EMBRACING A MULTI-PERSPECTIVE VIEW OF THERAPEUTIC ALLIANCE: A PROCESS-ORIENTED STUDY OF ALLIANCE FORMATION AND MANAGEMENT IN COUPLES THERAPY By Sara Elizabeth Timmons A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Human Development and Family Studies 2011 ABSTRACT EMBRACING A MULTI-PERSPECTIVE VIEW OF THERAPEUTIC ALLIANCE: A PROCESS-ORIENTED STUDY OF ALLIANCE FORMATION AND MANAGEMENT IN COUPLES THERAPY By Sara Elizabeth Timmons Therapeutic alliance is a widely researched topic in psychotherapy literature due to its positive relationship with treatment outcome. Recently, researchers have started to address therapeutic alliance in couple and family therapy, but have struggled to identify the essential elements and therapist and client behaviors that are influential of positive alliance formation with multiple family members. Review of the literature shows that research needs to measure alliance over time in order to see how it evolves in conjoint therapy, and needs to incorporate the four perspectives of alliance: the therapist, each member of the couple, and independent observer. This study adopted a process oriented approach to explore therapeutic alliance formation and maintenance with couples. Two doctoral level therapists were followed over the course of treatment with a total of 5 couples. Data were collected through semi-structured interviews and SOFTA (Friedlander, Escudero, & Heatherington, 2006a) observational and self-report measures to identify key themes and patterns of alliance development based on the experiences of each member of the therapeutic system and an independent observer. Thematic content analysis of the interview data identified that therapeutic factors pertaining to the training nature of the clinic and student status of the therapists had initial negative influences on the alliance formation. Additionally, findings suggested that the client factors related to level of relationship distress, interpersonal skills and gender contributed to the formation and progression of alliance over the course of therapy. Therapist characteristics such as personality and gender, as well as the interactive factors related to the couple-therapist goodness of fit and therapist skills also influenced the therapeutic relationship. The results from the SOFTA self-report data identified how alliance progressed throughout treatment for each couple. Alliance configurations depicted alliance patterns that stabilized around the fourth session followed by increased fluctuations after the sixth sessions. Overall, therapeutic alliance patterns were most heavily influenced by the level of relationship distress and strength of the within-couple alliance. How well therapists promoted individual alliances with both partners and the within-couple alliance contributed to the alliance ratings. The results of the study have important implications for the effective formation and management of alliance with couples. Discussion of the findings connects previous research to the current results to provide a greater understanding of how therapeutic alliance was formed with the couples in this study. Clinical and training implications are given for supervisors and therapists working with couples, as well as suggestions for future research on therapeutic alliance formation and management with couples. Copyright by SARA ELIZABETH TIMMONS 2011 DEDICATION If you raise your children to feel that they can accomplish any goal or task they decide upon, you will have succeeded as a parent and you will have given your children the greatest of all blessings. –Brian Tracy This dissertation is dedicated to my parents, Daniel and Lynne Timmons. Without them, this work would not be possible. Their continuous praise taught me to believe in myself, their unwavering guidance allowed me to identify my dreams, and their unconditional love supported me through this challenging, yet amazing journey. They sacrificed more than I will ever know to ensure that I achieved this remarkable accomplishment. For this, I am forever grateful. v ACKNOWLEDGEMENTS To get through the hardest journey we need take only one step at a time, but we must keep on stepping. –Chinese Proverb I have many people to thank for helping me to ―keep on stepping‖ throughout this indescribable journey. First and foremost, I would like to thank my Advisor, Dr. Tina Timm. Your countless hours of editing and talking me through my own madness was appreciated more than you will ever know. Thank you for helping me see the trees through the forest and find my thoughts through all of the data. Without you, I am not sure I would have ever stopped writing! I would also like to thank my other committee members for their guidance in this process. To Dr. Adrian Blow, thank you for providing me with your knowledge and resources on this topic that stimulated my thinking and influenced this study. To Dr. Esther Onaga, thank you for your support during the beginning of my doctoral studies. You introduced me to qualitative research and I so much enjoyed seeing your passion for pursuing community driven research. To Dr. John Mooradian, your friendly support and kind words were instrumental to this process. I hope to have continuing contact with you in the future. Thank you all, for your influence on my academic and research development. I would also like to thank the Graduate School and Department of Human Development and Family Studies for awarding me the Dissertation Completion Fellowship, as well as many other instrumental forms of support throughout my time here. Also, the Marriage and Family Therapy Program and Couples and Family Clinic had a pivotal role in my academic and clinical development. Thank you for allowing me to pursue this degree and giving me the resources to achieve it. vi To my fellow MFT friends. The memories we have shared during our time together will always be regarded fondly. The experience of becoming a therapist and ―Ph.D‖ is only understood by those who do it. The hours of sacrifice, hard work and continuous self-doubt both weaken and strengthen our spirits. It‘s a roller coaster of a ride, and I am glad I was on it with you. Thank you for the laughs, tailgates and experiences that I will never forget. To my family. Words cannot describe how much I appreciate and needed your support throughout my life. I mean this when I say it: This accomplishment and degree is just as much yours as it is mine, for without your confidence, encouragement and love for me, I would not have had the courage and ability to do this. To my brother, Christopher. I said it before, and I will say it again: I admire and respect the success you have achieved in such a short amount of time. I will continue to deny your claims that I am smarter than you as I know that we share the same level of commitment, motivation and hard work that have gotten us both to where we are today. Thank you for being my big brother and paving the way for me. I also want to thank you for finally giving me the sister I always wanted. And, thank you to Lisa, for taking care of my brother and adding another female voice to the family. To my parents. Every day I become increasingly thankful that I am your daughter. You both succeeded at the hardest job in life which is to raise children with confidence, compassion and perseverance to continually pursue the goals and purpose of their life. We learned this from watching you do the same repeatedly, even when it was easier to give up. Thank you for allowing me to witness your tenacity and determination to succeed and be the best you could be for us. I know it was not easy to let us go and watch as we developed into our ―selves,‖ but I thank you for doing it anyway even when it was most painful. To my ―Pops:‖ you taught me the meaning of character and respect, and how to ―lighten up‖ when I was taking myself too vii seriously. You taught me humor and the importance of being humble when it is easy to forget who I am. Your laugh makes me smile, and your smile makes me laugh. Thank you for being ―you.‖ To my Mom. I can‘t think of a time in my life when you didn‘t tell me how proud you are of me. I want to take this time to tell you how proud I am of you-For doing what others told you couldn‘t, for being a strong woman who fights for herself even when it seems impossible, and for raising me to know my mind and how to use it. Thank you for being my Mom when I needed you, my friend when I wanted you, and for ensuring that I accomplished all that you could not. I also want to thank my Grandmother for her legacy of strength and resiliency that continues to evolve throughout the generations. I am honored and thankful for the time I continue to have with such a remarkable woman. Thank you, for raising my Mom, and teaching us what it means to be women. To my Bryan. I truly believe that fate encouraged us to move here at the exact same time to begin the next phase in our lives. I am so incredibly thankful to have met you in that Fall of 2005, and to have had the opportunity to grow with you over the past few years. You have been here to support me, encourage me, take care of me (in oh so many ways), and handle things when I could not. Your kindness and patience are more than I deserve at times, and I want to thank you for loving me even when it is hard to do. I am a better person with you than I am without you, and I am so lucky that you ―chose‖ me. Thank you for ―feeling like home to me.‖ Oh, and thank you for all of the amazing parking passes, football and basketball tickets throughout the years. Those experiences helped to keep me sane! To the therapists who willingly participated in this research study. Thank you for allowing me to study your work even when it was an extra responsibility that you did not need (or want) to have. It was an invaluable experience to watch and learn from you. I hope that I am viii able to return the favor. Finally, to the couples who let me in on their journey as they faced difficult issues and challenges. Their words and experiences contributed to this study and will, hopefully, lead to more research that will positively impact all couples. ix TABLE OF CONTENTS LIST OF TABLES ....................................................................................................................... xvi LIST OF FIGURES…………………………………………………………………………….xvii CHAPTER I: INTRODUCTION .................................................................................................... 1 Statement of the Problem .................................................................................................... 3 Purpose of the Study……… ………………………………………………………...........3 Theoretical Framework…… ……………………………………………………………...4 Process-Oriented Approach to Methodology…… ………………………………..4 Conceptualization of Alliance in Couple and Family Therapy… ………………...6 Systems of Alliance in Couple and Family Therapy….…… .…………….9 Individual Members' Alliance with Therapist… ........ …………..10 Alliance between Family Members… ......................................... .11 Family's Collective Alliance… ..................................................... 12 System for Observing Family Therapy Alliances……....………………..13 Engagement in the Therapeutic Process..… ….………………….13 Emotional Connection to the Therapist..….… .………………….13 Safety within the Therapeutic System….…… ..................... ……14 Shared Sense of Purpose…..…..… ...... ………………………….14 CHAPTER II: REVIEW OF THE LITERATURE…… …..…………………………………….17 History of the Conceptual Development of Therapeutic Alliance….…………............... 17 Defining Alliance in Individual Psychotherapy…..…… ..... …………………….18 Review of Research on Therapeutic Alliance in Couple and Family Therapy………......20 Defining in Alliance in Couple and Family Therapy…...…. ……………………21 Influence of Therapeutic Elements on Alliance: Historical, Personal and Interactive Factors… ............................. ..……………………………………………………………22 Therapist Factors… ....... …………………………………………………………23 Interpersonal Skill Component… ........... ………………………………..24 Intrapersonal Dimension…… ........... ……………………………………25 Therapist Gender ....................................................................................... 26 Interactive Elements........... …….………………………………………..27 Client Factors……… ........... …………………………………………………….27 Current Level of Distress and Presenting Problem…… ........... …………28 Early Client Relationships… ........... …………………………………….29 Interpersonal Styles… ........... ……………………………………………30 Client Gender.………………………………… ........... …………………30 Split Alliances………………………… ....................... …………………………………32 Role of Therapeutic Alliance in Process and Outcome in Couples Therapy…… ........... .35 Perspectives of Therapeutic Alliance…….......... .……………………………………….36 Timing of Measurement… .......... ……….……………………………………………….39 x Therapeutic Alliance Measures…..….......... …………………………………………….44 Brief Overview of Alliance Scales for Individual Psychotherapy… ........ ………45 Couple and Family Therapy Alliance Scales……… ....... …….…………………45 Working Alliance Inventory-Couples………………..…………………..45 Integrative Alliance Scales……… ............. ..……………………………46 System for Observing Family Therapy Alliances…… .............. .………..47 Conclusions and Rationale for Current Study……… ............... ………………………...49 CHAPTER III: METHODOLOGY… .. …………………………………………………………50 Research Design…... …………………………………………………………………….50 Rationale for Qualitative Research Design…… ... ………………………………………51 Data Collection… ... ..……………………………………………………………………52 Research Site…… . ………………………………………………………………52 Recruitment Procedures… ... …………………………………………………….52 Therapists… .. ……………………………………………………………52 Couples…… …………………………………………………………….53 Informed Consent….. ……………………………………………………54 Compensation… . ………………………………………………………..54 Participants………………… ...... ………………………………………………..55 Therapists…… . ………………………………………………………….55 Couples…………… . ……………………………………………………55 Couple 1…….…… ….…………………………………………...56 Couple 2… ….……………………………………………………57 Couple 3…… .……………………………………………………57 Couple 4…. ….………………………………………….………..57 Couple 5…. ……………………………………………………....58 Procedure………… . …………………………………………………………….58 First Session…… . ……………………………………………………….58 Points of Data Collection…… .. …………………………………………59 Measures .............. ……………………………………………………………………….59 Quantitative Measures…… .. ……………………………………………………59 Revised Dyadic Adjustment Scale… ... ………………………………….60 System for Observing System for Observing Family Therapy Alliance Self-Report… ... ………………………………………………………….61 Qualitative Measures……………… .. …………………………………………..64 Observational Measure: SOFTA Observational Scale… ... ……………..64 Semi-Structured Interviews………………… .. …………………………66 Data Analysis………… .................. ……………………………………………………..68 Quantitative Measures: RDAS and SOFTA-s…… .. ……………………………68 Observational Measures: SOFTA-o…… .. ………………………………………68 Semi-Structured Interviews… .. …………………………………………………72 Coding… .. ……………………………………………………………….73 Trustworthiness……… ... ………………………………………………………………..74 Credibility…… . ………………………………………………………………....74 Reflexivity of the Researcher……….. …………………………………..75 Self-Reflective Journal…….. ……………………………………………75 xi Theoretical Sensitivity… . ……………………………………………….78 Examination of Previous Findings…………… ... ……………………….78 Triangulation………… . …………………………………………………79 Debriefing Sessions……… . …………………………………………….80 Transferability……… .............. ………………………………………………….80 Dependability…...…… …………………………………………………………..81 Confirmability……… .. ………………………………………………………….82 Audit Trail…… ………………………………………………………….83 CHAPTER IV: RESULTS…….. ………………………………………………………………..84 SOFTA-o Inter-Rater Reliability…… .. …………………………………………………84 Key Findings… ... …………..……………………………………………………………85 Research Question 1: What Therapeutic Components Contribute to Alliance in Couples Therapy?............................. ............... ...............................................................................87 Key Components of Alliance…………… ... …………………………………….87 Therapeutic Factors…. .. .……….………………….…………………….87 Training Clinic……………………………………… . ………….87 Pace of Initial Stage of Treatment… .. …………………………..89 Client Factors… .. ………………………………………………………..90 Expectations of Therapy/Therapist…..… ... ….………………….90 Relationship Distress… ... ……………………..………………...91 Interpersonal Styles and Past Experiences..….………… .... …….93 Client Gender…… .... ………………..…………………………..94 Therapist Factors… .... ……………..…………………………………….97 Characteristics… .... ……….………………………………….….97 Intrapersonal Dimension .............. ……………………………….97 Demographics…….. .............. …………………………………...99 Race/Ethnicity….……… ............ …………….………….99 Age…………….…… ........... ………………………..…101 Gender……………… ............ ….………………………102 Conceptualization of Therapeutic Alliance with Couples.. ...... ..104 Interactive Factors…..… ....... …………………………………………..105 Goodness of Fit: Style of the Therapist… ...... …………………105 Therapist 1… ..... ……………………………………….108 Therapist 2…… ..... …………………………………….113 Therapist Skills………………… ...... ………………...………..116 Therapist Competency…… ...... .……………………….116 Professionalism of Therapist……………………117 Meaningful Feedback……………….…………..119 Summary of Key Components of Alliances in Couples Therapy………………………121 Research Question 2: How do Therapists Manage the Multiple Systems of Alliance in Couples Therapy?......... ........... .......................................................................................121 Alliance with Individual Partners………… ………...………………………….122 Key Components of Alliance with Individual Partners…… ….………..122 Engagement Level of Partners……… .. ………………………..122 Therapist Self-Report and Observational Scores… .. …..125 xii Gender and Engagement Level…… . …………………..126 Split Alliances…… . ……………………………………………126 Emotional Connection to the Therapist……… .. ………127 Therapist Self-Report and Observational Scores ...... …..130 Gender and Emotional Connection to Therapist……… . 130 Couple Dynamics… .. …………………………………..131 Synergy Effect…… ……………………………………………133 Behaviors that Promote Individual Alliances…… …………….135 Promote Emotional Connection………… ……………..135 Promote Engagement…… ……………………………..137 Implement Individual Sessions………… .. …………….139 Within-Couple Alliance… ........................................... ………………………...141 Key components of Within-Couple Alliance…… .... …………………..141 Feelings of Safety………… . …………………………………..142 Therapist Self-Report and Observational Scores for Safety… .................................................................... …..144 Shared Sense of Purpose… ............. ……………………………146 Therapist Self-Report and Observational Scores for Shared Sense of Purpose ............................................................. 148 Gender and Shared Sense of Purpose…… ............ …….149 Behaviors that Promote Within-Couple Alliance… ... …………149 Promote Safety……………………………………….....149 Slowing the Process: Moving at Client's Pace .... 150 Provide a Safe Place ……………………………151 Effectively Manage Conflict… . ………………..152 Promote Shared Sense of Purpose……… . …………….154 Encourage Commitment: Establishing Shared Needs and Goals ................................................. 154 Emphasizing Common Feelings and Validating Perspectives......................................................... 155 Uniting the Couple……… .. ……………………157 Collective Couple Alliance…………………………… . ………………………158 Repairing the Alliances… . ……………………………………………………..161 Alliance Imbalances……… .. ………………………………………….162 Therapeutic Ruptures……………………………………… ............... ...163 Summary of the Multiple Systems of Alliance in Couples Therapy… ................. …….166 Research Question 3: How does Therapeutic Alliance Evolve over Time in Couples Therapy?.......................................................... ........................................................ .....1678 Within Case Analysis of Alliance Formation and Management…… ........... ….167 Couple 1…………………………………………………… ............ …..167 Extratherapeutic Factors… ........... ……………………………..168 Client Factors… ........... ……………………………...................169 Individual Dynamics………… ........... ……..…………………..170 Couple Dynamics… ........... …………………………………….170 Individual Alliances with Partners…… ........... …………...……170 Imbalance in Alliance… ........... ………………………..170 xiii Split Alliance… ………………………………………...171 Within Couple Alliance… ........... ………………………...……171 Couple 2…… ........... …………………………………………………...172 Extratherapeutic Factors… ........... ……………………………..173 Client Factors………… ……………………………………...…173 Interpersonal Dynamics… …………………………...…174 Couple Dynamics……………………………………….174 Focus on Progress… ........... …………………………………....174 Synergy Effect… ……………………………………………….175 Couple 3……… ... ……………………………………………….……..176 Client Factors……… …………………………………….……..177 Interpersonal Skills… …………………………….…….177 Couple Distress…… ……………………………………177 Individual Alliances with Therapist…………………………….178 Therapeutic Rupture…………………………………….178 Split Alliance… ……………………………………...…178 Synergy Effect…… …………………………………….179 Within-Couple Alliance…… …………………………………..179 Incongruent Perspectives…… …………………………………180 Couple 4…… …………………………………………………….……..180 Congruent Perspectives…………………………………………181 Within-Couple Alliance…… ……………………………...……182 Individual Alliances with Therapist………………………….…182 Split Alliance…………… ........... ………………………...……182 Synergy Effect……………… ............ …………………………183 Couple 5…… . ………………………………………………………….184 Strong Collective Couple Alliance…… ……………………….185 Intact Within-Couple Alliance……… …………………………186 Summary of Couple Alliance Ratings over Time………… ........... ……………………186 CHAPTER V: DISCUSSION…… …………………………………………………………….188 Research Question 1: Key Therapeutic Components of Therapeutic Alliance in Couples Therapy… ............... ………………………………………………………………..…..188 Therapeutic Factor: Nature of Training Clinic……… ………………………....189 Client Factors……… …………………………………………………………...190 Client Expectations of Therapy/Therapist…… …………………...……190 Relationship Distress …………………………………………………...192 Interpersonal Characteristics and Past Experiences…………………….193 Client Gender….. ……………………………...………………………..194 Therapist Factors… ........... ……………………………………………………..197 Intrapersonal Dimension…… …………………………………...……...198 Interactive Factors… ........... ……………………………………………………201 Goodness of Fit: Style of the Therapist… …………...…………………201 Therapist Skills…… ……………………………………………………202 Professionalism…… ……………………………………………202 Meaningful Feedback… ………………………………………..203 xiv Research Question 2: Managing the Multiple Systems of Alliance in Couples Therapy204 Individual Alliances with Partners…………………………………...…………205 Engagement and Emotional Connection… …………………………….206 Synergy Effect… ……………………………………………………….210 Within-Couple Alliance… ……………………………………………………...211 Safety………………………………………………………………...…211 Shared Sense of Purpose… …………………………………………….212 Collective Couple Alliance… ………………………………………………….213 Monitoring and Repairing Alliances…… . ……………………………………..213 Research Question 3: Couple Alliance Patterns over Time……… . …………………...215 Notable Patterns of Therapist and Observational Ratings… …………………..215 Alliance Patterns……………… ……………………………………………….217 Initial Alliance…… …………………………………………………….217 Alliance Fluctuations… …………………………………………...……217 Clinical Implications ....................... ……………………………………………………220 Implications for Therapist Training and Supervision… ........ ………………………….225 Limitations of the Current Study… ........... …………………………………………….227 Strengths of the Current Study…… ............ ……………………………………………227 Suggestions for Future Research… ........... …………………………………………….228 Conclusion……… ............. …………………………………………………………….230 APPENDICES ...………………………………………………………………………………..232 Appendix A: Script for Client Recruitment...… ……….……………………………….232 Appendix B: Therapist Consent for Participation in Research Study...... ……………..233 Appendix C: Consent for Participation in Research Study... …………………………..236 Appendix D: Confirmation of Compensation for Participation in Research Study…….240 Appendix E: Revised Dyadic Adjustment Scale… .. .……………………...…………..241 Appendix F: SOFTA-S (Client Version)……… .. .…………………………………….243 Appendix G: SOFTA-S (Therapist Version)…… .. ……………………………………244 Appendix H: SOFTA-O (Client Version)…… .. ……………………………………….245 Appendix I: SOFTA-O (Therapist Version)… .. ……………………………………….250 Appendix J: Therapist Interview………………………………………………………..255 Appendix K: Client Interview…………………………………………………………..256 REFERENCES…… …………………………………………………………………...……….258 xv LIST OF TABLES Table 3.1. Demographics of Therapists………………………………………………………….55 Table 3.2. Demographics of Couples ............................................................................................ 56 Table 3.3. Points of Data Collection ............................................................................................. 59 Table 3.4. Client Interview Questions Paired with Associated Research Questions and Alliance Components…… ... ……………………………………………………………….……………..67 Table 3.5. Client Interview Questions Paired with Associated Research Questions and Alliance Components .................................................................................................................................. 68 Table 4.1. Overview of Research Questions and Themes ............................................................ 85 Table 4.2. RDAS Scores and SOFTA-s Scores for Couples ........................................................ 91 Table 4.3. Mean Total SOFTA-s Scores According to Gender……………………………….....95 Table 4.4. Mean SOFTA-s Total Scores for Male and Female Partners… …………………...…94 Table 4.5. Client Mean SOFTA-s Ratings for Therapists…………… ………………………...105 Table 4.6 Therapist Mean SOFTA-s Ratings of Clients……….……………………………….105 Table 4.7 SOFTA Self-Report and Observational Mean Scores for Engagement… …………..121 Table 4.8 SOFTA Self-Report and Observational Mean Scores for Emotional Connection… ..126 Table 4.9 SOFTA Self-Report and Observational Mean Scores for Safety……………...