by Epstein, Baldwin, and Bishop (1983),(II) Family Satisfaction (FS) developed by Olson and Wilson (1982),(III) Family Environment Scale (FES) developed by Moos and Moos (1981),(IV) Family Adaptability and Cohesion Evaluation Scales III (FACES III) developed byOlson, Portner, and Lavee (1985).112I. Family Information Form (FIFThe Family Information Form is part of a family questionnaire known as theMcMaster Family Assessment Device (Epstein, Baldwin, & Bishop, 1983). The FIF gathersdemographic information about families. It was selected for this research to provide acomprehensive range of family related information, including names, family roles, age,gender, and medical, school, or psychiatric problems of each person living in the household.In addition, the FIF enquires into the marital records of both parents, as well as the totalfamily income and the ethnic and racial groups with which the family identifies. Finally, theFIF generates information regarding employment status, occupation for the heads of thehousehold, and family income.II. Family Satisfaction (FS)The marker variable selected for the family level of assessment in this study was theFS scale reviewed below. Olson and Wilson (1982) developed this instrument as a directmethod of assessing satisfaction with one’s family. Sharing the same conceptual roots asFACES III (Olson, Porter, & Lavee, 1985), the FS was designed with an appreciation that ifthe normative expectations of family members supported extreme qualities of familyinteraction, then a particular family can function smoothly so long as all family members aresatisfied with these expectations (Olson, 1986). Consequently, the FS taps level ofsatisfaction and generates sub-scales on the important dimensions of family cohesion andfamily adaptability in order to ascertain family members’ comfort with these central familyqualities.The FS is a 14 item instrument that asks respondents to rate how satisfied they arewith the family emotional bonds, coalitions, time, space, decision-making, interests andrecreational activities (cohesion dimension items) and family assertiveness, control,discipline, negation, roles and rules (adaptability dimension items). Items are scored on aS113point Likert scale ranging from 1 = dissatisfied to 5 = extremely satisfied. The mid-point of3 indicates generally satisfied and divides the scale into satisfied and dissatisfied responsefields.The FS was designed to assess one’s level of satisfaction in a valid and reliablemanner. The final instrument was derived from a 28 item questionnaire which was pilotedand later subjected to factor analysis using a varimax rotation of the principal axes. Thereduction in items that followed, left the final 14 item scale with each item loading morethan .50 on the first principal component.As a method of establishing concurrent validity of the FS, Caron and Olson (1984)compared the discrepancy between two administrations of FACES II with the FS test. Thediscrepancy scores were derived from establishing difference of scores between two FACESII administration, the first measuring perceived family and the second assessing respondents’views of an ideal family. A high negative correlation between the FS and the ideal-perceiveddiscrepancy was hypothesized and confirmed with the correlation on the cohesion dimension(r = -.58) and on the adaptability dimension (r = -.64) conforming to predicted relationships.Importantly, the FS has been shown to be a sensitive measure to therapeutic change. Borik(1984) assessed 20 alcoholic families before and after treatment and found significanttreatment effects on the FS scale.With respect to internal consistency, the Cronbach alpha for the instrument is .92.The test-retest reliability estimate which was based on two administrations with a 5 weektime interval separately, resulted in a Pearson correlation coefficient of .75.Norms for the FS have been generated and are based on a national survey in theUnited States. The sample consisted of 1,026 couples (n = 2,056 individuals) drawn from114families spanning the family life cycle and 412 adolescent children. With scores potentiallyranging from as low as 14 to as high as 70, the scores of 47.0 for parents and 45.0 foradolescents were established in the normative sample as the 50th percentile values. Whilethe total FS value was used as a marker variable in this study, the sub-scales of cohesion andadaptability sub-scales were included in the within systems analyses.III. Family Environment Scale (FES)The FES (Moos & Moos, 1981) has been widely employed to study both treatmentoutcomes and families affected by alcoholism (Abbott, 1976; Bader, 1976; Barry & Fleming,1990; Bromet & Moos, 1977; Christensen, 1977; Filstead, 1979; Filstead, Anderson, &McElfresh, 1989; Finney & Moos, 1979; Moos, Bromet, Tsu, & Moos, 1989). The FES iscomprised of 10 sub-scales, however, for the purpose of this investigation, 6 sub-scales werechosen to measure salient qualities of the family environment. The sub-scales selected forthis inquiry included all the sub-scales of both the relationship and system maintenancedimensions and one of the personal growth dimension sub-scales. More specifically, the sub-scales employed in this study tap:(1) Cohesion or the degree of commitment, help and support family membersprovide for one another,(2) Expressiveness or the extent to which family members are encouraged to actopenly and to share their feelings in a direct fashion,(3) Conflict or the amount of tension or openly expressed anger, aggression andhostility among family members,(4) Independence or the extent to which family members are assertive or selfsufficient and empowered with the ability to make their own decisions,(5) Organization or the degree of importance of clear lines of authority andstructure in planning family activities and responsibilities,115(6) Control or the extent to which set rules and procedures are adhered to andemployed to direct family life.Each sub-scale score is derived from responses on 9 items per sub-scale, bringing the totalnumber of FES items used in this study to a total of 54.The FES item development was carefully undertaken. The items on the FES wereconstructed from information gathered in structured interviews with members of differenttypes of families. Additional items were adapted from other Social Climate Scales (Moos,1974b). Several forms of the instrument were piloted and this procedure led to an initial 200item form of the FES.This initial form was then administered to over 1,000 people in 285 families thatrepresented a wide variety of types of families. Subsequent item reduction was based onpsychometric criteria. The overall item split needed to be close to 50-50 in order to avoiditems that were characteristic of only unusual families. Items were required to correlatemore closely with their own sub-scale than with any other sub-scale. In addition, the sub-scales were required to have low to moderate interactions and each item needed todiscriminate among families. This item criteria was met for all items in a variety of subsamples, including Caucasian, ethnic minority, and distressed families. This developmentresulted in the generation of the 90 item FES instrument that asked a respondent to scoreyes or no as to whether or not the family statement in each item applied to their own family.Norms have been developed for both normal and distressed families. While the normalfamily norms were based on 1125 families from across the United States, the distressedfamily norms were established on the responses of 500 families involved in a variety ofclinical settings, including psychiatric-oriented family clinics, probation and paroledepartments, alcoholic treatment centers, and psychiatric hospitals.The internal consistency estimates for the 6 sub-scales employed in this study are allacceptable. The Cronbach’s Alpha levels ranged from moderate for Independence, Control,116and Expressiveness (.61, .67, and .69 respectively) to substantial for Cohesion, Conflict, andOrganization (.78. .75, and .76 respectively). Test-retest reliability of the sub-scales werebased on a sample of 47 family members in 9 families who completed the instrument twice at8 week intervals. The 2 month test-retest reliability for the sub-scales were all acceptable,and ranged from .68 for the Independence sub-scale to .86 for the Cohesion sub-scale. The 4month test-retest stability estimates were also acceptable and went from moderate (r = .54for the Independence sub-scale) to very respectable (r = .78 for the Control sub-scale).IV. Family Adaptability and Cohesion Evaluation III (FACES III)This instrument is the third version in a series of FACES scales intended to assess thequalities of family cohesion and family adaptability. These two qualities are viewed as thecore orthogonal constructs which underlie a circumplex model which provides the theoreticalgrounding of the instrument (Olson, 1986). The revisions of FACES have been undertakenin order to increase the measure’s reliability, validity and clinical utility. Consequently,FACES III is an instrument that has resulted from considerable work in the area of familymeasurement.Family cohesion is defined as the emotional bonding that family members have withone another (Olson, 1989) and the specific concepts that are employed in the instrument toreflect family togetherness include emotional bonding, boundaries, coalitions, time, space,friends, decision making, interests and recreation. The cohesion dimension can besubdivided into four levels or qualities of togetherness, namely, disengaged, separate,connected, and enmeshed. Originally, Olson et al. (1985) suggested that optimallyfunctioning families would score in the separate and connected ranges of the dimensionwhile disturbed families would report extreme levels of the cohesion dimension falling ineither the disengaged or enmeshed range of the instrument.117Family adaptability has been defined by Olson (1989) as the ability of a family systemto alter or change important aspects of its identity. The concepts tapped to measure thisdimension are family power (assertiveness, control, discipline), negotiation styles, rolerelationships, and relational rules. The continuum of the adaptability dimension wasoriginally conceived of as ranging from the extremes of rigid to chaotic with the adaptabilityqualities of structured and flexible falling in the middle ranges of the dimension. As with thecohesion dimension, Olson et al. (1985) postulated that the moderate levels of adaptability(structured and flexible) were more optimal or conducive to family functioning, while thetwo extreme levels of adaptability (chaotic and rigid) were associated with problematicfamily organizations.Challenges to the postulated levels of the two dimensions by Green (1989) have ledto a revision of interpretation of FACES III. Importantly, Olson (1991) recently clarifiedthat on this instrument:high scores really measure Balanced family types and low scores measureextreme family types. More specifically, high scores on cohesion aremeasuring “connected families” (Balanced) and high scores on adaptability aremeasuring “flexible” families (Balanced). (p. 75)Accordingly, Olson (1991) directed users of the instrument to assume FACES III to be asimple linear measure and to interpret results in this fashion, Thus, high scores on cohesionare best interpreted as “very connected” rather than enmeshed and similarly, high scores onadaptability are best understood as “very flexible” rather than “chaotic”.FACES III is a 20 item scale containing 10 cohesion and 10 adaptability items. Thecorrelation between the two core constructs is very low (r = .03), indicating that the twoconstructs are indeed orthogonal. Olson et al. (1983) reported that factor analysis of theitems resulted in a two factor solution consistent with the underlying concepts of theinstrument. Both the convergent and discriminated validity of FACES III have been118reported by Edman, Cole, and Howard (1990), and Perosa and Perosa (1990). With respectto alcohol dependent individuals and their families, the original FACES measure was shownto discriminate between these distressed families and a non-distressed family comparisongroup (Killorin & Olson, 1984; Olson & Killorin, 1985).The internal consistency reliability of the instrument was established on a sample ofover 2,000 respondents. The Cronbach Alpha estimate indicated that the measure wasacceptably reliable (r = .77 on cohesion and r = .62 on adaptability). There are no test-retestreliability estimates provided for FACES III; however, the test-retest reliability for theearlier version of the instrument, FACES II, was very good. With a test-retest time intervalof 5 weeks, the 50 item FACES II resulted in a test-retest reliability estimate of .83 forcohesion and .80 for adaptability.Phases of Data Analysis and Operationalization of Research HypothesesThe following section maps out the data analysis plan and operationalizes the tworesearch hypotheses in terms of specific sub-hypotheses. There were four phases of dataanalysis required for this study and each phase built upon the results of the previousphase(s).Phase 1. Preliminary AnalysisThe first phase of analysis was concerned with generating means, standard deviationsand reliability estimates (internal consistency coefficients - Cronbach’s Alpha) forinstruments used in the study. The pre-treatment descriptive information was considered inlight of normative information in order to enable an appreciation of the clinical status ofparticipants prior to treatment. In addition, the first phase of analysis focused on both theexploration of the pre-treatment equivalency of participants randomly assigned to the threetreatment groups, and the comparability of the project participation subgroups detailedearlier in this document.119Phase 2. Eco-System AnalysisThe second phase of data analysis centered on testing the research hypotheses byanalysing data pertaining to various sub-hypotheses using the marker variables as dependentvariables.Marker VariablesIn order to facilitate the testing of eco-systemic research questions in a clear andefficient fashion, particular instruments were selected prior to data analysis which wouldserve as representatives of the entire instrument package. These instruments or markervariables were culled from the entire array of measures on the basis of their capacity toreflect general or summary characteristics of the level of assessment which it represented.Accordingly, four instruments, one from each assessment domain (alcohol, intrapersonal,couple, and family) were identified as marker variables. Each of the measures could besummarized by a single score and each marker variable revealed a global characteristic ofthe assessment level which it represented. The marker variables were:(1) ADD which measured the level of alcohol dependency,(2) SCL-90-R’s global symptomology index which measured the respondents’general level of intrapersonal distress,(3) DAS’s total score which measured the level of marital adjustment,(4) FS which measured the level of family satisfaction.These four marker variables were employed throughout tests which constituted theeco-system analysis phase of data analysis. For the analyses related to fathers’, data all fourmarker variables were used; however, in the analyses pertaining to mothers’, the alcoholmeasure was not applicable and therefore the remaining three marker variables wereemployed. Finally, in the eco-systems analyses regarding the eldest children’s perspective,only the family marker variable was relevant and used in the analyses.120Statistical ProcedureThe eco-systems analyses used a mixed model design with one between subjectsindependent factor (treatment) and one within subjects dependent factor (measurementoccasion). The treatment factor had two levels related to hypothesis 1 (ExST and SFT) andtwo levels related to hypothesis 2 (ExST-I and ExST-C). The measurement occasion factoralso had two levels regarding the first hypothesis (pre-test and post-test). To test the sub-hypotheses of the second hypothesis, two pairs of the measurement occasion factor wereconsidered. The two pairs of the measurement occasion for this hypothesis were pre-testand post-test, and post-test and follow-up.To test the two central hypotheses, all sub-hypotheses were analyzed separately. TheSPSS multivariate analysis of variance (MANOVA) program was used to test the sub-hypotheses related to fathers and mothers. Each marker variable identified in the sub-hypotheses statements served as dependent variables for the MANOVA runs. The analysesof the sub-hypotheses related to the eldest child in the family was based on data from onlyone marker variable. Consequently, the SPSS analysis of variance (ANOVA) program waschosen to test the sub-hypotheses connected to the child’s perspective. Thus, two separateMANOVA runs and one ANOVA run were required to test the first hypothesis and a totalof four separate MANOVA runs and two separate ANOVA runs were needed to test thepre/post and post/follow-up comparisons related to the second hypothesis.Phase 3. Within System AnalysisThe third phase of data analysis probed for differences within assessment levels(alcohol, intrapersonal, marital, family) as indicated by significant differences in the initialMANOVA and ANOVA runs used to test the sub-hypotheses of the two researchhypotheses. The MANOVA and ANOVA runs in the eco-systems analyses identifieddifferences based on data from each of the assessment levels. Within systems analyses were121conducted to further explore and elaborate on changes in the levels of assessment thatcontributed to the significant differences in the eco-systems analyses using the same mixedmodel design.A multivariate approach to analysis was taken at the within system analysis phase ofthe study. All instruments and their sub-scales for one level of assessment (as identified inthe instrumentation section of this dissertation) were employed to serve as dependentvariables for within system analyses. These MANOVA runs allowed for a more detailedexamination of the system in which change had been identified as having occurred as a resultof the experimental procedures. As a consequence of this approach to data analysis, thisstudy is able to provide a more comprehensive understanding of the nature of the changeswhich occurred than would normally have been possible had only the marker variables beenused as the dependent variables. It is important to stress that no within system analysis wasundertaken unless significant differences were observed in the hypothesis testing phase ofanalysis.Phase 4. Therapeutic Process ValidationThe fourth and final phase of data analysis involved a series of MANOVA runsconducted to determine the role that treatment variables may have played in the treatmentoutcomes. This phase of analysis focuses on whether or not there were significantdifferences between the two clinical sites (Surrey and Duncan), the participating therapistsand the gender of the therapists. Again the analyses were of mixed model design andfathers’ and mothers’ data were considered separately.Operationalization of Research Hypothesis 1The first hypothesis of this study states: When compared to the families in which thealcoholic father completed SF!’, the families in which the alcoholic father completed ExST122will report greater improvement on measures contained in the ecological assessmentpackage employed in the study.This hypothesis will be either supported or disconfirmed on the basis of theacceptance or rejection of sub-hypotheses which separately concern themselves withtreatment outcomes related to the father, mother and eldest child participants whoconstitute the families in this study. The following delineates the three sub-hypotheses of thefirst main hypothesis and operationalizes them in terms of statements that focus on eachlevel of ecological measurement.(la) When compared to alcoholics who completed the SFT, alcoholics who completedExST will report significantly greater improvement at post-test on measurescontained in the ecological assessment package employed in the study. Theoperationalized statements are:(i) The alcoholics who completed ExST will have improved with respect toalcohol to a significantly greater degree than the alcoholics who completed theSF!’ treatment as indicated by lower levels of alcohol dependency on the ADDat post-test.(ii) The alcoholics who completed ExST will have improved with respect to theirintrapersonal functioning to a significantly greater degree than the alcoholicswho completed the SF!’ treatment as indicated by lower levels of psychologicalsymptomology on the SCL-90-R at post-test.(iii) The alcoholics who completed ExST will have improved with respect to theirmarital relationship to a significantly greater degree than alcoholics who123completed the SFT treatment as indicated by higher levels of maritaladjustment on the DAS at post-test.(iv) The alcoholics who completed ExST will have improved with respect to theirfamily system to a significantly greater degree than alcoholics who completedthe SFT treatment as indicated by higher levels of family satisfaction on the FSscale at post-test.(ib) When compared to the wives whose husbands completed SF1’, wives whose husbandcompleted ExST will report significantly greater improvement at post-test onmeasures contained in the ecological assessment package employed in the study.This sub-hypothesis will be either supported or disconfirmed on the basis of thetesting of the following operationalized statements:(i) The wives whose alcoholic husbands completed ExST will have improved withrespect to their intrapersonal functioning to a significantly greater degree thanthe wives whose alcoholic husbands completed the SFT treatment as indicatedby lower levels of psychological symptomology on the SCL-90-R at post-test.(ii) The wives whose alcoholic husbands completed ExST will have improved withrespect to their marital relationship to a significantly greater degree than wiveswhose alcoholic husbands completed SF1’ as indicated by higher levels ofmarital adjustment on the DAS at post-test.(iii) The wives whose husbands completed ExST will have improved with respect totheir family system to a significantly greater degree than wives whose alcoholic124husbands completed SFT as indicated by higher levels of family satisfaction onthe FS scale at post-test.(ic) The children in families whose alcoholic fathers completed ExST will have improvedwith respect to their family system to a significantly greater degree than childrenwhose alcoholic fathers completed SFT as indicated by higher levels of familysatisfaction on the FS scale at post-test.Operationalization of Research Hypothesis 2The second hypothesis in this study is: When compared to the families in which thefather completed ExST-I, the families in which both father and mother completed ExST-Cwill report significantly greater improvement at post-test and/or follow-up as measured byselected instruments in the ecological assessment package used in the study.The second hypothesis will either be supported or disconfirmed on the basis of theacceptance or rejection of a series of sub-hypotheses which separately address the treatmentoutcomes related to the fathers, mothers and eldest children at post-test and follow-upmeasurement occasions. The following delineates the three sub-hypotheses for the pre/postcomparison and the three sub-hypotheses for the post-/follow-up contrast of the secondmain hypothesis and expresses them in terms of operationalized statements that focus oneach level of ecological assessment.(2a) When compared to the alcoholics who completed ExST-I, alcoholics who completedExST-C will report significantly greater improvement at post-test as measured byselected instruments in the ecological assessment package used in the study. Theoperationalized statements are presented below.125(i) The alcoholics who completed ExST-C will improve with respect to theiralcohol problem to a significantly greater degree than the alcoholics whocompleted ExST-I as indicated by lower levels of alcohol dependency asmeasured by the ADD at post-test.(ii) The alcoholics who completed ExST-C will improve with respect to theirintrapersonal functioning to a significantly greater degree than the alcoholicswho completed ExST-I as indicated by lower levels of psychologicalsymptomology as measured by the SCL-90-R at post-test.(iii) The alcoholics who completed ExST-C will improve with respect to theirmarital relationship to a significantly greater degree than alcoholics whocompleted ExST-I as indicated by higher levels of marital adjustment asmeasured by the DAS at post-test.(iv) The alcoholics who completed ExST-C will improve with respect to theirfamily to a significantly greater degree than alcoholics who completed ExST-Ias indicated by higher levels of family satisfaction as measured by the FS scaleat post-test.(2b) When compared to wives of alcoholics whose husbands completed ExST-I, wives whocompleted ExST-C with their husbands will report significantly greater improvementat post-test as measured by selected instruments in the ecological assessment packageused in the study. The operationalized statements for this sub-hypothesis read:(i) The wives of alcoholics who completed ExST-C along with their husbands, willimprove with respect to their intrapersonal functioning to a significantly126greater degree than the wives whose husbands completed ExST-I as indicatedby lower levels of psychological symptomology as measured by the SCL-90-Rat post-test.(ii) The wives of alcoholics who completed ExST-C along with their husbands, willimprove with respect to their marital relationship to a significantly greaterdegree than wives whose husbands completed ExST-I as indicated by higherlevels of marital adjustment as measured by the DAS at post-test.(iii) The wives of alcoholics who completed ExST-C along with their husbands, willimprove with respect to their family system to a significantly greater degreethan wives whose husbands completed ExST-I as indicated by higher levels offamily satisfaction as measured by the FS scale at post-test.(2c) When compared to the eldest children of alcoholics whose fathers completed ExST-I,the eldest children of alcoholics whose parents both completed ExST-C will improvewith respect to their family to a significantly greater degree as indicated by higherscores of family satisfaction as measured by the FS scale at post-test.(2d) When compared to the alcoholics who completed ExST-I, alcoholics who completedExST-C will report significantly greater improvement at follow-up as measured byselected instruments in the ecological assessment package used in the study.(i) The alcoholics who completed ExST-C will improve with respect to theiralcohol problem to a significantly greater degree than the alcoholics whocompleted ExST-I as indicated by lower levels of alcohol dependency asmeasured by the ADD at follow-up.127(ii) The alcoholics who completed the ExST-C will improve with respect to theirintrapersonal functioning to a significantly greater degree than the alcoholicswho completed ExST-I as indicated by lower levels of psychologicalsymptomology as measured by the SCL-90-R at follow-up.(iii) The alcoholics who completed the ExST-C will improve with respect to theirmarital relationship to a significantly greater degree than alcoholics whocompleted ExST-I as indicated by higher levels of marital adjustment asmeasured by the DAS at follow-up.(iv) The alcoholics who completed ExST-C will improve with respect to theirfamily to a significantly greater degree than alcoholics who completed ExST-Ias indicated by higher levels of family satisfaction as measured by the FS scaleat follow-up.(2e) When compared to wives of alcoholics whose husbands completed ExST-I, wives whocompleted ExST-C with their husbands will report significantly greater improvementat follow-up as measured by selection instruments in the ecological assessmentpackage used in the study.