…….142 Table 4.10 SOFTA Self-Report and Observational Mean Scores for Shared Sense of Purpose………………………………………………………………………………………….146 Table 4.11 SOFTA Self-Report and Observational Mean Scores for Collective Couple Alliance…………………………………………………………………………………………157 xvi LIST OF FIGURES Figure 1.1 Conceptual Framework of Therapeutic Alliance in Couples Therapy .........................16 Figure 4.1 Couple 1: SOFTA-s Scores over Time .......................................................................168 Figure 4.2 Couple 2: SOFTA-s Scores over Time .......................................................................172 Figure 4.3 Couple 3: SOFTA-s Scores over Time .......................................................................176 Figure 4.4 Couple 4: SOFTA-s Scores over Time .......................................................................181 Figure 4.5 Couple 5: SOFTA-s Scores over Time .......................................................................185 xvii CHAPTER I INTRODUCTION Therapeutic alliance is one of the most widely researched processes of therapy due to its essential role in treatment outcome (Brown & O‘Leary, 2000; Horvath & Bedi, 2002; Horvath & Symonds, 1991; Mamodhoussen, Wright, Tremblay, & Poitras-Wright, 2005; Martin, Garske, & Davis, 2000; Pinsof & Catherall, 1986; Sprenkle & Blow, 2004). Two reviews revealed that there are over 2,000 research documents published on this topic in the past four decades (Horvath & Bedi, 2000; Martin, 1998). The reason for this overwhelming interest has roots in the vast literature on evidenced based practices and the common factors involved in these treatments, which all point to the therapeutic relationship being the strongest predictor of individual client outcome (for review see Horvath & Bedi, 2002; Martin, Garske, & Davis, 2000; Sprenkle & Blow, 2004; 2007; Blow, Sprenkle, & Davis, 2007). Most studies of alliance in couple and family therapy (CFT) focus on change mechanisms and outcome-oriented research, rather than process-oriented research (Sexton, Robbins, Hollimon, Mease, & Mayorga, 2003). While it is important to identify which therapeutic models are effective, outcome-oriented research does not aid in identifying the specific elements influencing the change. The goal of process research is to discover processes of change in the interaction between systems. This includes all of the behaviors and experiences in the systems, internal and external to therapy sessions, which relate to the process of change (Pinsof, 1992). The methodological literature on therapeutic alliance is deficient in identifying the specific elements of the relationship between the therapist and client system that are influential predictors of positive therapeutic alliance. Researchers assent that the measurement of the alliance remains 1 to be one of the greatest topics of dissension in the literature (Knobloch-Fedders, Pinsof, & Mann, 2007) due to the inconsistency regarding the methods employed, timing of measurement, and important indicators of alliance. Despite the dissemination of theory-based measurements, little work has been accomplished to develop measures of the alliance in CFT (Sexton et al., 2003). To date, most studies rely on self-report measures, which could raise threats to social desirability, common method variance, and ignorance of multidimensional assessment (Friedlander, Lambert, & Muniz de la Pena, 2008). Additionally, research on the timing of alliance measurement depicts mixed results, indicating that a key element in the puzzle is actually discovering at which points in therapy alliance is actually developed, maintained, and influencing treatment (Crits-Christoph, Cooper, & Luborsky, 1988; Crits-Christoph et al., 2006; Friedlander et al., 2008; Gaston, Piper, Debbane, Bienvenu, & Garrant, 1994). Moreover, a general omission in the research of alliance measurement regards the importance of therapist and client characteristics in alliance. Whereas these variables may be elaborated on when discussing their importance to the therapeutic relationship, most studies fail to identify them as important aspects of assessment (Baldwin, Wampold, & Imel, 2007; Blow et al., 2007). Thus, the literature illustrates a very incomplete picture of alliance in CFT. The universal intent of these studies is to prove the relationship between therapeutic alliance and outcome exists, which has succeeded in providing the one clear truth in alliance research-that it predicts outcome. How to achieve that goal is ambiguous and requires an orientation based on process rather than outcome. Focusing on outcome minimizes the multidimensionality of alliance and neglects the multifarious research conceptualizations of process that could help describe the complexity of alliance (Pinsof & Wynne, 2000). Thus, what is now needed by research is 2 information about what creates strong alliances with couples and families (Friedlander, Escudero, Horvath, Heatherington & Cabero, 2006b). Statement of the Problem Many elements of the therapeutic alliance are still unclear in couple and family treatment. First, it is not clear how alliances are formed with multiple family members. Second, it is uncertain how alliance unfolds over the course of treatment with couples and families. Additionally, research needs to more deeply understand how alliance is managed between the multiple systems in family treatment (i.e., therapist-family alliance, therapist-individual member alliance, alliance between the family), and how this impacts overall alliance scores. Finally, most alliance research identifies individual and family level characteristics that may enhance or impede alliance development and maintenance, such as general demographics (age, gender, education, etc.), individual psychopathology, current level of distress, and family-of-origin experiences. While these are definite pieces to the alliance puzzle, attention needs to also be focused on the therapist characteristics, such as therapeutic skills and techniques that may influence how they build and maintain the alliance. Purpose of the Study The purpose of this study was to explore how therapists build and maintain therapeutic alliance with couples by utilizing a process oriented framework that incorporates on-going multidimensional assessment of therapeutic alliance in conjoint treatment. This study employed mixed methodology that included qualitative interviews, observational and self-report measures. This research sought to achieve the following objectives: (1) Identify key components that contribute to therapeutic alliance in couples therapy, (2) Understand how alliance is managed between the multiple systems in conjoint treatment (i.e., therapist-couple alliance, therapist- 3 individual spouse alliance, alliance between the couple) and (3) Explore how therapeutic alliance evolves over time in couples therapy. These objectives were achieved by collecting data via SOFTA (Friedlander et al., 2006a) self-reports (therapist and client versions) and observational measures, and semi-structured interviews with all members of the therapeutic system. This study provided an in-depth exploratory look into the process of alliance development and maintenance in conjoint therapy in order to further the understanding of the fundamental elements responsible for creating strong alliances with couples. Theoretical Framework Process-Oriented Approach to Methodology Review of the research on therapeutic alliance in CFT necessitates a shift from outcome oriented research to process driven inquiries. Research on the process of family therapy has grown over the past few decades as researchers have grasped the importance of context for identifying specific mechanisms of change and progress in therapy (Greenberg & Pinsof, 1986). Process research has important implications for what constitutes legitimate and valuable research on systems oriented therapy. Today, the scope of the field of marriage and family includes process variables that help define how, when and what changes occur in family therapy (Alexander, Holtzworth-Munroe, & Jameson, 1994; Blow, Morrison, Tamaren, Wright, Schaafsma, & Nadaud, 2009; Bourgeois, Sabourin, & Wright, 1990; Helmeke & Sprenkle, 2000). This movement of expanding research to incorporate process research has slowly diminished the process-outcome distinction as researchers embrace the importance of understanding how change is facilitated. As Pinsof (1992) debated, family therapy should not become lodged in the political or economic trap for implementing outcome studies with the sole 4 intention to show which therapy is effective. What is needed is evidence of what factors lead to successful outcomes. Research on the therapeutic alliance appears to be falling into the same trap of overly focusing on the relationship between alliance and outcome. This goal has created a heap of empirical evidence that alliance is a strong predictor of success in therapy in CFT (Bourgeois et al., 1990; Brown & O‘Leary, 2000; Holtzworth-Munroe, et al., 1989; Johnson & Talitman, 1997; Kaufman, 2000; Quinn, Dotson, & Jordan, 1997; Raytek, McGrady, Epstein, & Hirsch, 1999). While identifying key factors of change is quintessential to the field of family therapy, overly focusing in this realm creates a clear picture of what is needed in therapy, but disregards how to achieve these end results. Furthermore, therapeutic alliance is not a specific technique, but is rather a mutually created relationship based in the context of the here and now of therapy between the therapist and client systems (Bordin, 1979). The consensus of the research stating that therapeutic alliance is a significant agent of change now demands a shift of focus on how the various aspects of the alliance are formed. Pinsof and Wynne (2000) delineate a process research framework that incorporates process and outcome views into an integrated methodology that connects back to clinical practice. Family therapy process research integrates ideas from individual therapy process research with systemic concepts from family therapy. Pinsof (1988) put forth the following definition of family therapy process research: Family therapy process research studies the interaction between therapist and family systems. Its goal is to identify change process in the interaction between these systems. Its data include all of the behaviors and experiences of these systems and their subsystems, within and outside of the treatment sessions that pertain to changes in the 5 interaction between family members and in their individual and collective levels of functioning (p. 55). Pinsof (1986) presents a conceptual framework that identifies key methodological criteria to guide family therapy process research. He believes that a framework must be clear, comprehensive, and epistemologically adequate. In order to be clear, the framework should define terms and concepts, as well as the relationship between them. The second criteria, comprehensiveness, relates to the degree to which the framework incorporates and specifies the relevant variables within the domain to be studied. For family therapy, a conceptual framework must include the relevant therapist and family variables over time, including whole systems and subsystems, overt behavior, affect, and cognition (Pinsof, 1988). Finally, epistemological adequacy concerns the foundational scientific principles of family therapy and how well they are consistent with the conceptual framework. The theoretical conceptualization of therapeutic alliance in couple and family treatment for this study is described in the following sections. Conceptualization of Alliance in Couple and Family Therapy Therapy with multiple members of a family involves a conceptual shift from dyadic to triangular exchanges. This shift can utilize important elements from alliance conceptualization in individual therapy, but requires an expansion of theory to incorporate the dynamics intrinsic to CFT. Moreover, what research has shown is that the alliance is much more complex in family treatment, and studies indicate CFT alliances are distinct in nature, processes, and effects (Friedlander et al., 2006b). Therefore, alliance in CFT requires an expansion of the process described in individual therapy that includes these multiple distinctions. One such distinction involves the definition of alliance in CFT. Most alliance methods base their interpretation of the alliance on Bordin‘s (1979) model, thus indicating a mutual 6 consent that the alliance represents interactive and collaborative elements of the relationship in which the therapist and client engage in the tasks of therapy and agree on the goals of treatment in the context of an affective bond or positive attachment (Constantino et al., 2002). Specifically, Bordin (1979) posited the working alliance consists of: (a) agreement between therapist and client about the goals of treatment, (b) agreement about the therapy tasks needed to accomplish those goals, and (c) affective bonds necessary to sustain the hard work of therapeutic change. While the foundation of alliance in individual and family treatment includes the essential elements of bonds, tasks and goals, therapy with more than one individual requires a more detailed definition to properly describe what occurs between a therapist and multiple clients. Delineating the differences between individual psychotherapy and CFT has been a particular struggle in current research due to the complexity of alliance (Symonds & Horvath, 2004). Pinsof and Catherall (1986) presented the only formal definition of CFT alliance, and describe it as ―that aspect of the relationship between the therapeutic system and the patient system that pertains to their capacity to mutually invest in, and collaborate on, the therapy‖ (p. 139). This definition addresses the presence of multiple systems in therapy, but does not fully identify the complex nature of alliance with families. Therapy with couples and families poses unique challenges to the alliance due to the inclusion of additional clients in the room and the need for therapists to establish and maintain multiple alliances simultaneously (Friedlander et al., 2006a; Pinsof, 1994; Rait, 2000). Family therapists must concurrently formulate and manage these relationships with multiple family members who often initiate therapy due to some sort of conflict with each other. If therapists form too strong a working relationship with one family member, they may jeopardize relationships with other family members, resulting in unbalanced or split alliances. This imbalance may lead to poor retention and outcome in therapy (Robbins, et 7 al., 2006; Robbins, Turner, Alexander, & Perez, 2003). Therefore, couple and family therapy necessitates an expanded theory of alliance that differs from individual psychotherapy (Flicker, Turner, Waldron, Brody, & Ozechowski, 2008; Friedlander et al., 2006a; Sprenkle & Blow, 2004). Pinsof (1992) declared that family therapists have been simplistic in their views of the therapeutic relationship by plainly asserting the therapist joins the system. Family treatment necessitates a systemic understanding of therapeutic alliance to address the unique task of therapists to engage various systems simultaneously in order to establish mutual bonds that will aid in the success of therapy. A systemic framework demarcates the complexities by discussing the systems involved in CFT and how they impact therapeutic alliance. To address this, Pinsof (1992) put forth an integrative systems perspective of alliance in family therapy that identifies the multiple levels of alliance unique to this scope of treatment. CFT alliance develops in the context of multiple relationships of the compound systems in therapy. As Pinsof (1992) states, family therapy is the interaction between two systems, the therapist system and client system, which create the therapeutic system. This contrasts from the traditional view of therapeutic alliance that focuses solely on the bond and interaction between one client and therapist (Mamodhoussen et al., 2005; Pinsof & Catherall, 1986; Rait, 1995). An integrative systems model portrays therapeutic alliance as a bidirectional process that is a product of the client and therapist systems in therapy. It highlights the unique systemic influences particular to family therapy that require integration of the larger, indirect, external systems into the therapeutic system in order to initiate alliance formation. At times, the relationship between the therapist system and the client system can be broader than the alliance (Pinsof & Catherall, 1986). As such, it is imperative for family therapists to adapt this integrative perspective in order to work 8 effectively with clients. It is important to delineate the difference between a relationship and an alliance. A relationship is typically an emotional or other association, involvement or connection between two or more people that can be based in numerous factors. While a relationship is part of an alliance, the term alliance implies that there is an agreement between individuals that is made in order to advance common goals and secure common interests (Pinsof, 1992). The following sections identify the systems of alliance involved when working with families. Systems of alliance in CFT. The addition of multiple individuals in the room creates the reality of compound alliances occurring at various levels. CFT therapists have to integrate the similarities and differences in family members‘ feelings, values, beliefs, and objectives. Family members may be at differing developmental levels, holding secrets or hidden agendas, and have historical conflicts and misunderstandings that contribute to the degree in which a therapist can form and build multiple alliances with each member (Friedlander et al., 2006a). Specifically, therapists need to attend to the interrelationships of the multiple systems in treatment, including: a. alliance between therapist and each member of the family b. collective alliance between the therapist and the family c. alliance between family members (Friedlander et al., 2006a; Pinsof, 1988; Symonds & Horvath, 2004). The combination of all of these creates a clear picture of the interactions between the therapist and client systems and how they merge to form a systemic model of therapeutic alliance. Pinsof and Catherall (1986) described therapeutic alliance in conjoint treatment as commencing with each family member, the individual, and evolving to a higher order level with the whole family system becoming allied with the therapist. The role of the therapist, therefore, is to develop an alliance with each of the subsystems in a circular and reciprocal fashion. Thus, each person‘s 9 alliance, as well as the collective family‘s alliance with the therapist, and the alliance between family members must be considered conceptually, methodologically, and clinically (Friedlander et al., 2006a). Individual member’s alliance with therapist. Establishing bonds with each member of a family is a fundamental task for family therapists. This can be particularly challenging when members of a family are in conflict with each other, or in disagreement about the need or role of therapy (Friedlander et al., 2006a). Therapists must continually track the quality of the relationship with each client in the family, and how this impacts other members of the family (Pinsof, 1988). For instance, when individuals in a family are conflictual, aligning with one member can cause tension with another. Moreover, often members of the family are forced to attend therapy, which further impedes alliance building with each individual. Therapists must be careful to not alienate any member who is there unwillingly by overly aligning with the individual who is forcing them to attend. These issues have implications for alliance in conjoint therapy. Alliance building becomes a balancing act as increases in bonds with one member may create a weaker relationship with another, resulting in unbalanced alliances. Minuchin and Fishman (1981) sometimes used unbalancing as a technique to shift family dynamics but also discussed the possibility of an unbalanced alliance occurring when a therapist persistently and/or unconsciously aligns more strongly with one family over the others at any particular time in therapy. This latter type of imbalance may detrimentally affect the therapeutic alliance and treatment outcome (Pinsof, 1992). Additionally, Pinsof and Catherall (1986) proposed the ideas of split and intact alliances that occur when working with multiple family members. An alliance is considered intact when the collective alliances among a family are perceived similarly. 10 Conversely, a split alliance occurs when family members do not agree on the perceptions of the alliance.This idea becomes more stringent as research indicates that more than 10% of couples in couple therapy represent a split alliance (Mamodhoussen et al., 2005). Alliance between family members. Family members have relationships with each other that have been present long before they enter a therapy room. The state of these relationships tends to be the reason for seeking couple or family treatment, and the degree to which the family members have what Friedlander et al. (2006a) terms ―shared sense of purpose‖ appears to be an important variable in the alliance. Research indicates the strength of the alliance between family members is a more powerful predictor of outcome than the alliance of any one individual member (Friedlander et al., 2006b; Garfield, 2004; Mamodhoussen et al., 2005; Robbins et al., 2003; Symonds & Horvath, 2004). Marital distress at intake appears to most stringently influence the development of the couple‘s alliance with each other (Knobloch-Fedders et al., 2004). Symonds and Horvath (2004) found that the quality of allegiance in the couple‘s relationship is correlated with their achievement of a positive therapeutic alliance, and thus a positive outcome in treatment. They found that partners‘ mutual agreement about the strength of their alliances, not their individual assessments, was important in predicting positive outcomes in therapy. Likewise, when the partners were in mutual agreement about the direction of the alliance as therapy progressed, they were more likely to have positive therapeutic results (Symonds & Horvath, 2004). Similarly, a study by Blow and colleagues (2009) found that a couple tended to evaluate the therapeutic alliance similarly to how they rated their own relationship. Thus, if the partners rated their relationship as positive, they considered their alliance with the therapist as positive. Finally, Escudero and colleagues (2008) found that different stages of treatment have 11 implications for the various subsystems in CFT. Their binary logistic regression analysis showed that positive individual alliance behavior early in treatment, and positive within-alliance behavior in the sixth session permitted classification of seventy percent of the sample‘s outcomes as either improved or not improved. This research indicates that while individual alliance is important initially for retention, positive within-family alliance interactions are essential to successful family therapy (Escudero et al., 2008). Therefore, it appears it is essential for therapists to attune to the aspect of allegiance in the relationship of the couple in order to enhance the alliance process when working with multiple family members. Family’s collective alliance. The family‘s collective alliance pertains to the therapist‘s relationship with the whole family system. This is perhaps the most important level of alliance as the quality of this relationship grants the essential condition for successful therapy (Pinsof, 1988). If the collective family alliance is weak, therapy will not be successful. This collective alliance may be difficult to acquire in conjoint treatment. Most likely, couples and families initiate therapy in a state of conflict, or in a scenario in which one member or all members do not have faith in therapy or believe they are part of the problem or solution. When the members of the family are reluctant to attend therapy, therapists need to develop particular ways to engage those people without alienating them (Friedlander et al., 2006a). Research indicates that connecting with the most influential member of the family can aid in this dilemma and help engage other family members into the process of therapy (Friedlander et al., 2006a; Robbins et al., 2006; Robbins et al., 2003). By strengthening the family member with the most power in the family, the collective alliance of the family will increase. System for Observing Family Therapy Alliances (SOFTA; Friedlander et al. 2001; 2006a). While Pinsof‘s (1992) integrative systems view is essential to the understanding of the 12 multiple levels of alliance in therapy with families, it does not identify the one distinctive dimension of alliance in this context: safety. Additional members in the therapeutic process pose unique challenges to the development of safety, and this has implications for alliance formation. Also, the multiple levels of therapy require specific measurement that surpasses traditional measurement of Bordin‘s (1979) definition of alliance that incorporates the bonds, tasks, and goals for therapy. Research by Friedlander and associates developed the System for Observing Family Therapy Alliances (SOFTA; Friedlander et al., 2001) to address the gaps in the literature that necessitated a look into the actual in-session behavior of consistent findings about the effects of client conflict and negativity on CFT outcomes (Friedlander et al., 2001). This measure was inductively created from data and takes into account the unique positive and negative items reflecting different aspects and levels of client collaboration in CFT (shared sense of purpose and safety within the therapeutic setting), as well as Bordin‘s (1979) classic idea of mutual collaboration and bonds. The four dimensions of the scale adequately define and describe the essential elements of the therapeutic alliance in conjoint treatment (Friedlander et al., 2001). The following sections will delineate these concepts. Engagement in the therapeutic process. Engagement in the therapeutic process refers to viewing the treatment as meaningful and having a sense of involvement and working together with the therapist. Additionally, they feel that therapeutic goals and tasks in therapy can be discussed and negotiated with the therapist and that taking the process seriously is important. Most importantly, each client believes that that change is possible (Friedlander et al., 2001; 2006a). Emotional connection to the therapist. Emotional connection to the therapist is based on Bordin‘s (1979) dimension of bonds and implies that the therapeutic relationship is based on 13 affiliation, trust, caring, and concern. The clients view the therapist as an important person in her/his life, similar to a family member. They also feel that the therapist genuinely cares and ―is there‖ for the client and that he/she is on the same wavelength with the therapist (e.g., similar life perspectives, values). Finally, they believe that the therapist‘s wisdom and expertise are valuable (Friedlander et al., 2006a). Safety within the therapeutic system. Safety is a particularly important aspect in conjoint treatment as the increase in additional family members creates a context that has the potential to place risks and have adverse consequences (Christenson et al., 1998). Safety within the therapeutic system means that ―clients feel that they have a trusting connection to the therapist and do not fear repercussions for what is said and done in therapy,‖ (Christenson et al., 1998, p.183). Furthering this idea, Friedlander et al. (2006b) define safety as ―the client viewing therapy as a place to take risks, be open, flexible; a sense of comfort and an expectation that new experiences and learning will take place, that good can come from being in therapy, that conflict within the family can be handled without harm, that one need not be defensive (p. 216). Safety is an essential task in alliance development in all forms of therapy. However, in conjoint therapy, this element is not as easily achieved or controlled by the therapist due to the presence of multiple family members who have varying feelings, viewpoints, historical conflicts and misunderstandings about their relationship and reason for coming to therapy. This needs to be assessed for as it contributes to the extent of which a therapist can establish safety in the room (Friedlander et al., 2006a). Shared sense of purpose. Another unique aspect is the mutually agreed need for, purpose, and value of therapy (Friedlander et al., 2006a). This concept is similar to Pinsof‘s (1988) within system alliance and refers to consistency in family members‘ treatment goals as 14 well as to their cohesiveness as a family unit. Friedlander et al. (2006b) define a shared sense of purpose as ―Family members seeing themselves as working collaboratively to improve family relations and achieve common family goals; a sense of solidarity in relation to the therapy (―we‘re in this together‖); that they value their time with each other in therapy; essentially, a felt unity within the family in relation to the therapy,‖ (p. 216). This is an important concept in CFT as the family members often have varying views of the purpose and necessity of therapy. In fact, Friedlander, Lambert, & Muniz de la Pena (2008) found that the family‘s sense of a common purpose about the therapy seems to be the key ingredient for change in family therapy. As such, the challenging task for therapists is to engage all members of the family into the process of therapy, while systemically joining with their particular experiences. The present study adopts a theoretical framework of therapeutic alliance in CFT that incorporates the dimensions of the SOFTA (Friedlander et al., 2001). This framework identifies various dimensions of alliance with couples (individual alliance, between partners‘ alliance, and collective family alliance) , as well as the unique elements of safety and shared sense of purpose that are intrinsic to this form of treatment. Figure 1.1 presents a pictorial description of the model. 