(i) The wives of alcoholics who completed ExST-C along with their husbands, willimprove with respect to their intrapersonal functioning to a significantlygreater degree than the wives whose husbands completed ExST-I as indicatedby lower levels of psychological symptomology as measured by the SCL-90-Rat follow-up.128(ii) The wives of alcoholics who completed ExSt-C along with their husbands, willimprove with respect to their marital relationship to a significantly greaterdegree than wives whose husbands completed ExST-I as indicated by higherlevels of marital adjustment as measured by the DAS at follow-up.(iii) The wives of alcoholics who completed ExST-C along with their husbands, willimprove with respect to their family to a significantly greater degree thanwives whose husbands completed ExST-I as indicated by higher levels offamily satisfaction as measured by the FS scale at follow-up.(2f) When compared to the eldest children of alcoholics whose fathers completed ExST-I,the eldest children of alcoholics whose parents both completed ExST-C will improvewith respect to their family to a significantly greater degree as indicated by higherscores of family satisfaction as measured by the FS scale at follow-up.129CHAPTER IV: RESULTSIn this chapter the results of the data analyses are reported. The chapter is dividedinto four main sections which correspond to the four phases of data analyses outlined earlierincluding: (1) preliminary analyses, which focus on the participant descriptions at pretreatment and tests regarding group comparability, (2) eco-system analyses, which generateanswers to the hypotheses and sub-hypotheses of the study, (3) within system analyses, whichexpand upon the hypotheses testing results, and (4) therapeutic validation, which report onanalyses probing possible therapeutic confounds to the study.Preliminary AnalysisInstrument Overview and Pre-Treatment Participant DescriptionMeans, standard deviations and internal consistency reliability estimates (Cronbach’sAlpha) were calculated on data collected at screening and pre-treatment. The results arepresented below in Tables 3-7. Kolmogorov-Smirnov (K-S) tests were conducted for eachinstrument and sub-scale employed in the ecological battery. All K-S tests for the markervariables were normal. The K-S test results and the findings regarding kurtosis and skew forall scales and sub-scales used in this study may be found in Appendix H. A brief descriptionof the participants is provided for each level of assessment.AlcoholAs shown in Table 3, the men in this study clearly scored well above the suggestedcut-off value indicating alcoholism on the MAST (=31.54) and the mothers scored wellbelow the alcoholic threshold score (=2.14). As a group, the alcoholics’ scores on the ADDindicated that they were at the high end in the moderate dependency range of responding(X=56.05). The critical value for severe levels of alcohol dependency is 61 on thisinstrument.130The alcoholics in this study scored in the high risk range of scores on the IDS for allbut the physical discomfort sub-scale of the IDS (in which the participants scored in themoderate risk range of scores). The mean IDS values for the participants were compared tothe normative sample gathered on men who were entering treatment for a variety of alcohol-related problems. As a group, the alcoholic participants scored in the 64th percentile forunpleasant emotions, 1st percentile for physical discomfort, 79th percentile for pleasantemotions, 42nd percentile for testing personal control, 51st percentile for urges andtemptations, 41st percentile for conflict with others, 81st percentile for social pressure todrink, and 65th percentile for pleasant times with others.Table 3Means, Standard Deviations and Reliability Estimates (Cronbach’s Alpha) of Alcohol Measures for Fathers and Mothers at Pre-TestFathe__________________________Instrument Scale/Sub-Scale Mean SD Mean SDMAST Total 31.54 9.25 2.14 3.28 .79ADD Total 56.05 19.03 — — .97IDS Unpleasant emotions 55.34 25.36 — — .83Physical discomfort 28.58 24.49 — — .79Pleasant emotions 5821 24.73 — -- .84Personal control 47.90 30.31 -- — .87Urges/temptations 50.41 25.71 — — .74Conflict with others 45.57 24.49 — — .92Social pressure 59.53 29.26 — — .85Pleasant times 54.00 26.40 — — .88SCQ Unpleasant times 80.55 20.86 — — .93Physical discomfort 80.73 22.41 -- — .79Pleasant emotions 52.62 31.19 — — .87Personal control 64.70 25.55 — — .91Urges/temptations 73.53 22.23 — — .84Social problem/work 75.16 20.85 — — .87Social tensions 54.36 30.39 — — .91Pos. social situations 63.09 23.40 — — .97131With respect to the participants sense of confidence regarding their ability to resistdrinking as measured on the SCQ compared to the normative sample of men enteringtreatment for alcohol related problems, the participants scored in the 48th percentile forunpleasant emotions and frustrations, 45th percentile for physical discomfort, 44th percentilefor pleasant emotions, 38th percentile for personal control, 42nd percentile for urges andtemptations, 46th percentile for social problems at work, 51st percentile for social tensions,and 39th percentile for positive social situations at pre-treatment.IntrapersonalThe information in Table 4 reveals that the mean husband and wife performance on theSILS vocabulary sub-scale placed them just shy of the mean score for adults of their age (T= 49). In addition, the mean husband and wife abstraction scores placed participants withinthe first standard deviation of the normal population (T = 55 for men, T = 57 for women).Thus, it is safe to conclude that the participants verbal and mental abilities were within thenormal range and assume that deficits in these areas did not play a role in the study. Noestimates of internal consistency for the vocabulary or abstraction sub-scale are provided. Inaddition to being a timed test, the SILS is a power test which increases in difficultyincrementally from the first question to the last. Consequently, the internal consistencyreliability Cronbach’s Alpha is inappropriate for this measure. Test-retest and split-halfreliability estimates were presented earlier in the instrument section of this dissertation andranged from r = .60 to .89.The group means on the BDI placed both the father and the mother participants inthis study in the moderately depressed range of scores.The mean husband and wife scores on the nine sub-scales and the global distressindex of the SCL-90-R show that as a group, the participants were very distressed in terms of132psychological symptomology. The mean scores exceeded one standard deviation above themean of the non-patient comparison groups on all sub-scales for both husbandsTable4Means. Standard Deviations and Reliabilities (Cronbach’s Alyha of Intrapersonal Measures for Fathers and MothersInstrument Scale/Sub-scale Fathers__________________________SILS Vocab 30.90 5.09 30.82 30.82 —Abstract 28.67 6.76 30.66 6.15 —SCL-90-R Somatization 63.48 16.14 62.93 16.22 .83Obsessive/compulsive 73.24 17.04 70.78 17.54 .86Interpersonal sensitivity 75.84 20.14 70.03 19.35 .86Depression 78.56 19.19 79.60 19.74 .89Anxiety 74.58 20.50 70.64 21.57 .89Hostility 74.76 22.54 73.20 20.79 .84Phobic anxiety 60.65 18.75 56.29 18.94 .80Paranoid ideation 70.71 18.12 65.66 18.32 .78Psychoticism 80.09 24.22 68.99 22.18 .79GSI 78.91 19.33 74.64 20.38 .97BDI Total 14.32 19.33 14.83 8.61 .87and wives with the exception of wives phobic sub-scale score. The SCL-90-R data was recast against the in-patient psychiatric norm group. In order to compare the data to thepsychiatric norm group, the raw data was re-calibrated to fit the standard score mean andstandard deviation of the psychiatric norm. As the second graph in Figure 5 reveals, boththe husband and wife participants in this study scored in a fashion consistent withpsychiatrically hospitalized people. The participant comparison with both the non-patientand the in-patient norms are presented in Figure 5.133SCL-00-R Score SCL-90-R Score80 —Figure 5. Fathers’ and mothers’ mean pre-treatment SCL-90-R sub-scale scores comparedto normal non-patient and in-patient norms.CoupleThe means, standard deviations and reliability estimates for the instrument thecouples level of assessment appear in Table 5. As the Table shows, the mean EMCS valuefor fathers was 14.89 and for mothers was 6.67. While the instrument was not used to screenout participants (using a cut-off value of 20 or greater as indicative of an unacceptable levelof marital conventionality), the mean scores suggest that the tendency to misrepresent themarriage in an overly positive fashion was not a major issue for the participants in this study.Both husbands’ and wives’ mean DAS scores show that the couples in the study wereindeed maritally distressed and well below the critical value of 100. When compared tonormal contented couples on the DAS sub-scales, the couples in this research scored wellbelow the means on the sub-scales of dyadic satisfaction, dyadic cohesion and affectionalexpression. While the couples did scored below the normal average on the dyadic cohesionsub-scale, this difference was negligible.SCL-90-R Normal Contrast90807060504030SCL-90-R In-patient Contrast70I I I I3020I I I I I ISOM OBS NT DEP ANX HOS PHOB PAR PSYSCL-90-R Sub-ScalePtiveher -Mother‘ I I I I I I ISOM OBS INT DSP ANX HOS PHOB PAR PSYSCL-90-TR Sub-scale134Table 5Means. Standard Deviations and Reliability Estimates of Couples Measures for Fathers, Mothers and Couples at Pre-testFather Mother CoupleTnstrument Sub-Scale Mean SD Mean SD Mean SDEMCS Total 14.89 13.89 6.67 9.87 — — .75DAS Consensus 39.29 8.93 36.80 8.90 — — .85Satisfaction 30.45 6.37 27.26 6.43 — — .62Affection 6.52 2.70 5.94 2.68 — — .24Cohesion 12.60 4.05 11.06 4.14 — — .44Total 88.86 17.89 81.06 17.97 — — .65AC Agree 6.27 4.29 — — — — .62Disagree 5.97 4.15 — — — — .75Desired Change 24.03 16.28 — — — — .88Perceived Change 32.85 16.66 — — -_ — .89Agreement -- — 11.10 5.37 — — .75Disagree -- -- 5.79 3.69 -- — .64Desired Change — — 38.96 15.60 — — .87Perceived Change — — 28.46 14.90 — — .88Total Agree — — — — 17.37 7.96 .76Total Disagree — — — — 11.76 6.44 .79Desired Change — — — — 62.05 23.21 .86Perceived Change — — — — 60.07 23.76 .91Regarding the couples AC pre-treatment values, comparisons with distressed andnon-distressed normative groups available for some of the scales show that the fathers’ andmothers’ scores on desired change and perceived change fit within the distressed maritalranges. Thus, the amount of change that couples were asking for in this sample wereconsistent with the levels of change requested by couples in marital distress. While fathers’and mothers’ perceptual accuracy values were not within the distressed couple range, themean scores closely resemble the normative values for couples struggling with alcoholdependency provided by O’Farrell and Birchier (1987),135FamilyThe descriptive data related to the family assessment level are shown in Tables 6 and7. The level of parental satisfaction with their families was very low at pre-treatment. Thefathers’ mean score of 37.73 placed them in the 4th percentile, and the mothers’ mean scoreof 36.11 placed the women in the 1st percentile of scores on this instrument. The first bornchildren’s mean score of 40.09 suggested that they were less dissatisfied with their familiesthan were the parents. This mean score, although low when compared to the norm group,placed tthe eldest children in the 28th percentile and within the normal family range.Table 6Means Standard Deviations and Reliability Estimates of Family Measures for Fathers and Mothers at Pre-TestFathers_______________________Instrument Scale/Sub-scale Mean SD Mean SDFS Total 38.11 8.75 36.60 9.21 .89Cohesion 21.52 5.13 20.66 5.42 .81Adaptability 16.60 4.14 15.94 3.94 .82FES Cohesion 5.07 2.40 4.85 2.81 .64Expressiveness 4.55 2.11 4.46 2.09 .61Conflict 4.34 2.17 4.68 2.29 .73Independence 6.33 1.56 5.89 2.05 .46Organization 3.81 2.48 4.23 2.22 .70Control 4.29 1.96 4.84 2.06 .53FACES III Cohesion 32.42 7.37 34.47 7.61 .86Adaptability 2451 5.29 24.93 4.97 .67136Table 7Means, Standard Deviations and Reliability Estimates of Family Measures for Eldest Children at Pre-testInstrument Scale/Sub-scale SDFS Total 40.09 10.63 .87Cohesion 24.23 5.75 .79Adaptability 17.84 5.17 .79FES Cohesion 4.48 2.66 .64Expressiveness 3.62 1.95 .61Conflict 4.72 2.58 .74Independence 5.89 1.96 .46Organization 4.44 2.43 .69Control 4.28 2.41 .53FACES III Cohesion 29.55 7.36 .86Adaptability 25.67 6.37 .67The family participants’ FES scores were compared to the mean scores of thedistressed family norms provided by Moos and Moos (1981). The comparison between thefather, mother and eldest child pre-treatment data from the present study and the distressedfamily norm scores on the six sub-scales are presented below in Figure 6.PerspectiveDDistress lDFather Mother Eldeet ChildFigure 6. Participating family mean scores on FES at pre-treatment and distressed normFES ScoreCohesion Expressiveness IndependenceConflict Organization Controlcomparison.137This contrast suggested that the families in this project scored in a similar fashion toother distressed families. On the cohesion and expressive sub-scales, the eldest childrenreported scores somewhat lower than the distressed contrast group, however, the scores ofthe parents in the study were very similar to the comparison group. More conflict wasreported by both mothers and eldest children than the distressed norm group. While bothmothers and children scored identically to the norm group on independence, fathersreported somewhat higher levels than the distressed contrast group on this sub-scale. Allfamily members reported lower levels of organization than the distressed family norm.Finally, while both fathers and eldest children reported lower levels of control, mothersreported mean scores equal to the distressed norm value on this sub-scale. It should benoted that the control sub-scale does not differentiate distressed families from normalfamilies.The FACES III mean scores for father and mother were comparable. The difficultieswhich both parents reported were found in the cohesion sub-scale with fathers and mothersboth reporting that the families were somewhat disengaged. On the adaptability sub-scale,the parents’ score located the families in the mid-range of the scale indicating that from theirperspective the families were structured. The eldest children appear to agree with theirparents assessment of family cohesiveness rating it as disengaged. However, the eldestchildren differed with their parents in terms of how they rated the adaptability dimension ofthe family. While the parents viewed the families as structured, the eldest children reportedthat the families were more on the flexible side. In either case, these mean scores did notrepresent extreme scores and were not interpreted as clinically problematic.Group EquivalenceThe assumption of group equivalence was tested with respect to treatment groupmembership (ExST-I, ExST-C, and SF1’) and project participation (drop-out no treatment,138drop-out some treatment, complete treatment, and complete treatment with missing data).The analyses were based on data collected at screening and pre-treatment. The firstmultivariate analysis was conducted on a variety of demographic variables gathered atscreening on the FDF. The variables included number of people residing in the household,father’s marital record, mother’s marital record, father’s divorce rate, mother’s divorce rate,father’s number of previous marriages, mother’s number of previous marriages, presentmarital status, couples number of years living together, and father’s and mother’semployment status. The results of this multivariate analysis revealed no significantdifference between the three randomly assigned treatment groups, approximate F (S = 2,M=4, N=45.5) = 1.52, p = 0.07 or the project participation sub-groups, approximate F(S=3,M=3/12,N=45.5) = 1.10, p = 0.329.The second equivalency analysis was conducted on father’s age, census class number,socio-economic status, and income derived from the Blishen socio-economic index (Blishen,Carrol & Moore, 1982). This MANOVA showed that there was no significant differencebetween the 3 treatment groups, approximate F (S =2, M= 1, N=46) = 1.29, p = 0.238, orbetween the four project participation sub-groups, approximate F (S =5, M= -.5, N=46) =0.78, p = 0.768. Another MANOVA comparing the treatment and project participationgroups in terms of mother’s age and the Blishen index categories of census class number,socioeconomic status, and income also failed to establish any significant difference prior tothe commencement of treatment, approximate F (S =2, M= 1, N=39) = 0.93, p = 0.50 forthe treatment groups, and approximate F (S =3, M=.5, N=39) = 1.28, p = 0.216 for theproject participation groups.A third set of analyses were performed on fathers’ and mothers’ data on measureswhich were administered only at screening including the MAST, EMCS and SILS. Again,the MANOVA on fathers’ data did not indicate any significant pre-treatment difference139between the treatment groups, F (S =2, M=.5, N=49) = 0.69, p = 0.700 or the projectparticipation sub-groups, F (S =3, M=0, N=49) = 1.38, p = 0.176. Similarly, theMANOVA on mothers’ data did not detect any significant difference between the treatmentgroups, F (S =2, M=.5, N=49) = 0.69, p = 0.700 or the project participation sub-groups,approximate F (S =3, M=0, N=49) = 1.64, p = 0.08.Special attention was focused on the marker variables and the complete therapy subgroup equivalence because the bulk of the analyses performed was based on data from thegroup of participants that completed treatment and the hypotheses testing portion of theanalyses centered on these instruments. Consequently, a series of equivalency analysesfocused on the comparability of the complete treatment group in terms of treatment groupmembership (ExST-I, ExST-C, and SFT) using the marker variable indices (ADD, SCL-90-R, DAS, and FS) were conducted. The multivariate analysis of fathers’ scores did notindicate significant differences between the treatment groups, approximate F (S =2, M=.5,N=26) = 0.92, p = 0.51. Similarly, the MANOVA results of mothers’ data and theANOVA results of the eldest children’s scores showed no significant differences between theSFT, ExST-I, and ExST-C treatment groups at pre-treatment, approximate F (S=2,M=0,N=26.5) = 1.39, p = 0.22 for mothers, and F (30, 2) = 1.43, p = 0.256 for eldest children.The next line of equivalence analysis was concerned with project participation andexplored the assumption that participants in the group that completed treatment wereequivalent to those who either dropped out of the study or those whose incomplete data setsprecluded them from inclusion in the post-treatment analyses on the marker variables.Separate univariate analyses were performed on each of the marker variables.The ANOVA contrasting the complete treatment group with the treatment drop-outgroup and the missing data group on the ADD revealed no significant differences, F (3, 96)140= 0.66, p = 0.58. The pre-treatment drop out group could not be included in this analysis asthe ADD data was not collected at screening.The participation group equivalence analyses of the SCL-90-R revealed no significantdifference between the complete treatment group and the other three participation groups,F (3, 223) = 0.75, p = 0.52. The comparison between the two drop-out groups was also notstatistically significant on the SCL-90-R, F (82, 1) = 0.06, p = 0.81. The contrast betweenthe genders of parents also revealed no significant difference, however there was a trendtowards the men being more symptomatic than the women, F (1, 225) = 2.62, p = 0.10.The equivalence analyses of project participation group on the DAS revealed nosignificant difference between the four participation groups, F (3, 224) = 0.36, p = 0.78.Similarly, there was no significant difference between the two drop-out groups, F (1, 82) =0.86, p = 0.36. However, a significant difference between the husbands’ and the wives’scores on the DAS was identified revealed that the wives reported significantly more maritaldistress than their husbands, F (1, 226) = 10.78 p = 0.001.A statistically significant difference was found in the participation group contrast atthe family level of assessment for fathers and mothers. Since no family measure wasadministered at screening the pre-treatment drop out group was not considered in thisanalysis. Nonetheless, the ANOVA on fathers’ data was statistically significant, F (2, 82) =3.38, p =0.039. A review of the data revealed that the fathers who dropped out oftreatment were significantly less satisfied with their families than either the group thatcompleted treatment or the group that completed treatment but was missing data at posttreatment. Similarly, the mothers in the group that dropped out of treatment weresignificantly more dissatisfied with their families than were the women in either of the twogroups that completed treatment, F (2, 86) = 6.86, p = 0.002. The ANOVA comparing the141eldest children by participation group on the FS was not significant, F (2, 51) = 0.53, p =0.662.Summary of Equivalency TestsThe equivalency analyses indicated that the random assignment of clients totreatment groups had adequately controlled for chance bias. Based on the analyses of datafrom the entire study sample, and data on the families who completed treatment (the focusof the remaining analyses) it was shown that the treatment groups were indeed comparable.In the main, the project sub-group set of analyses failed to identify any uniquecharacteristics which differentiated those who completed treatment from those who did not.However, the comparability analyses identified one important factor in the study and thispertained to those in the project participation sub-groups that terminated treatmentprematurely. The analyses indicated that both fathers and mothers in this sub-groupreported significantly lower levels of family satisfaction than did the two sub-groups thatcompleted their course of treatment. Although the two remaining project participation subgroups were very dissatisfied with their families (pre-treatment mean score values placingthem on the 10th and 13th percentile), the parents that were extremely dissatisfied (meanscore in the 1st percentile) dropped out of the study before completing treatment. Thisresult would appear to be consistent with the findings of Bromet and Moos (1977) and Mooset al. (1979) who reported that better family environments were significantly correlated withbetter treatment outcomes for recovering alcoholics. In the present study, the less severelevels of family dissatisfaction appeared to support the treatment procedures in somefashion. In the cases of the most extreme levels of family dissatisfaction, the participantswere seemingly unable to honor their commitment to complete therapy.142Eco-System Analyses of HypothesesThe eco-system analyses generated the primary results for the two main researchhypotheses of the study. Testing the operationalized statements of the sub-hypotheses, theeco-system analyses results were used to point to further analytic elaboration in the withinsystems analyses section. The two hypotheses and their constituent sub-hypotheses arepresented in this section. All analyses were based upon data from the complete treatmentproject participation group and were conducted using SPSS statistical packages.First Research Hypothesis: Differential Efficacy of ExST and SFTThe first hypothesis was concerned with the differential treatment effects of ExSTand SF!’. This hypothesis was broken into three sub-hypotheses which separately addressedthe perspectives of the father, mother and children. The results from tests centered on eachsub-hypothesis are presented below.Sub-Hypothesis 1-aWhen compared to alcoholics who completed SFT, alcoholics who completed ExSTwill report significantly greater improvement at post-test on measures contained in theecological assessment package employed in the study.A multivariate analysis of varience was conducted on fathers’ pre-treatment and post-treatment scores on the marker variables to determine whether or not the two treatmentshad a differential treatment effect as predicted in hypothesis 1-a. The MANOVA analysis oftreatment and occasion interaction was not statistically significant, F (S= 1, M= 1, N=55.5) =0.038, p = 0.99. There was no evidence found that would support hypothesis 1-a andconsequently it was rejected.A highly significant occasion effect was found between pre-treatment and posttreatment scores, F (S= 1, M= 1, N=55.5) = 21.23, p = 0.001. This MANOVA result143indicates that both treatments were associated with important post-treatment changes andconsequently univariate F-tests were conducted on each marker variable inducted in theMANOVA. The ANOVA results are presented in Table 8. The pre-treatment and post-treatment means for the two treatments are also presented in Table 9 and are illustrated inFigure 7.Table 8Summar’ of Univariate Tests of Fathers’ Marker Variables for Pre-treatment/Post-Treatment Differences for ExST and SF1’Marker Variable F-Value ProbabilityADD 59.37 .001SCL,-90-R 48.75 .001DAS 16.69 .001FS 7.19 .008*Significant at alpha = .01Significant at alpha = .001Degrees of freedom = (1, 116)Table 9Fathers’ Pre-treatment and Post-treatment Means and Standard Deviations of Marker Variables for ExST and SF1’Pre-treatment Post-treatmentTreatment Marker Variable Mean SD Mean SDExST ADD 5488 18.38 23.03 23.90n=40 SCL-90-R 76.04 18.09 54.59 10.98DAS 87.52 1627 101.79 18.54FS 39.41 8.45 4332 947SF1’ ADD 53.65 19.75 22.50 21.57n=20 SCL-90.R 78.17 20.14 56.34 15.46DAS 86.78 18.81 101.17 19,99FS 39.59 9.44 44.82 8.62144SCL-90-R ScoreSymptomatic RangeNormal RangeFr.-treatrrientMeasirement OccasionTreatme--ExST--SFTIFigure 7. Fathers’ pre-treatment and post-treatment means of marker variable for ExST andSFT.Sub-hynothesis 1-BWhen compared to the wives whose husbands completed SFT, the wives whosehusbands completed ExST will report significantly greater improvement at post-test onmeasures contained in the ecological assessment package employed in the study.The MANOVA testing sub-hypothesis 1-b contrasted the pre-treatment and post-treatment scores of the wives in terms of the treatment their husbands received. TheMANOVA results show that there was no statistically significant interaction between themeasurement occasion and the spouses whose husbands completed SF!’ verses those whose807060DAS ScorePre-treatment Post-treatmentMeasurement Occasion110100Non-distressed Range--.,,..1Post-treatmentDistressed RangeRn-- .. - -494745434170Pre-treatment393735Measurement OccasionPost-treatmentMeasurement Occasion145husbands completed ExST, F (S= 1, M=.5, N=56) = 0.43, p = 0.729. In light of this finding,sub-hypothesis 1-b could not be viewed as tenable and was rejected.The MANOVA result concerned with measurement occasion was found to be highlysignificant and showed that the mothers’ scores on the instruments had dramaticallyimproved, F (S= 1, M=.5, N=56) = 9.77, p = 0.001. Subsequently, univariate F-tests wereperformed on each of the marker variables and these indicated that the treatments had asignificant effect on each measure. The ANOVA results for each marker variable is shownin Table 10.Table 10Summary of Univariate Tests of Mothers’ Marker Variables for Pre-treatment and Post-treatment Differences for ExST and SF1’Marker Variable F-Value ProbabilitySCL-90-R 19.60 .001*DAS 20.17 .001*FS 13.68 .001** Significant at alpha = .001Degrees of freedom = (1, 116)Table 11Mothers’ Pre-treatment and Post-treatment Means and Standard Deviations of Marker Variables for EXST and SF!’Pre-treatment Post-treatmentTreatment Marker Variable Mean SD Mean SDExST SCL-90.R 72.65 19.43 57.87 13.58n=40 DAS 83.18 19.42 98.64 20.45FS 37.63 8.66 44.29 8.02SF!’ SCL-90-R 74.50 16.04 63.76 16.44n=20 DAS 82.81 14.81 97.11 15.69FS 39.31 7.81 42.49 7.72146The pre-treatment and post-treatment means for the two treatments for mothers’scores presented in Table 11 show the level of change on the marker variables. The pretreatment and post-treatment mean scores for both treatments are illustrated in Figure 8.SCL-gO-R Score OAS Score1•10Non-distresaed RangeSymrtat,CRaflZZZ.- 1:...Normal Range-.. 80----- - ___._ ..I IPre-ireafl ent PuSt treatn ent Pre-treatment Post-treatmentMeasurement Occasion Measurement Occasionrso—___----SET36---30 IPm-treatment Post-treatmentMeasurement OccasionFigure 8. Mothers’ pre-treatment and post-treatment means of marker variables for ExSTand SFT.Sub-hypothesis 1-CThe eldest children in families whose alcoholic father completed ExST will haveimproved with respect to their family to a significantly greater degree than the eldestchildren in families whose alcoholic father completed SFT as indicated by higher levels offamily satisfaction on the FS scale at post-test.The ANOVA testing for differential treatment effects between ExST and SFT fromthe eldest children’s perspective showed no treatment by occasion interaction, F(1, 62) =0.11, p = 0.738. This result meant that the eldest children whose fathers received ExST had147not fared significantly better than the eldest children whose fathers completed SFT.Accordingly, sub-hypothesis 1-C was rejected.The analysis of the pre-treatment and post-treatment measurement occasion dataindicated that the SFT and ExST treatments had no statistically significant impact on thesiblings in the family from the eldest children’s perspective. Unlike the analyses of sub-hypothesis 1-a and 1-b, this analysis was not significant, F (1, 62) = 0.97, p = 0.330. Anexamination of the pre-treatment and post-treatment means found below in Table 12 andgraphed in Figure 9, reveals that there was indeed very little pre-treatment/post-treatmentchange at the family level reported by the eldest children.Table 12Eldest Children’s Pre-treatment and Post-treatment Means and Standard Deviations of Marker Variable for ExST and SF1’Treatment Marker Variable Pre-treatment Post-treatmentMean SD Mean SDExST FS 45.89 8.58 47.61 8.85n =23SF1’ FS 40.00 10.58 43.50 13.91n = 10Treatment-.- ExST—Sn.Figure 9. Eldest children’s pre-treatment and post-treatment means of marker variable forExST and SFT.Pre-treatment Post-treatmentMeasurement Oceasion148Second Research Hypothesis: Differential Efficacy of ExST-I and ExST-CThe second research hypothesis focused attention on the differential treatmenteffects of the two formats (individual and couples) of ExST. Six sub-hypotheses werearticulated for this hypothesis in which three dealt with pre-treatment/post-treatmentdifferences, and three focused on post-treatment/follow-up differences. The results fromthe analyses pertaining to each sub-hypothesis follow.Sub-hvnothesis 2-aWhen compared to alcoholics who completed ExST-I, alcoholics who completedExST-C will report significantly greater improvement at post-test as measured by selectedinstruments in the ecological assessment package used in this study.A multivariate analysis of variance using the pre-trcatment and post-treatment scoresas dependent factors and treatment format as independent factors was conducted to test thissub-hypothesis. Employing the ADD, SCL-90-R, DAS and FS as marker variables, theMANOVA results showed that there was no statistically significant interaction between thetwo treatment formats and the measurement occasions as far as the fathers’ were concerned,F (S= 1, M= 1, N=35.5) = 1.35, p = 0.261. Consequently, sub-hypothesis 2-a was rejected.A highly significant measurement occasion difference which contrasted the pretreatment and post-treatment scores also emerged as part of this analysis, F (S =1, M= 1, N=35.5) = 18.75, p = 0.001. The subsequent univariate analyses of fathers’ marker variablesrevealed that significant change had occurred at all levels of ecological assessment. TheANOVA results appear below in Table 13.149Table 13Summary of Univariate Tests of Fathers’ Marker Variables for Pre-Treatment and Post-Treatment Differences for ExST-I and ExST-CMarker Variable F-Value ProbabilityADD 45.61 .001SCL-90-R 41.63 .001DAS 14.42 .001FS 4.50 .037** Significant at alpha = .05Significant at alpha = .001Degrees of freedom = (1, 76)An examination of the pre-treatment and post-treatment means for the two treatmentformats of ExST shown below in Table 14 reveals the levels of change for each markerindex. The mean scores of the marker variables at pre-treatment and post-treatment forExST-I and ExST-C are found in graphic form in Figure 10.Table 14Fathers’ Pre-treatment and Post-treatment Means and Standard Deviations of Marker Variables for ExST-1 and ExST-CTreatment Marker Variables Pre-treatment Post-treatmentMean SD Mean SDExST-I ADD 56.40 18.46 16.82 23.31n =20 SCL-90-R 72.25 16.95 53.06 10.98DAS 93.18 13.78 106.75 17.33FS 41.70 8.16 46.50 7.92ExST-C ADD 53.36 18.64 29.24 23.41n=20 SCL-90-R 79.84 18.83 56.13 11.03DAS 81.87 16.91 96.84 18.81FS 37.11 8.29 40.54 10.13150ADD Score SCL9OR Score6050--403020I I I IPre-treatment Post-treatment Pre-treatment Post-treatmentMeasurement Occasion Measurement OccasionDAS Score FS Scorel1.. 50110——-.__...._Distressed Range85 ..80 30Pre-treatment Post-treatment Pro-treatment Post-treatmentMeasurement Occasion Treatment Measurement Occasion-*-EXST-I -.-ExST-CFigure 10. Fathers’ pre-treatment and post-treatment means of marker variables for ExST-Iand ExST-C.Sub-hypothesis 2-bWhen compared to the wives of alcoholics whose husbands completed the ExST-I,wives who completed ExST-C with their husbands will report significantly greaterimprovement at post-test as measured by selected instruments in the ecological assessmentbattery in this study.In keeping with the earlier analysis of mothers’ data, the SCL-90-R, DAS, and FSmarker variables were employed to test this sub-hypothesis. The multivariate analysiscontrasting the pre-test and post-test measurement occasions and the two treatment formatsof ExST relevant to sub-hypothesis 2-b showed no statistically significant interaction , F80Moderate Dependency :Mud5040-...Normat F4angeRanQe—A151(S 1, M=.5, N=36) = 0.98, p = 0.405. This finding rendered sub-hypothesis 2-b untenableand as a consequence it was rejected.The rejection of the sub-hypothesis was once again accompanied by a highlysignificant measurement occasion effect. The MANOVA results again indicated that bothtreatment formats had brought about eco-systemic change, F (S =1, M= .5, N=36) = 7.49, p= 0.001. Separate univariate F-tests were performed on each index included in the markervariable set and the results of these analyses are shown below in Table 15. The levels of thepre-treatment and post-treatment changes of the marker variables for both treatmentformats are found below in Table 16 and these are shown in graph form in Figure 11.Table 15nf T 1nirt. Tt nf Mrkr V,rhIp fr Pr..trp,tmpnt n1 Pnct-tn,,tmpnt ffpnc fnr Fvcl’.T P.vSTflMarker Variables F-Value ProbabilitySCL-90-R 15.55 .001*DAS 12.61 .001*FS 12.43 .001** Significant at alpha = .001Degrees of freedom = (1, 76)Table 16Mothers’ Pre-treatment and Post-treatment Means and Standard Deviations of Marker Variables for ExST-1 and ExST-CTreatment Marker Variable Pre-Treatment Post-TreatmentMean SD Mean SDExST-I SCL-90-R 68.86 12.77 58.44 12.45n=20 DAS 90.44 13.69 100.03 16.41FS 38.31 8.24 44.78 8.56ExST-C SCL-90-R 76.44 24.12 57.44 14.93n=20 DAS 75.92 21.79 97.25 24.19FS 36.96 9.23 43.79 7.64Measurement OccasionFigure 11. Mothers’ pre-treatment and post-treatment means of marker variables for ExST-Iand ExST-C.Sub-hypothesis 2-cWhen compared to the eldest children of alcoholics whose fathers completed ExST-I,the eldest children of alcoholics whose parents both completed ExST-C will improve withrespect to their family to a significantly greater degree as indicated by higher levels of familysatisfaction as measured by the FS scale at post-test.A univariate F test was performed to test the sub-hypothesis related to differentialtreatment format effects from the perspective of the eldest child using the FS marker as thedependent variable. The ANOVA results revealed no statistically significant interactionbetween the treatment formats and the measurement occasion from the children’s point of8070DAS ScoreSCL-90-R ScoreNormai Range5040110100908070Pre-treatment152Non-distassed RangePre-treatmentMeasurement OccaonPost-treatmentPost-treatmentMeasurement OccasionFS Scors50Normal RangeTreatment-‘.-ExST-I--ExST-CPre-treatment Post-treatment153view, F (1, 42) = 0.01, p = 0.989. In accordance with this finding, sub-hypothesis 2-c wasrejected.In keeping with the earlier analysis of the eldest children’s data, the ANOVA testverifying whether or not the treatment formats had any substantive impact was notstatistically significant, F (1, 42) = 0.19, p = 0.664. The eldest children’s pre-treatment andpost-treatment means on the marker variable for both ExST-I and ExST-C appear in Table12 and are charted in Figure 11. The data illustrates how little variation on the scores wereevident between measurement occasions.Table 17Eldest Children’s Pre-Treatment and Post-Treatment Means and Standard Deviations of Marker Variable for ExST-I and ExST-CTreatment Marker Variable Pre-treatment Post-treatmentMean SD Mean SDExST-I FS 46.47 4.38 48.22 8.76n = 11ExST-C FS 45.36 11.37 47.40 9.28n = 12tS ScoraNomaJRa11ge_______50 -—-•-—- ••---—-•—-••--••—---- —-...... Treatment40 ---—-------------—-——--—---—-———------——--—---‘---—---—-35 IPre-treatment Pt4reatmentMeasurement OccasionFigure 12. Eldest children’s pre-treatment and post-treatment means of marker variable forExST-I and ExST-C154Sub-hypothesis 2-dWhen compared to the alcoholics who completed ExST-I, alcoholics who completedExST-C will report significantly greater improvement at follow-up as measured by selectedinstruments in the ecological assessment battery used in the study.A multivariate analysis of variance was conducted on the fathers’ post-treatment andfollow-up data to test whether or not there was a differential treatment format effectdetectable at follow-up on the marker variables used in testing sub-hypothesis 2-a. Theresults of the MANOVA indicated that the treatment format and measurement occasioninteraction did not reach acceptable levels of significance, F (S= 1, M= 1, N= 11) = 1.15, p= 0.357. This finding necessitated the rejection of sub-hypothesis 2-d.The MANOVA analysis that was performed to test for a measurement occasion maineffect yielded no significant difference, F (S= 1, M= 1, N= 11) = 1.199, p = 0.337. Thisfinding indicated that the treatment gains noted at post-treatment were stable up to followup.A presentation of the mean score values of the marker variables for both posttreatment and follow-up occasions and the two treatment formats is found in Table 18. Itshould be noted that the sample in the post-treatment/follow-up contrast is smaller than thesample used in the treatment/post-treatment comparison. The mean values appearing inTable 18 are charted in Figure 13.155Table 18Fathers’ Post-Treatment and Follow-up Means and Standard Deviations of Marker Variables for ExST-I and ExST-CU IPost-treatment Follow-upMeasurement OccasionDAS ScoreI IPost-treatment Follow-upMeasurement Occasion Treatment Measurement Occasion-*-ExST-l-ExST-CFigure 13. Fathers’ post-treatment and follow-up means of marker variables for ExST-I andExST-C.Treatment Marker Variable Post-treatment Follow-upExST-In= 16ExST-Cn = 13ADDSCL-90-RDASFSADDSCL-90-RDASFSMean SD Mean SD16.83 21.54 23.06 28.6654.39 11.49 61.99 14.00103.63 15.62 100.84 18.9145.81 7.60 43.00 9 .0931.92 27.18 30.85 30.8658.45 12.99 58.15 19.5397.60 21.02 102.97 22.1141.14 9.69 40.85 11.22ADD Score SCL-90-R ScoreModerate Dependency_Mild Dependency - -Non-distressed Range- NcmialRange50 ————----—-—-—---———-—----—----—----—-—-----—--——--——--—--i.-4540Post-treatment Follow-upMeasurement OccasionFS Score5550Nomial Range45---..--------403530 I IPost-treatment Follow-up10095 --- ---.-...Distressed Range90 ------ -85 --- ------ --- ------no I I156Sub-Hypothesis 2-cWhen compared to wives of alcoholics whose husbands completed ExST-I, wives whocompleted ExST-C with their husbands will report significantly greater improvement atfollow-up as measured by selected instruments in the ecological assessment package used inthe study.The MANOVA conducted to test sub-hypothesis 2-e found that there was nosignificant treatment format and measurement occasion interaction differentiating thewomen in this study, F (S =1, M= .5, N= 12) = 0.236, p = 0.870. In view of this finding, thesub-hypothesis could not be held as tenable and was rejected.The durability of the significant treatment effects of both formats revealed in thetesting of sub-hypothesis 2-b were explored by the multivariate analysis testing sub-hypothesis 2-e. The MANOVA showed that the treatment effects had not deteriorated to astatistically significant degree, F (S =1, M= .5, N= 12) = 0.26, p = 0.852.The post-treatment and follow-up mean values for each marker variable andtreatment format group are found in Table 19. It should be noted that the posttreatment/follow-up analysis was conducted on a somewhat smaller sample size than thepre-treatment/post-treatment analysis. The mean values in Table 19 are illustrated inFigure 14.157Table 19Mothers’ Post-treatment and Follow-up Means and Standard Deviations of Marker Variables for ExST-I and ExSI’-CMean SD MeanSCL-90-R 57.06 11.98 56.22DAS 98.98 17.53 99.99FS 44.17 8.82 43.92SCL-90-R 55.51 11.99 54.03 13.57DAS 104.62 19.03 103.88 21.49FS 44.30 8.12 46.02 11.00SD13.3217.0810.00Symptomatic Range60-Aic . .- —I.95 -..-..--.--Distressed Rangeon -Figure 14. Mothers’ post-treatment and follow-means on marker variables for ExST-I andExST-C.Marker VariableTreatmentExST-In = 17ExSr-Cn = 13FnIlnw-iinSCL-90-R Score DAS SCORE110f rNon.dirssed RangePost-treatment Follow-upMeasurement Occasionoc)80Post-treatment Follow-upMeasurement OccasionTreatment-*-ExST-I-.-EXST-CPost-treatment Follow-upMeasurement Occasion158Sub-Hvnothesis 2-fWhen compared to the eldest children of alcoholics whose fathers completed ExST-I,the eldest children of alcoholics whose parents both completed ExST-C will improve withrespect to their family to a significantly greater degree as indicated by higher scores of familysatisfaction as measured by the FS scale at follow-up.The differential treatment format effect associated with the individual and couplesforms of ExST from the perspective of the eldest children was examined employing aunivariate analysis of variances. The ANOVA results based on data from the FS markervariable indicate that there was no significant interaction between the two treatment formatsand the measurement occasion as far as the eldest children were concerned, F (1, 15) = 0.33,p = 0.577. Accordingly, sub-hypothesis 2-f was rejected.The effect of the two formats were also shown to be not statistically significant in theANOVA employed for the analysis of the eldest children’s post-treatment follow-up data, F(1, 15) = 0.27, p = 0.6 12. Thus, it seems clear that the eldest children’s assessment of theirfamily was relatively unaffected by either treatment formats at post-test and follow-up. Thepost-treatment and follow-up means and standard deviations for the eldest childrenparticipants appear in Table 20 and are graphed in Figure 15.Table 20Eldest Children’s Post-treatment and Follow-up Means and Standard Deviations of Marker Variable for ExST-1 and ExST-C.Treatment Marker Variable Post-treatment Follow-upMean SD Mean SDExST-I FS 47.39 9.24 49.72 11.11n=9ExSr-C FS 44.69 9.31 44.58 9.15n =8.159FS Score€0— —______Treatment4 -—- —--- -— -.-ExSTINomialkange . .::..... . . ExST-C-‘-35 .30 IPost-treatment Follow-upMeasurement OccasionFigure 15. Eldest children’s post-treatment and follow-up means of marker variables forExST-I and ExST-CWithin Systems AnalysisThe within systems analyses were conducted as indicated by the eco-system analyses.While all of the differential treatment effect (ExST and SFT) and the differential treatmentformat effect (ExST-I and ExST-C) sub-hypotheses were rejected, the pre-treatment andpost-treatment contrasts for fathers and mothers for both treatment and format werecharacterized by highly significant occasion main effects across all of the marker variables.Consequently, within systems analyses were conducted for all of the levels of assessment forboth fathers and mothers in the pre-treatment/post-treatment contrast of all threeexperimental treatment groups.Within Alcohol SystemA MANOVA was conducted using the pre-treatment and post-treatment fathers’data. Included as dependent variables were the eight sub-scales of both the IDS and SCQ.Thus, a total of 16 sub-scales served as dependent measures in this analysis. The results ofthis multivariate test mirrored the eco-systems results and indicated that while there was nosignificant treatment and measurement occasion interaction, approximate F (S =1, M= 7.5,N= 18.5) = 0.96, p = 0.54, there was a substantial statistically significant difference found160between the pre-treatment and post-treatment scores, F (S =1, M= 7.5, N= 18.5) = 15.28, p= 0.001.Univariate F-tests performed on each of the sub-scales revealed that measurementoccasion differences on all but one of the sub-scales in SCQ were highly significant. Theresults of the ANOVA tests appear in Table 21Table 21Summary of Univariate Tests Within Alcohol Level of Assessment of Pre-treatment and Post-treatment DifferencesInstrument Sub-scale F ProbabilityIDS Unpleasant emotions 79.40 .001*Physical discomfort 40.03 .001*Pleasant emotions 86.89 .001*Testing personal control 40.28 .001’Urges and temptations 104.12 .001’Conflict with others 110.43 .001’Social pressure to drink %.94 .001’Pleasant times with others 122.38 .001’SCQ Unpleasant emotions 24.19 .001’Physical discomfort 20.89 .001’Pleasant emotions 2.91 .093Testing personal control 32.76 .001’Urges and temptations 26.35 .001’Social problems 47.39 .001’Social tensions 23.47 .001’Positive social situations 30.41 .001’‘Significant alpha = .001Degrees of freedom = (1, 55)The pre-treatment total sample means and post-treatment means for instrument sub-scales are charted below for the three treatment groups in Figure 16. As Figure 16 shows,161the pre-treatment total sample means and post-treatment values for all treatment conditionsshow considerable improvement.Within Alcohol SystemOS Score7060So4030 Moderate RistcKEY0 -s-GroupLit Pyi4i1 Pt PSfl Urg SOQ Pre-testEmia.s m$t Cnd Tt. Cn1i5 Prnuw. ThmlOS Sub-scale --ExST-lPost-test-.-EXST-CPost-test.-SFTFigure 16. Fathers’ total sample pre-treatment means and treatment group post-treatmentmeans for within alcohol system variables.Within Intrapersonal SystemThe indices involved in this within system analysis were the nine sub-scales from theSCL-90-R and the total depression score from the BDI. Thus, pre-treatment and post-treatment data from ten dependent variables served as data for the mixed model MANOVAconducted at the intrapersonal system level. Separate analyses were conducted on fathers’and mothers’ data.‘ISCO Sub-scales162The results from the fathers’ analyses were in keeping with earlier findings andshowed that there was no statistically significant interaction between the measurementoccasions and the treatments, approximate F (S =2, M=3.5, N=22.5) = 0.97, p 0.508.The MANOVA also expanded on the substantive client changes that was associated with thetreatments. Again, a highly significant difference was found between the pre-treatment andpost-treatment scores, F (S=1, M=4, N=22.5) = 10.43, p = 0.001. Separate univariate F-tests were conducted subsequent to this finding. As revealed in Table 22 there were largestatistically significant differences between each dependent variable at pre-treatment andpost-treatment.Pre-treatment total sample mean and post-treatment means for each treatmentcondition and for each dependent measure are compared to the non-patient norms and arecharted below in Figure 17.Table 22Summary of Univariate F-Tests of Fathers’ Within Intrapersonal Level of Assessment for Pre-treatment and Post-treatment DifferencesInstrument Sub-scale F-Value ProbabilitySCL-90-R Somatization 29.53 .001*Obsessive-compulsive 85.03 .001*Intrapersonal sensitivity 54.80 .001*Depression 65.45 .001*Anxiety 71.49 .001*Hostility 44.86 .001*Phobic anxiety 14.35 .001*Paranoid ideation 33.45 .001*Psychoticism 56.81 .001*BDI Total 35.94 .001** Significant at p = .001Degrees of freedom = (1, 56)163Within Intrapersonal SystemSCL-90-R ScoreKEY-eGroupPre-testExST-lPost-test- ExST-CPost-test•-SF1Post-testKEYDGroupPre-testExST4Post-testDExST-CPost-testIULISFTPost-testFigure 17. Fathers’ total sample pre-treatment means and treatment group post-treatmentmeans for within intrapersonal system variables.The multivariate within intrapersonal analysis of mothers’ pre-treatment and post-treatment data was consistent with fathers’ within intrapersonal system findings. Once again,the MANOVA, while establishing no significant measurement occasion and treatmentcondition interaction, F (S =2, M=3.5, N=23) = 0.67, p = 0.842, also indicated considerablestatistically significant measurement occasion main effect for all treatment groups, F (S=1,M=4, N=23) = 3.58, p = 0.001. Dependent variable ANOVA’s showed that there hadbeen marked significant improvement on all variables employed at this level of analysis.The results of the univariate F-tests appear below in Table 23 and the post-treatment dataare presented in Figure 18.SOM OBS NT DEP ANX HOS PHOB PAR PSYSCL-90-R Sub-scaleTreatment Group164Table 23Summary of Univariate Tests of Mothers Within Intrapersonal Ievel of Assessment for Pre-treatment and Post-treatment DifferencesInstrument Sub-scale F-Value ProbabilitySCL-90-R Somatization 14.21 .001*Obsessive-compulsive 26.41 .001’Interpersonal sensitivity 29.42 .001*Depression 27.81 .001”Anxiety 21.85 .001Hostility 11.21 .001*Phobic anxiety 4.48 .039*Paranoid ideation 17.84 .001*Psychoticism 10.07 .002**BDI Total 20.59 .001*SCL-90-R Score* Significant at alpha = .05Significant at alpha = .01Significant at alpha £01Degrees of freedom = (1, 57)Within Intrapersonal System8070605040BOl ScoreSCL-90-R Sub-scaleKEY-0-GroupPre-test—ExST-IPost-test-.-ExST-CPost-test--SFTPost-testKEYJGroupPre-testLIExST-lPost-testE]ExST-CPost-test!lSFTPost-testFigure 18. Mothers’ total samples pre-treatment means and treatment group post-treatmentTreatment Groupmeans for within personal intrapersonal system variables.165The treatment gains on these variables illustrated in Figure 18, show the threetreatment groups post-treatment means contrasted with the total samples pre-treatmentmeans for each sub-test. As the figure reveals, the improvement rate of the ExST treatmentsappear slightly higher than the SFT condition in a consistent fashion. While this differenceis not statistically significant, it would appear to be important on a clinical level as the ExSTvalues are within normal limits while the SFT values are outside of the normal ranges.Within Couple SystemThe within couple system analyses included three separate MANOVA runs includinganalyses focused on fathers’ mothers’ and couples data. The couples perspective analysiswas based on variables from the AC measure that require an aggregation of fathers’ andmothers’ scores. The analyses for both fathers and mothers included the four sub-scalevariables constituting the DAS and the five relevant sub-scales of the AC.The results from the MANOVA on fathers’ and mothers’ views reflected a similarpattern of results which found no treatment by measurement occasion interaction coupledwith substantive pre-treatment/post-treatment change. The values for fathers’ MANOVAwere, approximate F (S =2, M=3.5, N=23) = 90, p = 0.575 for the differential treatmentgroup contrast, and F (S=1, M=3.5, N=23) = 5.97, p = 0.001 for the measurementoccasion main effect. Similarly the values for mothers’ MANOVA were, F (S =2, M=3,N=23.5)= l.O4,p = 0.426, and F(S= 1, M=3.5, N=23) = 5.97, p = 0.001 for the twocontrasts respectively.The results from the subsequent ANOVA tests on each dependent variable forfathers and mothers are shown in Table 24.166Table 24Summary of Univariate Tests of Fathers’ and Mothers’ Within Couple Level of Assessment for Pre-treatment and Post-treatmentDifferencesFather MotherInstrument Sub-scale F-Value Probability F-Value ProbabilityAC Agreement 10.29 .002* 44.22 .001Disagreement 23.81 .001*1 2.26 .138Desired Change 8.49 .005* 34.29 .001Perceived Change 18.83 .0011* 21.19 .001*1Perceptual Accuracy 1.28 .262 .32 .574DAS Concensus 25.64 .001*1 29.62 .0011*Satisfaction 25.29 .0011* 36.33 .O01Affection 12.23 .0011* 8.53 .005*Cohesion 18.51 .001 16.88 .0011** Significant at alpha = .01Significant at alpha = .001Fathers’ analyses, degrees of freedom = (1, 56)Mothers’ analyses, degrees of freedom = (1, 57)The magnitude of the post-treatment changes for the fathers’ and mothers’ at the withincouples level of assessment are illustrated in Figures 19 and 20. In these figures, the totalsample pre-treatment mean scores are provided to serve as a contrast group to the treatmentcondition post-treatment scores for fathers and mothers on the DAS and AC instruments.16787654Within Couples SystemiJFigure 19. Fathers’ and mothers’ total sample pre-treatment means and treatment grouppost-treatment means for within couple system variable DAS.5045403530Satisfaction ScorePerspectiveFather MotherPerspectivePerspective PerspectiveKEYDGroup ExST4 DExST-C lSFTPre-test Post-test Post-test Post-test4035302520151021.50.5Within Couples SystemAccuracy ScoreFatherKEYDGroup 1ExST-I DExST-C DSFTPre-test Post-test Post-test Post-test168Figure 20. Fathers’ and mothers’ total sample pre-treatment means and treatment grouppost-treatment means for within couple system variables AC.Perspective PerspectivePerspectiveFather Mother Father MotherPerspective//Perspective169The MANOVA results on the couples sub-scales on the AC were consistent withthose found for the fathers’ and mothers’ within couple system analyses. The multivariateanalyses was performed on four dependent variables and showed that while no significantmeasurement and treatment condition interaction was detectable, approximate F (S =2,M= 1, N=25.5) = 1.52, p = 0.141, a large measure of change had occurred for all groups atpost-treatment, F (S= 1, M= 1.5, N=25.5) = 39.12, p = 0.001. The univariate analysescarried out subsequent to the MANOVA revealed that the pre-treatment/post-treatmentchanges were substantial across all four variables. A summary of the series of ANOVA’s isfound below in Table 25.Table 25Summary of Univarjate Tests of Couples Variables for Within Couples Level of Assessment for Pre-treatment and Post-treatmentDifferencesInstrument Sub-scale F-Value ProbabilityAC Total Agreement 44.72 .001Total Disagreement 16.52 .OOPDesired Change 35.59 .001Perceived Change 32.52 .001* Significant at alpha = .001Degrees of freedom = (1, 57)A visual comparison of the total samples pre-treatment means for each couplesvariable and the post-treatment values for each by treatment group is enabled by Figure 21.170Within Couples System756555453525//IFigure 21. Couples’ total sample pre-treatment means and treatment group post-treatmentmeans for within couple system variables.Within Family SystemThe final series of within systems analyses centered on the family level of assessmentand were based on the six sub-scales of the FES and the two sub-scales of both the FACESIII and the FS. Consequently, ten dependent variables were employed in each of theMANOVA runs for fathers’ and mothers’ perspectives.The pre-treatment/post-treatment MANOVA on fathers’ data, while closer thanother within system analyses, indicated that there was no statistically significantmeasurement occasion by treatment condition interaction, F (S =2, M=3.5, N=23) = 1.45, p= 0.118. The yield from the mixed model MANOVA on data from mothers’ perspectivesTotai Agreement ScoreKEYGroup 33ExST-I 1ExST-C EIISFTPre-test Posttest Post-test Post-testTreatment Group171was more definitive finding regarding the absence of differential treatment effects, F (S = 2,M=3.5, N=22.5) = 0.69, p = 0.8 18.Both fathers’ and mothers’ analyses echoed the previous within systems analyses interms of identifying highly significant measurement occasion main effects for all treatments,F(S=1,M=4,N=23) = 5.47, p = 0.O0lforfathers,andF(S=1,M=4,N=23) = 4.49, p =0.001 for mothers.The univariate F-tests that followed the MANOVA results showed that seven offathers’ ten and eight of mothers’ ten dependent variables were significant at the .05 level orgreater. The ANOVA findings are presented below in Table 26.Table 26Summary of Univariate Tests of Fathers’ and Mothers’ Within Family Level of Assessment for Pre-treatment -DifferencesFather MotherInstrument Sub-scale F-value Probability F-value ProbabilityFES Cohesion 17.68 .001* 12.48 .001***Expressions 21.69 .001* 20.95 .001Conflict 8.91 .004 14.41 .001*Independence .35 .558 8.14 .006**Organization 13.75 .001*** 4.35 .041*Control .87 .354 .24 .624FACES III Cohesion 15.12 .001* .01 .920Adaptability .002 .962 5.33 .025*FS Cohesion 19.19 .001 23.49 .001*Adaptability 10.66 .002 19.34 .001** Significant at alpha = .05** Significant at alpha = .01Significant at alpha = .001Degrees of freedom = (1,57)172The changes indicated by these analyses are illustrated below in Figures 22 and 23which chart fathers’ and mothers’ total sample pre-treatment means and the post-treatmentmeans for each dependent variable by treatment group.Within Family SystemFESKEY_______________________________________________________-e- GroupPre-test--ExST-lPost-test--ExST-CPost-test.