15 Figure 1.1. Conceptual Framework of Therapeutic Alliance in Couples Therapy Conceptual Framework of Therapist Therapeutic Alliance in Couples Therapy Emotional Engagement Connection Individual Alliance Collective Family Alliance Individual Alliance Shared Sense of Purpose Safety Partner Partner Within-Couple Alliance 16 CHAPTER II REVIEW OF THE LITERATURE The review of the literature will provide a summary of the development of therapeutic alliance as a theoretical and measurable concept, highlight the deficiencies in alliance conceptualization for conjoint treatment and outline a conceptual framework of alliance in CFT. Additionally, this chapter will also discuss its important role in therapeutic process and outcome. Finally, it will report important findings related to instrumentation of alliance and the appropriate timing and perspectives of measurement. History of the Conceptual Development of Therapeutic Alliance Research has not provided a clear theoretical conceptualization of therapeutic alliance due to the evolution of the term itself. The term therapeutic alliance has changed several times and has included the phrases of therapeutic relationship, working alliance, therapeutic bond, and helping alliance (Horvath & Luborsky, 1993). Additionally, therapeutic alliance has been considered as a single construct (Martin et al., 2000), multidimensional (Bordin, 1979; Pinsof, 1994), and existing in a continuously changing state (Safran and Muran, 2000). Generally, therapeutic alliance is described as the therapist and client‘s commitment to exploring problems, establishing a mutual trust, and working together to achieve therapeutic goals (Bordin, 1979). Enhancing that description, Pinsof (1988) posited that therapeutic alliance is an integral part of the interpersonal relationship between the therapist and client systems in which therapists hold the responsibility to structure therapy and help the couple understand and relate to the treatment process. Other theorists explored various components of the alliance in terms of the therapist/client relationship and identified trust, rapport and resistance as viable processes that 17 are intrinsic to alliance building (Horvath & Bedi, 2002). While trust is an essential element that strengthens therapeutic alliance, how alliance builds is still unclear and dependent on individual therapists and modes of therapy. Defining Alliance in Individual Psychotherapy Despite the extensive research on the alliance, there remains lack of consensus regarding the definition of alliance (Friedlander, Escudero, & Heatherington, 2006a; Horvath & Bedi, 2002). This issue is of utmost importance when discussing the measurement and methods of alliance as many instruments are designed according to the researcher‘s concept of the alliance. Conversely, many studies define the alliance according to the instrument they use (Horvath & Bedi, 2002). Thus, it is important to discuss the various definitions in order to distinguish the alliance-outcome relationship (Crits-Christoph, Gibbons, & Hearon, 2006). Alliance in therapy has historical roots in psychoanalysis based in Freud‘s (1940) work that differentiated alliance from transference, and gradually evolved to encompass the processes of both intrapersonal and interpersonal elements (Friedlander, et al., 2006a; Horvath & Bedi, 2002). Its progress included other modifiers for the alliance, including ego alliance, working alliance, and finally therapeutic alliance. Greenson (1965) added to this evolving process by differentiating working alliance (the client‘s ability to align with the tasks of therapy) from therapeutic alliance (the ability of the therapist and client to form a personal connection). Luborsky (1976) posited that the alliance developed in two phases, which he referred to as Type I and Type II alliance. Type I regards the therapist‘s ability to be warm, supportive, and caring, as well as the client‘s belief in the therapist as a valuable source of help. Type II involves the client‘s faith in the therapeutic process, willingness to commit to therapy, and shared ownership of the therapist process. Luborsky is responsible for implementing alliance assessment methods 18 for raters by utilizing transcripts or live versions of sessions to analyze in-session indicators of alliance (Horvath & Bedi, 2002). Finally, Bordin (1979) put forth the most influential model of therapeutic alliance for individual and conjoint therapy by proposing that the working alliance consists of the three components of goals, tasks, and the bond between therapist and client. Bordin (1979) defined the goals aspect of the alliance as the collaborative negotiation of the desired outcomes of treatment by the therapist and client system. Tasks are the steps, behaviors, and cognitions that occur within the therapeutic process to achieve the specified goals. The bond dimension defines the relationship that exists in therapy between the therapist system and the client system, which includes the affective bond and the interpersonal relationship between the two systems that are necessary to sustain the hard work of the therapeutic change (Pinsof & Catherall, 1986). Bordin‘s (1979; 1994) departed from traditional ideas of alliance that focused on the therapist‘s contributions or unconscious distortions of the therapeutic relationship by emphasizing the collaboration and interaction between therapist and client. Combining these ideas allows us to expand into a comprehensive description of the therapeutic alliance as: The quality and strength of the collaborative relationship between client and therapist. It is inclusive of the positive affective bonds between client and therapist, such as mutual trust, liking, respect, and caring. It is a consensus about, and active commitment to, the goals of therapy and to the means by which these goals can be reached. It is a sense of partnership (Horvath & Bedi, 2002, p.41). Most alliance methods base their interpretation of the alliance on Bordin‘s (1979) model, thus indicating a mutual consent that the alliance represents interactive and collaborative elements of the relationship in which the therapist and client engage in the tasks of therapy and 19 agree on the goals of treatment in the context of an affective bond or positive attachment (Constantino et al., 2010). Research commonly uses Bordin‘s (1979) stable definition of the alliance; however there are other ways of defining the alliance, such as the more flowing, changing aspects of the client-therapist relationship, as indicated by the alliance ruptures described by Safran and Muran (1998). Review of Research on Therapeutic Alliance in Couple and Family Therapy Research on the alliance in conjoint treatment is considerably deficient in relation to the abundant literature describing alliance in individual therapy (Sexton et al., 2003). However, the past few decades have produced numerous studies that show the importance of alliance to progress (Knobloch-Fedders et al., 2007) and outcome in couple and family therapy (Friedlander et al., 2006b; Hotltzworth-Munroe et al., 1989; Johnson & Talitman, 1997; Pinsof & Catherall, 1986). Johnson and Talitman (1997) found that the quality of alliance seems to be a more powerful and general predictor of success than initial distress level for couples, which has not been found to be an important predictor of long-term success in Emotion Focused Therapy. This is noteworthy as initial distress level has been historically viewed as the best predictor of longterm success in couple therapy (Glebova et al., 2011; Knerr et al., 2011; Ligiero & Gelso, 2002; Symonds & Horvath, 2004). In addition, therapeutic alliance is found to be differentially important to treatment retention in Hispanic and Anglo families (Flicker et al., 2008). Therapeutic alliance theories for CFT are even more uncertain than for individual psychotherapy. How alliance builds is particularly unclear in conjoint treatment and dependent on individual therapists and modes of therapy. Structural therapists may align with one spouse to act as a co-therapist in the session while other therapists insist on treating each client equally (Rait, 2000). Pinsof and Catherall (1986) advocate that alliance in family therapy begins with 20 each family member and then evolves to a higher order level with the entire family system forming an alliance with the therapist. Rugel (1997) believed that therapeutic alliance in couple therapy involved multiple factors in which therapists must join with the couple, the nature of their distress and problem, and show acceptance, involvement, empathy, empathic probing, and competence with each spouse‘s relationship reality. Defining Alliance in CFT Definitions of the alliance in CFT have lagged behind in relation to individual psychotherapy. While most CFT researchers and measures base their ideas and scales on Bordin‘s (1979) conceptualization, there are unique aspects of conjoint alliance that are not included in his definition. In response to this, Pinsof and Catherall (1986) offered an integrative systems definition of alliance that incorporates the unique features of alliances. They describe it as ―that aspect of the relationship between the therapeutic system and the patient system that pertains to their capacity to mutually invest in, and collaborate on, the therapy‖ (p. 139). Pinsof (1988) proposed that there is an integrative process to alliance in CFT. This alliance includes three components: alliance between the therapist and each individual family member, alliance between the family members, and the collective family alliance. In order to build strong alliances, family therapists must establish bonds with each of these systems. Friedlander et al. (2006a) furthered the evolution of the therapeutic alliance definition in CFT through their extensive research. They found that the mere inclusion of multiple clients in conjoint treatment also adds elements to the therapeutic alliance. CFT is similar to individual treatment in the need for the three aspects of emotional bond, mutual goals and tasks, however there are two additional aspects of conjoint treatment that influence alliance building: safety and mutually agreed need for, purpose, and value of therapy (Christensen, Russell, Miller, & 21 Peterson, 1998; Friedlander et al., 2006b; Pinsof, 1994). Maintaining and establishing alliances with multiple family members is challenging due to the varying perspectives of the problem, conflict with each other, and individual willingness to participate therapy, but also because discussing these various issues in the presence of others can cause consequences after therapy (Friedlander et al., 2006a). Therefore, safety is an essential element when working with relational dyads and families. Another unique aspect is the mutually agreed need for, purpose, and value of therapy (Friedlander et al., 2006a). This is an important concept in CFT as the family members often have varying views of the purpose and necessity of therapy. In fact, Friedlander, Lambert, & Muniz de la Pena (2008) found that the family‘s sense of a common purpose about the therapy seems to be the key ingredient for change in family therapy. As such, the challenging task for therapists is to engage all members of the family into the process of therapy, while systemically joining with their particular experiences. For purposes of this study, the definition of therapeutic alliance in conjoint treatment put forth by Friedlander et al. (2006a) will be utilized as the underlying foundation for the contextual framework of alliance with couples. Influence of Therapeutic Elements on Alliance: Historical, Personal and Interactive Factors The therapist and client systems are the core players that interact and co-construct the therapeutic alliance. Their relationship, however, includes elements that symbolize dynamic mechanisms of past relationships that both client and therapist bring to the present interaction (Horvath & Bedi, 2002). Historical elements are therefore important to identify in therapeutic alliance. Family-of-origin experience has long been established as having significant influences on establishing important relationships (Bowlby, 1969; Kerr & Bowen, 1988; Skowron & Friedlander, 1998), and this transmits into the therapeutic relationship as well. 22 Research is beginning to address the family-of-origin issues that influence the abilities of both therapists and clients to form therapeutic alliances (e.g., Eames & Roth, 2004; Garfield, 2004; Knobloch-Fedders et al., 2004). Horvath and Bedi (2002) state that an essential element of conceptualizing the alliance is the understanding that the alliance is a present, in the moment, conscious concept, yet is impacted by prior relational history at various levels. The therapeutic relationship is inclusive of elements that represent elements of past relationships. In fact, research has demonstrated that when historical factors are controlled, the impact of the alliance is significant (Gaston et al., 1994; Henry, Strupp, Schacht, & Gaston, 1993). Therefore, identifying historical factors will provide a useful overarching model of the therapeutically active ingredients of the relationship (Horvath & Bedi, 2002). Echoing this idea, Gelso and Carter (1994) stated that the alliance is created in the intricate interaction between the therapist and client, each of whom conveys his or her own characteristics, personality, and history into therapy. The following sections will describe the past and present elements of both the therapist and client systems. Therapist Factors Although there is building consensus about the pivotal role therapists play in therapeutic alliance, research is generally deficient in clearly establishing which therapist characteristics influence alliance. Nevertheless, the therapist system is proving to be one of the most influential factors in therapeutic treatment. Outcome research consistently supports the finding that treatment progress and success is more closely related to the personal characteristics of therapists than to any specific intervention or approach (Angus & Kagan, 2007; Baldwin et al., 2007; Castonguay, Constantino, & Holtforth, 2006; Crits-Christoph, Gibbons & Hearon, 2006; Wampold, 2001). Research indicates that five to ten percent of the total variability in outcomes is 23 attributable to between-therapist differences (Crits-Christoph et al., 1991; Elkin, Falconnier, Martinovich, & Mahoney, 2006; Kim, Wampold, & Bolt, 2006; Wampold & Brown, 2005). As a number of researchers contend, the person who provides treatment is much more essential than the specific components of the model (Blow et al., 2007; Wampold, 2001). While clinicians are often processing information through their particular lens, the therapist‘s ability to assess client expectations and present therapy as consistent and congruent with client expectations is important to treatment outcome (Blow et al., 2007). Research on therapist influences on alliance is not as extensive, but studies are beginning to show that therapists are significant sources of variability in the alliance. Baldwin et al. (2007) explored the relative importance of client and therapist variability in the alliance as they relate to outcomes. Therapists who form stronger alliances with their clients show statistically significant better outcomes than therapists who do not form as strong of alliances (Baldwin et al., 2007). Research illustrates that the certain key ingredients essential to positive alliance building are divided into three expansive categories: interpersonal skill component, intrapersonal element, and interactive components (Horvath & Bedi, 2002). Interpersonal skill component. The interpersonal skill dimension identifies the reality facets of the therapeutic experience in that it addresses the therapist‘s ability to possess and express sensitivity to clients‘ needs, be responsive, and foster hope. (Horvath, 2001; Horvath & Bedi, 2002). Research suggests that effective therapists can build and enhance the alliance by nurturing hope and providing a treatment that is coherent and responsive to clients‘ needs (Frank & Frank, 1991). Additionally, this interpersonal component requires the therapist to react to in the moment processes, such as responding to any challenges and reactions by the client. This dimension is particularly important to alliance. Henry and Strupp (1994) studied moment-to- 24 moment progressions of the alliance and found that the therapies with poor alliances and outcomes had a destructive interpersonal process in which the therapist displayed hostility, controlling responses toward challenging clients, and dissociative with clients. Effective therapists are able to engage patients in collaborative, purposive work, whereas ineffective therapists may be less able to do so. Effective therapists displayed genuineness and empathy towards their clients (Ackerman & Hilsenroth, 2003). Communication-related skills are also important elements in this component. Effective therapists foster an expectation among their clients that treatment can and will help by providing them with articulate and relevant explanations of their problems and delivering a treatment consistent with such explanations (Horvath & Bedi, 2002). This may help clients feel understood and collaborate on the tasks and goals of therapy, thus enhancing the probability building a strong alliance. Assessing the therapeutic relationship from the client‘s perspective is another facet of this dimension. Communicating with the client about helpful aspects of the alliance, as well as exploring how they feel therapy is progressing fosters a strong alliance (Horvath, 2001). Intrapersonal dimension. The intrapersonal dimension contains the therapist‘s qualities that he or she brings into the therapy. This system includes personality traits, mannerisms, characteristics, temperament, and attachment. Traditionally, this concept has been regarded as countertransference, which refers to the idea that the therapist‘s responses to the client system are determined by his or her personal history (Pinsof, 1988). As such, although the therapist is responding to the present actions and words from the client system, his or her responses are consequences of how these aspects are filtered through the therapist‘s personal reaction system. This process may be the most important aspect of therapists as it guides their ability to form alliances and portray the characteristics of empathy, attunement, and consistency that are 25 quintessential to the therapeutic relationship and alliance (Angus & Kagan, 2007). Luborsky et al. (1985) stated that ―the major agent of effective psychotherapy is the personality of the therapist, particularly the ability to form a warm and supportive relationship‖ (p. 609). Despite the established importance of this, research has yielded minimal information about what these personal characteristics could be. Therapist‘s technical skills and training are also considered part of this dimension. Research displays mixed findings in this realm; some research posits that therapist training is not a significant predictor of therapeutic alliance (Horvath, 1994), while others postulate clients value the expert position of the therapist within a context of support, mutuality, and empathetic understanding (Bischoff & McBride, 1996). Bischoff and McBride (1996) concluded that a hierarchical relationship between the therapist and client system is expected and desired by couples. Couples reported feeling frustrated when the therapist did not appear to be in control of the direction or process of therapy. However, the therapist‘s expert role is only trusted when the clients sense empathy, understanding, and mutuality in the relationship (Bischoff & McBride, 1996). It appears that the bond dimension of the alliance serves as a framework for therapy such that the interpersonal skills of showing empathy, understanding, and support provide a starting point for therapy. However, this is not the only essential element to successful alliance in that therapists training and skills become more important as clients progress and work in therapy (Bischoff & McBride, 1996). Therapist gender. There is also mixed evidence that suggests the gender of a therapist plays a role in the therapeutic relationship. Some research has suggested that both male and female clients prefer women therapists (Johnson, 2005), while others conclude that the gender of the therapist did not matter (Blow, Timm & Cox, 2008). These gender related preferences appear 26 to relate to the socially ascribed roles of women and men. As such, both men and women believe that female therapists are more knowledgeable about relationship issues and are, therefore, more comfortable to express their emotions and vulnerabilities with female therapists (Johnson, 2005; Scher, 2005). In terms of alliance, research has demonstrated that men who are highly identified with the male gender role will disclose less and have weaker alliances than other men who identify less with the male gender role and even women (Dailey, 2004). Interactive elements. The interactive elements include therapist-client complementarity and collaboration (Horvath & Bedi, 2002). Complementarity refers to the idea that harmonious interactions are evident in positive relationships, rather than negativity or hostility. Higher alliance ratings are associated with friendly and autonomy-enhancing relationships rather than competitive, hostile, or controlling interactions (Henry & Strupp, 1994). Thus, it seems that hostile, negative, or competing behaviors are not elements of a strong alliance. Collaboration refers to the idea that the therapist and client are mutually working together to accomplish the tasks and goals of therapy. Collaboration is thought to be the foundation for strong therapeutic alliances (Bordin, 1979). Indeed, research suggests that collaboration and cooperation are conducive to stronger alliances, and in turn, better outcomes (Herman, 1998). Client Factors The client system consists of any individual system that is involved in the maintenance or resolution of the presenting problems (Pinsof, 1988). The direct client system includes the members of the family that are physically present in therapy. Evidence suggests that the client system greatly impacts therapeutic alliance (Garfield, 2004; Knobloch-Fedders et al., 2004), and that client characteristics are one of the sources of variability in the relationship between therapeutic alliance and outcome (Horvath & Bedi, 2002; Martin, Garske, & Davis, 2000). These 27 characteristics can include fixed factors, such as age, gender, and race, or other factors, including the ability of clients to form an alliance, involvement in therapy, and personality issues (Barber, Connolly, Crits-Christoph, Gladis, & Siqueland, 2000; Krupnick et al., 1996; Sprenkle & Blow, 2004). Current level of distress and presenting problem. Perhaps one of the most influential factors of clients is their current level of distress or symptomology. Kaufman‘s (2000) study of the effects of therapist overt self-monitoring on therapy alliance and outcome given the client‘s level of ―health-sickness‖ found that the strongest predictor of outcome was initial level of client functioning. In the past, research has shown that highly reactive clients tend to have a poor alliance with their therapist (Knerr, Bartle-Haring, McDowell, Adkins, Delaney, & Gangamma, 2011Tyron & Kane, 1993). These clients often present challenges to therapists, including resistance following therapist directives, increased symptom severity, and a greater number of premature terminations (Knerr et al., 2011; Ligiero & Gelso, 2002; Symonds & Horvath, 2004). This has significant implications for CFT. In conjoint treatment, couples and families often initiate therapy due to conflict or issues between the individual members. The level of conflict, or distress, can have significant impact on the alliance (Knerr et al., 2011; KnoblochFedders et al., 2004). Creating solid bonds with all family members is relatively easy to accomplish when motivation is high and when family members have an agreed upon purpose of therapy (Friedlander et al., 2006b). However, the challenge for family therapists arises when members are at odds with each other. Thus, alliance in conjoint therapy can be particularly daunting if the degree of distress in the family is high. Research on marital therapy found that three variables were significant in accounting for the drop-out of couples: having less than two children, having a male intake clinician, and the 28 presenting problem relating to only one partner (Allgood & Crane, 1991). They also found that when the male partners experienced high anxiety and the presenting problem was associated with parenting, the couples were more likely to drop out of treatment. Likewise, couples were more likely (17%) to drop out of therapy than continue (4%) if they had an individual dysfunction. The authors concluded that men may have a significant role in the decision to remain in treatment since their level of anxiety was associated with dropping out of treatment. These findings suggest that the distress level and nature of the presenting problem for couples pose certain challenges to treatment retention. Thus, couple and family therapists need to understand how these dynamics affect the development and maintenance of therapeutic alliance. Early client relationships and alliance. Research indicates that both the client‘s early relationship experiences and the quality of their current relationships correlate reliably with the capacity to develop a good alliance in the early phases of treatment (Horvath, 1994; Mallinckrodt, Coble, & Grantt, 1995). Indeed, attachment styles have specifically been targeted as factors that could affect their ability to foster a strong alliance with their therapist (Mallinckrodt, 2000). Moreover, the issue of transference, in which the client perceives the therapist as possessing aspects of past and present relationships, plays a role in alliance formation and growth (Pinsof, 1988). Gelso and Carter (1994) believe that transference can have differential effects on alliance. Specifically, positive transference (when the therapist is perceived as possessing positive aspects of past and present relationships of the client) can help foster alliance early in treatment. However, if that alliance is not fostered, negative transference in which the therapist is believed to possess any negative aspects, could cause ruptures in the alliance. 29 Two studies found that partners who reported early family distress were more likely to have difficulty forming positive alliances in couples therapy (Eames & Roth, 2000; KnoblochFedders et al., 2004). Knobloch-Fedders et al. (2004) found that for women, family-of-origin experience appears to be a powerful predictor of the tendency to form a split alliance, both early and later in treatment. For men, higher levels of self-reported family-of-origin distress were related to their ability to form a strong therapeutic relationship with the therapist, as well as their wives early in treatment. The opposite is true for women in that family-of-origin distress is related to poorer alliances in later treatment. Interpersonal styles. Previous research has also shown there are specific client interpersonal styles that may be important to alliance formation. Friendly submissive interpersonal styles, the ability to trust, and comfort with intimacy have been positively related to the alliance (Kivilighan, Patton, & Foote, 1998). Conversely, hostility, fear of abandonment, defensiveness, and perfectionism have been found to be negatively related to the alliance (Kivilighan et al., 1998). Client gender. Gender seems to be another puzzling aspect of alliance. Few studies have directly addressed the influence of gender on therapeutic alliance (Bourgeois, Sabourin, & Wright 1990; Knobloch-Fedders et al., 2004). Most conclude that the influences of therapeutic alliance differ by gender (Thomas, Werner-Wilson, & Murphy, 2005). Research suggests that alliance tends to be a stronger predictor of outcome for men in couple and family therapy than for woman (Bourgeois et al., 1990; Knobloch-Fedders et al., 2004; Mamodhoussen, et al., 2005). Bourgeois et al.(1990) determined that the strength of the males‘ alliance was a more powerful predictor of outcome than that of their female partners in a study of marital skills training. Similarly, Symonds and Horvath (2004) found that when the male partner‘s alliance was greater 30 than the female‘s alliance, and when the male‘s alliance was increasing over time, the relationship between alliance and outcome was strong. These results can be explained by the clinical literature that shows men are less likely than women to talk to others about their problems (Berger, 1979) and men tend to be the ones who cancel therapy sessions (Berg & Rosenblum, 1977). Furthermore, women tend to desire and initiate therapy more often than men (Rait, 2000). Research by Garfield (2004) suggests that there may be a disadvantage and power differential for men due to the nature of therapy that typically focuses on emotions and relationships. Men typically have power over resources while women hold more emotional power and control over the intimate aspects of the relationship (Symonds & Horvath, 2004). Thus, it seems that men will be more successful in therapy if they are engaged into the process earlier. Symonds and Horvath (2004) found that more positive outcomes are likely when men have a stronger alliance than women at the onset of therapy and they both continue together with a positive alliance. These findings suggest that a balanced therapeutic alliance necessitates therapists to strongly engage the male partner immediately and continue to stay engaged with the female partner throughout therapy (Garfield, 2004). Quinn and colleagues (1997) provide conflicting findings that state the association between alliance and outcome is stronger for women than men in couples therapy. They found that the outcome of therapy is more positive when the women‘s alliance is higher than the men‘s alliance. Interestingly, they found that positive outcomes were more likely associated when women reported a higher task alliance, while more negative outcomes resulted when men reported a higher task alliance. Gellhaus Thomas, Werner-Wilson, and Murphy (2005) studied the influence of therapist 31 and client behaviors on alliance and found distinct differences between male and female partners. For men, negative statements made by their partner were a consistent harmful predictor of all dimensions of therapeutic alliance. Additionally, men‘s alliance scores were positively associated with therapists who challenge them and offer advice, as well as when their partners were challenged. This implies that men expect therapists to protect them from criticisms and balance challenging statements. For female partners, challenging statements by therapists positively influenced the tasks and bond dimensions of the alliance, while challenging statements from their partner negatively impacted these two dimensions. These findings support Friedlander‘s et al., (2006b) finding that an important aspect of CFT alliance is the development of safety. Moreover, this research suggests that therapists need to form strong, immediate alliances with men to accomplish successful outcomes. These historical, personal, and interactive elements that both therapists and clients bring to the therapeutic relationship are instrumental in alliance formation. As such, they are important elements to consider when assessing alliance. The relationship between clients and therapists is an evolving, in the moment interactive bond that is influenced by historical and personal dimensions. Split Alliances Split alliances are a phenomenon unique to therapy with couples and families. The definition of split alliance was originally put forth by Pinsof and Catherall (1986) and refers to significant differences in family members‘ attitudes toward the therapy or the therapist in terms of the goals or tasks of therapy or the bonds with the therapist. The definition of split alliance has evolved over the years and is believed to occur when family members disagree about the effectiveness, competency, or degree to which the partners believe that the therapist cares about 32 the family (Friedlander et al., 2006a). Due to the expansion of the definition of a split alliance, most researchers consider split alliances to result from both within-family and therapist-client relational aspects of the therapeutic alliance. Disparity between family members about the value or direction of treatment results in a poor shared sense of purpose. These two processes overlap, however. When family members begin treatment with polarized views about the purpose of therapy, some members can become more emotionally available to the therapist than others. On the other hand, the weak within-system alliance can cause the therapeutic relationship to resemble the couple relationship in which one partner is in conflict with the therapist. Split alliances can also be a result of the client-therapist relationship and is most commonly evident in the Emotional Connection to the Therapist dimension of alliance. When all members are in agreement about the purpose and value of therapy, different perspectives can emerge over time about the therapist‘s skills, neutrality, or personality that could result in fragmented views about the therapist (Friedlander et al., 2006a). Individual family members‘ emotional connection to the therapist can be influenced by other dimensions of the alliance, such disagreements about the appropriate goals and tasks of treatment (engagement level) and the within-system alliance (Friedlander et al., 2006a). However, these differences may not always result from disagreements about the particular goals and tasks of therapy. Thus, split alliances can occur as a result of a weak within-system alliance where the therapeutic relationship begins to resemble the family dynamic, or over the course of treatment due to specific interactions with the therapist. Researchers commonly identify split alliances by comparing self-reported scores of alliance and have found interesting findings about the development and role of split alliances in 33 couple and family therapy. Two studies (Heatherington & Friedlander et al., 1990; Mamodhoussen et al., 2005) found that a significant proportion of couples and families had split alliances after the third session due to diverging views about how the therapist was conducting therapy and the quality of their relationships with the therapist. Knobloch-Fedders and colleagues (2004) found that couples who report split alliances with the therapist after Session 1 were significantly more likely to have a history of distressed family-of-origin relationships. In the same study, split alliances occurring after Session 8 were most frequent when the wives continued to view the marriage as distressed. Interestingly, outcomes were better when husbands‘ alliances with the therapist were stronger than those with their wives‘ (KnoblochFedders et al., 2004). Mamoudhoussen et al. (2005) discovered that split alliances occurred more frequently in younger couples who recently married and when men were highly distressed about the marriage and their wives had few mental health symptoms. Split alliances have also been found to play a role in treatment outcome. Research shows that poorer retention rates and outcomes consistently occur when couples and family members have differing views of the therapeutic relationship and context (Robbins et al., 2003; Symonds & Horvath, 2004). This has considerable implications for couple and family therapists in their effort to build strong alliances with all members involved in treatment. The necessity for therapists to monitor and address split alliances is supported by these findings. Attempts to balance the alliance can increase retention and rates of positive outcomes for couples. Detection of split alliances might not always be beneficial, however. A study by Hight (1997) found that split alliances were highly correlated with therapist ratings of alliance and beliefs about the success of the therapy, even more than the couples‘ ratings of alliance. Thus, therapists‘ knowledge of a split alliance affects therapy by having a negative effect on therapist 34 alliance and evaluation of treatment. If the therapist has lower alliance, he or she may be less engaged in treatment, which would decrease the likelihood of positive treatment outcomes. To place all the responsibility on therapists, however, would be too simplistic. Split alliances are not only due to the therapist‘s skill, experience, and interpersonal dynamics. As discussed in the previous section, clients have varying characteristics, motivation, presenting problems, and levels of distress that contribute to the formation of intact or split alliances. Role of Therapeutic Alliance in Process and Outcome of Couples Therapy In the spirit of combining theory, research, and practice, many researchers have approached the empirical studies of client change in therapeutic settings in order to determine which factors contribute to this process. The systematic study of this complex phenomenon has led to a debate about the necessary elements of change. Much of the debate has focused on which theories are most effective, however the prevailing literature depicts that there are a number of common factors in clinically successful therapy that are present in all modes of treatment (Sprenkle & Blow, 2004). This finding has stimulated many therapists due to its basic premise that most empirically evidenced therapies, such as cognitive-behavioral and emotion-focused, have similar outcomes as other theories that have not been shown to effect change consistently. While therapeutic change is relative to the client and therapist, there appears to be one factor that is common to all therapies, and is a key predictor of success in therapy- the therapeutic alliance (Brown & O‘Leary, 2000; Mamodhoussen et al., 2005; Pinsof & Catherall, 1986; Sprenkle & Blow, 2004). Horvath and Symonds (1991) found an effect size of .26 in their study of therapeutic alliance across diverse theoretical models in 24 studies. An additional analysis by Martin et al. (2000) found an average of .22 between alliance and outcome in 68 studies. The most recent 35 meta-analysis by Horvath and Bedi (2002) reported a median effect size of .25 between alliance and outcome. These studies demonstrate that therapeutic alliance influences treatment outcome in individual psychotherapy. Additionally, contemporary therapeutic alliance research demonstrates that alliance predicts outcome in couples therapy across diverse treatment orientations and modalities (Bourgeouis, Sabourin, & Wright, 1990; Brown & O‘Leary, 2000; Holtzworth-Munroe, Jacobson, Delyen, & Whisman, 1989; Johnson & Talitman, 1997). This research proves that the common factor status of alliance in individual therapy extends to couple and family therapy (Blow et al., 2007; Sprenkle & Blow, 2004). Perspectives of Therapeutic Alliance A consistent inquiry in therapeutic alliance measurement regards whose perspective matters (Horvath & Symonds, 2006; Knobloch-Fedders, Pinsof, & Mann, 2004; Kramer et al., 2008). There is great discrepancy in the field regarding whose perspective is the best in which studies generally differentiate the three standpoints of the client‘s, the therapist‘s, and independent observers view of the alliance. Research empirically validates that the client‘s perspective of the alliance is most explanatory of the correlation between individual alliance and outcome (Horvath et al., 1993; Horvath, 1994; Kaufman, 2000; Knobloch-Fedders et al., 2004). Conversely, Kramer and colleagues (2008) found that therapist‘s patterns of alliance construction were most predictive of positive outcome. Adding more complexity to this issue, Horvath (2001) conducted a meta-analysis of empirical studies on therapeutic alliance and concluded that the size or magnitude of the relationship between alliance and outcome is unrelated to the mode of therapy or the point-of-view from which it is assessed. Moreover, Kivlighan and Shaughnessy (1995) have shown that the more the client and the therapist agree on the quality of their relationship at the end of therapy, the better the outcome. On the contrary, Fitzpatrick et al.‘s 36 (2005) study did not reveal such convergence. Finally, research by Fenton and associates (2001) determined that observer-rated instruments were associated with treatment outcome. Recent research adapts a more convergent view of alliance by assessing both client and therapist perspectives (Kivlighan, 2007; Kramer et al., 2008). The APA Presidential Task Force on evidence-based practice (2006) directly calls for increased multidimensional assessment of the therapeutic relationship that combines client, therapist, and observational assessments (i.e., therapist, client, or external rater). Kivlighan (2007) argues that the current controversy over which perspective matters most causes conceptual and measurement problems in the alliance literature. Instead, he posits that embracing both perspectives of client and therapist provides a more dyadic perspective of therapeutic alliance that reflects Bordin‘s (1979) definition of alliance as a partnership between the client and therapist based in bonds and collaboration on the goals and tasks of therapy. As such, the therapist and client are joining and co-creating the alliance together (Gelso & Hayes, 1998), which indicates that the alliance is a shared perception and reality based in the context of the therapeutic relationship (Kivlighan, 2007). This research suggests the nature of therapeutic alliance would be best depicted by assessing both perspectives of the client and therapist. However, evidence also indicates that relying on the dyadic assessment of alliance may not be sufficient in understanding alliance. Kivlighan (2007) found that therapists have a significant influence on their clients‘ perception of sessions, which does not appear to be reciprocal. Not surprisingly, research about perspectives of alliance in CFT is deficient in relation to individual psychotherapy. Current research indicates that there are unique aspects of alliance in conjoint treatment that need to be included in proper assessment. Research by Pinsof (1995) revealed that there are various subsystems in CFT that influence the overall alliance with the 37 therapist. As discussed earlier, conjoint treatment involves three alliances: the alliance between individual members and the therapist, the alliance between family members, and the collective family‘s alliance (Pinsof, 1992). Research has shown that consideration of each client‘s alliance independently does not provide a useful picture of the therapeutic relationship (Bourgeois et al., 1990; Catherall, 1984; Pinsof & Catherall, 1986). Thus, each person‘s alliance, as well as the collective family‘s alliance with the therapist, and the alliance between family members must be considered conceptually, methodologically, and clinically in order to fully capture the therapeutic alliance in couple and family therapy (Friedlander et al., 2006a). Escudero, Friedlander, Varlea, and Abascal (2008) argue that observation of alliancerelated behavior is superior to self-reported perceptions of the alliance due to the complexity of a multi-person treatment context. They posit that this context creates difficulties for family members to report about the within family alliance. Additionally, they state therapeutic alliance is an ―interpersonal phenomenon and intrapersonal process‖ that requires observation of the interpersonal aspects of behavior (Escudero et al., 2008, p.195). The prevailing research indicates that embracing a multi-perspective view of therapeutic alliance is the preferred method for assessing alliance. According to the APA Task Force (2006), researchers are encouraged to evade a therapist centered view of the therapy relationship and to study the contributions of both clients and therapists to the relationship. An observational perspective is a fundamental method to be employed in future studies of the therapy relationship. These principles would also benefit research specifically addressing therapeutic alliance, especially in the field of marriage and family therapy. The research suggests that examining multiple perspectives will allow for a clearer picture of the alliance. Moreover, this approach identifies with traditional conceptualizations of alliance being dyadic in nature (Bordin, 1979; 38 Kivlighan, 2007). Therapeutic alliance is created by therapists and clients joining together, cocreating a partnership and collaboration that aims to fulfill the tasks and goals of therapy. As such, the alliance is the shared awareness and reality of the client and therapist. As Kivlighan (2007) states, ―using alliance ratings from only the client or therapist, even if these ratings are of perceptions of collaboration, misses the dyadic and interactional nature of the therapeutic alliance‖ (p. 424). This method is especially necessary for family therapists who partake in multiple relationships in conjoint treatment. Embracing a both/and position, rather than an either/or stance, will allow clinicians and researchers to assess the multiple alliances occurring in family therapy. Pinsof (1992) describes the various alliances in family therapy that include individual member alliance, alliance between family members, and collective family alliance. All three of these alliances have been influential in the relationship between alliance and treatment progress and outcome (Pinsof, 1992; Safran & Muran, 1998; Symonds & Horvath, 2004), indicating a need to assess each perspective. Observational methods facilitate insight into the multiple subsystems of alliance and behaviors between the various members in therapy. However, relying completely on observation would lose the voice of the clients, whose perspective have been particularly influential in rating the therapeutic alliance (Horvath et al., 1993; Horvath, 1994; Kaufman, 2000; Knobloch-Fedders et al., 2004), as well as the perspective of the therapist. Therapists are an important part of the therapeutic relationship, and embracing a systemic view of alliance mandates inclusion of all perspectives in therapy. Timing of Measurement To date, there is conflicting evidence about the timing of measurement of therapeutic alliance as indicative of outcome. A review of the research shows that there is lack of consensus 39 about the timing or frequency in which alliance should be measured, although the research does show promising results for particular and multiple time assessments. There are three common time points in which alliance is assessed. Research studies measure the alliance early in treatment (within the first third of treatment), mid-phase of therapy, and late phase, although traditional alliance measures define the level of alliance as a mean alliance measured at the third session (Horvath & Symonds, 1991). Many studies identify an important time frame between the third and fifth sessions, as alliance measured in that time is a consistent predictor of final outcome in treatment (Barber et al., 1999; Horvath & Symonds, 1991; Knobloch-Fedders, et al., 2004). Typically, therapeutic alliance has been measured after three sessions in marital therapy (Bourgeois et al., 1990; Heatherington & Friedlander, 1990; Mamodhoussen et al., 2005; Knobloch-Fedders et al., 2004). In a review of studies assessing the relationship between alliance and outcome, Horvath and Bedi (2002) found that the majority of research depicting alliance as a predictor of outcome is measured early before the fifth session (N=130), with an average effect size of .22 for these studies. Alliance assessed midway through treatment (N=38) had an effect size of .19, and late alliance scored .25 in its relationship to outcome. Some studies that assessed alliance at multiple time points and averaged the scores across treatments (N=68) had the highest effect size of .27. Thus, alliances assessed in the middle phases of treatment have the weakest association with outcome, while early and late phases are moderate predictors of alliance. Most importantly, alliance measured periodically throughout treatment seems to have a strong relationship to therapeutic outcome. Researchers advocate for establishing a strong alliance early in therapy due to evidence suggesting that the alliance is particularly predictive of outcome when measured early in treatment (Horvath & Symonds, 1991; Raytek et al., 1999). In fact, a few researchers agree that 40 early assessment of the alliance is a better predictor of outcome than later alliance (Gaston et al., 1998; Horvath & Bedi, 2002; Knobloch-Fedders et al., 2004). Martin et al.‘s (2000) metaanalysis of the relationship between therapeutic alliance and outcome found that clients are more likely to view the alliance as positive at termination if their initial assessment was positive. Early alliance is also associated with higher rates of retention in therapy. Research shows that couples who remain in therapy report higher alliance in early treatment than those who report low alliance initially in therapy (Knobloch-Fedders et al., 2004; Raytek et al., 1999). Moreover, investigators advocate that measuring alliance during the first session is imperative as the alliance begins developing in the initial client-therapist contact (Castonguay, Constantino, & Holtforth, 2006; Friedlander et al., 2008).This idea is supported by Sexton, Hembre, and Kvarme (1996) who found that alliance in individual psychotherapy evolved rapidly in the first session. Furthermore, researchers argue that observing alliance behavior immediately before improvements take place, allows research to assess the primacy of the alliance as a mechanism of change, which helps differentiate the reciprocal relationship between alliance and therapeutic progress. This facilitates minimization of potential alternative explanations of the allianceoutcome relation, such as the reverse causation rationalization that suggests clinical improvement causes a positive alliance (Crits-Christoph et al., 2006; Friedlander et al., 2008). Early improvement occurring before the alliance is measured in therapy might be associated with both the alliance and eventual treatment outcome. As such, early improvement can be conceptualized as a third variable in the relationship between alliance and outcome, and measuring alliance instantly can help eliminate this confounding factor. Other studies report that alliance measured in early phases is not the best predictor of change in treatment (Florsheim et al., 2000). Knobloch-Fedders, Pinsof, and Mann (2004) found 41 in their longitudinal investigation that mid-treatment alliance predicts outcome beyond that accounted for in early alliance scores. Interestingly, variables that predict the development of the alliance into the middle phase of psychotherapy have yet to be identified despite the findings that alliance measured in this phase of treatment has a significant impact on outcome (KnoblochFedders et al., 2004). Some explanations for the negative findings in this phase of treatment may be attributed to the middle phase being a particularly difficult time for clients in therapy. Clients are in the midst of working on issues that may produce strong feelings of anxiety and stress, which could impact alliance ratings (Horvath & Bedi, 2002). This reason is precisely why Knobloch-Fedders et al. (2007) advocate for immediate assessment of alliance. They believe that when alliance is assessed in the midst of therapy by client self-reports, they are most likely reporting about their cumulative thoughts and feelings to date, rather than their initial reactions to the therapeutic alliance. Therefore, research indicates that timing of alliance reflects various aspects of the alliance, and has differential influences on treatment outcome (Knobloch-Fedders et al., 2007). What this research depicts, then, is a conflicting view of the role timing plays in alliance measurement, and that alliance is not a continuous process that is discernable at any one point in time. Both individual and family therapists have suggested that it is unlikely that the strength of the alliance will remain constant throughout therapy (Horvath & Marx, 1991; Pinsof, 1994). The research of Gelso and Carter (1994) can help explain the discrepancies in timing of measurement. Their studies propose that alliance develops in stages. The initial stage of treatment is when clients are mobilized and hopeful about their work in therapy. This stage is followed by a middle phase plagued with ambivalence and skepticism about what therapy can provide. This idea is supported by Golden and Robbins (1990) who postulated that clients 42 typically experience a period in mid-treatment of increased negative affect, attitudes, and behavior. Additionally, two more studies indicate a fluctuating alliance pattern, with low levels associated with this sense of middle-stage angst. Kivlighan and Shaughnessy (2000) examined patterns of alliance development in 79 therapist-client dyads in a four-session therapeutic process, and found three distinctive patterns of alliance formation: stable alliance, linear alliance growth, and quadratic alliance growth. The quadratic growth, or U-shaped pattern, is most predictive of outcome. A study by Stiles et al. (2004) demonstrated that V-shaped deflections in the alliance over time were associated with greater therapeutic gains. This shape depicts the repair-rupture sequences and highlights the therapeutic benefits of reparative processes over the course of treatment. This research suggests that the alliance fluctuates during the process of treatment, and that measurement at any one point and time may not provide an adequate picture of the alliance. Review of the research identifies a need for longitudinal investigation in which the alliance is measured over time, rather than at the beginning, middle, or end phases of treatment (Knobloch-Fedders et al., 2007; Kramer et al., 2008). Consistent measurement of the alliance can enhance understanding of the evolving nature of the alliance. For instance, Kivlighan and Shaughnessy (1995) found that the correlation between client and counselor ratings increase over time in therapy. Additionally, some research indicates that alliance measurements taken during different phases of individual therapy do not correlate with each other (Crits-Christoph et al., 1988; Gaston et al., 1994). Finally, more recently, Escudero and colleagues (2008) found that different stages of treatment have implications for the various subsystems in CFT. Their binary logistic regression analysis showed that positive individual alliance behavior early in treatment and positive within-alliance behavior in session six, permitted classification of seventy percent of 43 the sample‘s outcomes as either improved or not improved. This research indicates that while individual alliance is important initially for retention, positive within-family alliance interactions are essential to successful family therapy (Escudero et al., 2008). More importantly, it is still unclear whether the alliance creates the same patterns or remains relatively stable over the course of treatment in conjoint therapy (Knobloch-Fedders et al., 2004). Alliance at various phases of therapy may be unique in their relationship to outcome (Horvath, 1995; Knobloch-Fedders et al., 2007). For instance, Horvath (1994) posited that alliance in early phases of treatment is similar across different treatments because it involves the development of a collaborative framework and negotiation of immediate and final goals. Later alliance may include elements that are specific to the therapeutic tasks, length of treatment, and degree of final outcome. Thus, research indicates that ―timing is of the essence‖ in therapeutic alliance and needs to extend across the course of treatment, especially for conjoint treatment. Therapeutic Alliance Measures Employing a multi-perspective assessment of therapeutic alliance in CFT requires scales that will incorporate the important elements of alliance in conjoint treatment, as well as include the appropriate versions of observational and self-report measures. The past few decades have introduced a myriad of alliance measures that incorporate observational, client and therapist selfreport techniques to assess therapeutic alliance. There are close to thirty alliance scales in use by researchers, with only three designated specifically for work with couples and families (Horvath & Bedi, 2002). The following section will introduce the history of alliance measurements with a brief review of alliance scales for individuals. 