-SFTPost-testFES Standard ScoreCohesion Expressive. Conflict Indep. Organization ControlFES Sub-scaieFACES III FSCohesion Adaptability Cohesion AdaptabilityFACES Iii Sub-scale FS Sub-scaleKEYC]Group ExST-l C]ExST-C IIIIIIISFTPre-test Post-test Post-test Post-testFigure 22. Fathers’ total sample pre-treatment group post-treatment means for within familysystem variables.Within Family SystemFESKEYDGitup ExST-1 EJExST-C tISFTPre-test Posttest Posttest Post-testKEY-& GroupPre-test--ExST-lPost-test--ExST-CPost-test—SF1Post-test173Figure 23. Mothers’ total sample pre-treatment means and treatment group post-treatmentmeans for within family system variables.Therapeutic Process ValidationIn this section the data was analyzed to explore for the possibility of treatment effectsassociated with clinic site, gender of therapist and therapist.Clinic SiteA MANOVA was conducted comparing the pre-treatment and post-treatment databy clinic site on the marker variables for fathers and mothers separately. The MANOVAresult based on data from the fathers’ perspective indicated that there was no statisticallyFES Standard Score7060So4030FACES IIICohesion Expressrve. Conflict Indep. Organization ControlFES Sub-scaleFS Score4035302520isES2723: f:1-iCohesionCohesion AdaptabilityFACES Ill Sub-scaleI .-7AdaptabilityFS Sub-scale174significant interaction between the measurement occasions (pre-test and post-test) and clinicsite, F (S= 1, M= 1, N=26.5) = 0.324, p = 0.860. The clinic site MANOVA based onmothers’ data was also not statistically significant, F (S= 1, M= 5, N=27) = 0.257, p =0.855.Therapist GenderThe hypothesis that the gender of the therapist played a major role in the results ofthe study was explored by comparing the pre-treatment and post-treatment scores on markervariables for fathers and mothers seen by the male and female therapists. The MANOVAcontrast for therapist gender and measurement occasion for fathers was not statisticallysignificant, F (S = 1, M= 1, N=26.5) = 0.59, p 0.671. Similarly, the gender of therapist andmeasurement occasion multivariate analysis based on mothers’ data showed no significantinteraction and indicated that this variable had no substantial impact on the results, F (S=1,M=.5,N=27) = 0.400, p = 0.753.TherapistThe multivariate analysis test meant to verify whether or not the fathers’ results hadbeen substantially effected by the individual therapist that had provided the treatment wasconducted using the pre-treatment and post-treatment marker variable data. Only thetherapists who had completed four or more cases were included in this particular contrast(n = 8). The MANOVA results indicate that there was no statistically significant interactionbetween the therapists and the measurement occasions, approximate F (S=4, M= .5, N=22)= 0.858, p = 0.658. The MANOVA testing for therapist effects based on mothers’ dataalso yielded no significant difference, approximate F ( S=3, M= 1, N=22.5) = 1.09, p =0.362.175CHAPTER V: DISCUSSIONIn this chapter the results of the four phases of data analysis are discussed. Since theresearch questions centered on the differential efficacy issues, the results of the hypothesestesting Phase 2 will be presented first before the within systems and therapeutic processvalidation phases of analysis are considered. The limitations and the generalizability of thestudy are addressed prior to the conclusions of the research and future directions sectionswhich complete the discussion chapter.Eco-System AnalysesThis portion of the data analysis challenged the tenability of the two researchhypotheses. The first research hypothesis which asserted the ExST treatment would be morepotent than the SFT comparison treatment was not supported by the study. The multivariateanalyses of the marker variables indicated that there were no statistically significantdifference between the two treatments and a review of the means and standard deviationsrevealed that there was indeed very little separating ExST and SFT at post-treatment.The second research hypothesis that the couple treatment format of ExST would besuperior to the individual treatment format of ExST at post-test and at follow-up was notsupported. At post-treatment there was no indication that either of the formats outperformed the other. Furthermore, no differences between the formats emerged at followup. These findings are expanded upon in more detail below.First Research Hypothesis: Differential Efficacy of ExST and SETOne of the aims of this study was to test the efficacy of ExST. The ExST model hadbeen developed for the treatment of alcohol dependency in response to both the calls forinnovation in treatment in the area by researcher and the pressing needs of therapists in thefield. The integrative qualities of ExST gave the model a novel treatment orientation that176allowed therapist to tailor therapy programs to meet the unique needs of individual clients.Highly favorable results from informal case studies by therapists implementing ExST inclinical settings motivated the developers of the model to further examine the effectivenessof the treatment.The study was originally designed to test the efficacy of ExST compared to a wait-listcontrol group, however, the design was altered for pragmatic reasons. Accordingly, SFT wasdeveloped and implemented as a comparison treatment condition. The intent behind thegeneration of SFT was to develop a treatment condition that was distinctly different fromExST in a number ways and at the same time, control for a variety of variables includingtherapeutic involvement, attention, and gains arising as a consequence of the repeatedmeasurement of participants with an extensive array of instruments. Nonetheless, theresults suggest that the SFT treatment functioned as a potent form of intervention.The non-significant differential treatment effect results are best understood in view ofthe magnitude of pre-treatment/post-treatment changes associated with both treatmentconditions. The fathers and mothers involved in both ExST and SFT reported significantamounts of change at all levels of eco-systemic assessment. Consequently, the reason therewas no differential treatment effect was not due to a poor performance of the ExSTtreatment. Indeed, the ExST treatment outcomes were statistically significant and clinicallyimpressive. Rather, the non-significant finding was the result of the SFT treatmentperforming in a fashion that far exceeded expectations. The unpredicted potency of SF1’raises questions regarding how this quasi-control comparison treatment function so well. Itis necessary to consider several factors to explain the success of SFT.The first consideration that is noteworthy with respect to the SFT treatment pertainsto the broad-spectrum quality of this therapy. The assumption that underlies broadspectrum models of treatment is that the drinking behavior is functionally related to a177variety of other problems in a person’s life. Broad-spectrum treatment approaches whichaddress a host of difficulties in the alcoholics life rather than focusing on the drinkingbehaviour alone have been shown to be effective in treating dependency problems (Miller& Hester, 1986). SFT focused on a considerable breadth of life problems with clients andwas based on the Weekly Situation Diaries which were carefully designed to measure majorareas of the clients’ ecological context terms of change, satisfaction and proximity to theideal. Accordingly, it may be appropriate to think of SFT as a broad-spectrum systemicmodel of treatment.A second feature in explaining SFTs success is concerned with technique. One ofthe technical cornerstones of SFT is the behaviour monitoring aspect of the therapy.Behaviour monitoring is a component of therapy that has been used in treating a variety ofproblems including alcohol dependency. Self-monitoring has a long-standing impressiveclinical record (e.g., Alden, 1988; Beck, Rush, Shaw, & Emery, 1979; Connors, Tarbox ,&Faillace, 1992; Elkin, Shea, Watkins, Imber, Sotsley, Collins, Glass, Pilleonis, Leber,Docherty, Fiester, & Parloff, 1989; Heather, Robertson, MacPherson, Ailsop, & Fulton,1987; Sanchez-Craig, Annis, Bornet, & MacDonald, 1984). Self-monitoring has recentlybeen used successfully as a key component in the psychological treatments of a wide array ofproblems including depression. (Beck et al., 1979; Elkin et al., 1989; McKnight, Nelson-Grey& Barnhill, 1992), panic attacks (Salkovskis, Clark, & Hackmann, 1991), bulimia (Wilson,Eldredge, Smith, & Niles, 1991), inflammatory bowel disease (Schwartz & Blanchard, 1991),chronic lower back pain (Nicholas, Wilson, & Goyern, 1991), and alcohol dependency(Connors, Tarbox, & Faillace, 1992). Given that behaviour monitoring has been associatedwith positive treatment outcomes, it is not surprising that SFT fared so well. In addition,because the scope of the self-monitoring procedure used in this study had a systemicorientation that addressed a broad- spectrum of relational contexts, it is possible that the178positive effects typically associated with behaviour monitoring procedures were amplified inthe case of SFT.A third point in connection to the behaviour monitoring and charting process, is thatthe procedure served to assist the clients in deconstructing their lives (Foucault, 1980; White,1992). This is to say that clients in SFT learned to differentiate various domains (individual,couple, family, friends, and work) of their lives that were usually fused at the onset oftreatment. Typically, when clients would appear at their first SF!’ sessions, the evaluation ofthe various systems they were monitoring would be identical and undifferentiated. Within 2or 3 session, the assessments at each level would routinely begin to differentiate and thevarious domains of life would begin to take on a quality of independence from one another.Accordingly, clients learned to find areas of life that were improving even as others were notand this aspect of the feedback seemed particularly encouraging and helpful to clients.A fourth factor important in explaining the potency of SFT is the charting componentof the approach. The wall charts were designed to allow the person’s own narrative toappear before them and thereby, as Grigg et al. (1991) noted, “enable client’s own livingprocess to speak directly to them” (pg. 1). In this way, many aspects of the client’s life andtheir problems were made external to them. The process of externalization has recentlybecome an important feature to many approaches in family therapy (e.g., White & Epston,1991) and is thought to be an important technique in enabling clients to generate newsolutions to their problems.Another factor connected to the discussion of SFI”s potency is concerned with theareas of overlap that were shared by both SFT and ExST. Both treatments aimed atestablishing a therapeutic relationship between client and therapist that was empathic,caring, non-judgmental, genuine, and warm. Both treatments conveyed a respectful and non-179coercive valuing of clients that welcomed clients’ problems and honesty. Additionally, thetherapies were based on the attainment of clearly defined goals that were established at thebeginning of the treatments and both treatments functioned as forms of brief therapy with aclearly defined time of termination. These general factors of the treatments are themselvesimportant features of therapy which no doubt contribute substantively to the improvementsreported by clients directly involved in the treatments.In addition to the treatment factors that help explain the impressive performance ofSFT, two issues related to the research design are germane to the discussion. The first ofthese issues relates to the context of post-treatment responding for SFT participants. WhenSFT families entered the study, they did so with the understanding that after the 16 weeks ofSF!’ treatment was finished, they would be asked to complete the post-treatmentquestionnaires and would then be welcome to continue with additional treatment. In fact,68.75% of the SET participants continued with some form of treatment after the post-treatment questionnaires were collected, and 90% of these people initiated couples therapywith clinicians trained in ExST. In view of this fact, it is clear that many of the SETparticipants were anticipating further treatment at the time of post-treatment questionnairecompletion. In contrast, the ExST participants were facing a 3 month follow-up period at thetime they responded to the post-test battery. This difference would seem to weigh in favor ofpositive post-treatment assessments for the SET treatment condition. Since most of the SEThusbands had successfully completed one course of treatment and were committed toadditional therapy at the time of post-test, the assessments done by SET families may wellhave included measures of hope and enthusiasm that were not available to ExSTparticipants. In this connection, it should be pointed out that since no follow-up informationregarding SET was generated by this study, the stability of the treatment effects of this newmodel of therapy remains unsubstantiated.180The second connection regarding the efficacy of SFT and research design addressesthe attention and measurement effects that are components of an outcome study of this type.At present, it is impossible to determine the role these factors played in the results. Onemust assume that some portion of the overall change variance for both treatments wasattributable to this aspect of research and presumably both treatments were equally affectedby them. The magnitude of improvements of clients cannot be accounted for by such thingsas Hawthorne effect or statistical regression to the mean. Nevertheless, measurementeffects must be included when considering the post-treatment outcome of both the SFT andExST treatments.In view of the methodological limitations in the implementation of SFT in thisresearch study, excitement regarding the efficacy of the treatment must be somewhattempered by caution. The results appear to support the position that SF1 is an effectivetreatment, however, until its efficacy is more fully evaluated a more definitive assertionwould be premature.The ExST treatment outcomes at post-treatment consistently showed highlystatistically significant and clinically relevant improvements across all levels of the ecologicalassessment for both fathers and mothers in this study. While it would be desirable tocontrast the improvement rates of the treatments in this study, with comparable studies inthe area of alcoholism treatment, such a comparison is hampered by the unavailability ofinvestigations with similar assessment strategies and instrument selection. Of the welldesigned studies reviewed earlier, only one study implemented the same measure of maritaladjustment that was used in this study. Zweben et al. (1988) reported significantimprovements on the DAS for both treatment conditons in their study. The significantimprovements in marital adjustment were based on group mean change values of 7-8 points(approximate effect size .36 to .42) on the DAS for both husbands and wives, The group181mean improvements of couples in this study are slightly higher and range from 14 to 16points on the DAS (effect size ranging from .87 for fathers and .79 for mothers).Additionally, couples in this study began treatment in states of considerably more maritaldistress than was the case in Zweben et al. (1988). While the mean scores of couples at pretreatment in this study were well within the distressed range ( = 88.86 for husbands, and I= 81.06 for wives), the couples in the Zweben et al. initiated treatment with considerablyless marital distress. In fact, on average both identified clients and spouses in Zweben et al.scored in the non-distressed range of the measure (DAS total score of 103.35 for identifiedpatients, and 104.45 for spouses). Accordingly, couples in the present study improvedsomewhat more in terms of marital adjustment than participants in Zweben et al. (1988).Additionally, the changes in this study were qualitatively different, since they involvedmoving from levels of marital distress to non-distressed ranges of scores. In the Zweben etal. study, the changes at the couples level were matters of increasing the level of maritaladjustment of couples who were non-distressed at the onset of treatment.Similar comparisons of the rates of improvements with other studies are impossible atthe present. Few of the stEdies have employed measures relevant to the individual andfamily domains of assessment, and none of the studies reviewed earlier employed the samemeasure of alcohol dependency. Consequently, the clinical meaning of the rates ofimprovement are best interpreted in light of the available normative information providedfor each instrument.While the clinical profile of the parents at pre-treatment was highly distressed, thesituation had improved substantively at post-treatment. The marker variable treatmentgains included: (1) a decrease in level of alcohol dependency from the extreme end of themoderate dependency range to a level well within the mild dependency range, (2) animprovement in intrapersonal psychological symptomology for both fathers and mothers thatwent from levels clearly within the psychiatric in-patient norms to tolerable levels within thenon-patient norms, (3) an increase in marital adjustment for both fathers and mothers that182improved from levels in the distressed range to levels within the normal range for fathersand slightly below the normal range for mothers, (4) an improvement in levels of familysatisfaction for both fathers and mothers that went from levels below the normal range tolevels well within the normal range. In view of these results, it is legitimate to assert thatExST has shown itself to be an effective form of treatment for fathers and mothers in theamelioration of alcohol dependency problems.The SFT and ExST treatment conditions implemented in this investigation did nothave any statistically significant effect on the eldest children’s assessments of familysatisfaction. In discussing this finding it is important to mention that the pre-treatment meanassessment of family satisfaction from the eldest child perspective fell within the normalrange before the treatment had commenced. Consequently, a ceiling effect may have playedsome role in the post-treatment outcomes. It should be recalled that the mean age of theseproject participants was in the teen years which suggests that the developmental focus of theeldest children was on separation from their families and this may have also influenced theresults. However, while the teenage concern of differentiation may have in part accountedfor why the eldest children’s family satisfaction rating did not change along with their parentsassessments, it is unlikely that this alone accounts for the lack of improvement. It wouldseem more reasonable to argue that the eldest children, having grown up in familiestroubled by the presence of parental alcohol dependency, were already hardened to changesin their families and suspicious of any shifts might have emerged as a consequence oftreatment. Accordingly, they would be unlikely to recognize or acknowledge any familialimprovement that had not withstood the test of time.A final point regarding the eldest children’s family satisfaction result is that thisfinding should not be interpreted as indicating the treatments had no effect on the children.Of all the instruments selected as marker variables, the family satisfaction index was theleast affected by the treatments for fathers and mothers. In addition, it is possible that183important changes in the eldest children that may have emerged as a result of thetreatments (e.g., self-esteem), were not identified because they were not measured in theecological assessment procedure used in this study.With respect to the many systematic differences between the two therapies which hadbeen built into the design of SF!’, this research appears to have shown that a number ofaspects intrinsic to the ExST model were not necessarily critical to bringing abouttherapeutic change in alcohol dependent clients. In particular, the inquiry has demonstratedthat the unstructured, individually tailored treatment approach of ExST was not superior tothe more structured uniformed procedures of SF!’. Accordingly, the notion that a treatmentwhich was flexible enough to address unique client concerns would be more effective than astructured treatment was not supported by this research.The experiential and symbolic orientation of ExST did not show itself to be moreeffective than the cognitive and behavioural focus of SFT. These dimensions of treatmentappear to have functioned equally effectively for the two therapies ,and it would seem thatboth perspectives are powerful means of bringing about therapeutic changes. The changeconstructs which underlie the two therapies of atypical responding and learning functionedequally well within treatment models. Thus, the here-and-now intense way of working whichcharacterized ExST was not any more effective than the there-and-then less intense learningformat which typified SF!’.The two treatments also differed in terms of the distance between therapists andclients assumed to be optimally therapeutic. The results of this study showed that the closeprofessionally intimate therapist and client relationship of ExST was no more curative thanthe more distant and less involved therapist and client relationship of SF!’. Consequently, itcan be understood that therapist distance may vary considerably from one treatment toanother without necessarily implying that one is superior over another.184In conclusion, the systematic between therapy model differences built into the designof this study have revealed that these features are not in and of themselves the crucialelements of transformation with alcohol dependent clients. Rather it would appear that themany shared aspects of the treatments detailed earlier in this section were the mostimportant features to the amounts of change reflected in the ecological assessment. This isnot to say that the between model differences were trivial components to the treatmentoutcomes. These differences are critical within treatment model elements which give themodels their own flavour or style. While such aesthetic aspects may not play a measurablerole in the pragmatics of between therapy treatment outcomes, they nonetheless areessential to the processes of change unique to each model.Second Research Hypothesis: Differential Efficacy of ExST-I and ExST-CThe second goal of the study was to evaluate the differential effects of the individualand couples formats of the ExST approach. There were competing ideas that lay beneaththis question. On the one hand, there was the view that ExST was believed to bring aboutsystemic change when conducted in an individual format. On the other hand, there was theperspective that the couples format would have the added benefit of allowing directtherapeutic involvement with the marital relationship, and therefore should perform in asuperior manner. The perspective that the couples format would be more powerful than theindividual format had some empirical support and accordingly, the research hypothesis wasframed in directional terms which favored the couples format. The directionality of thehypothesis was not supported by data analyses.It would appear that both treatment formats of ExST were equally effective. Again,the lack of differential treatment format effect must be understood in terms of by thestatistically significant pre-treatment/post-treatment changes that were noted in the data185analyses related to fathers and mothers. Both treatments were associated with highlysignificant eco-systemic change for both fathers and mothers. In addition, the treatmentgains established at post-treatment were found to be stable at the 3 month follow-up with nosignificant occasion difference between post-treatment and follow-up being identified. Itshould also be recalled that neither format appeared to have a measurable effect on theeldest children’s assessments of family satisfaction.These results would seem to question the contributing role often ascribed to non-alcohol abusing spouses in the persistence of alcohol dependency in their mates. Evenwithout receiving direct treatment the participating women in this study, whose husbandsreceived individual therapy, improved to a remarkable degree. This would seem to suggestthat in many cases, the non-alcoholic spouses’ difficulties are directly related to thehardships related to their partners’ abuse of alcohol. Accordingly, the results would seem tosupport the perspective offered by Zweben (1986) who argued that spousal problems aremore often than not the result of the alcoholic’s impaired behavior rather than the cause. Inany case, the present study established that therapy with alcoholics can lead to theamelioration of psychological symptomology, marital distress, and familial dissatisfaction inspouses not directly involved in treatment.The results related to the second hypothesis are different from the results of previousresearch in the area of treatment format (Foley et al., 1989; Jacobson et al., 1991; O’Leary &Beach, 1990). All previous research to date on treatment format had found evidence whichsupported the view that marital treatment formats had certain advantages over individualtreatment formats. The present study results stand in contrast to this perspective andindicate that there was no particular added benefit to receiving the couples form oftreatment. In interpreting this result it is germane to bear in mind that unlike the previousefforts, the individual treatment format implemented in this study was oriented systemically186in underlying theory. The individual formats employed in the earlier research efforts wereindividual treatments in both format and theoretical orientation. Accordingly, the individualtreatments of earlier research were not designed to address systemic issues in the clientsinterpersonal lives. In contrast, the individual format of ExST was meant to concern itselfwith the interpersonal contexts of the individual clients. This difference in underlyingtheoretical orientation would seem to best explain the differences between the findings ofthe previous research on treatment format and the present investigation.The results pertaining to this research question are particularly surprising withrespect to the wives whose husbands completed the individual form of ExST. Not only didthese women change as much across the three marker variables as their couples formatcounterparts at post-treatment, these gains (which might have been thought of as lessdurable) were also retained equally well at follow-up. It would seem that the therapeuticefforts with the fathers in the individual format had profound and lasting effects on themothers such that mothers’ assessments of themselves, their marriages and their familieswere markedly improved. This result is in keeping with the study by Sisson and Azrin (1986)which showed that people not specifically attending therapy could nonetheless experienceimportant improvements