44 Brief Review of Alliance Scales for Individual Psychotherapy There are six families of instruments that are generally used in the vast majority of empirical studies (Horvath & Bedi, 2000). These scales attempt to address individual therapeutic alliance based on various theories and were the impetuses for group, couple, and family scales (Martin et al., 2000). These measures include the Penn Helping Alliance Scale (Luborsky, CritsChristoph, Alexander, Margolis, & Cohen, 1983), the Vanderbilt Therapeutic Alliance Scale (Hartley & Strupp, 1983), the Toronto Scales (TARS; Marziali et al., 1981), the California Psychotherapy Alliance Scales (CALPAS; Marmer, Gaston, Gallagher, & Thompson, 1989), the Therapeutic Bond Scales (TBS; Saunders et al., 1989), and the Working Alliance Inventory (WAI; Horvath, 1981). These scales are designed based on varying theoretical conceptualizations of the alliance, thus creating unique measures that seem to address differential aspects of the therapeutic alliance in psychotherapy. The Working Alliance Inventory also has a couples version that will be explored below. Couple and Family Therapy Alliance Scales The following scales build upon the ideas of therapeutic alliance in individual therapy. All of the scales designed for CFT embrace Bordin‘s (1979) tripartite conceptualization and Pinsof and Catherall‘s (1986) integrative systems approach to alliance. Working alliance inventory-couples (WAI-CO; Symonds, 1998). This scale was based on the Working Alliance Inventory originally developed by Horvath (1981) to measure alliance in individual therapy. The WAI-Co has 63 items, each rated on a 7-point Likert scale. Its development has been described previously (Symonds, 1998). Both therapist and client versions of the WAI-Co are administered. The client version has three subscales of Bond, Task, and Goal (Bordin, 1979) and each item is rated from the perspective of Self, Partner, or Couple (Catherall, 45 1984; Pinsof, 1994). The items on the WAI-Co are worded to capture the raters‘ judgment of the alliance between client and therapist. The following item appears, for example, as a Bond item: ‗‗The therapist and I trust each other.‘‘ The therapist version contains items parallel to the client version, and asks the therapists to rate their alliance with the female partner, male partner, and couple in three separate sections of the instrument. An item from the client version, such as ‗‗The counselor and I have an understanding about what we are trying to accomplish in therapy,‘‘ would appear on the therapist version, female client section, as ‗‗She and I have an understanding about what we are trying to accomplish in therapy‖ (Symonds, 1998). Integrative alliance scales (IPAS; Pinsof & Catherall, 1986). Pinsof and Catherall‘s (1986) integrative Couple and Family Therapy Alliance Scales is a client self-report instrument based on the integrative systems perspective of the therapeutic alliance, which views the alliance as involving significant people within the therapist and client systems who influence therapy to varying degrees (Pinsof, 1994). These scales have 40 items and are self-report of clients‘ perceptions of the therapeutic alliance. The Integrative Alliance Scales measure the three components of the alliance identified by Bordin (1979) (tasks, bonds, and goals) for each of the 4 possible alliance subsystems in conjoint therapy. These subscales are labeled: (1) self-therapist (the ―self‖ subscale; (―I trust the therapist‖; 11 items), (2) partner-therapist (the ―Other‖ subscale; ―Some of the other members of my family are not in agreement with the therapist about the goals for their therapy‖;11 items), (3) couple/family-therapist (the ―Group‖ subscale; ―The members of my family are not satisfied with the therapy‖; 7 items), and (4) self-partner (the ―Within‖ subscale; 11 items). The integrative scales also have an individual scale designed for individual clients. The wording changes to adapt this, for instance, an example in the Group-Therapist dimension would be ―The therapist cares about my important relationships,‖ (Pinsof, 1988). 46 Items are rated on a 7-point Likert scales, from completely agree (1) to completely disagree (7). With reverse scoring on selected items, the ratings are summed. Higher ratings reflect more favorable perceptions. Studies indicate that the individual (ITAS), couple (CTAS), and family (FTAS) scales have adequate test-retest reliability (.72 to .83) for the total scores and subscales (.48 to .92), except the self- scale on the family scale (.35) (Pinsof & Catherall, 1986). Catherall (1984) found adequate predictive validity (p＜.05) on correlations between the overall ITAS score and client progress. Additionally, split alliances can be identified using these alliance scales by comparing standard deviations of responses (Heatherington & Friedlander, 1990; Pinsof & Catherall, 1986). System for observing family therapy alliances (SOFTA; Friedlander, Escudero, & Heatherington, 2001). SOFTA (Friedlander et al., 2001) was developed to address the gaps in the literature that necessitated a look into the actual in-session behavior of consistent findings about the effects of client conflict and negativity on CFT outcomes (Friedlander et al., 2001). This measure takes into account the unique positive and negative items reflecting different aspects and levels of client collaboration in CFT (shared sense of purpose and safety within the therapeutic setting), as well as Bordin‘s (1979) classic idea of mutual collaboration and bonds. The SOFTA is a set of tools that include observational assessments (SOFTA-o) and self-report (SOFTA-s). There are four dimensions to the scale: (1) ―emotional connection with the therapist‖ (a sense that the relationship is based on affiliation, trust, caring, and concern; 10 items), (2) ―engagement in the therapeutic process‖ (viewing the treatment as meaningful and having a sense of involvement and working together with the therapist; 11 items), (3) ―shared sense of purpose within the family‖ (family viewing themselves as working collaboratively to improve 47 family relation and achieve goals; 11 items), and (4) ―safety within the therapeutic system‖ (client viewing therapy as a place to take risks, sense of comfort; 12 items). The last two dimensions, shared sense of purpose and safety in the therapeutic system, reflect the uniqueness of CFT. Bordin‘s (1979) scales of goals and tasks in therapy were combined into the dimension labeled engagement in the therapeutic process because of their high intercorrelation (.92) as reported in the self-report of the WAI (Friedlander et al., 2006b). Ratings are not summed, because the variability across dimensions is meaningful and depicts the varying experiences of the family members (Friedlander et al., 2006b). Relations among the four alliance dimensions of the SOFTA range from .18 to .75 (Friedlander et al, 2006b). The two most closely associated dimensions are engagement and emotional connection (r=.75, p ＜.0001), while the two dimensions of safety and shared purpose (r=.18, p ＜.025). This low correlation reflects the focus of these dimensions, individual behavior in context (safety) versus family interactions (shared purpose). All other intercorrelations were moderate (rSD=.30-.47, ps ≤ .0001) (Friedlander et al., 2006b). Review of alliance measures indicates that available measures in CFT are severely lacking in relation to individual psychotherapy. There are only three scales designed to assess alliance in family therapy (WAI-CO, Symonds, 1998; IPAS, Pinsof & Catherall, 1986; SOFTA, Friedlander et al., 2001). The SOFTA by Friedlander et al., (2001) estimates the strength of the alliance from observable behavior, as well as self-report measure of client and therapist. SOFTA (Friedlander et al., 2001) allows researchers and clinicians to identify the quality of the family members‘ interactions with each other as well as each individual‘s interaction with the therapist. Additionally, this assessment can inform practice, training, and future family research, as it 48 incorporates an exploratory process that can help illuminate important behaviors and aspects of the alliance. Conclusions and Rationale for Current Study The review of the literature depicts several important implications for the study of therapeutic alliance with couples. First, research needs to embrace a process oriented framework to measuring alliance in order to expound upon research in individual psychotherapy and define the unique and important elements of alliance as it pertains to conjoint treatment. Second, a multidimensional assessment is needed that incorporates the three perspectives that matter in alliance: observer, therapist, and client. The System for Observing Family Alliance (SOFTA; Friedlander et al., 2001) is a useful tool that embraces these three references points, and offers a way to assess behavioral analysis of in-session interactions between the multiple subsystems of alliance in family therapy. Third, data collection over time is necessary in order to identify how alliance changes in couple and family therapy. Lastly, inclusion of therapist and client characteristics is recommended to help identify the historical and personal factors of each system that influences alliance. These suggestions were derived from the empirical research on alliance in individual and family therapy and inform this current study. The present study attempted to address the gaps in the literature about therapeutic alliance in conjoint treatment by utilizing a process oriented, multidimensional assessment of the four perspectives of alliance-both members of the couple, therapist, and independent observer. The purpose of this research was to identify important elements of therapeutic alliance formation and maintenance in couples therapy over time. 49 CHAPTER III METHODOLOGY The purpose of this study was to explore the formation and maintenance of therapeutic alliance in couples therapy. Informed by the extensive empirical literature on CFT alliance, this study aimed to address the deficit in previous research that is unable to provide a clear picture of how alliance evolves throughout treatment. This study adapted a process-oriented framework and collected multiple sources of data during the course of treatment to assess therapeutic alliance via the clients, therapist, and independent observer‘s perspective. A multidimensional study over time is important to the study of therapeutic alliance in couples therapy in three ways: 1) it provides an in-depth analysis of therapeutic alliance formation and maintenance in couples therapy, 2) it identifies idiosyncratic elements that are important to alliance with couples as identified by the perspectives of the individual members of the couple, therapist and independent observer, and 3) it reveals how alliance changes over time in couples therapy. This study was guided by the following research questions: 1) What therapeutic components contribute to alliances in couples therapy? 2) How do therapists manage the multiple systems of therapeutic alliance in couples therapy? 3) How does therapeutic alliance evolve over time in couples therapy? Research Design This study employed a process-oriented approach in order to provide an in-depth examination of therapeutic alliance with couples. Empirical research declares that proper assessment of the alliance necessitates a triangled approach that includes the four perspectives of the therapist, both partners of the couple, and independent observer (Friedlander et al., 2006a). 50 To achieve this, the current study utilized a triangulation of multiple sources of data that includes quantitative and qualitative data. Data in this study were collected through semi-structured interviews, observations and self-report measures to discover and identify key themes and patterns of alliance formation based on the experiences of each member of the therapeutic system and an independent observer. The SOFTA (Friedlander et al., 2001) self-report measures were used to measure alliance scores for therapist and each member of the couple at each session. The SOFTA (Friedlander et al., 2001) observational scale was used to provide an independent observer view of the alliance for therapist and members of the couple during each session. Finally, semi-structured interviews were incorporated at specific points during treatment to identify important dimensions of alliance formation from all three perspectives of therapist and each member of the couple. The combination of self-report, observational and qualitative interviews created an inductive approach that allowed concepts and themes to emerge from patterns in the data (Patton, 2002). Rationale for a Mixed-Method Research Design Prevailing research emphasizes the need for a more in-depth look at therapeutic alliance to gain a deeper understanding of its nature and formation in conjoint treatment. Adapting a mixed-method research design can achieve this goal because it enables a deeper understanding of participants‘ experience. Quantitative measures can provide statistics to describe how the therapeutic alliance changes over the course of treatment, while qualitative inquiry can perhaps identify themes and patterns in the data that explain why the changes occurred. Together, they can produce a clearer picture of therapeutic alliance in conjoint treatment (Marshall & Rossman, 2006). Additionally, the goal of this study was to develop a comprehensive understanding of how therapeutic alliance forms with couples by adopting a process oriented approach. Mixed 51 methodology enhances process research because it elicits idiosyncratic elements of a phenomenon. Therefore, utilizing mixed methods enabled this study to achieve a degree of depth and detail that is not possible through a singular methodology. Data Collection Research Site The present study was conducted in a university affiliated outpatient clinic at a large Midwestern campus. The clinic was staffed by Master‘s and Doctoral level student therapists who were in the Commission on Accreditation for Marriage and Family Therapy Education (COAMFTE) accredited marriage and family therapy doctoral program. All therapists were supervised by licensed, Ph.D. level supervisors. Due to the training nature of the clinic, all therapy sessions were taped for proper supervision. Additionally, each treatment room is connected to a control room by close circuit monitoring to allow for live supervision with supervisors and other training therapists. The clinic provides therapeutic services to couples, families, and individuals throughout the community. Couples commonly seek services for premarital/marital issues, remarriage adjustment, family communication, parenting concerns, relationship building, and coping with stress and life transitions. Recruitment Procedures Therapists. Two therapists were recruited by the researcher from a variety of therapists in the training clinic. It was decided to recruit two therapists in order to secure enough couples needed for the research and to capture more than one style of therapist to increase the credibility of the results. For purposes of this study, participating therapists had to meet the following requirements: hold a Master‘s degree in MFT, able to take new couples at the clinic, and willingness to be involved in the research. The decision to focus on doctoral level student 52 therapists was derived after careful consideration about how skill and educational level of therapists may influence the study. While alliance development is largely based on individualities of therapists, this study sought to at least eliminate the variable of level of education. Several other doctoral level therapists in the clinic that met the above criteria were not taking new clients either due to having a large caseload or because they were phasing out of the clinic. This selection process potentially has bias in that a therapist with openings may or may not have them due to issues of retention. This was not thought to be the case with these two therapists. While the willingness to participate may also have bias, the climate in the clinic is one of collaboration with research and would not be considered unusual. Couples. Couples were recruited when they were assigned to either of the two participating therapists. It is important to note that client assignment procedures were practiced as usual during the time of the research study. Couples were assigned to the two participating therapists based on scheduling and availability. At the time of the initial intake call in which the couple was assigned to either of the participating therapists, the clinic coordinator would read a script (See Appendix A) that invited them to participate in the research project. If they indicated interest, the clinic coordinator told them they would be contacted by the researcher to discuss the details of the study and their participation once they set an appointment with their therapist. Once the couple was contacted by the treating therapist to arrange the time of the initial session, the researcher called them to explain the details of the project. At the time of the phone call, the nature of the research and the requirements of their participation was explicitly detailed. Couples were also assured that declining the invitation would not affect their ability to receive counseling services. After the couple gave verbal consent on the phone, they were asked to arrive to their session 10-15 minutes early in order to meet with the researcher alone. 53 Informed Consent. The researcher met with each participant to obtain informed consent. Separate informed consents were created for the therapist and client participants (See Appendices B and C). Therapists were approached immediately after the researcher attained approval from the University‘s institutional review board for human subjects to explain the research study and the details of their participation. Verbal consent was necessary in order to identify which cases would participate in the study. Written consent was not obtained, however, until they were assigned a couple who agreed to participate in the study. Official informed consent (see Appendix B), therefore, was collected from the therapists at the time of the initial intake session. On the date of the initial intake session, the researcher greeted the couple in the waiting room and gave them a few minutes to review the informed consent. The researcher then sat down with them in a confidential therapy room to read the consent to them and inquire about any concerns or questions. The consent form (see Appendices C) described the research study and identified criteria for participation, the data collection procedures, possible risks and benefits, aspects of confidentiality, handling of data, and their rights as research participants. The consent emphasized that their participation is voluntary and they withhold the right to revoke their consent to participate at any time during the study. They were also assured that their refusal to participate in the study, or revocation of their involvement during the study would not affect their ability to participate in treatment at the clinic. Compensation. Each member of the couple received a Visa gift card for $25 any time they participated in an interview. Thus, when a couple completed an interview, they received a total of $50 in gift cards. Each member of the couple signed a receipt indicating they received the compensation (see Appendix D). The therapists in the study were not compensated. 54 Participants Therapists. The two therapists who agreed to participate in the study had Master‘s degrees in Marriage and Family Therapy and were both in their second year of the doctoral program. Therapist 1 was a Hispanic/Caucasian male and Therapist 2 was an African-American female. Table 3.1 depicts the demographics and experiences of the two therapists and assigns a pseudonym to protect the confidentiality of the therapists involved. Table 3.1. Demographics of Therapists Gender Age Ethnicity T1 Chad Male 29 Hispanic and Caucasian T2 Leah Female 27 AfricanAmerican Degree/ License M.S. in MFT M.A. in MFT Limited Licensed MFT Experience Theory nd Primary theory is 2 year Bowenian/ Doctoral Student with Transgenerational 4 years of clinical experience nd Primary theory is 2 year Bowenian/ Doctoral Student with Transgenerational 4 years of clinical experience Couples. Six couples consented to participate in the study, but only five couples were included in the study. One couple only completed one session and was eliminated from the study. One couple dropped out of treatment after only five sessions. The researcher decided to include this couple because they completed one interview, and thought that they could still provide meaningful data to the study. The age of the individual partners of the couples at the time of the study ranged from 26 to 50 with a mean age of 35 (SD=8.53). Table 3.2 provides the demographics of each couple and their respective pseudonyms, as well as the total number of 55 sessions and interviews. The table is followed by brief narrative descriptions of each couple. Table 3.2. Demographics of Couples Couple Couple 1 Bob Age 46 Number of Sessions 10 total sessions 8 couple sessions Number of Interviews Caucasian 8 couple sessions 2 Caucasian 5 couple sessions 1 10 couple sessions 2 9 couple sessions 2 Race/Ethnicity Caucasian Therapist 1 Susan 50 Couple 2 Brad Anna 40 39 Mark 36 Heather 30 Adam 28 Erin 27 Caucasian Joe 28 African-American Christina 26 Caucasian 1 Therapist 1 Couple 3 Therapist 2 Couple 4 Therapist 2 Couple 5 Therapist 2 Couple 1. This couple was Bob and Susan. They were a middle-aged married couple selfreferred to the clinic for marital counseling. Their primary concern that brought them to therapy was Bob‘s struggle with hoarding. Their goals for therapy were to increase communication and resolve conflict due to the impact of the hoarding. They completed a total of 10 sessions and one interview. Their participation in therapy was sporadic and inconsistent. They would often arrive late for sessions, not show for scheduled appointments, and cancel numerous sessions in between visits. For instance, there were seven weeks between the fourth and fifth session. During the tenth session, the couple and therapist decided that it would be best for the couple to do a couple communication program that they could do at home due to their busy schedules. After they completed the program, they were to call to schedule a session with the therapist. The couple did 56 not call to return to therapy and did not respond to any attempts on part of the therapist or researcher to contact them to schedule a time to meet. Their case was eventually closed after two months of no contact. Couple 2. This couple was Brad and Anna. They were a married couple who sought therapy to increase communication and strengthen their relationship after two years of separation. They completed a total of eight sessions and two interviews. In the midst of treatment, they found out that Brad needed to relocate out of state for his employment. Their completion of treatment, therefore, was not due to successful achievement of their goals, but rather due to the timeline of their relocation. Couple 3. This couple was Mark and Heather. This unmarried, not living together couple sought therapy to attain relationship counseling and guidance. Heather had a disability that impaired some motor, speech and interpersonal functioning. Her disability was a continuous factor in the couples‘ dynamics, although she was quite independent. Heather was a also a single mother of a school-aged daughter. This couple completed 5 sessions of therapy and one interview. At the end of the fifth session, Heather became very angry with Mark while they were filling out the paperwork. She left angry and then called the therapist to express her anger toward her and the treatment. They did not return to therapy and would not respond to the researcher‘s attempts to contact them to discuss what happened. Their case was closed after a month of no contact. Couple 4. This couple was Adam and Erin. They were an unmarried, cohabitating couple who had been dating for 7 years. Adam had children with other partners, and the couple had one young child together. They sought therapy because they were unsure of whether to stay together or separate due to their high level of distress and uncertainty about the future of their 57 relationship. The couple completed 10 sessions and 2 interviews. The couple dropped out of treatment after the tenth session. The couple did not return to therapy nor responded to either Leah‘s or the researcher‘s attempts to contact them. Their case was closed after a month of no contact. Couple 5. This couple was Joe and Christina. They were an unmarried, cohabitating couple with a blended family of five children who had been dating for 5 years. They sought therapy due to constant conflict and trust issues. They wanted to strengthen their relationship by learning new tools for communication and conflict resolution. They completed 9 sessions and 2 interviews. They did not show for their 10th session, and did not return Leah‘s phone calls or the researcher‘s attempts to contact them. Their case was eventually closed after no contact was made for two months. Procedure First Session. As discussed previously, details of the research study were given to the couple at the time of the first session. The informed consent, goals of project, and nature of their involvement were explained. The couple consented to completing questionnaires at the end of each session, as well as interviews with the researcher during the beginning, middle, and late phases of treatment. The couple were required to complete a general family background questionnaire and the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995; See Appendix A) which are included in the initial intake forms as part of standard clinic protocol. Additionally, their involvement in services at the clinic required that they consent to having their sessions video recorded. These videos were used in the research project as a source of data and were coded by using the observational form of SOFTA (Friedlander et al., 2001). The first 58 session was recorded as this marks the first point in data collection. The couple and therapist also completed the self-report SOFTA during this session (Friedlander et al., 2001). Points of data collection. There were multiple points of data collection throughout the therapeutic process that will include both quantitative and qualitative measures. Each session was video recorded for observational ratings of the alliance. The couple and therapist completed the appropriate self-report forms of SOFTA following each session (Friedlander et al., 2001). Qualitative interviews with the therapists and couples were individually conducted after the fourth, eighth, and twelfth or final session (whichever came first). Table 3.3 provides an outline of the various data points. Table 3.3. Points of Data Collection Family Data Information RDAS SOFTA-O Each Session Therapists Clients SOFTA-S Intake Session Intake Session Each Session Long Interview After 4, 8, and 12 Couple sessions (or last session whichever comes first) After 4, 8, and 12 Couple sessions(or last session whichever comes first) Researcher Each Session Measures Quantitative Measures Both members of each couple completed the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995) in its entirety at the initial intake session. This is a standard assessment that is 59 required for every couple participating in treatment to complete at the onset of therapy. The purpose of utilizing the RDAS (Busby et al.; 1995) for this study was to help identify any elements, such as level of relationship functioning, that may influence the formation and level of alliance. Following each session, both the therapist and the individual members of the couple completed the System for Observing Family Therapy Alliances (SOFTA-s; Friedlander et al., 2001) self-report forms beginning with the initial intake session and continuing until the twelfth or final session (whichever came first). These two instruments will be described below. Revised-dyadic adjustment scale (RDAS; Busby et al., 1995). The RDAS (Bugsby et al., 1995) is a 14-item self-report measure of marital and relationship quality (See Appendix E). Both partners of the couple rated the extent to which they agree or disagree with their partner on specific issues and how often they engage in various behaviors on a 6-point Likert scale ranging from 0 (never) to 5 (always). The RDAS encompasses the three areas of functioning: dyadic consensus, dyadic satisfaction, and dyadic cohesion. Dyadic consensus is the extent of agreement between partners in relation to money, religion, recreation, friends, household friends, and time spent together. Dyadic satisfaction is the amount of tension in the relationship and the extent to which the individual has considered ending the relationship. Finally, the dyadic cohesion is the common interests and activities shared by the couple. An example of an item in this area is, ―Do you and your mate engage in outside interests together?‖ All items were scored and summed to form a total marital satisfaction score. Individual scores are compared to clinical and non-clinical averages, and have a range of scores from 0 to 69. Low scores (47 and below) indicate the presence of problems. High scores (48 and above) are indicative of the absence of a problem. The established population mean for non-distressed couples is 52.3 (SD=6.60), while the mean score for distressed couples is 41.6 (SD=8.20). 60 Internal consistency reliability for the RDAS reports at .90 (Busby et al., 1995). In addition, the subscales have fair to excellent internal consistency with the dyadic satisfaction (DS) at .94, dyadic cohesion (DCoh) at .81, and dyadic consensus (DCon) at .90 (Busby et al., 1995). The RDAS has acceptable levels of construct validity as demonstrated by several confirmatory factor analyses with more than one sample. Despite the brevity of the RDAS, it is as successful at discriminating between distressed and non-distressed individuals, which illustrates known-groups validity (Busby et al., 1995). There is also evidence that the RDAS is as highly correlated with DAS and the Locke-Wallace Marital Adjustment Scale (Busby et al., 1995). System for observing family therapy alliances self-report (SOFTA-s; Friedlander, Escudero, & Heatherington, 2001). SOFTA-s questionnaires assess the clients‘ and therapist‘s rating of alliance in each session (See Appendix G and H). SOFTA (Friedlander et al., 2001) was developed to address the gaps in the literature that necessitated a look into the actual in-session behavior of consistent findings about the effects of client conflict and negativity on CFT outcomes (Friedlander et al., 2001). This measure takes into account the unique positive and negative items reflecting different aspects and levels of client collaboration in CFT (shared sense of purpose and safety within the therapeutic setting), as well as Bordin‘s (1979) classic idea of mutual collaboration and bonds. The SOFTA is a set of tools that include observational assessments (SOFTA-o) and self-report (SOFTA-s). There are four dimensions of the scale that are depicted on both (client and therapist) versions of the SOFTA-s for a total of 16 statements. Thus, the two versions measure the same aspects of the alliance, however the statements are changed to reflect the client or therapist perspective. Each dimension has four items. The first dimension is ―emotional connection with 61 the therapist‖ (a sense that the relationship is based on affiliation, trust, caring and concern). An example of a positive connection on the client form is ―The therapist understands me.‖ The corresponding statement on the therapist form is ―I understand this family.‖ The second dimension is ―engagement in the therapeutic process‖ (viewing the treatment as meaningful and having a sense of involvement and working together with the therapist). Bordin‘s (1979) scales of goals and tasks in therapy were combined into this dimension because of their high intercorrelation (.92) as reported in the self-report of the WAI (Friedlander et al., 2006b). ―The therapist and I work together as a team‖ is a statement on the client form that represents positive engagement. On the therapist version, this statement reads as ―The family and I are working together as a team.‖ The third dimension is ―safety within the therapeutic system‖ (client viewing therapy as a place to take risks; having a sense of comfort). An example of a statement that indicates negative safety on the client version is ―There are some topics I am afraid to discuss in therapy.‖ The equivalent statement on the therapist version is ―There are some topics that the family members are afraid to discuss in therapy.‖ The final dimension of the SOFTA-s is ―shared sense of purpose within the family‖ (family viewing themselves as working collaboratively to improve family relation and achieve goals). An illustration negative shared purpose on the client and therapist versions of the scale is ―Some members of the family don‘t agree with others about the goals of therapy.‖ The last two dimensions, shared sense of purpose and safety in the therapeutic system, reflect the uniqueness of CFT. Each version of the SOFTA-s has 16 positive and negative items to be rated on a 5-point Likert scale (1 = not at all, 5 = very much). Ratings for all four dimensions can be summed for a total score that indicates alliance overall, as well as derived for each subscale to show variability across dimensions in order to depict the varying experiences of the family members (Friedlander 62 et al., 2006b). The total score can range from 16 to 80, while the four subscale scores range from 4 to 20. Relations among the four dimensions of Engagement, Emotional Connection, Safety, and Shared Purpose range from .18 to .75 (Friedlander et al., 2006). The two most closely associated dimensions are engagement and emotional connection (r =.75, p ＜.0001), while the dimensions of safety and shared purpose are the least closely associated (r =.18, p ＜.025). This low correlation reflects the differing focus of these dimensions. Safety focuses on individual behavior in context, while Shared Purpose concentrates on family interactions. All other intercorrelations are moderate (rSD=.30-.47, ps ≤ .0001) (Friedlander et al., 2006b). The internal consistency reliability of the 16-item SOFTA-s is α =.87 (client form; English) and .95 (therapists form; English). The reliabilities for Engagement, Emotional Connection, and Shared Purpose are α = .62 (client form; English) and α =.88 (therapist form; English). The SOFTA-s has demonstrated sound concurrent and predictive validity. Concurrent validity is supported in three ways. First, significant associations were found between some observational and self-report scales. Clients‘ Emotional Connection and Shared Purpose-related behaviors (SOFTA-o ratings) in Session 6 were significantly associated with their self-reported perceptions of the family‘s Shared Sense of Purpose (SOFTA-s) obtained immediately after that session. These results support the supposition that alliance-related behaviors in the SOFTA-o reflect clients‘ thoughts and feelings about the alliance with the therapist (Friedlander et al, 2006a). Also, a significant association was found between client-and therapist-rated SOFTA-s Shared Purpose scores after Session 6. Finally, significant associations were found between some SOFTA-s score with the Penn Helping Alliance Questionnaire (Luborsky et al., 1983) and the Working Alliance Inventory-Couples (Horvath et al., 2003). 63 Predictive validity for the SOFTA-s is also demonstrated for SOFTA-s. Two studies have shown significant positive associations were found between the SOFTA measures and postsession scores on the Stiles and Snow‘s (1984) Session Evaluation Questionnaire (Friedlander, Escudero, Haar, & Higham, 2005; Friedlander et al., 2006b). Additionally, significant positive associations were found with therapist-rated (SOFTA-s) scores on all four dimensions after the sixth session and their ―estimate of improvement so far‖ on the Pen Helping Alliance Questionnaire (Luborsky et al., 1983). Qualitative Measures Observational measure: System for observing family therapy alliances observational scale (SOFTA-O; Friedlander et al., 2001). The SOFTA-o was used by the researcher to assess therapeutic alliance behaviors during each session of all couples participating in the research project. It is comprised of two versions (client and therapist) that provide an observer rating scale of client behaviors and therapist contributions to the alliance (See Appendixes I and J). Each version has a list of positive and negative behaviors that exemplify the same four dimensions as the self-report measures: (1) ―Engagement in the Therapeutic Process (e.g., ―Client complies with therapist‘s requests for enactments‖ and ―Therapist explains how therapy works‖); (2) Emotional Connection (―Client verbalizes trust in the therapist‖ and ―Therapist expresses caring or touches clients affectionately yet appropriately); (3) Safety with the Therapeutic System (―Client shows vulnerability‖ and ―Therapist provides structure and guidelines for safety and confidentiality); (4) Shared Sense of Purpose (―Family members ask each other for their perspective‖ and ―Therapist praises clients to for respecting each other‘s point of view‖). Assessment includes recording the presence of these specific items in the entire session, and making global ratings for each SOFTA-o dimension on a -3 (extremely problematic) to +3 64 (extremely strong) ordinal scale, where 0 = unremarkable or neutral. Specific guidelines help raters determine the appropriate rating based on the frequency, intensity, and context of the observed behaviors. In the client version, individual family members are rated separately on Engagement, Emotional Connection, and Safety, while the entire couple of family unit is rated on Shared Sense of Purpose. In the therapist version, the therapist is rated on each dimension (Friedlander et al., 2006a). The SOFTA-o (Friedlander et al., 2001) has moderate to high intraclass correlations in interrater reliability tests (0.72 to 0.95). Additionally, sound evidence of known-groups, factorial, concurrent, and predictive validity have been demonstrated for the SOFTA-o. In terms of known-groups validity, statistically significant comparisons of participants‘ ratings for strong and weak alliances support the construct validity of each dimension of the SOFTA-o and indicate that the measure accurately reflects clinically meaningful alliance-related behavior (Friedlander et al., 2006a). To measure factorial validity, an exploratory factor analysis to identify any higher order factors in the client SOFTA-o showed that a single factor accounted for roughly half the variance. This indicates that the four SOFTA dimensions are various aspects of a single construct-therapeutic alliance (Friedlander et al., 2006a). Concurrent validity was demonstrated through associations between some observational and self-report scales (see previous discussion of these above under validity of SOFTA-s). Additionally, qualitative themes in post-session interviews about the alliance with individual family members were congruent with their SOFTAo ratings, and the families‘ SOFTA-o Shared Purpose ratings were consistent with clients‘ selfreported within-system alliances on Pinsof‘s (1999) Family Therapy Alliance Scale-Revised (Friedlander et al., 2006a). Finally, predictive validity was supported by two studies. In one study, significant associations were identified between SOFTA-o ratings and adolescent-therapist 65 interactions reflecting verbal relational control (Friedlander et al., 2005). An analysis of four case studies shows that low SOFTA-o Safety and Shared Purpose ratings early in therapy were observed only in less successful families and favorable ratings on all four SOFTA-o dimensions were associated with more successful families early in treatment (Beck, Friedlander, & Escudero, 2006). Semi-structured interviews. Post-session Interviews were conducted with both members of the couple and therapist independently during treatment. These interviews were semi-structured and include questions that pertain to the content of their sessions that demonstrate important influences of alliance (See Appendixes J and K). The interviews were scheduled with clients in advance to be conducted after the fourth, eighth, and twelfth couple sessions, unless treatment ended earlier. Thus, the interviews did not necessarily take place after the designated sessions if there were individual sessions during that phase of treatment. The routine semi-structured interviews consisted of questions designed to elicit information regarding the couple and therapist‘s experience of the therapeutic alliance. A semistructured interview is commonly used in qualitative research due to the emphasis on preventing preexisting conceptions and theory from limiting the information gathered and allowing concepts and hypotheses to emerge inductively from the data (Strauss and Corbin, 1998). The interview questions were derived from the research questions and any observed changes in the alliance as indicated via the quantitative measurement of the SOFTA-s and the observational measurement of the SOFTA-o as treatment unfolds (See Tables 3.4 and 3.5). The investigator developed a set of open-ended questions based on the existing theoretical concepts of therapeutic alliance in conjoint treatment. The questions for clients and therapists differ slightly in terms of point of reference as well as theoretical ideas. Clients were asked to evaluate and discuss their 66 experiences of the therapeutic alliance from their perspective, while therapists were asked to report on the same from their reference, as well as some general questions of their understanding of the alliance. Tables 3.4 and 3.5 delineate the interview questions and their corresponding research questions. Table 3.4. Client Interview Questions Paired with Associated Research Questions and Alliance Components Interview Questions 1. How do you feel about the relationship that you and your therapist have established? 2. What type of therapist behaviors/skills/characteristics do you feel were instrumental in developing a relationship with your therapist? Research Question (RQ)/ Alliance Process Component General/Open-ended Question RQ #1: What specific therapeutic elements and techniques are effective in building therapeutic alliance in couple‘s therapy? 3. Can you tell me about any times that either of you feel more or less connected with your therapist as individuals? As a couple? RQ #2: How do therapists manage the therapeutic alliance between the multiple systems in conjoint treatment? 4. How do you think your relationship with your therapist changes from session to session? RQ #3: How does therapeutic alliance evolve over time in couple‘s therapy? 67 Table 3.5. Therapist Interview Questions Paired with Associated Research Questions and Alliance Components Research Question (RQ)/ Alliance Process Interview Questions Component 1. What is your understanding of the General/Open-ended Question to elicit therapeutic alliance? understanding of alliance 2. How do you feel about the relationship that General/Open-ended Question you and your clients have established? 3. What specific therapeutic elements and RQ #1: What specific therapeutic elements and techniques have you used to build therapeutic techniques are effective in building therapeutic alliance with your couple? alliance in couple‘s therapy? 4. How do you build the relationship with both individuals of the couple as well as the couple as a whole? 5. How do you think your relationship with your clients changes from session to session? RQ #2: How do therapists manage the therapeutic alliance between the multiple systems in conjoint treatment? RQ #3: How does therapeutic alliance evolve over time in couple‘s therapy? Data Analysis This study utilized multiple sources of data that included quantitative, observational, and qualitative measures. Each session provided four perspectives of data: self-report of alliance by each individual member of the couple, self-report of alliance by therapist, and observation of alliance by researcher. This allowed for triangulation of data in order to assess congruency between all perspectives of alliance. Triangulation refers to combining methods and/or sources of data as a way to enhance understanding of the phenomena or people being studied (Taylor and Bogden, 1998). This section will address how each of these sources were analyzed. Quantitative Measures: RDAS and SOFTA-s All quantitative self-assessments including the RDAS (Busby et al., 1995), and SOFTA-s (Friedlander et al., 2006a) (client and therapist versions) were scored by the researcher. The results of the quantitative measures were analyzed in the Statistical Package for Social Sciences 68 (SPSS; Version 16.0) to provide descriptive data for the study, as well as ensure validity of the qualitative measures. The RDAS scores provided quantitative support to the observed couple dynamics that influenced the therapeutic progress and overall alliance between the couples and therapists. SOFTA-s ratings were used as primary indicators of the clients‘ perspectives of the alliance and analyzed to assess how alliance progressed through treatment. Observational Measure: SOFTA-o The researcher observed and coded each therapy session using the SOFTA-o (Friedlander et al., 2006a). The SOFTA-o is designed to be completed by trained raters while observing a videotaped couple/family therapy session. The researcher was trained through a tutorial computer program available online (www.softa-soatif.net). Sixteen training vignettes were downloaded that depicted brief excerpts of sessions showing several positive and negative behaviors of alliance. The researcher rated each vignette and compared the ratings with the ―Solution‖ (also available online), which explained each marking of the individual items for the vignettes. When there were discrepancies between the researcher‘s ratings and those depicted in the ―Solution‖, the researcher consulted the operational definitions of the items for review. The researcher first focused on the therapist version to learn how to code sessions because it was easier to focus on one individual‘s behavior at a time. Once the researcher‘s results on the vignettes were comparable to those in the ―Solution,‖ she began training on the client version. She repeated the process until her ratings were similar to the ―solutions‖ for the client version. At this point, she was ready to begin coding the sessions. Since Friedlander et al. (2006a) do not recommend rating both clients and therapists at the same time, the researcher decided that she would need to watch sessions at least two times in order to focus on both the client and therapist behaviors. Thus, the initial observation for each session was focused on 69 therapist behaviors. The second review of the session was then concentrated on the behaviors of the clients. This process was repeated for all 42 therapeutic sessions in the research project. In order to achieve good interrater reliability, the researcher invited a graduate student who was an intern at the clinic to rate sessions. This graduate student was a doctoral level therapist, and had a strong understanding of alliance in couple and family therapy. The intern was given a description of the research project, as well as various articles about SOFTA and alliance in couple and family therapy in order to gain greater understanding of the theoretical foundation of the instrument. After she was familiar with the dimensions of SOFTA, there were two meetings to discuss the operational definitions of each behavior and their discriminating features. The researcher and the graduate student felt confident about their ability to identify features of the domains of Engagement and Shared Sense of Purpose, but occasionally struggled with Emotional Connection and Safety due to their affective nature. This challenge was expected as Friedlander et al. (2006a) reported that these emotion-laden dimensions tend to be more challenging for raters. Therefore, the second meeting was focused on these dimensions so that both raters felt more comfortable in their ability to properly identify the behaviors. The graduate student then downloaded the training vignettes and completed the tutorial for both the client and therapist versions. She consulted the training manual and operational definitions when her results were inconsistent with the ―solutions.‖ The researcher then picked 15 sessions that were representative of the data set and provided examples of negative and positive Engagement, Emotional Connection, Safety, and Shared Purpose. The raters focused on the client for five of those sessions, and the therapist for ten of the sessions. This specific number was chosen due to a recommendation by Friedlander et al. (2006) that posits that this number of sessions would increase the strength of interrater reliability. The researcher and graduate student 70 reviewed two sessions together and openly discussed their observations of the behaviors on each dimension until they mutually decided on ratings. For one session, they focused on the therapist, and for the other, they focused on the clients. Interestingly, the ratings of the researcher and graduate student were generally consistent for the clients. Conversely, they had discrepancies with the ratings of the therapist. This appeared to be due to their own experiences as therapists and assumptions about the treating therapists‘ behaviors. After extensive discussions about this, it was decided to review another session together while focusing on the therapist. During this rating session, the two raters increased their reliability and felt they were ready to advance in the rating process. To assess interrater reliability, the researcher and graduate student observed 12 more sessions independently. Eight sessions were focused on the therapist while the other four were focused on the clients. They met weekly to discuss and compare the results of the ratings. The behavioral tallies for each dimension were compared, as well as the global score for each person. Friedlander et al.(2006a) suggest that raters can compute reliabilities once they reach a point where their dimensional ratings do not differ by more than a single scale point for at least 90% of the time. Thus, reviewing of sessions occurred any time there was a difference in the ratings of more than one point. This occurred twice during the reviewing process. Reviewing ceased when the raters were able to reach consensus or have scores that differed no more than one point. The researcher then performed an interrater reliability analysis using Krippendorf‘s Alpha (Krippendorf, 1980; 2004) for the entire scale and Intraclass Correlation (ICC) to determine the consistency between raters for each SOFTA dimension. 71 Semi-Structured Interviews Data analysis of the interviews was based on qualitative procedures universal to qualitative research (Gilgun, 1992; Strauss & Corbin, 1990). All interviews were transcribed from digital video discs by the researcher. The interview data were used to explore factors, characteristics, and circumstances that contributed to the therapeutic alliance development for each member of the therapeutic system. Transcripts were inductively evaluated utilizing thematic content analysis to identify the recurring patterns or common themes that cut across the data. This form of analysis portrays the thematic content of interview transcripts by identifying common themes and patterns of living and/or behavior (Benner, 1985; Leininger, 1985; Krippendorff, 2005; Taylor & Bogdan, 1998). Thematic analysis consists of exploration of themes that emerge as being important to the description of the phenomenon (Creswell, 2003). It is a form of pattern recognition within the data, where emerging themes become the categories for analysis. The task of the researcher is to search for the themes that emerge from the interview data through careful examination and re-examination of data. Thematic content analysis is the most foundational of qualitative analytic procedures and in some way informs all qualitative methods (Creswell, 2003). This method of analysis complemented the aims of the study by allowing for theory driven research questions and measures while enabling themes to emerge directly from the data through inductive coding. Themes are the concepts identified by the researcher before, during, and after data collection (Lincoln & Guba, 2000). They are defined as units derived from patterns such as "conversation topics, vocabulary, recurring activities, meanings, feelings, or folk sayings and proverbs" (Taylor & Bogdan, 1998, p.131). Leininger (1985) posited that themes are identified by combining components or fragments of ideas or experiences that would otherwise 72 be meaningless when observed alone. Themes that emerge from the participants‘ experiences are combined to generate an inclusive representation of their collective experience. How these experiences are woven together, however, is based in the eyes of the analyst. As Leininger (1985) states, ―the coherence of ideas rests with the analyst who has rigorously studied how different ideas or components fit together in a meaningful way when linked together" (p. 60). Thus, it is important that the process of the researcher is explicitly delineated so that it can be discernable at any given time by any given reviewer. Coding. The coding process involved identifying important themes and encoding them prior to the interpretation process (Cresswell, 2003). The goal is to identify codes that capture the qualitative richness of the phenomenon. Encoding the information arranges the data in a way that allows the researcher to identify and develop themes (Benner,1985; Krippendorff, 2005; Leininger, 1985; Taylor & Bogdan, 1998). From the transcribed interviews (n=24), themes of experiences of alliance were analyzed and developed using the four step process put forth by Creswell (2003). First, the interviews were reviewed to increase familiarity with the data. Next, notations of emerging themes were made in the transcripts alongside the raw data. After this was completed for each interview, coding began by identifying these patterns and/or themes that were significant to the experiences of the participants and the research questions. These themes were listed using either direct quotations or paraphrased common ideas. The next step was to identify all data that relate to the already classified themes. Thus, the patterns are expounded upon by detailing all the corresponding themes. These related themes were combined and categorized into subthemes. Subthemes help to give a comprehensive view of the information that aids in recognizing pattern developments (Creswell, 2003). 