as a consequence of relationally focused interventions. Certainlythis treatment potential would seem to have been realized by the ExST-I mothers who didnot attend a single therapy session.The observation that the wives of the men who attended ExST-I changed in such apositive way is at odds with the first-order cybernetic view of alcohol dependency. Thehomeostatic assumption of first-order family therapy models asserts that symptoms play animportant role to the entire family system. As a result, the family was predictably seen toresists improvement in the symptomatic behaviour in an effort to preserve its sense ofidentity and equilibrium (Killorin & Olson, 1984). Such a dynamic would not appear tohave been operating in any important fashion for the fathers and mothers participating in187this study. To the contrary, the women who did not attend therapy appear to have enjoyedthe benefits of their husbands’ improvement and in fact, reported improvements inthemselves, their marriages and their families. Similarly, the husbands, who did not have theadvantage of having their wives present in treatment were nevertheless able to makesubstantial and lasting gains with respect to their drinking, themselves, their marriages andtheir families. Additionally, the inclusion of the alcoholics’ spouses in treatment did notprove to significantly add to the post-treatment gains made by the men who participated inthe couples treatment when compared to their counterparts involved in the individualtreatment format.The follow-up results would seem to lend further support to the view that bothtreatment formats of ExST performed equally well for fathers and mothers. In view of thisfinding, it would seem safe to assert that the treatment effects for both formats were bothcomparable at post-treatment and durable up to 3 months after the cessation of therapy.One would assume that if the kinds of systemic dynamics postulated by the first-order familymodels were at work, then the post-treatment gains of the individual treatment conditionwould have deteriorated substantially at follow-up. This of course was not the case in thepresent investigation. To the contrary, the individual treatment gains appeared to be asstable as the couple treatment condition.The result related to the second hypothesis support the view that systemic familytreatment can be successfully accomplished when only the alcohol dependent individual isseen in therapy. The benefits of individual treatment need not necessarily be vieweddisparagingly as only concerned with individual intrapsychic processes. Rather, the resultsfrom this study suggest that the improvements associated with the systemic treatment ofindividuals can generalize to spouses not attending treatment. This finding highlights theneed for a clear distinction between format and theory in the field of family therapy.188It was noted earlier that for many family therapists, systemic therapy means workingwith couples and families in conjoint treatment. This position essentially fuses format andtheory and the distinction between the two is lost. If the distinction between format andtheory is blurred, and the fact that both the individual and couples formats in this study weresystemically based is ignored, then the format of treatment would define the therapeuticprotocol. In this case, the findings of this research would appear to challenge the veryassumptions which led to the splitting of marriage and family therapy from the individuallyoriented mainstream of psychiatry and psychology. This is to say that the commonassumption in family therapy that working with larger systems (such as the couple) is a moreeffective and efficient way of treating problems in comparison to working with individualswas not supported by the empirical evidence of this study.Accordingly, a clear recognition of the difference between format and theoreticalorientation becomes key to the interpretation of the results of the present investigation.This study did not pit an individually oriented therapy against a systemically orientedtreatment. Rather, it established a comparison between the individual and couples formatsof a single systemically grounded model of therapy. Consequently, while the findings may beseen as relevant to the historical assumptions of the field, they are better understood ascontributing empirical support for the systemic treatment of individuals. The present studysuggests that family systems models that integrate individual processes into their theoreticalformulations can be as successfully employed with individuals as they are with couples in thetreatment of the myriad of family problems associated with alcohol dependency.Within Systems ComparisonsThe within systems analyses were congruent with the eco-systems analyses andaugmented the understanding of the breadth of improvements that had occurred in thisstudy. Pre-treatment/post-treatment within system analyses were undertaken for bothfathers and mothers as indicated by the earlier eco-systems analyses. The results of the189within systems analyses will be discussed with reference to normative data in order toconsider the significant findings within a clinically relevant frame.Within Alcohol SystemThe results from the within alcohol system analyses indicated that there was nosignificant differential treatment effect coupled with a highly significant level of changebetween pre-treatment and post-treatment scores. In fact, the alcoholics in this study whoreceived ExST-I, ExST-C and SFT had improved to a highly statistically significant degreeon 16 of the 17 dependent variables used in these analyses. Because there were nosignificant between treatment group differences, the results will be discussed in terms of theentire sample of participants.The alcoholics in this study dropped from their pre-treatment level of dependencythat was just shy of the severe dependency cut-off value in the moderate range, to a level ofmild dependency. This substantial improvement in dependency level indicates that animportant shift in the participants relationship to alcohol had occurred over the course oftreatment.The results from the IDS elaborates on this point. At pre-treatment the participatingfathers had mean scores on the 8 sub-scales that placed them in the high risk range for allbut one of the sub-scales which fell in the moderate category. At post-test the mean scoresfor alcoholics in this study were lowered to the moderate risk level across all 8 sub-scales.Furthermore, with respect to the norms based on people entering treatment for alcoholrelated problems, the mean scores for each sub-scale at post-treatment was found to bebelow the first percentile across all variables.190The SCQ findings further support the view that the alcoholic participants relationshipto alcohol had changed in important ways. On 7 of the 8 variables included in the analyses,the mean scores for the fathers were significantly different from pre-treatment at post-treatment. While the mean total SCQ score at pre-treatment indicated that the alcoholic’swere 68.31% (43rd percentile) confident they could resist drinking heavily in the array ofsituations covered in the instrument, the mean total assessment of confidence value hadincreased to 84.25% (65th percentile) at post-treatment.Within Intrapersonal SystemThe within intrapersonal system analyses added to the understanding of how thefathers and mothers had individually changed. The significant changes in terms of theparents intrapersonal measures showed that while there was no statistically significantdifference between the three treatment groups, the fathers and mothers had improvedconsiderably by post-treatment. Importantly, the post-treatment improvements representedshifts in individual symptoms that moved from considerably distressed magnitudes to levelsthat were for the most part within normal ranges of behaviour. In a few instances, thedecreased levels of symptomology require some additional comment.The BDI scores indicated that at pre-treatment both husbands and wives weremoderately depressed, however by post-treatment only the women whose husbands hadattended SFT remained in the moderate range of depression. All fathers’ and ExSTmothers’ scores were within the asymptomatic range at post-test. This result whileinteresting, is best seen in light of the pattern of responding on the depression sub-scale onthe SCL-90-R. On this measure of depression, the participants had reported pre-treatmentlevels of depression that were around the average values reported by in-patient psychiatricpatients. By post-treatment both husbands and wives in all treatment groups had improvedsignificantly, however the mean values for both husbands and wives remained slightly above191one standard deviation from the mean of the non-patient norm group. In view of the findingon the SCL-90-R depression sub-scale, it would seem necessary to temper the BDI resultswith a measure of caution and assert that the symptom of depression remained somewhat ofa clinical concern for both husbands and wives connected to the study.Two other symptoms contained in the within intrapersonal level of assessment seemnoteworthy in that the scores across treatments uniformly hovered around one standarddeviation from the mean of the non-patient norm group. Fathers’ and mothers’ psychotismratings remained inflated at post-treatment indicating considerable levels of interpersonalalienation (Derogatis, 1983). In addition, both husbands’ and wives’ levels of anger asmeasured by the hostility sub-scale of the SCL-90-R remained at levels higher than average.Thus, some areas of clinical concern remain within the intrapersonal domain for both fathersand mothers.Notwithstanding the areas where further therapeutic effort was indicated, the chargesat the individual level of assessment were dramatic for all treatment groups and reflected areturn to near normal levels of individual functioning for the participants in this study atpost-treatment. This finding further supports the view that all of the treatments under studyhave proven themselves to be effective in moderating the difficulties associated with alcoholdependency.Within Couples SystemThe within couples pre-treatment and post-treatment analyses revealed that whilethere were no significant differential treatment effects, there were significant improvementsin a variety of facets of the relationship from fathers’, mothers’ and couples’ perspectives.Both husbands and wives improved significantly on all of the areas measured by theDAS including dyadic consensus satisfaction, affectional expression and cohesion.Interpretation of the cohesion and affection sub-scales of this instrument should be viewedwith a degree of caution in view of the low levels of internal consistency found in the first192phase of data analysis. Nonetheless, the couples relationships seem to have been greatlyaided by the treatments they received. This is particularly true of the measure of dyadiccohesion which reached normal levels for all treatment groups for both husbands and wives.Despite the considerable change which was reported for all sub-scales by both members ofthe couple, the mean values on the sub-scales related to dyadic consensus, satisfaction andaffectional expression suggested that further couples enrichment was required. The DAStotal score means for the husbands and wives in all treatments were scattered about thevalue of 100 which has been suggested as the cut-off score separating distressed from non-distressed couples. Thus, the average couple in the study appears to have been beginning tofunction in an adjusted fashion at post-treatment.The improvements in the marital relationships were also reflected by the ACmeasure. In keeping with past studies (e.g., Margolin et al., 1983), the assessments ofdesired change and perceived change from both husbands’ and wives’ perspectives droppedsignificantly. It would appear that the marital relationship had improved by post-treatmentand both partners recognized their own decrease in wishes for change and their partners’decrease in desire for improvement. Clearly the levels of desired and perceived changeindicated that more work on the marital relationship was indicated, however considerablyless change was being sought by the couple at post-treatment when compared to the pretreatment levels.The perceptual accuracy sub-scale results showed that both fathers’ and mothers’levels of agreement with their spouses at post-treatment had significantly decreased. At thesame time, the disagreement estimates for fathers also decreased at post-treatment.However, mothers’ disagreement scores did not change over the course of treatment. Withrespect to the perceptual accuracy ratios of the husbands and wives, (which reflects theextent to which spouses are aware of which specific behaviors are meant to be changed andin which direction the change is desired), the couples in this study scored relatively low at193pre-treatment and did not change in any significant way at post-treatment. This finding issomewhat different to the findings of Margolin et al. (1983) who reported a positiverelationship between marital distress and increased levels of perceptual accuracy. In thepresent study, increases in marital adjustment were not found to be accompanied by aconcomitant decrease in perceptual accuracy by either spouse. In this connection, it wouldseem most likely that two factors affected the perceptual accuracy scores. Firstly, inMargolin et al. (1983) the men in the maritally distressed relationships were found to havenear perfect perceptual accuracy scores, while the alcoholics in the present study did notpresent with this degree of understanding at pre-treatment. Secondly, in the present study,the treatments may well have played mediating roles in the perceptual accuracy of thecouples such that the accuracy with which the couples perceived the others desires forchange did not deteriorate significantly over the course of therapy.The results from the couples variables of the AC were consistent with the fathers’ andmothers’ AC results in as much as they showed no significant differential treatment effectscoupled with highly significant pre-treatment/post-treatment change. The univariateanalyses were similar to the results related to fathers and mothers, with couples total desiredchange, total perceived change, and total agreement decreasing significantly over the courseof the therapeutic interventions. The total disagreement result was in keeping with thefathers’ perspectives and in contrast with the mothers’ perspectives in as much as it too haddecreased to a statistically significantly degree after treatment at post-test. It would seemthat the weight of the changes in fathers’ disagreement estimates were substantial enough tocarry the couples disagreement values over the line of statistical significance.Taken together, the couples within systems analyses match the marker variablehypotheses testing findings and elaborate on the breadth of change that was reported bycouples involved in this study. All the aspects of marital adjustment measured in this study194improved and the need for couple change decreased. In this way, the couples whocompleted the study were able to make important steps towards more harmonious andricher marital relationships as a consequence of the treatments offered.Within Family SystemThe fathers’ and mothers’ within family system analyses were congruent with thepattern in previous within systems analyses, finding no differential treatment effects pairedwith highly significant pre-treatment/post-treatment change. The univariate tests on eachvariable revealed that both fathers’ and mothers’ perspectives had shifted substantially withrespect to their families by post-treatment.The family environments had clearly improved by post-treatment for both parents onmany of the sub-scales employed in the study. The levels of cohesion, expressiveness, andorganization had all increased significantly, while the levels of conflict had significantlydecreased in the families according to both fathers and mothers. Importantly the averagevalue of scores for parents in this study now matched the norms for normal families on thecohesion, expressiveness, and conflict sub-scales. In this study, the independence sub-scale’sinternal consistency estimates was somewhat low, suggesting tentativeness of viewing thisresult. In any case, the mothers’ independence rating had significantly improved andnormalized. However, fathers’ assessment of family independence, though improvedsomewhat, had not changed significantly and remained below the normal family contrastnorm. This result would seem to be important clinically as it suggested that mothers’ senseof family autonomy had increased by post-test, leaving the families under her direction lesssusceptible to changes in fathers’ condition. Neither parents’ assessments of control in thefamily changed significantly between pre-treatment and post-treatment but scores on thissub-scale had been in the normal range prior to treatment. The low level of reliability forthis sub-scale shown in the first phase of data analysis, suggest this result should be viewedwith some caution.195With respect to parents’ assessments of the adaptability and cohesiveness of thefamilies, the results of fathers’ and mothers’ analyses complemented one another. Whilefathers’ assessments of family cohesiveness and unity increased significantly, mothers’reported significant increases in family adaptability. It would appear that as the fathers’sense of family closeness grew stronger, the mothers’ experience of the family’s ability toaccommodate to change also increased.The analyses of the parents’ assessments of their satisfaction with the level of familycohesiveness and adaptability showed that the families had significantly improved from theperspectives of the fathers and mothers. The average post-treatment values on thesevariables had increased substantially from the pre-treatment levels of the 7th and 13thpercentile on cohesiveness and adaptability for fathers, and 4th and 7th percentiles on thesame sub-scales for mothers. The results suggested that both parents at post-treatment sawtheir families as somewhat low on the cohesion sub-scale (scoring in the 28th percentile), butaverage in terms of adaptability (scoring near the 50th percentile) at post-treatment. Both ofthese scores were within the normal range of scoring.The within family system results expanded the understanding of the ways in which thefamily had been affected by the treatments from the parents’ points of view. Clearly, theparents believed that the therapies had helped their families in many important ways.However, the results also suggested that there remained some areas of concern in thefamilies that could benefit from further therapeutic attention.Within Systems SummaryThe within systems analyses were performed on the systems of alcohol, intrapersonal,couple and family from both fathers’ and mothers’ perspectives on pre-treatment/posttreatment data as indicated by the eco-systemic analyses. The within systems analysesexpanded upon the hypotheses testing findings, sharpening the scope of analyses upon each196level of assessment. Taken together, the within system results demonstrated the far reachingeffects of the treatments and detailed the many areas of the participants’ lives which hadbeen improved as a consequence of receiving one of the therapies offered. The withinsystem results were consistent with the pre-treatment/post-treatment eco-systemic approachtaken in the hypotheses testing phase of data analysis and can be seen as lending furtherweight to the conclusions regarding the two central research hypotheses.Therapeutic ValidationThe results from the analyses included in the therapeutic validation section tested forpossible moderating variables in the treatment outcomes including clinic site, therapistgender and therapist. There were no statistically significant moderating variables found inthe results. Consequently, it can be asserted that the treatments were equally effective inboth the urban and the rural clinical contexts. The treatments performed comparablyregardless of the gender of the therapist providing the treatment. In addition, no therapistwas identified as being associated with significantly better or worse treatment effects.Limitations of the StudyIn the design of this study, the variable of therapist was partially nested and crossed.Therapists were nested with respect to treatment (SFT and ExST) and crossed in connectionto ExST treatment format (ExST-I and ExST-C). Consequently, it is impossible to separatethe effects of therapists from the effects of treatment in the SFT and ExST contrasts(Jacobson, 1985; Kasdin, 1986). This limitation of design was a consequence of the changein research designs noted earlier that resulted in the generation and implementation of SF1’in the place of a wait-list control group. By the time that this decision was made,commitments to the ExST therapists had been made that precluded introducing additionrequirements on the practitioners (as would have been necessary had they also delivered theSET treatment). In any case, the benefits of having the same therapists perform both the197ExST and SFT treatments would most certainly have been complicated by the potential forthe bleeding through of ExST procedures into the SFT treatment protocol.The limitation of nesting therapists within treatments was softened by the relativelylarge numbers of therapist per treatment (5 ExST therapists and 7 SFT therapists). Therandom assignment of this number of therapists with clients minimized the likelihood ofmean differential therapist skills between treatments affecting the results. In addition, inview of the finding that no single therapist, regardless of treatment, was associated withgreater or lesser treatment effects, the limitation of nesting would seem to be less of aconcern. The crossing of therapists in the ExST format conditions was a strong designfeature of the study. Because therapists were crossed, the confound of therapist withtreatment was eliminated in the ExST format comparisons. Since the ExST therapists wereperforming the same therapy in both formats, there was little concern that personal biases ofthe clinicians influenced the results.A second limitation to the study is the reliance on self-report measures for data. Ithas been argued that combining both self-report and direct observational sources of data is apreferred strategy in marital and family therapy research (Cline, Jackson, Klein, Mejia &Turner, 1987; Gurman & Kniskern 1981; Wynne, 1988). However, direct observationalsources of data have not been shown to be particularly sensitive in identifying relationshipchanges (Jacobson, Follette, & Elwood, 1984; Jacobson et al., 1991). In addition, directobservational data have been found to be very expensive to generate. Accordingly, theutility of adding direct observational data was questionable and unjustifiable from amonetary point of view.In connection to the issue of instrumentation, the study is limited with respect to whatcan be said about the treatment effects upon the eldest children. Without having a more198complete ecological assessment of the eldest children, it is not legitimate to conclude thatthe treatments had no effect upon them. Rather the conclusions must be limited to assertingthat in terms of their assessment of family satisfaction, the treatments did not have ameasurable impact upon the eldest children.Limitations of Field-based ResearchSome limits to the research arose as a consequence to the field-based nature of thestudy. Of these limits, the loss of the wait-list control group and the implementation of thecomparison treatment condition was the most impactful. As mentioned earlier, this changelimited the clarity of the results regarding the efficacy of ExST since the changes associatedwith the repeated testing, maturation, and regression to the mean could not be controlled.This change in design was necessary in order for the research to accommodate to the needsof the clinics and their clients. However, the implementation of SFT also broughtlimitations. As discussed earlier, the situation of the post-treatment responding for SFTparticipants was not the same as it was for the ExST participants. An ideal comparativetreatment design would have seen the SFT participants involved in the follow-up portion ofthe study. Again due to the clinical circumstances of the investigation, this option was notavailable. Because SFT was an entirely experimental treatment with no history ofimplementation, the researcher was left with no ethical option but to exclude the SFTparticipants from the follow-up and to offer additional treatment at the conclusion of thetreatment. Judging by the percentage of those in SFT who did continue with additionaltherapy, this precaution appears to have been warranted. However, the design unfortunatelylimits the results with respect to what might have been said about the SFT treatment had thesame pattern of results been found and the SFT participants been involved in the 3 monthfollow-up.199Other factors connected to the study being conducted in the field also affected theinvestigation. For example, some participants were lost and other treatments were greatlyaffected when therapist were ethically compelled to contact the child protection authoritiesdue to information revealed over the course of therapy. In some situations, the participantshad to be dropped from the study when the constraints of the study in terms of treatmentformat were clearly at odds with the needs of the participants (e.g., suicidal children, spousesneed for therapeutic care). In others, participants were dropped from the study due toincarceration for wife assault, hospitalization, and suicide. In this way, the study wasconstrained by the legitimate needs of the agencies involved and the day-to-day operation ofdrug and alcohol clinics. Nonetheless, the limits of field based research studies are balancedby the clinical applicability of such investigations. Issues regarding the generalizability of theresearch are presented in the next section.GeneralizabilityThere are several considerations of note in connection to the generalizability of thestudy. The present research was a field based experiment that was conducted in twooutpatient clinical sites. The participants were paid volunteers who were randomly assignedto both treatments and therapists. The therapists were regular therapists employed by theparticipating centers. No treatment effects regarding clinic site, therapist or therapist genderwere evident and accordingly, the results can be generalized to both urban and rural outpatient clinics and therapists trained in the treatments that were studied.The participating families came to the study from two main sources including normalclinic intake procedures, and responses to the public attention given to TARP by the localmedia. In addition, the participating families in this study had to meet the inclusioncharacteristics of the study which included an alcoholic father who had consumed alcohol inthe last 3 months and a non-alcohol or drug abusing mother who were maritally distressed200and had at least one child living at home with them. As such, the study participantsrepresent a particular sub-set of the total alcohol dependent population and generalizationsbeyond this sub-set of alcoholics and their families are not justified.It has been noted by Kasdin (1986b), that even when volunteers and clinicalpopulations share the same presenting problems, generalizing from research subjects toclinical populations is an open question which must be established empirically because ofpotential differences related to volunteerism. Therefore, generalizations of this studybeyond the volunteer participants to the larger alcoholic populations should be restrictedand extended with a modicum of caution.The loss of the participants who were extremely dissatisfied with their families priorto treatment completion is another factor of concern regarding the generalizability of thestudy. Although familial dissatisfaction was not an inclusion characteristic of the study, it isnonetheless true to state that as a group, the participants in this research were verydissatisfied with family life. In light of the observation that most extremely dissatisfiedfamilies dropped out of treatment prior to its completion, the findings of the investigationregarding treatment outcomes should not be extended to the most severely distressedfamilies with alcoholic dependent fathers.The final consideration with regards to external validity is the extensive testing thatwas an essential feature of this investigation. The questionnaires included in the ecologicalassessment that were repeatedly administered to participants cannot be considered astherapeutically inert. Since all treatment groups were exposed to equal amounts of testing, itis valid to assert that testing was not a factor contributing to the no differential treatmentresult. However, it is not possible to know how much a role the testing played incontributing to the extensive improvements which were found for all treatment groups. In201addition, it is impossible to assess the role that artifacts of research of this kind includingHawthorne effects and statistical regression to the mean may have played in the overallimprovements found for the treatments. Thus it would seem best to assume that the effectsof testing, measurement, research participation and treatment are compounded in thefindings of significant changes for fathers and mothers in this study. At the same time, theconsistency and magnitude of the improvements across instruments, supports the view thatthe treatments were nevertheless effective in helping participants in a variety of ways.Accordingly, the outcome results of the study may be extended to on untested populationwith the proviso that treatment gains may be somewhat reduced in the absence of repeatedtesting.ConclusionsIn this study, a multidimensional ecological assessment approach for measuringfathers, mothers and eldest children was taken to test for the effects of three treatments.The scope of assessment used in this investigation was considerably wider than typicallyemployed in alcohol treatment outcome studies. This approach to measurement has beendemonstrated as an effective method of identifying an array of important changes connectedto treatment that would otherwise have gone undetected. Accordingly, it would seemdesirable for future alcohol treatment outcome studies to adopt a similar measurementstrategy.The ExST model was developed in response to the practical needs of clinicians in thefield and the call for therapeutic innovation in the area of alcohol dependency (Institute ofMedicine, 1992; Jacobson, Munroe, & Schmaling, 1989; Miller & Hester, 1986; Nathan &Skinstad, 1987). The present research contributes to the clinical body of knowledge bycarefully testing the empirical efficacy of this integrative treatment. This investigation hasshown that the ExST treatment approach was effective in generating an array ofimprovements for the fathers and mothers that completed the study. This research suggest202that ExST can be added to the list of therapeutic approaches for alcoholism that have anempirically based claim to efficacy.The ExST model did not prove to be more effective than SF!’. The SF!’ model oftherapy which was designed as a quasi treatment/control group exceeded expectations andappears to have performed as well as the ExST treatment. Due to methodologicallimitations, the clarity of the efficacy of SFT remains somewhat clouded. Nonetheless, SF!’was associated with highly significant post-treatment improvements in alcohol dependency,intrapersonal functioning, marital adjustment and family satisfaction and has shown promiseas an effective form of treatment for alcohol dependency.The present study is an important response to the need for the evaluation ofsystemically oriented therapies that was issued by McCrady (1989), Jacobson et al. (1989)and O’Farrell (1992). These authors rightly noted that the efficacy of behavioral maritaltherapy in the treatment of alcohol dependency had been generalized to untested systemicmodels of marital. This investigation has established empirical support for ExST, a second-ordered systemic therapy in the treatment of alcohol dependency, and consequently, addssome credibility to the use of systemically based treatments in working with alcoholism.Finally, the study has contributed to the knowledge base of psychotherapeutic formatresearch. Unlike previous investigations, this research has shown the performance of theindividual and couples formats of treatment to be comparable. The systemic orientation ofthe individual treatment implemented in this study differs from the individually orientedtreatments used in previous format research. This study has established the individual formof a systemically oriented treatment to be a valuable means of perturbing eco-systemicchanges that are comparable to the changes brought about by a couples treatment format.203In the final analysis, this study has been about finding ways to better help familiesnegotiate the difficult road of alcohol recovery. Each of the participants could well haveended up as another member of the skid-row community. However, they have chosen adifferent avenue. Hidden within the many statistics, tables, and graphs are the courageousstories of people struggling to find better lives for themselves. Without their commitment tohealth, it is unlikely that any therapy could succeed. The contribution to science made bythe participants in this study has been a deeply personal one and it seems only fitting to endthe investigation with an acknowledgment of that fact.Future ResearchThe first direction of future research would seem to point to an additional follow-upof participants in this investigation at a 2-4 year time interval. Such a study would beessential in testing for the long term effects of the treatments and would be useful incontinuing to probe for differential treatment effects that might become evident much laterafter the termination of therapy.The second line of inquiry that is called for involves a more detailed look at the datain an effort to unearth whether or not particular client characteristics may have beenpredictive of how the participants faired in treatment. By sub-dividing the sample in termsof those who improved, deteriorated or stayed the same, the wealth of informationgenerated by the ecological assessment might serve to help illuminate factors useful inmatching client needs and therapeutic modality.In view of the finding that the eldest children’s assessments of their families wasunaffected by the treatments, an additional outcome study that included a family therapyformat of ExST would be valuable. Very few studies have been conducted using family204therapy for problems of adult alcohol dependency and such a study would appear to be alogical direction in which to proceed.Additional study is indicated to establish a more solid empirical base for SFT and toexplore the utility of this treatment. Ideally, a Solomon four-group design study could beconducted which would control for testing factors as well as other threats to generalizability.In terms of the ExST model, two lines of further efficacy study are called for. Firstly,it would be helpful to contrast ExST and another form of marital or family therapy. In thisconnection, behavioural marital therapy would seem to be the contrast treatment of choice,since it has already established a strong treatment track record with couples struggling withalcohol dependency. Secondly, it would seem important to take ExST out of the alcoholtreatment field and to test it with other clinical populations such as depression. Such a lineof work would be helpful in establishing whether or not the treatment efficacy of ExST canbe generalized to other clinical problems.Finally, a most promising direction of inquiry is the undertaking of process research.This study has established that the individual and couples formats of ExST were effective inbringing about eco-systemic improvements. However, further research is required to betterunderstand the means by which these changes come about. The tape recordings of eachsession of treatment made over the course of this investigation provide a rich source ofprocess related data waiting to be explored more completely.205REFERENCESAbbott, D. (1976). The effects of open forum family counseling on perceived familyenvironments and on behaviour change of the child. (Doctoral dissertation, BrighamYoung University, 1975). Dissertation Abstracts International,, 8335A.Alden, L. E. (1988). Behavioral self-management controlled-drinking strategies in a contextof secondary presentation. Journal of Consulting and Clinical Psychology, .5, 280-286.Allen, R. E. (1990). The concise Oxford dictionary of current English - 8th edition. Oxford:Oxford University Press.American Psychiatric Association (1987). Diagnostic and statistical manual of mentaldisorders (3rd ed., rev.). Washington, D.C.: Author.American Psychiatric Association (1980). Diagnostic and statistical manual of mentaldisorders (3rd ed.). Washington, DC: Author.Amir, M., & Eldar, P. (1978). An experiment in the treatment of alcoholics in Israel. DrugForum, .7, 105-119.Anderson, H., & Goolish, H. (1988). Human systems as linguistic systems: Preliminary andevolving ideas about implications for clinical theory. Family Process, .7, 371-393.Annis, H. M. (1982). Inventory of Drinking Situations. Toronto: Addiction ResearchFoundation of Ontario.Annis, H. M. (1986). A relapse prevention model for treatment of alcoholics. In W. R.Miller & N. Heather (Eds.), Treating addictive behaviours: Process of change (pp.407-433). New York: Plenum Press.Annis, H. M., & Graham, J. M. (1988). Situational Confidence questionnaire (SCO): Usersguide. Toronto: Addiction Research Foundation of Ontario.Annis, H. M., Graham, J. M., & Davis, C. S. (1987). Inventory of Drinking Situations (IDS):User’s guide. Toronto: Addiction Research Foundation of Ontario.Antill, J. K., & Cotton, S. (1982). Spanier’s Dyadic Adjustment Scale: Some confirmatoryanalyses. Australian Psychologist, .12, 181-189.Auerswald, E. H. (1985). Thinking about thinking in family therapy. Family Process, 24, 1-12.Azrin, N. H. (1976). Improvements in the community-reinforcement approach toalcoholism. Behavior. Research and Therapy, j4, 339-348.Azrin, N. H., Naster, B. J., & Jones, R. (1973). Reciprocity counseling: A rapid learning-based procedure for marital counseling. Behavior. Research, and Therapy, jj, 365-382.206Bader, E. (1976). Redecisions in family therapy: A study of change in an intensive familytherapy workshop. (Doctoral dissertation, California School of ProfessionalPsychology, 1976). Dissertation Abstracts International, .Z, 249 lB.Bailey, M. B. (1961). Alcoholism and marriage: A review of research and professionalliterature, quarterly Journal of Studies on Alcohol, 22, 8 1-97.Banathy, B. H. (1987). The characteristics and acquisition of evolutionary competence.World Futures,2, 123-144.Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.Psychological Review,.4, 191-214.Barlow, D. H., Mavisskalin, M., & Hay, L. R. (1981). Couples treatment of agoraphobia:Changes in marital satisfaction. Behaviour Research and Therapy, j9, 245-256.Barry, K. L., & Fleming, M. F. (1990). Family cohesion, expressiveness and conflict inalcoholic families. British Journal of Addiction,., 81-87.Bateson, G. (1979). Mind and nature: A necessary unity. New York: Bantam Books.Baucom, D. H. (1982). A comparison of behavioural contracting and problemsolving/communication training in behavioral marital therapy. Behavior Therapy, i,162-174.Baucom, D. H., & Hoffman, J. A. (1988). The effectiveness of marital therapy: Currentstatus and application to the clinical setting. In N. S. Jacobson & A. S. Gurman(Eds.), Clinical handbook of marital therapy (pp. 599-620). New York: GuilfordPress.Beach, S. R. H., & O’Leary, K. D. (1986). The treatment of depression occurring in thecontext of marital discord. Behavior Therapy, ,fl, 43-49.Beck, A. T., & Beamesderfer, A. (1974). Assess of depression: The Depression Inventory.In P. Pichot (Ed.), Psychological measurements in psychopharmacology, modernproblems in pharmacopsvchiatrv (Vol. 7) (pp. 15 1-169). Basel, Switzerland: Karger.Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression.New York: Guilford Press.Beck, A. T., & Steer, R. A. (1984). Internal consistencies of the original and revised BeckDepression Inventory. Journal of Clinical Psychology, .4., 1365-1367.Beck, A., & Steer, R. A. (1987). Beck Depression Inventory manual. New York: ThePsychological Corporation. Harcourt, Brace, Jovanovich, Inc.Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the BeckDepression Inventory: Twenty-five years of evaluation. Clinical Psychology Review,., 77-100.Beck, A., Ward, C. Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory formeasuring depression. Archives in General Psychiatry, .4, 55-63.Becker, J. V., & Miller, P. M. (1976). Verbal and nonverbal marital interaction patterns ofalcoholics and nonalcoholics. Journal of Studies on Alcohol, .j, 1616-1624.207Becvar, D. S., & Becvar, R. J. (1988). Family therapy: A systemic integration. NeadhamHeights, MA: Allyn and Bacon.Bednar, R. L., Burlingame, G. M., & Masters, K. S. (1988). Systems of family treatment:Substance or semantics? Annual Review of Psychology, .39, 401-434.Bennett, L. A., Wolin, S. J., Reiss, D., & Teitelbaum, M. A. (1987). Couples at risk fortransmission of alcoholism: Protective influences. Family Process,2, 111-129.Bennun, I. (1985). Two approaches to family therapy with alcoholics: Problem-solving andsystemic therapy. Journal of Substance Abuse Treatment, 2, 19-26.Bennun, I. (1986). Evaluating family therapy: A comparison of the Milan and ProblemSolving approaches. Journal of Family Therapy, ., 225-242.Billings, A. G., Kessler, M., Gomberg, C. A., & Weiner, S. (1979). Marital conflictresolution of alcoholic and nonalcoholic couples during drinking and nondrinkingsessions. Journal of Studies on Alcohol,.4Q, 183-195.Billings, A. G., & Moos, R. H. (1982). Family environments and adoption: A clinically-applicable topology. American Journal of Family Therapy, IQ, 213-237.Billings, A. G., & Moos, R. H. (1983). Psychosocial processes of recovery among alcoholicsand their families: Implications for clinicians and program evaluators. AddictiveBehaviors, ., 205-218.Blishen, B. R., Carrol, W. K., & Moore, C. (1982). The 1981 socioeconomic index foroccupations in Canada. Canadian Review of Sociology and Anthropology, 24, 465-488.Birchler, G. R., & Webb, L. J. (1977). Discriminating interaction behaviors in happy andunhappy marriages. Journal of Consulting and Clinical Psychology, .4, 494-495.Bogdan, J. L. (1984). Family organization as an ecology of ideas: An alternative to thereification of family systems. Family Process, 2.3, 375-388.Bonk, J. (1984). Perceptions of psychodynamics during a transition period as reported byfamilies affected by alcoholism. Unpublished doctoral dissertation. University ofArizona, Tucson, Arizona.Bowen, M. (1974). A family systems approach to alcoholism. Addictions, .Zj, 28-39.Bowen, M. (1976). Theory in practice of psychotherapy. In P. J. Gurein (Ed.), Familytherapy: Theory and practice (pp. 42-90). New York: Gardner Press.Bowers, T. G., & Al-Redha, M. (1990). A comparison of outcome with group/marital andstandard/individual therapies with alcoholics. Journal of Studies on Alcohol, 5j, 301-309.Boyd, J. H., Weissman, M. M., Thompson, W. D., & Myers, J. K. (1983). Differentdefinitions of alcoholism: I. Impact of seven definitions of prevalence rates in acommunity survey. American Journal of Psychiatry, j4, 1309-1313.208Bradshaw, J. (1988). Bradshaw on: The family. Deerfield Beach, FL: HealthCommunications Incorporated.Broderick, C. B., & Schrader, S. S. (1981). The history of professional marriage and familytherapy. In A. S. Gurman and D. P. Kniskern (Eds.), Handbook of family therapy(pp. 5-35). New York: Brunner/Mazel.Bromet, E., & Moos, R. (1977). Environmental resources and the post-treatmentfunctioning of alcoholic patients. Journal of Health and Social Behavior, j, 326-335.Bronfenbrenner, U. (1977). Toward an experimental ecology of human development.American Psychologist, .2, 513-53 1.Bronfenbrenner, U. (1979). The ecology of human development: Experiments by natureand design. Cambridge, MA: Harvard University Press.Brophy, C. J., Norell, N. K., & Kiluk, D. J. (1988). An examination of the factor structureand convergent and discriminant validity of the SCL-90-R in an outpatient clinicpopulation. Journal of Personality Assessment, .2, 334-340.Burger, A. L., & Jacobson, N. S. (1979). The relationship between sex role characteristics,couple satisfaction, and problem solving skills. American Journal of Family Therapy,2, 52-61.Burton, 0. (1962). Group counseling with alcoholic husbands and their wives. Marriage andFamily Living, 24, 56-61.Burton, G., & Kaplan, H. M. (1968). Marriage counseling with alcoholics and their spouses -II: The correlation of excessive drinking behavior with family pathology and socialdeterioration. British Journal of Addiction, ., 161-170.Burton, G., Kaplan, H. M., & Hudd, E. H. (1968). Marriage counseling with alcoholics andtheir spouses - I: A critic of the methodology of a follow-up study. British Journal ofAddictions,, 15 1-160.Cadogan, D. A. (1973). Marital group therapy in the treatment of alcoholism. OuarterlyJournal of Studies on Alcohol, .34, 1187-1194.Cahalan, D. (1970). Problem drinkers: A national survey. San Francisco: Jossey-Bass.Camacho-Salinas, R. L., O’Farrell, T. J., Jones, W. C., & Cutter, H. S. G. (1984). Servicesfor the families of alcoholics: A national survey of Veterans Administrationalcoholism treatment programs. Unpublished manuscript.Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs forresearch. Chicago: Rand McNally.Caron, W., & Olson, D. H. (1984). Family satisfaction and perceived-ideal discrepancy ofFACES II. St. Paul, MN: Family Social Science, University of Minnesota.Cecchin, 0., Lane, G., & Ray, W. A. (1993). From strategizing to nonintervention: Towardirreverence in systemic practice. Journal of Marital and Family Therapy, j9, 125-136.209Chapman, P. L H., & Huygens, I. (1988). An evaluation of three treatment programmes foralcoholism: An experimental study with 6- and 8-month follow-ups. British Journalof Addiction, , 67-8 1.Chiavzzi, E. J. (1991). Preventing relapse in the addictions: A biopsychosocial approach.New York: Pergamon Press.Chiles, J. A., Stauss, F. S., & Benjamin, L. S. (1980). Marital conflict and sexual dysfunctionin alcoholic and non-alcoholic couples. British Journal of Psychiatry,j, 266-273.Christensen, B. (1977). A family systems treatment program for families of delinquentadolescent boys. (Doctoral dissertation, Brigham Young University, 1976).Dissertation Abstracts International, 7, 6092A.Christie, N., & Bruun, K. (1969). Alcohol problems: The conceptual framework.International Congress on Alcohol and Alcoholism, Lect. 28th, 65-73.Collins, R. L. (1990). Family treatment of alcohol abuse: Behavioral and systemsperspectives. In R. L. Collins, K. E. Leonard, & J. S. Searles (Eds.), Alcohol and thefamily: Research and clinical perspectives (pp. 285-308). New York: Guilford.Conger, R. D. (1981). The assessment of dysfunctional family systems. In B. B. Lakey & A.E. Kazdin (Eds.), Advances in clinical child psychology (Vol. 4, pp. 199-242). NewYork: Plenum Press.Connors, G. J., O’Farrell, T. J., & Pelcovits, M. A. (1988). Drinking outcome expectanciesamong male alcoholics during relapse situations. British Journal of Addiction, .,561-566.Connors, G. J., Tarbox, A. T., & Faillance, L. A. (1992). Achieving and maintaining gainsamong problem drinkers: Process and outcome results. Behavior Therapy, 2, 449-474.Constantinopole, A. (1969). An Ericksonian measure of personality development in collegestudents. Developmental Psychology, j, 357-372.Corder, B. F., Corder, R. F., & Laidlaw, N. D. (1972). An intensive treatment program foralcoholics and their wives. Ouarterly Journal of Studies on Alcohol,, 1144-1146.Crane, D. R., & Meed, D. E. (1980). The Marital Status Inventory: Some preliminary dataon an instrument to measure marital dissolution potential. The American Journal ofFamily Therapy, ., 31-35.Cyr, J. J., Doxey, N. C., & Vigna, C. M. (19881). Factorial composition of the SCL-90-R.Journal of Social Behavior and Personality, ., 245-252.Davenpost, Y. B., & Mathiasen, E. H. (1988). Couples psychotherapy group: Treatment ofthe married alcoholic. Group, 32, 67-75.Davidson, R. J., Bunting, B., & Raistrick, D. (1990). The homogeneity of the alcoholdependence syndrome: A factorial analysis of the SADD questionnaire. BritishJournal of Addiction, .S4, 907-915.Davidson, R. J., & Raistrick, D. (1986). The validity of the Short Alcohol Dependence Data(SADD) questionnaire. British Journal of Addiction, .j, 2 17-222.210Davis. D. I. (1987). Alcoholism treatment: An integrative family and individual approach.New York: Gardner Press.Davis, D. I., Berenson, D., Steinglass, P., & Davis, S. (1974). The adaptive consequences ofdrinking. Psychiatry, 3..7, 209-215.Dell, P. F. (1985). Understanding Bateson and Maturana: Toward a biological foundationfor the social sciences. Journal of Marital and Family Therapy, .fl, 1-20.Derogatis, L. R. (1983). SCL-90-R: Administrative, scoring and procedures manual - II forthe revised version and other instruments of psychopathology rating scale series.Towson, MD: Clinical Psychometric Research.Derogatis, L. R., & Cleary, P. (1977). Confirmation of the dimensional structure of the SCL90: A study in construct validation. Journal of Clinical Psychology,, 98 1-989.Derogatis, L. R., Rickels, K., & Rock, A. (1976). The SCL-90 and the MMPI: A step in thevalidation of a new self-report scale. British Journal of Psychiatry, i2, 280-289.Donovan, D. M. (1988). Assessment of addictive behaviors: Implications of an emergingbiopsychosocial model. In D. M. Donovan & G. A. Marlatt (Eds.), AddictiveBehaviors (pp. 3-48). New York: Guilford Press.Downing, C. (in press). Three models of treatment for alcoholism: AA, the MinnesotaModel, and the CENAPS Biopsychosocial Model. Journal of Studies on Alcohol.Eddy, N. B., Halbach, H., Isbell, H., & Seevers, M. H. (1965). Drug dependence: Itssignificance and characteristics. WHO Bulletin, .2, 72 1-733.Edman, S. 0., Cole, D. A., & Howard, G. S. (1990). Convergent and discriminant validity ofFACES III: Family adaptability and cohesion. Family Process, 29, 95-103.Edmonds, V. H. (1967). Marital conventionalization: Definition and measurement. Journalof Marriage and the Family, 29, 681-688.Edmonds, V. H., Withers, G., & Dibatista, B. (1972). Adjustment, conservatism, and maritalconventionalization. Journal of Marriage and the Family, 29, 96-103.Edwards, G., & Gross, M. M. (1976). Alcohol dependence: Provisional description of aclinical syndrome. British Medical Journal, , 1058-106 1.Edwards, 0., Gross, M. M., Keller, M., & Moser, J. (1976). Alcohol-related problems in thedisability perspective: A summary of the consensus of the WHO group ofinvestigators on criteria for identifying and classifying disabilities related to alcoholconsumption. Journal of Studies on Alcohol, ., 1360-1382.Edwards, G., Orford, J., Egert, S., Guthrie, S., Hawker, A., Hensman, C., Mitcheson, M.,Oppenheimer, E., & Taylor, C. (1977). Alcoholism: A controlled trial of “treatment”and “advice”. Journal of Studies on Mcohol,, 1004-1031.Edwards, P., Harvey, C., & Whitehead, P. C. (1973). Wives of alcoholics: A critical reviewand analysis. Ouarterly Journal of Studies on Alcohol, .4, 112-132.211Elkin, I., Shea, T., Watkins, J. T., Imber, S., Sotsky, S., Collins, J., Glass, D., Pilkonis, P. A.,L.eber, W., Docherty, J., Fiester, S., & Parloff, M. (1989). National Institute ofMental Health treatment of depression collaborative research program. Archives ofGeneral Psychiatry, .4, 97 1-982.Emrick, C. D., & Hansen, J. (1983). Assertions regarding effectiveness of treatment foralcoholism. American Psychologist, ., 1078-1088.Epstein, N. B., Baldwin, L. M., & Bishop, D. S. (1983). The McMaster family assessmentdevice. Journal of Marital and Family Therapy, .9, 171-180.Erickson, E. H. (1963). Childhood and society. New York: Norton.Eron, J. B., & Lund, T. W. (1993). How problems evolve and dissolve: Integrating narrativeand strategic concepts. Family Process, .32, 29 1-309.Esser, P. H. (1968). Conjoint family therapy for alcoholics. British Journal of Addiction, .3,177-182.Esser, P. H. (1970). Conjoint family therapy with alcoholics: A new approach. BritishJournal of Addiction, .4, 275-286.Esser, P. H. (1971). Evaluation of family therapy with alcoholics. British Journal ofAddiction, ., 25 1-255.Ewing, 3., & Fox, R. E. (1968). Family therapy of alcoholics. In 3. Masserman (Ed.),Current Psychiatric Therapies,.S, 86-91.Ewing, 3. A., Long, V., & Wenzel, G. G. (1961). Concurrent group psychotherapy ofalcoholics and their wives. International Journal of Group Psychotherapy, jj, 329-338.Feighner, J., Robins, E., Guze, S., Woodruff, R., Winokur, G., & Munoz, R. (1972).Diagnostic criteria for use in psychiatric research. Archives of General Psychiatry, 2,57-63.Filstead, W. (1979). Comparing the family environments of alcoholic and “normal” families.Unpublished monograph. Alcoholism Treatment Center, Lutheran GeneralHospital, Park Ridge, Illinois.Filstead, W. J., Anderson, C., & McElfresh, 0. (1980). An examination of male and femalealcoholics’ perception of their present and ideal family environments. In M. Galanter(Ed.), Currents in alcoholism. Vol. 17. Recent advances in research and treatment(pp. 435-445). New York: Grune & Stratton.Fiimey, J. W., & Moos, R. H. (1979). Treatment and outcome for empirical subtypes ofalcoholics. Journal of Consulting and Clinical Psychology, .42, 25-38.Finney, J. W., Moos, R. H., Cronkite, R. C., & Gamble, W. (1983). A conceptual model ofthe functioning of married persons with impaired partners: Spouses of alcoholicpatients. Journal of Marriage and Family, 4, 23-34.Fisher, L., Kokes, R. F., Ransom, D. C., Phillips, S. L., Rudd, P. (1985). Alternativestrategies for creating “rational” family data. Family Process, 24, 213-224.212Floyd, F. J. (1983). Observational biases in spouse observation: Toward acognitive/behavior model of marriage. Journal of Consulting and ClinicalPsychology, j, 450-457.Foley, S. H., Rounsaville, B. J., Weissman, M. M., Sholomskas, D., & Chevron, E. (1989).Individual versus conjoint interpersonal psychotherapy for depressed patients withmarital disputes. International Journal of Family Psychology, jQ, 29-42.Foucault, M. (1980). Power/knowledge: Selected interviews and other writings. New York:Pantheon Books.Frankenstein, W., Hay, W. M., & Nathan, P. E. (1985). Effects of intoxication on alcoholics’marital communication and problem solving. Journal of Studies on Alcohol,, 1-6.Friedrich, W., & Lonftsgard, S. 0. (1978). Comparison of two alcoholism scales withalcoholics and their wives. Journal of Clinical Psychology, .4, 784-786.Friesen, J. D. (1983). An ecological approach to family counselling. Canadian Counsellor,12, 98-104.Friesen, J. D., Grigg, D. N., & Newman, 3. A. (1991). Experiential systemic therapy: Anoverview. Unpublished manuscript. Department of Counselling Psychology,University of British Columbia, Vancouver, B.C.Friesen, J. F., Grigg, D. N., Newman, J. A., & Peel, C. P. (1989). Experiential systemictherapy. Vancouver, B.C.: Western Family Learning Institute.Gacic, B. (1986). An ecosystemic approach to alcoholism: Theory and practice.Contemporary Family Therapy, ., 264-278.Galizio, M., & Maisto, S. A. (1985). Toward a biopsychosocial theory of substance abuse. InM. Galizio & S. A. Maisto (Eds.), Determinants of substance abuse: Biological.psychological and environmental (pp. 425-429). New York: Plenum Press.Gallant, D. M., Rich, A., Bey, E., & Terranora, L. (1970). Group psychotherapy withmarried couples: A successful technique in New Orleans alcoholism clinic patients.Journal of the Louisiana State Medical Society,j, 4 1-44.Geiss, S. K., & O’Leary, K. D. (1981). Therapist ratings of frequency and seventy of maritalproblems: Implications for research. Journal of Marital and Family Therapy, .7, 515-520.Gergen, K. (1991). The saturated self. New York: Basic Books.Gibbs, J. C. (1979). The meaning of ecologically oriented inquiry in contemporarypsychology. American Psychologist, 4, 127-140.Gibbs, L. E. (1983). Validity and reliability of the Michigan Alcoholism Screening Test: Areview. Drug and Alcohol Dependence, 12, 279-285.Gliedman, L. H. (1957). Concurrent and combined group treatment of chronic alcoholicsand their wives. International Journal of Grout, Psychotherapy, .7, 414-424.213Gliedman, L. H., Rosenthal, D., Frank, J. D., & Nash, H. T. (1956). Group therapy ofalcoholics with concurrent group meetings of their wives. Ouarterly Journal ofStudies on Alcohol, fl, 655-670.Goldner, V. (1988). Generation and gender: Normative and covert hierarchies. FamilyProcess, 22, 13-31.Goolishian, H. A., & Anderson, H. (1992). Strategy and interventino versusnonintervention: A matter of theory? Journal of Marital and Family Therapy, iS, 5-15.Gorad, S. L. (1971). Communication styles and interaction of alcoholics and their wives.Family Process, jQ, 475-489.Green, R. G. (1989). Choosing family measurement devices for practice and research: SF!and FACES III. Social Review,, 304-320.Grigg, D. N., Friesen, J. D., & Conry, R. F. (1993). The Alcohol Recovery Project: Atreatment outcome study. Poster presented at the American Association of Maritaland Family Therapy, 50th Annual Conference, Anaheim: CA.Grigg, D. N., Friesen, 3. D., & Sheppy, M. I. (1989). Family patterns associated withanorexia nervosa. Journal of Marital and Family Therapy, 15, 29-42.Grigg, D. N., Friesen, J. D., Weir, W., & Bate, C. (1991). Supported feedback therapymanual. Vancouver: Western Family Learning Institute.Guerney, B. G., Jr. (1977). Relationship enhancement: Skill training programs for therapy.problem prevention, and enrichment. San Francisco: Jossey-Bass.Guerin, P. J., Jr. (1976). Family therapy: The first twenty-five years. In P. J. Guerin, Jr.