73 Broad themes were then developed based on the research and interview questions. These themes were reviewed and subjected to more extensive analysis in order to develop more defining categories. This critical analysis established the rationale for choosing the themes based on prevailing literature. Supporting the themes with established research strengthens the findings and increases the credibility of the researcher. Additionally, a well-developed narrative facilitates comprehension of the process, understanding, and motivation of the researcher (Creswell, 2003). The final template identified broad categories and the explicit themes produced from the interviews. Subthemes were then applied to those themes that required more description. Trustworthiness Qualitative inquiry entails alternative methods from those common in quantitative research to establish reliability and validity in the research design and findings. Thus, researchers must utilize models suitable to qualitative designs to ensure rigor without sacrificing the relevance of qualitative research. In quantitative research, investigators provide evidence of internal validity, external validity, reliability and objectivity to demonstrate the credibility of their study. Qualitative research can provide confirmation of trustworthiness by adapting Guba‘s (1981) model for evaluation of qualitative research that employs corresponding constructs to the aforementioned quantitative methods of fidelity. The present study adapts this model to ensure trustworthiness by demonstrating: a) credibility (internal validity); b) transferability (external validity); c) dependability (reliability); and d) confirmability (objectivity) (Guba, 1981). Credibility Lincoln and Guba (1985) argue that demonstrating credibility is one of the most significant indicators of trustworthiness .Credibility refers to the confidence that the study has accurately recorded the phenomena under examination. This study established credibility by 74 disclosing reflexivity, establishing theoretical sensitivity, examining research findings, triangulation, and debriefing sessions. Reflexivity: Self of the researcher. Qualitative studies require researchers to interact with the empirical materials, wrestle with data, and interpret the results. The entire research process is influenced by the past experiences, values, and beliefs of the investigators conducting the study. Thus, neither the data nor the ideas are passively observed and compiled (Holstein & Gubrium, 1995). Furthermore, what researchers know shapes what they find (Denzin, 1994). Therefore, a qualitative researcher is a tool of the research. He or she acts as an interpreter or translator of the lived experiences of people‘s lives. These interpretations are ―filtered through the humanness of the researcher who, draws on their own experiences, knowledge, theoretical dispositions, and collected data to present their understanding of the other‘s word (Glesne & Peshkin, 1992, p. 153). The challenge then, is for the researcher to demonstrate that his or her personal interest will not bias the study (Marshall & Rossman, 2006). This can be achieved through reflexivity. Reflexivity is defined as ―self-awareness and the agency within that self-awareness‖ (Rennie, 2004, p. 183). It refers to the importance of making covert biases and assumptions overt to self and others. Two ways to exhibit reflexivity are to utilize a self- reflective journal and disclose the researcher‘s background and experiences in order to identify any potential biases that may impact the analysis. The researcher is a Caucasian female in her early thirties. She has a Master of Science in Marriage and Family Therapy and works in private practice. She has 8 years of clinical experience working with individuals, couples, and families. She particularly enjoys working with couples and has focused the majority of her academic career on strengthening her ability to work 75 with them. This commitment to improving her treatment of couples is influenced by her passion for the subject. Therefore, she is particularly interested in the theories and factors of couples therapy that are correlated with successful treatment outcomes. One of those factors is therapeutic alliance. While she has been familiar with therapeutic alliance throughout her clinical experience, she was frustrated with the lack of training and understanding of how alliance is developed and maintained with couples and families in therapy. One of her greatest strengths as a clinician is her ability to effectively build a trusting relationship with her clients, yet she was aware of how the context of conjoint treatment can pose challenges to building this relationship when there are multiple clients in the room. Additionally, she also continuously questioned what roles the therapist and clients have in building the therapeutic relationship. As she researched these different issues, she noticed the lack of research focusing on therapeutic alliance with couples and families. As she progressed through the doctoral program, she decided to focus her comprehensive exams and dissertation on this topic by conducting a study that monitored alliance in couples therapy over the course of therapy. As a result, the researcher has awareness, knowledge, and beliefs about the phenomenon being studied (Creswell, 2003). The researcher chose to conduct this study at the training clinic in her department. It is important to disclose that this site is also where she worked as a doctoral student, as well as where she completed her internship. Therefore, she was familiar with the clinic and the other student therapists that provided services to clients. Due to the experience and involvement of the researcher, biases must be addressed (Yin, 2009). Acknowledgement of biases is defined by Creswell (2003) as ―the capacity or quality of the researcher to think in terms about the emerging data based on personal and professional experiences, empirical and theoretical knowledge, and personal insights and understandings‖ (p. 76 183). The researcher monitored her biases throughout the research process and continuously challenged herself to understand how they influenced the data collection, analysis, and discussion of findings. While she appreciated her knowledge about the intricate dimensions of alliance in couples therapy, she also questioned her reflections every step of the way to ensure the results were products of the data and not her hypotheses or expectations. This task was daunting due to her clinical experience as a therapist. At times, her experience as a therapist inclined her to observe the therapist‘s behaviors with a questioning or critical stance rather than observing from a place of curiosity. Likewise, her clinical experience influenced her to switch to the therapist role instead of remaining the observer while assessing the clients‘ behaviors in sessions. She utilized debriefing sessions (discussed later in this section) and the use of selfreflective journals to balance her biases and maintain objectivity. Self-reflective journal. Keeping a self-reflective journal is one of the most valuable strategies for sustaining reflexivity (Morrow, 2005). It is a means to monitor potential biases based in the experiences and beliefs of the researcher (Creswell, 2003). There were two ways the researcher monitored biases in this study. First, a self-reflective journal was used to document thoughts, reactions, and decisions made at important points during the research process. For instance, the researcher consistently tracked any thoughts related to the study during the phases of data collection and analysis when there was a decision or interpretation made. This helped to document the process as it evolved and provided explanation for why and how decisions were made. Another way the researcher maintained a journal of reflections was to incorporate thoughts, reactions, and questions on the observational documents. Thus, when the researcher reviewed a therapy session, she would detail her reactions alongside the behavioral notations of 77 the therapist and couple. This allowed her to keep an organized tracking of her feelings as they pertained to each session. Theoretical sensitivity. Theoretical sensitivity is one way to ensure credibility by applying the findings to existing literature. Theoretical sensitivity can be derived from a variety of sources, including professional literature, professional experiences, and personal experiences (Strauss & Corbin, 1990). This study is guided by the theoretical framework of therapeutic alliance in conjoint treatment that is based in research and supported in current literature. Prior to the commencement of the present study, the researcher extensively studied the history of therapeutic alliance in individual and conjoint treatment. The concept and methodology of this study were developed through careful and extensive review of the theoretical development and successful measurement of therapeutic alliance with couples. Chapter I presents the theoretical framework of this study based on current literature that supports the conceptualization of therapeutic alliance in conjoint treatment put forth by Friedlander et al. (2006a). Chapter II is a compilation of past and current literature depicting the evolution of therapeutic alliance and its measurement in individual and couples therapy. Chapter III describes the methodology of this study that employs the successful strategies and methods of past research for measuring alliance in conjoint treatment. Therefore, this study systematically demonstrates theoretical sensitivity through the rich detailed descriptions of each chapter. Examination of previous research findings. Silverman (2000) argues that the ability of the researcher to relate his or her findings to an existing body of knowledge is a key criterion for evaluating works of qualitative inquiry. Chapter V discusses the results of the current study and the degree to which they are congruent with those of past studies. It also discusses how the 78 current study furthers the current literature on therapeutic alliance in couples therapy, as well as any limitations of the study or its methodology. Triangulation. Triangulation refers to the use of multiple methods or sources of data to study the phenomenon in question (Patton, 2002). Triangulation amplifies the credibility of the study by optimizing proper interpretation of the data. The concurrent use of different methods exploits their benefits and compensates for their limitations (Guba, 1981). To achieve a comprehensive theory of therapeutic alliance, this study utilizes multiple sources of data. It incorporates interviews, observations and self-report assessments to measure therapeutic alliance in conjoint treatment. Data from these multiple sources were converged into the analysis process rather than handled individually in order to strengthen the findings. Each source of data serves to explain, support or contradict the findings from the varying sources. This study also provides three perspectives of alliance in couples therapy by collecting data from both members of the couple, therapist, and observer. This comprehensive collection and integration of the varying perspectives of the therapeutic relationship facilitates a deeper holistic understanding of therapeutic alliance with couples. The individual viewpoints and experiences can be compared and verified with one another to create an understanding of alliance that encompasses the dyadic nature of the relationship. Debriefing sessions. Guba (1981) recommends researchers have routine meetings with members of the research team or superiors to discuss the process of the research as it develops. These sessions allow the investigator to openly discuss the findings and elicit feedback that helps to maintain objectivity (Creswell, 2003). The researcher of the current study met consistently with her committee chair during the research proposal and data collection periods. The committee chair was a Licensed Marriage and Family Therapist with 18 years of clinical 79 experience. Her clinical and research experience in couples therapy enabled her to guide the researcher through the conceptualization, development and implementation phases of the project. The meetings were beneficial to the researcher because it allowed her to openly discuss developing ideas and interpretations. Additionally, the involvement of an outside reviewer helped the researcher to identify and recognize her own biases about therapeutic alliance and couples therapy. She also assisted in problem solving and addressing any unexpected issues that arose during the research process, as well as provided an outside perspective to critically examine the emerging data. This outside scrutiny helped the investigator recognize when she was too entrenched in the data and was unable to maintain an open and objective perspective. Finally, these meetings were a prominent source of support to the researcher during the analysis and writing phases of the study by providing organizational strategies and recommendations for the overall presentation of the results. Transferability Transferability corresponds to external validity in quantitative research, and addresses the extent to which the findings of a study can be applied to other situations or to a larger population (Lincoln & Guba, 1985). Critics of qualitative research posit that it is impossible to demonstrate that findings are applicable to other situations or populations because the results are specific to a small number of participants or distinct contexts. Qualitative researchers generally propose that while transferability should be pursued with caution, it does not necessarily need to be automatically rejected. They state that while each case may be unique, it is also an example within a broader group, and therefore can achieve transferability (Lincoln & Guba, 1985; Stake, 1995; Yin, 2009). 80 In order to achieve some degree of transferability, researchers must provide sufficient contextual information about the fieldwork sites and exhaustive description of the phenomenon being studied (Guba, 1981). This allows readers to have proper understanding of the variables, which enables them to compare the instances of the phenomenon described with those that they study. This is challenging because factors that are considered by the researcher to be insignificant, may be critical in the eyes of other researchers. The researcher should demonstrate how, in terms of contextual data, the case study location compares with other locations, as well as provide the boundaries of the study (Denscombe, 1998). The researcher addressed each of these inquiries in earlier sections of this chapter. Please review the section entitled, ―Data Collection Methods‖ for information about each of these boundaries. It is important to note that the research setting was a training clinic, and results from the study may not be representative of typical alliance processes of couples in other clinical settings. Dependability Dependability is important to demonstrate in qualitative research in order to ensure comparable results will be found if the same methods were repeated in an equivalent context with similar participants. Lincoln and Guba (1985) argue that dependability is closely related to credibility, and that if researchers demonstrate one, they ensure the other. Thus, there are various ways to demonstrate dependability. One way is to use overlapping methods, or triangulation of data. The present study incorporates multiple sources of data to offer various measures of alliance for each participant. The researcher utilized self-report questionnaires, interviews, and observational data to assess therapeutic alliance in conjoint treatment. Additionally, the research study should provide a detailed report of the processes, steps, and methods to enable other researchers to repeat the work (Lincoln & Guba, 1985; Yin, 2009). 81 This extensive description allows proper assessment of the extent to which appropriate research practices have been followed. The present study provides the research design and describes each step that was planned and executed throughout the data collection process. The researcher provides an operational detail of the data gathering by addressing the procedures of recruitment, points of data collection, sources of data at each point of collection, and steps in the analysis. This provides a chain of evidence that allows the reader to walk through the analysis from operationalizing concepts to methodological procedures and results (Yin, 2009). Confirmability Confirmablilty refers to promoting objectivity in research. Objectivity in any research is difficult to ensure because the instruments are designed by humans and inclusion of the researcher‘s biases is inexorable (Patton, 2002). However, steps must be taken to verify to the greatest extent that the findings are results of the experiences and ideas of the informants and not those of the researcher. The current study used triangulation and disclosure of investigator bias to increase objectivity of the results. Triangulation reduces the effects of investigator bias by providing multiple indicators of the phenomenon. Each indicator of alliance (self-report, observation, and interviews) served to ―check‖ the other in order to ensure the results are correct. Finally, the researcher disclosed her background and experiences that could influence investigator bias. Miles and Huberman (1994) consider this step to be a key criterion for confirmability. The researcher also details the reasoning for selecting the specific instruments and adopting the specific methodology in the literature review to demonstrate transparency of her motive for choosing the research design. Once more, the researcher provides a detailed methodological description to allow the reader to determine how well the data and results can be accepted. 82 Audit trail. An audit trail is critical to the process of confirmability because it allows others to follow the progression of the research step-by-step via the decisions made and procedures described. The audit trail for this study provided a detailed account of all decisions made throughout the study. During the data gathering stage, the researcher created documents that correlated with each therapeutic session that described the observations, questions, hypotheses, and explanations for the behavior recorded. These documents also incorporated the quantitative scores from the self-reports to provide the perspectives of the therapists and clients alongside the observational data. This allowed the researcher to organize and analyze the data as it was collected through all four viewpoints. These documents also identified additional questions to ask during the interviews based on the noted behaviors and scores of alliance. During the analysis stage, the researcher created multiple documents that detailed the coding process as it was created, refined, and finalized to represent the within-case and cross-case themes. 83 CHAPTER IV RESULTS The general purpose of the present study was to obtain a greater understanding of therapeutic alliance in couples therapy by examining it through the perspectives of the clients, therapists and observer. The study utilized quantitative and qualitative measures to analyze alliance over the course of therapy for five couples. This chapter identifies the key findings obtained from the self-report, observational and interview data. Results are organized according to the research questions and include all sources of data from the multiple perspectives of therapeutic alliance. Scores from the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995), System for Observing Family Therapy Alliances (SOFTA; Friedlander et al., 2006a) selfreport and observational measures, as well as samples of quotations from the client and therapist interviews are used throughout the chapter to support each finding. Before discussing the specific findings, the results from the internal consistency analysis are presented to demonstrate that there was strong inter-rater reliability among the raters of the SOFTA-o measure. SOFTA-o: Inter-rater Reliability An internal consistency analysis performed on all four dimensions of the SOFTA-o produced a Cronbach‘s coefficient alpha of .94, suggesting that the four subscales of Engagement, Emotional Connection, Safety and Shared Sense of Purpose measure the underlying construct of alliance. These results suggest that the SOFTA-o is a reliable measure of therapist-couple alliance for this sample. Consistent with prior research on the SOFTA-o, raters were able to achieve a high degree of inter-rater reliability. A coefficient of .80 or higher is considered highly acceptable for inter-rater reliability (Krippendorf, 2004). Raters achieved a mean Interclass Correlation (ICC) of .92 for scale as a whole. Since ICC does not take into 84 account the likelihood of chance agreement between raters and is likely to inflate agreement percentages, Krippendorf‘s alpha (Krippendorf, 1980; 2004) was also used to assess inter-rater reliability. Results reported a score of .94. Thus, the results of these two tests indicate that the raters achieved between 92% and 94% degree of reliability in their coding. Intraclass correlations were also run for each dimension of SOFTA-o to identify whether certain dimensions had higher or lower reliability between the raters. A high degree of inter-rater reliability was shown for each dimension. The highest degrees of reliability were achieved on the dimensions of Shared Sense of Purpose (.98) and Engagement (.96). Raters achieved slightly lower scores on Safety (.92) and Emotional Connection (.91), which is consistent with past research by Friedlander et al. (2006) that suggests these affect-related dimensions pose greater problems to consensus among raters. Key Findings This section provides the key findings of the study organized according to the research questions. Table 4.1 presents an overall summary of the research questions and the corresponding results displayed according to major themes and sub-themes. The first column restates the research questions that guided the study. The second column identifies the major themes that evolved based on the research and interview questions. The third and fourth columns represent the subthemes that were identified after the initial phase of coding. These subthemes substantiate the major themes that addressed the research questions. 85 Table 4.1. Overview of Research Questions and Themes Research Question 1. What Therapeutic Components Contribute to Alliances in Couples Therapy? Major Themes Key Components of Alliance Subtheme #1 Therapeutic factors Subtheme #2 Training clinic Pace of early phase of treatment Client factors Expectations of therapy/therapist Relationship distress Interpersonal styles and past relationships Gender Characteristics Demographics Concept of alliance Therapist factors Interactive factors 2. How do Therapists Manage the Multiple Systems of Alliance in Couples Therapy? Managing Multiple Systems in Couples Therapy Alliance with Individual Partners Within-couple alliance Goodness of fit Therapist skills Engagement level Split alliance Synergy effect Strategies to promote individual alliances Feelings of safety Shared sense of purpose Strategies to promote within-couple alliance Collective couple alliance 3. How does therapeutic alliance evolve over time? Within-Case Analysis Couple scores Repairing the alliances Couple 1 Couple 2 Couple 3 Couple 4 Couple 5 Imbalanced alliances Therapeutic ruptures Key components for each alliance pattern 86 Research Question 1: What Therapeutic Components Contribute to Therapeutic Alliances in Couples Therapy? Data from the SOFTA self-report scales, observational measures, and interview transcripts were analyzed to identify important elements and characteristics of therapists that are influential to alliance formation with couples. The triangulation of the data identified themes that were organized into the broad category of Key Components of Alliance. This category has multiple subthemes that incorporate important elements, skills, and dynamics that were influential to building either a strong alliance or weak alliance. Key Components of Alliance When asked to describe their feelings about the therapeutic relationship, participants identified a wide variety of therapeutic elements that influenced their ratings of the therapeutic alliance. Alliance ratings for the sample in this study appeared to be related to factors specific to the clients and therapists, as well as general alliance building techniques and overall feelings regarding the fit between the therapist and couple. The sub-themes for this category are 1) Therapeutic Factors 2) Client factors and 3) Therapist factors. Therapeutic factors This subtheme refers to the therapeutic context in which the sessions took place and the unique processes in the initial phase of treatment. These factors impacted the initial alliance formation due to their challenges to safety, emotional connection and engagement. Training clinic. While all the clients understood the nature of the clinic which mandated the recording of all sessions and therapist participation in supervision with other graduate students and supervisors, this element did play a role in the alliance formation for some clients, particularly in terms of safety. It was apparent that some clients (n = 4; 40%) were apprehensive of the video cameras in the room as evidenced by their periodic glances toward the camera and 87 inquisitions about who would have access to the videos. This enhanced the typical discomfort about confidentiality and personal disclosure that the initial stage of therapy arouses by the exposure of these matters to other strangers: There‘s a limit to the privacy. I still think that there were a few times that maybe I would have said or done something a little less guarded had it really been more contained. It‘s like this strange feeling when people are looking into our closets and cabinets. So, I think there were a few times that my openness was affected because there is still this tape running that you know about. I think it compromised it somewhat. (Anna, C2, Interview #2) Additionally, the knowledge of the therapists role as graduate level clinicians impacted the initial impression of the therapist for these clients. The therapists had to overcome their student status in order to establish respect as a professional clinician: I don‘t know how much it has to do with him being a student. That was one of my main concerns coming here. So, yea, he is a student and it does make me question what he does in session and why he does it. (Susan, C1, Interview #1) Most of the clients (n = 6; 60%) were not skeptical of the student status because they were aware that they would be treated by student therapists if they chose to participate in treatment at the clinic. Christina and Erin found it comforting that Leah was a student because it allowed them to easily identify with her. Christina was a student herself pursuing a career in psychology, and Erin thought Leah‘s student status allowed her to have a fresh perspective. Also, the therapists in this study were doctoral level students, thus the clients knew that they completed their master level training and had a few years of experience. Moreover, they seemed to appreciate the resources the therapists had at their disposal as graduate students, such as access to classes and supervision with advanced level therapists. Finally, Bob thought that the training element of the clinic provided mutual benefits in which the clients were helping the therapist as much as the therapist was helping them: 88 He has classes, he has professors. You are all in this field for a reason, you like to help people and you are learning how to help people. This might be a challenging dynamic for him, but it helps him learn stuff, too, so in that respect it‘s also that we are helping him as much as he is helping us. So, it is a safe give and take. (Bob, C1, Interview #1) The training nature of the clinic was a therapeutic element that contributed to alliance formation for some clients. Most clients appreciated the availability of this type of clinic due to the reduced fees and ability to have therapists who are involved in continuous training. While this element was an important factor, it did not appear to be significantly detrimental to the alliance in this study. Pace of initial stage of treatment. The early phase of therapy typically involved assessment and formulation of goals. The therapists‘ main goals for the initial sessions were to learn about the couple, as well as form a connection. The need for historical background appeared to negatively affect the alliance ratings for some clients (n = 3; 30%): In the beginning, there is a lot of background information and there were maybe 1 or 2 sessions that were ok, but we didn‘t come away with a feeling gratified in that we learned something and this was really meaningful. (Brad, C2, Interview #1) The focus on historical background information could have been due to the therapists‘ theoretical orientation. They both utilized a transgenerational perspective which focuses on how the history of the client impacts their present functioning. Nonetheless, it appeared that the second and third sessions were often viewed as ―alright‖ for these clients, even if they understood this as part of the process of therapy. This suggests that they wanted more in terms of overall engagement and progress rather than working on forming a bond with their therapist. Thus, it suggests that the decline in alliance ratings during the initial sessions can be explained by feelings of disengagement. This may be more likely to occur if the clients perceived the first session to be particularly meaningful and rated the alliance high. 89 Client Factors Client factors referred to the dynamics that were specific to the individual partners or the couple as a whole. Analysis indicated that there were four general sub-themes of client factors that were influential of alliance ratings: 1) Client expectations of therapy/therapist, 2) Relationship Distress and 3) Interpersonal characteristics and past experiences and 4) Gender. Expectations of therapy/therapist. This sub-theme was identified primarily from the interviews with the clients. Clients had preconceptions or expectations of therapy that were present before they ever stepped in the door of the therapy room. These beliefs may be due to past experiences in therapy, or just general assumptions about the therapeutic process. The degree to which the therapists and therapeutic process met these expectations contributed to alliance formation. Overall, all clients discussed their expectation of confidentiality and anonymity. These elements were important for all of the client participants due to the vulnerable nature of therapy: Well, I think that goes back to my preconceptions coming in. I knew that there is patientdoctor confidentiality, and he can‘t use what I say here anywhere else against us. So there is that safety and anonymity that makes him safe. He‘s safe. So, being that, it‘s comfortable to talk to him about that kind of stuff. But, I knew that coming in. I trusted that coming in. (Bob, C1, Interview #1) Additionally, 60% of participants discussed the expectation of therapy being a safe place to discuss the serious matters that were influencing their relationship distress. For all couples, discussing these matters on their own would cause significant conflict or disconnection. Partners discussed how their struggles to resolve their relationship issues were related to their inability to view each other‘s perspectives without judgment and anger. They expected therapists to provide an unbiased and neutral perspective. They wanted an outside perspective that would provide nonjudgmental, diplomatic guidance and feedback as they worked towards their goals: 90 Well, I expected the therapist to give us feedback without making judgments about who was wrong or right, or what behavior is good or bad. And, she does that. She also pointed out things to us that maybe we didn‘t get a chance to see. She was digging in deeper to what we were saying, so she was understanding things more than a person who was just listening. (Christina, C5, Interview #1) Lastly, three participants (30%) reported that they had specific expectations for their therapist. For instance, the female partners of Couple 4 and Couple 5 indicated that they had certain preferences for the type of therapist they wanted. They both had previous therapeutic experiences and knew they were looking for someone who they could connect with on an interpersonal level, and one who did not hold a hierarchical stance with them: I wanted someone that was caring, that was nonjudgmental, and that didn‘t come across to me as being better than me. Because sometimes with the other therapists, they would try to be mothering, or they are trying to tell you how to live your life. And, I just kinda wanted someone who was going to be a friend who is on the same level as me that‘s gonna give me input because they know. (Christina, C4, Interview #1) Relationship distress. The degree of couple distress influenced the overall therapeutic process, as well as alliance ratings over time. Each partner completed the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995) at the intake session as a standard procedure of the clinic. These scores were combined with the average mean total scores from the SOFTA selfreport measures and observations in order to assess how alliance was influenced by the level of couple distress. Table 4.2 displays the RDAS and SOFTA-s scores for each partner of the five couples. 