(Ed.), Family therapy: Theory and practice (pp. 2-22). New York: Garner.Gurman, A. S., & Kniskern, D. P. (1981). Family therapy outcome research: Knowns andunknowns. In A. S. Gurman & D. P. Kniskern (Eds.), Handbook of family therapy(pp. 742-775). New York: Brunner/Mazel.Guze, S. B., Tuason, V. A., Steward, M. A., & Picken, B. (1963). The drinking history: Acomparison of reports by subjects and their relatives. Ouarterly Journal of Studies onAlcohol, 24, 249-260.Hazelrigg, M. D., Cooper, H. M., & Bordvin, C. M. (1987). Evaluating the effectiveness offamily therapies: An integrative review and analysis. Psychological Bulletin, ji,428-442.Heather, V., Robertson, I., MacPherson, B., Ailsop, S., & Fulton, A. (1987). Effectiveness ofa controlled drinking self-help manual: One year follow-up results. British Journal ofClinical Psychologv, ., 279-287.Hedberg, A. G., & Campbell, L. M. (1974). A comparison of four behavioral treatmentapproaches to alcoholism. Journal of Behavioral Therapy and ExperimentalPsychiatry, .5, 25 1-256.Hedlund, 3. L., & Viewag, B. W. (1984). The Michigan Alcohol Screening Test (MAST): Acomprehensive review. Journal of Operational Psychiatry, 15, 55-65.214Heizer, J. E. (1987). Epidemiology of alcoholism. Journal of Consulting and ClinicalPsychology,, 284-292.Hersen, M., Miller, P. M., & Eisler, R. M. (1973). Interactions between alcoholics and theirwives: A descriptive analysis of verbal and nonverbal behavior. Ouarterly Journal ofthe Study of Alcohol, .4, 516-520.Hoffman, L. (1981). Foundations of family therapy. New York: Basic Books.Hoffman, L. (1986). Beyond power and control: Toward a second-order cybernetics.Family Systems Medicine, .4, 38 1-396.Hoffman, L. (1990). Constructing realities: An art of lenses. Family Process, .29, 1-12.Holon, S. D., Waskow, I. E., Evan, M., & Lowery, H. A. (1984, May). Systems for ratingtherapies for depression. Paper presented at the annual convention of the AmericanPsychiatric Association, Los Angeles, California.Hore, B. D. (1971a). Life events and alcohol relapse. British Journal of Addiction,, 83-88.Hore, B. D. (1971b). Factors in alcoholic relapse. British Journal of Addiction, ., 89-96.Humphrey, L. L., & Benjamin, L. 5. (1986). Using Structural Analysis of Social Behavior toassess critical but elusive family processes. American Psychologist, .4.1, 979-989.Humphrey, L. L., Apple, R. F., & Kirschenbaum, D. S. (1986). Differentiating bulimicanorexic from normal families using interpersonal and behavioral observationsystems. Journal of Consulting and Clinical Psychology, .4, 190-195.Institute of Medicine (1992). Prevention and treatment of alcohol-related problems:Research opportunities. Journal of Studies on Alcohol,, 5-16.Jackson, D. D. (1959). Family interaction, family homeostasis, and some implications forconjoint family therapy. In J. Masserman (Ed.), Individual and family dynamics (pp.93-127). New York: Grune and Stratton.Jacob, T., Dunn, N. J., & Leonard, K. (1983). Patterns of alcohol abuse and family stability.Alcoholism: Clinical and Experimental Research, .7, 382-385.Jacob, T., & Krahn, G. L. (1988). Marital interactions of alcoholic couples: Comparisonwith depressed and nondistressed couples. Journal of Consulting and ClinicalPsychology, ., 73-79.Jacob, T., Ritchey, D., Cvitkovic, J. F., & Blane, H. T. (1981). Communication styles ofalcoholic and nonalcoholic families when drinking and not drinking. Journal ofStudies on Alcohol, .4., 466-482.Jacob, T., & Seilhamer, R .A. (1982). The impact of alcoholism on spouses and how theycope. In J. Orford and J. Harwin (Eds.), Alcohol and the family (pp. 114-126).London: Crown Helm.215Jacob, T., & Seilhamer, R. A. (1987). Alcoholism and family interaction. In T. Jacob (Ed.),Family interaction and pathology: Theories, methods and findings (pp. 535-580).New York: Plenum Press.Jacobson, N. S. (1978). A review of the research on the effectiveness of marital therapy. InT. L. Paolino & B. S. McCrady (Eds.), Marriage and marital therapy (pp. 395-444).New York: Brunner/Mazel.Jacobson, N. S. (1985). Family therapy outcome research: Potential pitfalls and prospects.Journal of Marital and Family Therapy, .jj, 149-158.Jacobson, N. S., & Bussod, N. (1983). Marital and family therapy. In M. Hersen, A. E.Kazdin, & A. S. Bellack (Eds.), The clinical psychological handbook (pp. 611-630).New York: Pergamon.Jacobson, N. S., Dobson, K., Fruzzetti, A. E., Schmaling, K. B., & Salvsky, S. (1991). Maritaltherapy as a treatment for depression. Journal of Consulting and Clinical Psychology,..9, 547-557.Jacobson, N. S., Follette, W. C., & Elwood, R. W. (1984). Outcome research on behavioralmarital therapy: A methodological and conceptual reappraisal. In K. Hahleveg & N.S. Jacobson (Eds.), Marital interaction: Analysis and modification (pp. 113-129).New York: Guilford.Jacobson, N. S., Munroe, A. H., & Schmaling, K. B. (1989). Marital therapy and spouseinvolvement in the treatment of depression, agoraphobia, and alcoholism. Journal ofConsulting and Clinical Psychology, 57, 5-10.Jellinek, E. M. (1960). The disease concept of alcoholism. New Brunswick, NJ: HilihousePress.Joanning, H., Quinn, W., Thomas, F., & Mullen, T. (1992). Treating adolescent drug abuse:A comparison of family systems therapy, group therapy, and family drug education.Journal of Marital and Family Therapy, 18, 345-356.Johnson, S. M., & Greenberg, L. S. (1985). Differential effects of experiential and problem-solving interventions in resolving marital conflict. Journal of Consulting and ClinicalPsychology, .53, 175-184.Jolly, P. A., Fleece, E. L., Galanos, A .N., Milby, J. B., & Ritter, S. C. (1983). DSM-III:Alcohol abuse and alcohol dependence; Inter-rater diagnostic reliability. AddictiveBehaviors, ., 20 1-204.Kasdin, A. E. (1986a). Comparative outcome studies of psychotherapy: Methodologicalissues and strategies. Journal of Consulting and Clinical Psychology, .54, 95-105.Kasdin, A. E. (1986b). Evaluation of psychotherapy: Research design and methodology. InS. L. Garfield and A. E. Bergin (Eds.), Handbook of psychotherapy and behaviorchange (pp. 28-68). New York: Wiley and Sons.Kaufman, E., & Kaufman, P. N. (1979). Family therapy of drug and alcohol abuse. NewYork: Gardner Press.Kaufman, E., & Pattison, E. M. (1981). Differential methods of family therapy in thetreatment of alcoholism. Journal of Studies on Alcohol, .42, 95 1-971.216Kozak, A. E., Jarmas, A., & Snitzer, L. (1988). The assessment of marital satisfaction: Anevaluation of the Dyadic Adjustment Scale. Journal of Family Psychology, 2, 82-91.Keeney, B. P. (1983). Aesthetics of change. New York: Guilford Press.Keller, M. (Ed.) (1974). Trends in treatment of alcoholism. In Second special report to theU.S. Congress on Alcohol and Health. Washington, DC: Department of Health,Education and Welfare.Keller, M., McCormick, M., & Efron, V. (1982). A dictionary of words about alcohol (2nded.). New Brunswick, NJ: Rutgers University, Center of Alcohol Studies.Killorin, E., & Olson, D. H. (1984). The chaotic flippers in treatment. In E. Kaufman (Ed.),Power to change (pp. 99-130). New York: Gardner Press.Kissin, B. (1983). The disease concept of alcoholism. Research Advances in Alcohol andDrug Problems, 2, 93-126.Kissin, B., & Hanson, M. (1982). The bio-psychosocial perspective in alcoholism. In J.Solomon (Ed.), Alcoholism and clinical psychiatry (pp. 1-19). New York: PlenumPress.Lawson, G., Peterson, J. S., & Lawson, A. (1983). Alcoholism and the family: A guide totreatment and prevention. Rockville, MD: Aspen Publications.Lazelere, R. E., & Huston, T. E. (1980). The Dyadic Trust Scale: Toward understandinginterpersonal trust in close relationships. Journal of Marriage and Family, .4, 595-604.Lewis, R. A., Piercy, F. P., Sprenkle, D. H., & Trepper, T. 5. (1991). The Purdue BriefFamily Therapy Model for adolescent substance abusers. In T. C. Todd & M. D.Selekman (Eds.), Family therapy approaches with adolescent substance abusers (pp.29-48). Boston: Allyn & Bacon.Liddle, H. A., Dakof, G. A., Porter, K., Garcia, R., Diamond, G., & Barrett, K. (1991).Effectiveness of family therapy with adolescent drug abusers. Paper presented at theannual meeting of the American Psychological Association, San Francisco, CA.Litman, G. K., Stapleton, J., Oppenheim, A. N., Peleg, M., & Jackson, P. (1983). Situationsrelated to alcoholism relapse. British Journal of Addiction,., 381-389.Macdonald, D. E. (1958). Group psychotherapy with wives of alcoholics. Ouarterly Journalof Studies on Alcohol, j9, 125-132.Maisto, S. A., Sobell, L. C., & Sobell, M. B. (1979). Comparison of alcoholics’ self-reports ofdrinking behavior with reports of collateral informants. Journal of Consulting andClinical Psychology, .42, 106-112.Margolin, G., Talovic, S., & Weinstein, C. D. (1983). Areas of Change Questionnaire: Apractical approach to marital assessment. Journal of Consulting and ClinicalPsychology, .j, 920-93 1.217Margolin, 0., & Wampold, B. E. (1981). Sequential analysis of conflict and accord indistressed and nondistressed marital partners. Journal of Consulting and ClinicalPsychologv, .49, 544-567.Margolin, G., & Weiss, R. A. (1978). A comparative evaluation of therapeutic componentsassociated with behavioral marital therapy. Journal of Consulting and ClinicalPsychology, .4, 1476-1486.Marlatt, 0. A. (1976). The drinking profile: A questionnaire for the behavioral assessmentof alcoholism. In E. Mash, & L. Terdel (Eds.), Behavior therapy assessment:Diagnosis design and evaluation (pp. 121-137). New York: Springer.Marlatt, G. A. (1978). Craving for alcohol, loss of control, and relapse: A cognitive-behavioral analysis. In P. E. Nathan, 0. A. Marlatt, & T. L.oberg (Eds.), Alcoholism:New directions in behavioral research and treatment (pp. 271-3 14). New York:Plenum Press.Marlatt, G. A. (1979). Alcohol use and problem drinking: A cognitive behavioral analysis.In P. C. Kendall & S. D. Hollon (Eds.), Cognitive-behavioral interventions: Theory.research and procedures (pp. 3 19-355). New York: Academic Press.Marlatt, G. A., & Gordon, J. F. (1978). Determinants of relapse: Implications for themaintenance of behavior change. Alcohol and Drug Institute, Technical Report No.78-07, University of Washington.Marlatt, G. A., & Gordon, J. F. (1980). Determinants of relapse: Implications for themaintenance of behavior change. In P. Davidson & S. Davidson (Eds.), Behavioralmedicine: Changing health lifestyles (pp. 410-452). New York: Brunner/Mazel.Maturana, H., & Varela, F. (1980). Autopsies and cognition: The realization of living.Dordrecht, Netherlands: D. Reidl.McCrady, B. S. (1982). Conjoint behavioral treatment of an alcoholic and his spouse. In W.M. Hay and P. E. Nathan (Eds.), Clinical case studies in the behavioral treatment ofalcoholism (pp. 127-156). New York: Plenum Press.McCrady, B. S. (1986). The family in the change process. In W. R. Miller & N. Heather(Eds.), Treating addictive behaviors: Processes of change (pp. 305-3 18). New York:Plenum Press.McCrady, B. S. (1989). Outcomes of family-involved alcoholism treatment. In M. Gallanter(Ed.), Recent developments in alcoholism: Vol. 7. Treatment research (pp. 165-182). New York: Plenum.McCrady, B. S., Moreau, J., Paolino, T. J., & Longabaugh, R. (1982). Joint hospitalizationand couples therapy for alcoholism: A four year follow-up. Journal of Studies onAlcohol, 4, 1244-1250.McCrady, B. S., Noel, N. E., Abrams, D. B., Stout, R. L., Nelson, H. F., & Hay, W. M.(1986). Comparative effectiveness of three types of spousal involvement in outpatientbehavioral alcoholism treatment. Journal of Studies on Alcohol, .42, 459-467.McCrady, B. S., Paolino, T. J., Longabough, R., & Rossi, J. (1979). Effects of joint hospitaladmission and couples treatment for hospitalized alcoholics: A pilot study. AddictiveBehaviors,.4, 155-165.218McKnight, D. L., Nelson-Greg, R. 0., & Barnhill, J. (1992). Dexamethasone suppressiontest and response to cognitive therapy and antidepressant medication. BehaviorTherapy,2, 99-111.McLellan, A. T., Luborsky, L., Cacciola, J., Griffith, J., Evans, F., Barr, H. L., & O’Brien, C.P. (1985). New data from the addiction severity index: Reliability and validity inthree centers. Journal of Nervous and Mental Disorder, 17.3 , 412-423.Mead, D. E., Vatcher, G. M., Wyne, B. A., & Roberts, S. L. (1990). The comprehensiveareas of change questionnaire: Assessing marital couples’ presenting complaints.The American Journal of Family Therapy, is, 65-79.Mead, G. H. (1934). Mind, self, and society. Chicago: University of Chicago Press.Meeks, D. E., & Kelly, C. (1970). Family therapy with the families of recovering alcoholics.Ouarterly Journal of Studies on Alcohol, 3j, 399-413.Midanik, L. (1982). The validity of self-reported alcohol consumption and alcohol problems:A literature review. British Journal of Addiction, .27, 357-382.Miller, J. G. (1978). Living systems. New York: McGraw-Hill Inc.Miller, P. 3., Ross, S. M., Emmerson, R. Y., & Todt, E. H. (1987). Self-efficacy in alcoholics:Clinical validation of the Situational Confidence Questionnaire. Addictive Behaviors,14, 217-224.Miller, W. R., Crawford, L., & Taylor, C. (1979). Significant others as corroborative sourcesfor problem drinkers. Addictive Behavior, .4, 67-70.Miller, W. R., & Hester, R. K. (1980). Treating the problem drinker: Modern approaches.In W. R. Miller (Ed.), The addictive behaviors (pp. 11-141). New York: PergamonPress.Miller, W. R., & Hester, R. K. (1986). The effectiveness of alcoholism treatment methods:What research reveals. In W. R. Miller & N. Heather (Eds.), Treating addictivebehaviors: Processes of change (pp. 121-179). New York: Plenum Press.Miller, W. R., & Hester, R. K. (1989). Treating alcohol problems: Toward an informedeclecticism. In R. K. Hester and W. R. Miller (Eds.), Handbook of alcoholismtreatment approaches: Effective alternatives (pp. 3-14). New York: PergamonPress.Moore, R. A. (1972). The diagnosis of alcoholism in a psychiatric hospital: A trial of theMichigan Alcoholism Screening Test (MAST). American Journal of Psychiatry, j5,115-119.Moos, R. (1974). The Social Climate Scale: An overview. Palo Alto: ConsultingPsychologist Press.Moos, R. H., Bromet, E., Tsu, V., & Moos, B. (1979). Family characteristics and theoutcome of treatment for alcoholism. Journal of Studies on Alcohol, .4Q, 78-88.Moos, R. H., & Moos, B. S. (1981). Family Environment Scale: Manual. Palo Alto, CA:Consulting Psychologists Press.219Murray, H.A. (1938). Explorations in personality. New York: Oxford University Press.Nathan, P. E., & Skinstad, A. (1987). Outcomes of treatment for alcohol problems: Currentmethods, problems, and results. Journal of Consulting and Clinical Psychology, 55,332-340.Newman, J. A. (1990). Relational novelty: The Experiential Systemic means to individualand couples change. Unpublished manuscript, Department of CounsellingPsychology, University of British Columbia.Newman, J. A., Friesen, J. D., & Grigg, D. N. (1991). The supervision of ExperientialSystemic Therapy for individuals, couples and families. The Clinical Counsellor, j,12-20.Nicholas, M. K., Wilson, P.H., & Goyen, J. (1991). Operant-behavioural and cognitivebehavioural treatment for chronic low back pain. Behaviour Research and Therapy,29, 225-238.Nichols, M. P. (1984). Family therapy: Concepts and methods. New York: Gardner Press.O’Farrell, T. J. (1986). Marital therapy in the treatment of alcoholism. In N. S. Jacobson, &A. S. Gurman (Eds.), Clinical handbook of marital therapy (pp. 5 13-535). New York:Guilford.O’Farrell, T. J. (1992). Families and alcohol problems: An overview of treatment research.Journal of Family Psychology, .5, 339-359.O’Farrell, TJ., & Birchier, G. R. (1987). Marital relationships of alcoholic, conflicted, andnonconflicted couples. Journal of Marital and Family Therapy, j, 259-274.O’Farrell, T. J., & Cowles, K. (1989). Marital and family therapy. In R. Hester & W. R.Miller (Eds.), Comprehensive handbook of alcoholism treatment approaches (pp.183-205). New York: Pergamon.O’Farrell, T. J., & Cutter, H. S. G. (1982, November). Effects of adding a behavioral or aninteractional couples group to individual outpatient alcoholism counseling. In T. J.O’Farrell (Chair), Spouse-involved treatment for alcohol abuse. Symposiumconducted at the Sixteenth Annual Convention of the Association for Advancementof Behavior Therapy, Los Angeles.O’Farrell, T. J., & Cutter, H. S. (1984). Behavioral marital therapy couples groups for malealcoholics and their wives. Journal of Substance Abuse Treatment, j, 19 1-204.O’Farrell, T. J., Cutter, H. S., & Floyd, F. J. (1985). Evaluating behavioral marital therapyfor male alcoholics: Effects on marital adjustment and communication before andafter treatment. Behavior Therapy, j, 147-167.O’Farrell, T. J., Cutter, H. S., Choquette, K. A., Floyd, F. J., & Bayog, R. D. (1992).Behavioral marital therapy for male alcoholics: Marital and drinking adjustmentduring the two years after treatment. Behavior Therapy, 23, 529-549.O’Leary, K. D., & Beach, S. R. (1990). Marital therapy: A viable treatment for depressionand mental discord. American Journal of Psychiatry, j47, 183-186.220Olson, D. H. (1970). Marital and family therapy: Integrative review and critique. Journal ofMarriage and the Family,, 501-538.Olson, D. H. (1986). Circumplex Model VII: Validation studies and FACES III. FamilyProcess, 2, 337-35 1.Olson, D. H. (1989). Circumplex model of family systems VIII: Family assessment andintervention. In D. H. Olson, C. S. Russell, & D. H. Sprenkle (Eds.), Circumplexmodel: Systemic assessment and treatment of families (pp. 7-49). New York:Haworth Press.Olson, D. H. (1991). Commentary: Three dimensional circumplex model and revisedscoring of FACES III. Family Process, Q, 74-79.Olson, D. H., & Killorin, E. (1985). Chemically dependent families and the circumplexmodel. Unpublished manuscript. Family Social Science, University of Minnesota, St.Paul, Minnesota.Olson, D. H., Portner, J., & Lavee, Y. (1985). FACES III. St. Paul, MI: University ofMinnesota Press.Olson, D. H., & Wilson, M. (1982). Family satisfaction. St. Paul, MI: Family Social Science,University of Minnesota.Orford, J. (1975). Alcoholism and marriage: The argument against specialism. Journal ofStudies on Mcohol,, 1537-1563.Orford, J. (1990). Alcohol and the family: An international review of the literature withimplications for research and practice. In L. T. Kozlowski et al. (Eds.), Researchadvances in alcohol and drug problems (Vol. 10, pp. 81-155). New York: Plenum.Orford, J., & Harwin, 3. (Eds.) (1982). Alcohol and the family. London: Crown Helm.Orford, S., Oppenheimer, E., Egert, S., Hensman, C., & Guthrie, S. (1976). Thecohesiveness of alcoholism-complicated marriages and its influence on treatmentoutcome. British Journal of Psychiatry, j., 318-339.Paolino, J., & McCrady, B. S. (1976). Joint admission as a treatment modality for problemdrinkers: A case report. American Journal of Psychiatry, j..3., 222-224.Paolino, T. 3., & McCrady, B. S. (1977). The alcoholic marriage: Alternative perspectives.New York: Grune and Stratton.Pattison, E. M., & Kaufman, E. (1982). The alcoholism syndrome: Definitions and models.In E. M. Pattison & E. Kaufman (Eds.), Encyclopedic handbook of alcoholism (pp. 3-30). New York: Gardner Press.Paolhos, D. L. (1984). Two-component models of socially desirable responding. Journal ofPersonality and Social Psychologv,, 598-609.Perosa,, L. M., & Perosa, S. L. (1990). Convergent and discriminant validity for family selfreport measures. Educational and Psychological Measurement, ., 855-868.Pinsof, W. M. (1983). Integrative problem-centered therapy: Toward the synthesis of familyand individual psychotherapies. Journal of Marital and Family Therapy, .9, 19-35.221Pinsof, W. M. (1992). Toward a scientific paradigm for family psychology: The integrativeprocess systems perspective. Journal of Family Psychology, 5, 432-447.Pinsof, W. M., & Catherall, D. R. (1986). The integrative psychotherapy alliance: Family,couple, and individual therapy scales. Journal of Mental and Family Therapy, J2,137-151.Piotrowski, C., Sherry, D., & Keller, J. W. (1985). Psychodiagnostic test usage: A survey ofthe society for personality assessment. Journal of Personality Assessment, A9, 115-119.Rabin, C., Margolis, G., Safir, M., Talovic, S., & Sadeh, I. (1986). The Areas of ChangeQuestionnaire: A cross-cultural comparison of Israeli and American distressed andnondistressed couples. American Journal of Family Therapy, j4, 324-335.Raistrick, D., Dunbar, G., & Davidson, R. (1983). Development of a questionnaire tomeasure alcohol dependence. British Journal of Addiction, Z.S, 89-95.Regan, 3. M., Connors, G. J., O’Farrell, T. 3., & Jones, W. C. (1983). Services to the familiesof alcoholics: A survey of treatment agencies in Massachusetts. Journal of Studies onAlcohol, 44, 1072-1082.Rogers, C. A. (1951). Client-centered therapy. Boston: Houghton Mifflin.Rogers, C. A. (1961). On becoming a person. Boston: Houghton Mifflin.Rogers, C. A. (1972). Becoming partners. New York: Dell.Rubin, Q. (1970). Measurement of romantic love. Journal of Personality and SocialPsychology, J, 265-273.Rychtarik, R. G., Tarnowski, K. J., & St. Lawrence, J. S. (1989). Impact of social desirabilityresponse sets on the self-report of marital adjustment in alcoholics. Journal ofStudies on Alcohol, 5, 24-29.Salkovskis, P. M., Clark, D. M., & Hackmann, A. (1991). Treatment of panic attack usingcognitive therapy without exposure or breathing retraining. Behaviour Research andTherapy, 29, 16 1-166.Sanchez-Craig, M., Annis, H. M., Bornet, A. R., & MacDonald, K. R. (1984). Randomassignment to abstinence and controlled drinking: Evaluation of a cognitive-behavioral program for problem drinkers. Journal of Consulting and ClinicalPsychology, 5, 390-403.Sands, P. M., & Hansen, P. G. (1971). Psychotherapeutic groups for alcoholics and relativesin an outpatient setting. International Journal of Group Psychotherapy, 21, 23-33.Schaap, C. (1984). A comparison of the interaction of distressed and nondistressed marriedcouples in a laboratory situation: Literature survey, methodological issues and anempirical investigation. In K. Hahlweg & N. S. Jacobson (Eds.), Marital interaction:Analysis and modification (pp. 133-158). New York: Guilford Press.Schaefer, E. S. (1965). Configurational analysis of children’s reports of parent behavior.Journal of Consulting Psychology, 29, 552-557.222Schag, C. C., Heinrich, R. L., & Ganz, P. A. (1983). Cancer inventory of problem situations:An instrument for assessing cancer patients rehabilitation needs. Journal ofPsychosocial Oncology, j, 11-24.Schumm, W. R., Boliman, S. R., & Jurich, A. P. (1980). Marital communication or maritalconventionality? A brief report on the relationship inventory. Psychological Reports,, 1171-1174.Schwartz, G. E. (1982). Testing the biopsychosocial model: The ultimate challenge facingbehavioral medicine. Journal of Consulting and Clinical Psychology,, 1040-1053.Schwartz, R. (1987a). Our multiple selves. Family Therapy Networker, jj(2), 24-3 1, 80-83.Schwartz, R. (198Th). Working with “internal and external” families in the treatment ofbulimia. Family Relations, 3, 242-245.Schwartz, R. (1988). Know theyselves. Family Therapy Newsletter, .12(6), 21-29.Schwartz, R., & Grace, P. (1989). The systemic treatment of bulimia. Journal ofPsychotherapy and the Family, ., 89-105.Selekman, M. D. (1993). Pathways to change: Brief therapy solutions with difficultadolescents. New York: Guilford Press.Scharz, S. P., & Blanchard, E. B. (1991). Evaluation of a psychological treatment forinflammatory bowel disease. Behaviour Research and Therapy,29, 167-177.Seizer, M. L. (1971). The Michigan Alcoholism Screening Test: The quest for a newdiagnostic instrument. American Journal of Psychiatry, IZZ, 1653-1658.Shaffer, H. 3., & Milkman, H. B. (1985). Introduction: Crisis and conflict in the addictions.In H. B. Milkman & H. J. Shaffer (Eds.), The addictions: Multidisciplinaryperspectives and treatments (pp. ix-xviii). Lexington, MA: Lexington Books.Sharpley, C. F., & Cross, D. G. (1982). A psychometric evaluation of the Spanier DyadicAdjustment Scale. Journal of Marriage and the Family, .44, 739-747.Simon, G. M. (1992). Having a second-order mind while doing first-order therapy. Journalof Marital and Family Therapy, j, 377-387.Sisson, R. W., & Azrin, N. H. (1986). Family member involvement to initiate and promotetreatment of problem drinkers. Journal of Behavioral Therapy and ExperimentalPsychiatry, jJ, 15-21.Skinner, H .A. (1979). A multivariate evaluation of the MAST. Journal of Studies onAlcohol,.4.Q, 831-844.Skinner, H. A., & Horn, J. L. (1984). Alcohol dependence scale (ADS): User’s guide.Toronto: Addiction Research Foundation of Ontario.Skinner, H. A., & Sheu, W. J. (1982). Reliability of alcohol use indices: The lifetimedrinking history and the MAST. Journal of Studies on Alcohol,.4, 1157-1170.Sluzki, C. E. (1985). A minimal map of cybernectics. Family Therapy Networker, .9(2), 26.223Smith, C. J. (1969). Alcoholics: Their treatment and their wives. British Journal ofPsychiatry, 11.5, 1039-1042.Snyder, D. K., & Wills, R. M. (1989). Behavioral versus insight-oriented marital therapy:Effects on individual and interpersonal functioning. Journal of Consulting andClinical Psychology, 17, 39-46.Sobell, L., & Sobell, M. (1975). Outpatient alcoholics give valid self-reports. Journal ofNervous and Mental Disease, jj, 32-42.Sobell, C., Sobell, M. B., Riley, D. M., Schuller, R., Pavan, D. S., Cancilla, A., Klajner, F., &Leo, G. I. (1988). The reliability of alcohol abusers’ self-reports of drinking and lifeevents that occurred in the distant past. Journal of Studies on Alcohol, 49, 225-232.Sobell, M., Sobell, L., & Samuels, F. (1974). Validity of self-reports of alcohol-relatedarrests by alcoholics. Quarterly Journal of Studies on Aichol, .5, 276-280.Sokal, R. J., Miller, S. I., & Debanne, S. (1981). The Cleveland NIAAA perspective alcohol-in-pregnancy study: The first year. Neurobehavioral Toxical Teratol, ., 203-209.Solomon, K. E., & Annis, H. M. (1988). Outcome and efficacy expectancy in the predictionof post-treatment drinking behavior. British Journal of Addiction, ., 659-665.Spanier, G. B. (1976). Meausring dyadic adjustment: New scales for assessing the quality ofmarriage and other similar dyads. Journal of Marriage and the Family, 3, 15-28.Spanier, G. B., & Filsinger, E. E. (1983). The dyadic adjustment scale. In E.E. Filsinger(Ed.), Marriage and family assessment (pp. 155-168). Beverly Hills, CA: Sage.Spanier, G. B., & Thompson, L. (1982). A confirmatory analysis of the Dyadic AdjustmentScale. Journal of Marriage and the Family, .44, 731-738.Spicer, J. (1980). Outcome evaluation: How to do it. Center City, NM: Hazleton.Stalling, D. L, & Oncken, G. R. (1977). A relative change index in evaluating alcoholismtreatment outcome. Journal of Studies on Alcohol, 3S, 457-464.Stanley, D. T., & Campbell, J. C. (1963). Experimental and quasi-experimental designs forresearch. Boston: Houghton Mifflin Company.Stanton, M. D., & Todd, T. C. (1982). The family therapy of drug abuse and addiction. NewYork: Guilford.Steer, R. A., Beck, A. T., & Garrison, B. (1985). Applications of the Beck DepressionInventory. In N. Sartomus & T. A. Ban (Eds.), Assessment of depression (pp. 121-142). New York: Springer-Verlag.Steiner, C. M. (1969). The alcoholic game. Ouarterly Journal of Studies on Alcohol, .,920-938.Steinglass, P. (1976). Experimenting with family treatment approaches to alcoholism, 1950-1975: A review. Family Process, .15, 97-123.224Steinglass, P. (1979). An experimental treatment program for alcoholic couples. Journal ofStudies on Alcoho1,.4, 159-182.Steinglass, P. (1979). The alcoholic family in the interaction laboratory. The Journal ofNervous and Mental Disease, j, 428-436.Steinglass, P. (1980). A life history model of the alcoholic family. Family Process, 19, 211-226.Steinglass, P. (1981). The impact of alcoholism on the family. Journal of Studies onAlcohol,.4, 288-303.Steinglass, P. (1985). Family systems approach to alcoholism. Journal of Substance Abuse,2, 161-167.Steinglass, P. (1991). An editorial: Finding a place for the individual in family therapy.Family Process, 3..Q, 267-269.Steinglass, P., Bennett, L. A., Wolin, S. J., & Reiss, D. (1987). The alcoholic family. NewYork: Basic Books.Steinglass, P., Davis, D. I., & Berenson, D. (1977). Observations of conjointly hospitalized“alcoholic couples” during sobriety and intoxication: Implications for theory andtherapy. Family Process, j, 1-16.Steinglass, P., & Robertson, A. (1983). The alcoholic family. In B. Kissin & H. Begleiter(Eds.), The biology of alcoholism: Vol. 6. The pathogenesis of alcoholism:Psychosocial factors (pp. 243-255). New York: Plenum Press.Sternberger, L. G., & Leonard, B. G. (1990). Obessions and compulsions: Psychometricproperties of the Padua Inventory with an American college population. BehaviourResearch and Therapy, 2, 34 1-345.Straws, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics(CT) Scales. Journal of Marriage and Family,.4j, 75-88.Stuart, R. B. (1969). Operant-interpersonal treatment for marital discord. Journal ofConsulting and Clinical Psychology, 3., 675-682.Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: Norton.Taggart, M. (1985). The feminist critique in epistemological perspectives: Questions ofcontext in family therapy. Journal of Marital and Family Therapy, jJ, 113-126.Tamkin, A. S., Carson, M. F., Nixon, D. H., & Hyer, L.A. (1985). A comparison amongsome measures of depression in alcoholics. IRCS Medical Science Psychology andPsychiatry, .i, 23 1-235.Thomas, E. 3., & Santa, C. A. (1982). Unilateral family therapy for alcohol abuse: Aworking conception. American Journal of Family Therapy, .jQ, 49-58.Thompson, J., Friesen, J. D., Grigg, D. N., Weir, W., & Mitchell, D. (1993). The AdherenceRating Scale. University of British Columbia: Vancouver, British Columbia.225Todtman, D. A., Friesen, J. D., Newman, J. A., & Grigg, D. N. (1993). Experientialexternalization: Evoking the bottle in session. Paper presented at the AmericanAssociation of Marital and Family Therapy 50th Annual Conference, Anaheim, CA.Treadway, D. (1989). Before it’s too late. New York: Norton.Usher, M. L., Jay, J., & Glass, D. R. (1982). Family therapy as a treatment modality foralcoholism. Journal of Studies on Alcohol,, 927-938.Vaillant, G. E. (1983). The natural history of alcoholism. Cambridge, MA: HarvardUniversity Press.Vannicelli, M., Gingerich, S., & Ryback, R. (1983). Family problems related to thetreatment and outcome of alcoholic patients. British Journal of Addiction, .2, 193-204.Varela, F. J. (1979). Principles of biological autonomy. New York: Elsevier North Holland.Vogel, S. (1957). Some aspects of group psychotherapy with alcoholics. InternationalJournal of Group Psychotherapy, 2, 302-310.Von Bertalanffy, L. (1968). The meaning of general system theory in General SystemsTheory. New York: Braziller.Waldo, M., & Guerney, B. G. (1983). Marital relationship enhancement therapy in thetreatment of alcoholism. Journal of Marital and Family Therapy, .9, 321-323.Wallace, J. (1985). Predicting the onset of compulsive drinking in alcoholics: Abiopsychosocial model. Alcohol, 2, 589-595.Walter, G. A., & Marks, S. E. (1981). Experiential learning and change. New York: Wiley& Sons.Wanberg, K. W., & Horn, 3. L. (1983). Assessment of alcohol use with multidimensionalconcepts and measures. American Psychologist, ., 1055-1069.Watzlawick, P., Beavin, J., & Jackson, D. D. (1967). Pragmatics of human communication.New York: Norton.Watzlawick, P., Weakiand, 3., & Fisch, R. (1974). Change. New York: Norton.Wegscheider, S. (1981). Another chance: Hope and health for the alcoholic family. PaloAlto: Science and Behavior Books.Weiner, N. (1948). Cybernetics or control and communication in the animal and in themachine. New York: Wiley.Weiss, R. L., & Birchier, G. R. (1975). Areas of change. Unpublished manuscript.University of Oregon, Eugene, Oregon.Weiss, R. L., & Cerreto, M. C. (1980). The Marital STatus Inventory: Development of amesure of dissolution potential. American Journal of Family Therapy, .S, 80-85.Weiss, R. L., Hops, H., & Patterson, G. R. (1973). A framework for conceptualizing maritalconflict, a technology for altering it, some data for evaluating it. In L. A.226Hamerlynck, L. E. Handy, & E. J. Mash (Eds.), Critical issues in research andpractice: Proceedings of the fourth Banff International Conference on behavioralmodification (pp. 309-342). Champaign, IL: Research Press.Weissman, M. M., Sholmskas, D., Pottenger, M., Prusoff, B.A., & Locke, B. Z. (1977).Assessing depressive symptoms in five psychiatric populations: A validation study.American Journal of Epidemiology, jQ, 203-2 14.Whalen, T. (1953). Wives of alcoholics: Four types observed in a family service agency.Ouarterly Journal of Studies on Alcohol, j4, 532-64 1.Whitaker, C. A., & Keith, D. V. (1981). Symbolic-experiential family therapy. In A. S.Gurman & D. P. Kniskern (Eds.), Handbook of family therapy (pp. 226-266). NewYork: Brunner/Mazel.White, M. (1992). Deconstruction and therapy. In D. Epston & M. White (Eds.),Experience, contradiction, narrative and imagination (pp. 109-15 1). Adelaide, SouthAustralia: Duiwich Center Publications.White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York: Norton.Whittingham, B. (Ed.) (1987). Report of the task force on alcohol and drub abuse in theworkplace. Victoria, B.C.: Government of British Columbia.Williams, A. M., & Miller, W. R. (1981). Evaluation and research on marital therapy. In G.P. Sholevar (Ed.), The handbook of marriage and marital therapy (pp. 373-4 15). NewYork: Spectrum Publications.Wilson, G. T., Edredge, K. L., Smith, D., & Niles, B. (1991). Cognitive-behaviouraltreatment with and without response prevention for bulimia. Behavioural Researchand Therapy, 29, 575-583.Wolin, S. J., Steinglass, P., Sendroff, P., Davis, D. I., & Berenson, D. (1975). Maritalinteraction during experimental intoxication and the relationship to family history. InM. Gross (Ed.), Alcohol intoxification and withdraw/experimental studies (pp. 645-653). New York: Plenum.Wynne, L. C. (1988). An overview of the state of the art: What should be expected incurrent family therapy research. In L.C. Wynne (Ed.), The state of the art in familytherapy research: Controversies and recommendations (pp. 249-266). New York:Family Process Press.Zachery, R. A. (1986). Shipley Institute of Living Scale: Revised manual. Los Angeles, CA:Western Psychological Services.Zucker, R. A., & Gomberg, E. S. L. (1986). Etiology of alcoholism reconsidered: The casefor a biopsychosocial process. American Psychologist, .4j, 783-793.Zung, B. 3. (1980a). Unidimensionality of the Michigan Alcoholism Screening Test. BritishJournal of Addiction,., 389-391.Zung, B. J. (1980b). Factor structure of the Michigan Alcoholism Screening test (MAST).Journal of Clinical Psychology, ., 1024-1030.227Zung, B. J. (1982). Evaluation of the Michigan Alcoholism Screening Test (MAST) inassessing lifetime and recent problems. Journal of Clinical Psychology, ., 425-439.Zweben, A. (1986). Problem drinking and marital adjustment. Journal of Studies onAlcohol,.42, 167-172.Zweben, A., & Pearlman, S. (1983). Evaluating the effectiveness of conjoint treatment ofalcohol-complicated marriages: Clinical and methodological issues. Journal ofMarital and Family Therapy, 9, 6 1-72.Zweben, A., Pearlman, S., & Li, S. (1988). A comparison of brief advice and conjointtherapy in the treatment of alcohol abuse: The results of the Marital Systems study.British Journal of Addiction, ., 899-9 16.228APPENDIX ASchedule of ExST Training Events andAlcohol and Drug ProgramTherapist Appraisals of ExST Training Experiences229Schedule of Training ActivitiesFebruary, 1987 — Prince George2 day presentation to Northern Region Alcohol and Drug Program Directors onMarital and Family Therapy and the relationship between alcohol dependencyand marital and family systems.May, 1987 — Prince George4 day training workshop to 22 alcohol and drug counsellors in level 1,Western Family Learning Institute training commencement.July, 1987 — Prince George4 day training workshop to 22 alcohol and drug counsellors in level 1,Western Family Learning Institute training continues.September, 1987 — Prince George4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training completed, Certificate awarded.September, 1987 — to March, 1988 — Prince George4 counsellors on Level II, Western Family Learning Institute training.Submission of 15 video—tapes for supervision and feedback.January, 1988 — Kelowna2 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute Training commencement.February, 1988 — Victoria2 day training workshop to 18 alcohol and drug counsellors in level 1,Western Family Learning Institute training commencement.March, 1988 — Vancouver2 day training workshop to 25 alcohol and drug counsellors (exploratoryworkshop).September, 1988 — Kelowna4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training continues.October, 1988 — Victoria4 day training workshop to 18 alcohol and drug counsellors in level 1,Western Family Learning Institute training continues.230November, 1988 — Kelowna4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training completed, certificates awarded.November, 1988 — present — KelownaLevel II training continuing to interested counsellorsDecember, 1988 — Victoria4 day training workshop to 18 alcohol and drug counsellors in level 1,Western Family Learning Institute training completed, certificates awarded.January, 1989 — Maple Ridge4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training commencement.March, 1989 — Maple Ridge4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training continues.May, 1989 — Maple Ridge4 day training workshop to 20 alcohol and drug counsellors in level 1,Western Family Learning Institute training completed.October, 1989 — VancouverAddress to the British Columbia Psychological Association. Presentation ofthe Experiential Systemic Treatment of Alcoholic Families at the annualconference.Vancouver has requested Level I training for its alcohol and drugcounsellors but because of our interest in research, we have not entered into acontract.Summit clinic in Duncan are implementing levels 2, 3 and 4 training forsenior staff who gould then become training associates of the Western FamilyLearning Institute. These senior staff would train and supervise their ownstaff at the Summit Clinic.231Participant Evaluations of Experiential Systemic Therapy Training: SelectedHighlights— I’ve been in practice for 15 years and have never had training to comparewith it. All practising counsellors should have opportunities to take it.— The training really brought my understanding of family systems into concreteexperience.— This training has brought me a whole view concept of therapy, the tools touse and a greater understanding of what the tools are used for and where.— I gained a new awareness of myself [which] I will surely benefit my activityas a therapist. I would highly recommend this training.— This model and training gave me new and creative tools for working withcouples. I have renewed excitement about working with families.— This training integrates existing skills into systemic therapy.— The training demonstrates and teaches counselling skills that focuses oninterconnections between people.— Presents a “valuable and meaningful approach to therapy, adjustment andchange.— The model enhances effectiveness and efficiency in client work.— The approach provides a frame work enabling a client and therapist toexperience change.— Easily incorporated into existing therapy styles with families, couples andindividuals.— Provides many effective ways to improve my therapy.— I already use these techniques and materials in all of my family sessions atwork.— The training has offered me specific skills and a theoretical foundation forthose skills to be used.— I have learned valuable ‘landmarks’ in therapy sessions so that I don’t feelso much a helpless onlooker in family therapy as things develop. I havegained control and in so doing, I have been able to relinquish meaningless‘control’ devices that I used to employ.— The theoretical and practical applications are a very important part of thisprogram.—Many of the therapies I have used in the past are now under one frameworkwith the techniques and tools of the trade being utilized to the utmostclarity.232— I enjoyed the workshops because you get a chance to learn as your clientslearn.— Encouraged to use the blocks our clients present and change the process inour sessions.— My scope of therapy is enlarged due to new ways of thinking and doing.— This model works for individuals, couples and families.— I have renewed enthusiasm for family systems work.— What an excellent model!! Hard to improve this.— Program offers a balance between learning theory and applying principles on apractical level.— Although its focus is professional, personal growth is inevitable due to theexperiential components.— The training offers a solid knowledge base in systemic family therapy.— Much needed skill building and opportunity to practice.— Opens the heart as well as the mind.APPENDIX BExperiential Systemic Therapy Overview233234EXPERIENTIAL SYSTEMIC THERAPYAn Overview(revised March, 1991)ByJohn D Friesen, Ph.D.Darryl N Grigg, M.A.Jennifer A Newman, M.A.Dept. of Counselling PsychologyUniversity of British ColumbiaVancouver, B.C.235Experiential Systemic TherapyExperiential Systemic Therapy (ExST) is an integrativetreatment approach and paradigm of psychotherapy that synthesizesindividual and family therapy concepts and techniques. ExST isbased on a unified set of assumptions and concepts which applyequally to individuals, couples and families. The model is not apatching together of different theories of psychotherapy, but isa broadly based approach to change and health. It attends to themultiple layers of experience such that theory and practice arewoven together. As a therapy, ExST emphasizes the importance ofpreserving the contextual nature of relationships and avoidsstripping events from their contexts through the use ofreductionistic approaches to life situations.Basic ConceptsThe primary component around which all the various conceptsof ExST revolve is that of relationships. As such, ExST viewsrelationships as the bedrock of human existence and understandsthe human condition as an intricate web of systemic connections.ExST is concerned with how relationships manifest in theintrapersonal domain as well as the interpersonal and largercontextual fields. In this respect, ExST shares with the Britishobject relations theorists ( Kernberg, 1976; Klein, 1991; Kohut,1977; Mahier, 1975) an emphasis on intrapersonal relations. ExSTalso maintains close conceptual ties with interpersonal andexistential theorists (Homey, 1939; Kiesler, 1982; May, 1967;Sullivan, 1954; Yalom, 1980) and systemic and ecological236advocates (Bateson, 1972, 1979; Bronfenbrenner, 1979, Keeney,1983; Whitaker & Keith, 1981). In agreement with the views ofthese major theoretical positions, ExST holds that the humanexperience is relational in nature. Human beings from infancyonward form, revise and reform their identities based on theirrecurrent relationships with others.ExST concurs with Attachment Theory (Bowiby, 1969, 1973,1988) asserting that optimal human personality development andfunctioning is predicated upon the formation of a healthyattachment experience in infancy and childhood. The earlycaretaking relationship with mother and father is of immensephysical, emotional, intellectual and social importance since itinfluences both the child’s intrapersonal and interpersonal lifejourney. These early experiences with caregivers areparticularly significant as they are the foundation upon whichrelational rules and codes of conduct continue to develop.Caregiver/child relationships are internalized by the child andplay major roles in the ensuing development of substantiverelational themes which integrate a person’s ongoing relationalexperiences in a meaningful fashion. That is, childhood andlater life experiences combine to contribute to dynamicinterpersonal and intrapersonal relational patterns that have astheir undercurrents the themes of love and influence.As a therapy, ExST consists of three interlockingdimensions: the experiential, the symbolic, and the systemic.The systemic dimension goes far beyond the common view of systemsas referring mainly to couples and families. Our view includes237progressively larger and more complex systems such ascommunities, nations and the globe. In addition, we construe thehuman being as a system of parts including the psychological(consisting of the cognitive, affective and behaviouraldimensions) and physiological (involving the neural, respiratory,digestive, circulatory and immune systems). These systems andsubsystems may be meaningfully dissected even further to includethe molecular, atomic and subatomic levels and beyond. At eachlevel of analysis, any given system represents a synthesis ofsmaller constituent subsystems. In turn, each system combines toform systems larger than itself. These systems and subsystemsare interdependent and fit coherently together such that theintegrity of the whole is preserved. The universe is thusconceived of as an all encompassing collection of isomorphicallyrelated systems. We agree with Bateson (1979) that all livingbiological creatures are connected by what constitutes a “sacredunity of the biosphere” (p. 19). This eco-systemic perspectiveprovides the framework for ExST and contributes to its broadtheoretical range and integrative potential.The symbolic dimension of ExST adds an element of solemnformality and playful curiosity to the therapeutic experience.Symbols are metaphors and possess meaning at multiple levels. Assuch, they may be words, actions and objects, as well asfeelings, thoughts and creations. We view symbols as the basicbuilding blocks of human experience and communication.The use of symbols in therapy is an important attempt toanalogically access perceptual experience. In this regard,238Bateson (1972, 1979) noted that linguistic thought is structuredin a digital code and perceptual experience in an analogic codeand is communicated in the form of models, metaphors, analogies,stories and rituals. Consequently, human beings relate almostentirely through analogic forms.ExST views the act of therapy as symbolic. Therapy isrecognized as a culturally sanctioned change ritual or rite ofpassage and all activities connected to it are viewed asintrinsically meaningful within a symbolic frame. In therapy,actual symbolic objects are often employed to provide the vehicleby which the inner world of the client is explored.Alternatively, symbols may be helpful in describing theexperience of interpersonal relationships between groups ofpeople. Symbols may also be employed in therapy to representsymptoms such as depression or alcohol dependency. Similarly,they may be used to refer to relationships that clients have withinanimate objects, institutions, culture or ways of life.Symbols provide a means to present ideas at a more experientialand indirect level, making them easier for people to accept andassimilate. At one level, that which is enacted in therapy isviewed as symbolic of that which is enacted in other lifesituations. Relational changes experienced directly andsymbolically in the therapeutic arena translate into changesbeyond the therapeutic setting.In ExST sessions, we have successfully used a variety ofsymbols such as deeply burnt out hulks of candles to representinner emptiness and personal bankruptcy; logging trucks burdened239with heavy logs to symbolize the pressures of life; a doormat asa symbol for a mother being trampled by her son or spouse; a cupand saucer to represent a relationship pattern between a man anda woman; rubber snakes to represent vile and underhandedbehaviours; a beer bottle to symbolize drinking behaviour; saltto represent curative activity; a burning candle to representuniversal light; an owl as a symbol of being open to informationand wise; a song bird to symbolize cheerfulness and spontaneity;and finally a donkey to represent slowness and lack offriendship.The third dimension of ExST is the experiential focus.Along with Frieda Froimu-Reichmann (May 1967), and Yalom (1980),ExST insists that clients do not need an explanation, they needan experience. The deepest and the most profound form of knowingresults from experience rather than dialogue or didacticinstruction. ExST observes that people are not only thinking andverbal beings, they are also acting and feeling beings.Consequently, the potential intensity of the therapeuticexperience is greatly limited by a passive physical stanceprovided for by verbal discourse alone. ExST is an activetherapy which engages the client’s entire body in affectivelyladen and cognitively significant movement thereby increasing theintensity of the client’s therapeutic experience.In ExST, clients externalize aspects of their internalizedrelational worlds, explore and change their substantiverelational themes, and transform their relationship patterns. Bydramatizing aspects of self-in—relationship, systemic experience240is intensified leading to greater awareness of self andalternative ways of being. Ultimately, the purpose of theexperience is to provide the opportunity and motivation for theclient to engage in relational novelty which involves the actualtransformation of particular relationships in the here and now.This process engages the client in spontaneous and creativeaction in which the therapist acts as facilitator or processguide. In therapy, the client is encouraged to be self-attentiveand to listen to the inner voice and the heart. Some therapeuticprocedures are similar to the dramatic technique of the two chairdeveloped by Pens (1973) in which the client enacts andinteracts with different aspects of self. In this technique, theclient moves from chair to chair and addresses aspects of self indialogue with other aspects of self and others. ExST extendsthis technique to include the externalization of aspects of selfin relationship to such issues as symptoms, problems, relationalthemes or relationship patterns. It is common in ExST to sculpt,to use empty chairs, to conduct role reversals, and to enactreciprocal metaphors. Clients are also often invited to imagine,to draw, make collages, paint and dance if they so desire. Theyare asked to engage in activities quite unlike their every dayworld. They are given a context to explore new ways of being, tochallenge old recursive patterns and to emancipate themselvesfrom the shackles that restrain them.In ExST, we are concerned not only with intrapsychicexploration, but also with the exploration of couple and familydynamics including intergenerational legacies. This is in sharp241contrast to the work of Pens (1973) who focussed solely on theindividual’s intrapsychic process in order to increase awarenessand contact with the world. ExST is a systemically sensitivetherapy that pays considerable attention to the between-peopleprocess and the social conditions in which the client functions.As such, changes in the relationship are contextualized andunderstood as isomorphic to other levels of the system. Issuesrelating to social, interpersonal, intrapersonal and spiritualchange are woven into the unfolding therapeutic story.PersonalityWe agree with Sullivan (1953) that personality can best beunderstood as the result of the “relatively enduring pattern ofrecurrent interpersonal situations which characterize a humanlife” (p.11). It is within this interactive relational contextthat humans construct ways of being in the world. These ways ofbeing arise and develop over time and are nourished byexperiences with others to form personal identity. It is ouridentity which helps us to understand events, make plans and haveexpectations.In addition to viewing personality relationally, ExST takesa wholistic perspective with respect to human functioning.Arguments over the primacy of cognition, affect and behaviour areavoided. The parcelling of human experience into fragmentedparts and the resulting reductionistic therapy practice to whichthis compartmentalization gives rise, does little to integratepeople’s frequently disjointed existence (May, 1967). ExST242maintains that all three domains of experience namely cognition,affect and behaviour offer a wealth of information to awareness.From our perspective, streams of thought, feeling and behaviourflow concurrently. They are valued in concert and are equallyemphasized in therapy. ExST works toward a harmonic balance ofthe domains that constitute experience. A peaceful co—existenceof thoughts, feelings and behaviour is sought, while neglect,denial, suspicion and repression of any stream of experience isgrist for therapy.ExST holds that a person’s personality structure must beflexible or plastic in order to both grow along with andinfluence changes in the environment. In regard to therelationship between structure and environment, Maturana (1978)has coined the term “structural coupling” to refer to the processof ongoing relationship building with the environment. Maturana(1978) has referred to the interlocking conduct of individualsresulting from mutual structural coupling, such as occurs infamilies, as a “consensual domain”. It is in the consensualdomain that people learn about self, the world around them, andthe meaning of behaviour. In this way, people continue to learnhow to punctuate the flow of experience. This punctuationultimately determines the “reality” that is lived, Consequently,ones sense of personhood arises with experiences of others.These experiences inform each of us about who and what we are,and what we can and can not do without challenging the integrityof our sense of identity.243The Role of SymptomsIn ExST, symptoms are viewed as indications of relationshipproblems. Symptomatic relationships are generally characterizedby restrictive, rigid and repetitive patterns of interaction.When using the term “relationship” in this context, we do notnecessarily mean to infer only living human relations, althoughthis may be the case. Rather, we mean to imply that theidentified symptom-bearer has a relationship with the “problem”as do other members of the symptom bearer’s family or community.For example a “problem drinker” has a particular relationshipwith alcohol as does his/her spouse, family,