91 Table 4.2. RDAS Scores and SOFTA-s scores for Couples Clients Couple 1 Couple 2 Couple 3 Couple 4 Couple 5 1 RDAS Bob Susan Brad Anna Mark Heather Adam Erin Joe Christina SOFTA-s Total Scores 1 Mean 63.60 60.25 62.88 64.13 59.00 50.20 60.44 72.89 70.77 70.89 46 54 46 45 38 37 21 17 49 47 SD 5.17 6.49 1.81 2.83 2.30 3.08 7.00 4.73 4.55 3.29 ≥48 Non-Distressed, ≤47=Distressed Results of the RDAS for the individual partners indicated that most (n = 8; 80%) clients scored in the distressed level on the RDAS. Overall, the partners of each couple reported similar levels of distress with the exception of Couple 1. Additionally, observational coding revealed that Couples 1, 3 and 4 had the most relationship distress. This distress was due to the level of conflict between the couple, their perspectives about the problems in the relationship, and commitment to the making the relationship work. When the RDAS scores are paired with the SOFTA-s mean total scores and observational data, results indicate that level of couple distress has an influence on overall alliance ratings during treatment. For example, lower level of distress can translate into higher alliance ratings. Couple 5 were the least distressed as demonstrated by the RDAS scores (Joe-49, Christina-47) and the observational data. As a couple, they also had the highest mean alliance scores. Joe‘s mean total alliance score was 70.77 (SD = 4.55), while Christina‘s mean total score was 70.89 (SD = 3.29). Conversely, RDAS scores for Couple 3 indicated significant distress (Mark-38, Heather-37). Observational data also revealed that the partners of Couple 3 differed in their 92 views of what the goals of therapy were due to Heather‘s belief that Mark needed to change. Their overall couple dynamics were conflictual during the sessions. This couple had the lowest alliance scores in the sample,(Mark: m=59,00; SD=2.30; Heather: m=50.20; SD=3.03) and dropped out after the fifth session. The relationship gets more complex when there is a difference in the level of distress reported by the partners. RDAS scores for Couple 1 revealed that Bob (46) viewed their relationship as more distressed than Susan (54). Through observation of the sessions, it was apparent that the difference in scores was somewhat indicative of their conflicting perspectives about the problems in the relationship. Susan believed their distress was due to Bob‘s issues. She was often attacking and blaming towards him in the sessions which caused his alliance ratings to decline. These behaviors occurred during Session 1 (initial rating-no change), Session 6 (scores decreased from 63 to 57) and Session 9 (scores decreased from 70 to 63). Couple 4 scored the lowest on the RDAS (Adam-21, Erin-17). Their significant distress was due to the lack of trust in the relationship. Session content depicted Erin as angry, hurt, and attacking, and Adam as disconnected. Instead of translating into low alliance scores for both (as with Couple 3), their alliance scores differ by 12.45 points. Erin‘s mean alliance score of 72.89 (SD = 4.73) was the highest of all participants in the study, while Adam‘s score was significantly lower (m = 60.44; SD = 7.00). The difference in the alliance ratings for these partners indicated that they did not feel similarly aligned with the therapist, which is problematic in conjoint treatment. It was hypothesized that the variance in their alliance ratings may have contributed to their decision to drop-out of treatment after the tenth session. Interpersonal styles and past experiences. The ability and/or willingness to connect with others is present in varying degrees in people. This can influence the strength of alliance 93 they have with therapists as evidenced by the differences in the overt signs of engagement, such as amount of talking in sessions, as well as feelings of safety. For all couples, there was at least one partner who tended to dominate the sessions. This was the female partner for all couples except Couple 3, in which the male partner was more talkative in session. However, it is important to note that Heather, the female partner in Couple 3, had a physical disability that contributed to her interpersonal functioning. She was not as talkative in sessions and her interpersonal skills made it difficult to assess her level of connection to her partner and therapist. Conversely, Mark talked about how it is easy for him to talk with people, even strangers, in his interview: I think it is easier for me than it might be for others. I‘m an engineered socialite, I guess I would call it, so I tend to be able to engage people in conversation easier, so it is also easier for me to open up to people as well. (Mark, C3, Interview #1) Additionally, three (30%) clients discussed their ability to form a relationship based on past experiences. This could mean that the therapist reminded them of someone they know, or themselves, that made it easier for the clients to connect with the therapist. The familiarity of the therapist allowed them to feel at ease: I feel comfortable because she reminds me of someone I used to work with. Very similar in appearance. Same facial expressions, and what‘s weird is I think the other woman‘s name was also similar. So, I think that helped to make it more comfortable to me. The familiarity. (Joe, C4, Interview #1) Client gender. Data were analyzed to ascertain whether there were differences in alliance ratings according to gender. Table 4.3 provides the mean total SOFTA-s alliance scores for the entire sample according to gender. 94 Table 4.3.SOFTA-S Mean Total Scores According to Gender Gender Men Women Mean 63.33 64.63 SD 4.54 7.40 Averaging the total alliance scores (n = 42 ratings were used for each group) for the men (n = 5) and women (n = 5) in the study showed that the male partners reported slightly lower mean alliance ratings than the female partners. The mean rating for total alliance scores for all the male partners was 63.33 (SD = 4.54), while the female partners averaged 64.63 (SD = 7.40) in their total scores. In order to assess whether all male clients scored lower than their female partners, total SOFTA-s and SOFTA-o alliance ratings were also compared between the male and female partner of each couple. Table 4.4 details the results according to each couple. 95 Table 4.4.SOFTA-s Mean Total Scores for Male and Female Partners Participants 1 Couple 1 Bob (n =10) Susan (n=8) SOFTA Self-Report Mean SD 63.60 5.17 SOFTA Observational Mean SD 0.84 1.40 60.25 6.49 0.62 1.94 Couple 2 Brad (n=8) Anna (n=8) 62.88 64.13 1.81 2.83 1.50 1.50 0.53 0.53 Couple 3 Mark (n=5) Heather (n=5) 59.00 50.20 2.30 3.08 0.75 0.65 1.25 1.26 Couple 4 Adam (n=9) 60.44 2 7.00 1.97 1.54 Erin (n=9) 72.89 2 4.73 2.07 1.72 Joe (n=9) Christina (n=9) 70.77 70.89 4.55 3.29 1.77 2.33 1.92 0.86 Couple 5 1 n = number of sessions 2 Score is based on 10 sessions because clients did not complete SOFTA-s for final session, but the session was included for the observational scores In this sample, male partners in Couple 1 and 3 had higher mean self-report and observational total alliance scores than their female counterparts. For Couple 1, Bob‘s average alliance rating is 63.60 (SD = 5.17), while Susan‘s mean score is 60.25 (SD = 6.49). Additionally, Mark‘s mean alliance rating is higher (m = 59.0; SD = 2.30) than Heather‘s average score (m = 50.20, SD = 3.08). For the remaining three couples, the female partners had higher average alliance ratings than the males. For couples 2 and 5, the difference between partners was not substantial. Couple 4 had the largest difference in their alliance ratings out of all the couples. Adam‘s mean rating for total alliance was 60.44 (SD = 7.00), while Erin‘s mean rating was 72.89 (SD = 4.73). Thus, it appeared that for this sample, alliance varied by gender with no clear pattern of men or women being consistently higher. 96 Therapist Factors There were also factors specific to therapists that were influential in how the participants aligned with their therapist in the study. Therapist factors were any elements that were distinctive to the therapist. The subthemes identified under this were: 1) Characteristics and 2) Conceptualization of Therapeutic Alliance. Characteristics. This category consists of any qualities the clients indicated were influential of their alliance development that were not related to skills. These are the idiosyncratic traits or characteristics of the therapist such as personal style and demographics that represent the overall ―person‖ of the therapist. The subthemes of this theme are Intrapersonal Dimension and Demographics. Intrapersonal dimension. This subtheme pertains to the therapist‘s qualities that he or she brings into the therapy, such as personality traits, mannerisms and characteristics. When asked what qualities of the therapist helped build their relationship, clients consistently spoke about specific mannerisms and aspects of their personalities that represented the overall intrapersonal style of the therapists. Three clients (30%) referred to the facial expressions, posture, eye contact and gestures that helped make them feel comfortable in the room: Like I said, just overall, the way she presents herself. She actually looks like a normal person, she looks like she‘s not trying to be above us. So, the way she sits, she faces us, she has eye contact, it is not like she is looking off. She has a more soft spoken voice. We can tell by her posture, and just her. By just her. It‘s the way she is. (Christina, C4, Interview #1) Others (n = 4, 40%) discussed how they felt comfortable if the therapist was someone that they would typically be drawn to in other situations. This meant that there was something about the therapist they could relate to in terms of personality or appearance: 97 I guess his demeanor, or just the way he projects himself, I can relate to in a sense. And, in that sense, he is someone I would be drawn to, I suppose, in a social setting. (Bob, C1, Interview #1) Most (n = 6, 60%) clients reported that the therapist‘s willingness to show his or her personality influenced their alliance. Clients discussed how seeing their personality allowed them to relate to therapist on an interpersonal level: It‘s her expressions, her body movements. She‘s real, just like I said, I can tell. I know it‘s her job, but at the same time she has to bring something real into it of herself, and she does that. I know that she is normal, and not just being a robot doctor. (Adam, C4, Interview #1) Participants also revealed that seeing the therapist‘s ―real‖ personality offsets the traditional ―professional-patient‖ dynamic that is often intimidating to clients. This seemed to be a quality that helped them acclimate to the therapeutic process. Two clients in the study discussed past therapeutic experiences with therapists were ―too cold and analytic‖ and believed it was because they held on too tightly to their role as a professional. Generally, the clients indicated that they felt more connected when they were able to see their therapist as a person. Indeed, when aspects of the therapists‘ personalities, experiences and lives were appropriately shared, the clients felt more connected: I think it‘s nice when we step out of the room, and he does very small chit-chat talk, and he kind of softens his persona a little bit. He will kinda laugh and smile more than he does in the session, and may exchange some story about our kids who are about the same age, and then that‘s it. I think that it helps knowing he is more of a person. (Anna, C2, Interview #2) Overall, results indicated that clients want to relate to their therapists. They look for general mannerisms, aspects of personality, and expressions that help them connect on a personal level and show that the therapist is real. These characteristics are idiosyncratic to therapists as evidenced by one client‘s statement about how she was able to connect with Leah: Her hair. I think it shows creativity, and confidence once again…open to fun, energetic, I guess I get that feeling that she is not afraid to be real. (Erin, C4, Interview #1) 98 Demographics. The demographics of the therapist appeared to play a role in how the alliance was formed for the participants. The aspects of race, gender, and age were identified as subthemes through data from both client and therapist interviews. These characteristics added an additional element to the therapeutic process, and influenced the therapeutic relationship. Race/Ethnicity. Nine out of 10 clients in the study were Caucasian, and Couple 5 was a biracial couple. In terms of the race/ethnicity of the therapist, Chad was Hispanic/Caucasian and Leah was African-American. While none of Chad‘s clients discussed how his mixed race influenced their relationship, this subtheme emerged with one of Leah‘s clients. When talking about past positive therapeutic experiences, Erin (Couple 4) discussed how she believed the race/ethnicity of the therapist influences the connection she feels with them: Both times I had an African-American girl, and they are fantastic. I usually get, you know, white female or a male, and I‘m just not feeling a connection. She just seems fresh. I think African-American women are great at this! (Erin, C4, Interview #1) Erin stated that she believed Leah‘s experience as an African-American woman enabled her to have more compassion and understanding. Due to how Leah asked questions instead of making statements, she believed that Leah was able to attend to individual experiences rather than work from assumptions based on stereotypes, readings or theories. This increased her connection to Leah because she felt heard and validated: She is still pretty open to gaining a sense of real life. I can understand that things can get general, and she just seems to kinda step away from that. I think it may be due to her culture. She kinda has more compassion and I like to think that. My family does, too, so she is really comforting. (Erin, C4, Interview #1) Additionally, the researcher directly asked Leah and Couple 5 in their individual interviews about whether or not Leah‘s race had an influence on their relationship with her. This couple was a multi-racial couple in which Joe was African-American and Christina was 99 Caucasian. During sessions, Leah openly inquired about their experience as an interracial couple in several sessions. They both were comfortable talking about it, and they stated that it was not a particular struggle for them. However, in the interview, Leah discussed her curiosity about Joe‘s ability to be open in therapy due to research that shows African-Americans are not as comfortable being in therapy. I was just wondering what being in low SES and with his Black ethnicity and everything together, as far as them trusting in me. And, I was really thinking about when you think about African-Americans in therapy, which is not really common. How much are they willing to put out there to me. If you look at what research says is that there are certain questions you don‘t ask right way until you build that trust. (Leah, T2, Interview #1) Leah also felt confident that race was not a factor for Christina because it appeared that she was comfortable around people of different races. She did not directly ask the couple how this dynamic influenced how they felt with her because she did not want to put pressure on them to respond in a certain way. However, she was conscious of that being a potential barrier to their feelings of comfort and ability to completely share everything with her. In the interviews, both members of Couple 5 discussed the dynamic that race brought into their relationship with Leah. Christina disclosed that she was initially apprehensive about Leah being African-American based on how she has been treated in the past: Because a lot of Black women have not been the kindest to me because of that. That‘s ok, I have dealt with that and I try to show them differently. But, yea, when we first came, that was one of the first things I thought about. I wonder how this is going to go. I wondered deep down, maybe she does feel that way. I didn‘t know her at first. (Christina, C5, Interview #2) As therapy progressed, however, Christina felt that Leah‘s race would actually help her partner feel more comfortable in therapy: I actually think it is good because I think it helps him open up more. As stupid as that sounds, but race is a big issue nowadays. And, I really think her being a Black woman can help him open up more, and in turn, open up more to me and really help our 100 relationship because I think it is not as intimidating for him. So, I think it has helped a lot. I think it is a good thing. (Christina, C5, Interview #2) Additionally, Christina thought that Leah being the same race as her partner actually provided a balance of demographics. While she could relate to Leah as a woman, Joe could relate to her as a member of his racial group. Her partner, Joe, stated in the interviews that he did not really think that race affected how he connected with Leah. He admitted it might make it more comfortable, but that it was not a factor in how he evaluates his relationship with therapists: I would say more comfortable in a way. That is not a big factor for me, not what your race is, but how qualified you are. It is different because I had other therapists. She is my second female. But, I also had a white male and a black male. It hasn‘t played a factor. It‘s kind of good in a way to see that it‘s not a huge factor, but it‘s there (Joe, C5, Interview #2) Joe also mentioned that the issue of race was discussed during sessions in terms of how it affected their relationship or his role as an African-American man. He stated that the ability and willingness of Leah to raise those issues was comforting because those conversations ―need to happen.‖ The fact that they could discuss the issue in a positive manner increased his connection with her. Age. Age was also a demographic that clients mentioned when they spoke about the qualities that influenced their relationship with their therapist. This was a particularly important dynamic given the nature of the clinic in which the data was collected. Clients knew it was a training clinic for graduate level students which would suggest that some therapists would be younger than others. Overall, this did not seem to affect the ability for clients to connect with their therapist. Three out of the five couples were similar in age to their therapists. Partners in these couples discussed how this actually helped them build a relationship with the therapist because they were someone they could relate to: 101 I think because he seems to be in our same age range, kind of similar type of person, I don‘t know how to explain this. Similar status, there‘s not this gap you know where you are coming out and someone who is much older, with traditional methods, and who is talking down to you. (Brad, C2, Interview #1) Additionally, it seemed that they believed older therapists were more demeaning and had preferred methods of working that were based more on their overall biases and assumptions than the individual needs of the client. The participants believed that the therapists‘ involvement in classes, training, and supervision enabled them to adapt to their needs and address their issues with a fresh perspective. Gender. Finally, gender of the therapist was another element that influenced how comfortable clients feel in the room. Two themes developed from the interview data related to gender: 1) both men and women clients feel more comfortable with a female therapist and 2) male clients tend to feel intimidated when there is a gender imbalance in the room. The first theme was supported by three men and one woman in the study. Overall, these participants discussed how the gender of their therapist influenced their comfort level during sessions. They believed that they would feel more comfortable with a female therapist. This belief appeared to be embedded in traditional gender roles that influence people to feel more comfortable discussing emotional issues with women: I think the gender is important because I can relate to her more on the gender level. I think I would be more uncomfortable with a man. And, I think my partner would be more uncomfortable with a white man, and even more so with a black man because men don‘t open up like that to each other. You usually see a lot of male therapists don‘t see men, they usually see women and children. I think that is just who we are as people. (Christina, C5, Interview #2). Christina‘s quote also identified how culture can also influence preference for a certain gender of therapist based on a person‘s racial or ethnic background. She believed that her partner, who was 102 African-American, would feel uncomfortable with a White male therapist, but even more so with a Black male therapist due to the stigma about therapy in his culture. Additionally, Brad discussed how gender was an issue for him despite the fact that he was the same gender as his therapist. He discussed how he felt more reserved in the room with Chad and attributed this to his mother being his main caregiver: I don‘t know if it‘s because I was raised by mother, but even though I feel very comfortable with him and I think he‘s fine, I don‘t know if I‘d have the same connection as I would if it were a woman therapist. I think because I was raised by my mother. I‘ve always been more comfortable around women, and I think it comes out in therapy. There‘s some little resistance or something. I don‘t think I could overcome that. (Brad, C2, Interview #1) Brad‘s family-of-origin dynamics impacted his relationship with Chad. While his reservations about being open with men did not hinder his participation in therapy, it did impede his ability to connect with Chad. This suggests that a client‘s ability to form alliances can be impacted by a specific characteristic of a therapist, such as gender, when it evokes feelings related to past experiences. The second theme pertained to the gender imbalance that couples therapy creates. While this dynamic is inevitable, it can have implications for the therapeutic alliance. The individual alliance with the partner of the opposite gender can be challenged if he or she does not feel that their perspective is validated and respected. This dynamic was discussed by two male clients in the study. They both revealed that the imbalance creates intimidation and can hamper their willingness to share honest feelings. The men believed that this could be counteracted if the therapist directly addresses the dynamic and actively supports their experience. This was best exemplified by Joe‘s (Couple 5) response that discussed his experience as the only male in the room and how Leah‘s ability to directly address these dynamics increased his confidence in her: 103 I am outnumbered. And, we have discussed it. It can be a little uncomfortable, but she made me feel more at ease by addressing it, and assuring that she was aware of how it was uncomfortable for me. Which has made it easier for me to discuss things and I am feeling more comfortable now because I see her continuing to make sure I don‘t feel ganged up on or irrelevant. Her ability to talk about that stuff increased my confidence in her. She did a good job about asking how I feel and about my role just as being a man, an African-American man, so. (Joe, C5, Interview #2) Results indicated that the demographics of the therapists contributed to the overall aspects of the ―self‖ of the therapist. While they are undeniable characteristics of the therapist, they did not necessarily play a significant role in how the couples formed an alliance with their therapists. The most significant demographic was gender due to its contribution to an inevitable imbalance in therapy that could negatively impact the alliance. However, as with all evident dynamics, the clients appreciated it when the therapists openly addressed these issues. The transparency of the therapist helped to establish a dialogue about any intimidating elements. It also increased feelings of safety and confidence in the therapist. Conceptualization of therapeutic alliance with couples. This theme is important for understanding the results of this study because it influenced how the two therapists built the relationships with their couples. Both therapists had a strong understanding of therapeutic alliance and its role in all therapeutic contexts. They believed therapeutic alliance is the initial relationship that therapists build with their clients that encompasses the expectations and goals of the clients, as well the overall feelings of safety and trust in the relationship. They had a high regard for the alliance and its influence on the overall therapeutic process. [It] sets the foundation of how joined you are with your clients as far as the relationship you all have built, and if they feel safe enough to really open up and be themselves in session. I guess from my perspective, the alliance determines whether they will stay in treatment. They may stay even if it is not that strong, but I don‘t know how open they would be. I think the therapeutic alliance sets the foundation for how joined you are with your clients as far as the relationship you all have built, and if they feel safe enough to really open up and be themselves in session. (Leah, T2, Interview #1) 104 The therapists also believed therapeutic alliance is a tool for change: The alliance is the foremost tool to help change. As long as you have a strong alliance, you can help the client or family move towards change in the direction that either you feel is best, or they feel as best. (Chad, T1, Interview #1) In terms of how much they focus on the alliance with their couples, both therapists indicated that they are continuously assessing how it is developing and changing throughout treatment. They may not be actively attending to it during each session, but are aware of it at all times. While both therapists had similar understandings and views of therapeutic alliance in couples therapy, there was a defining difference in how they built relationships with their couples, which is discussed in the following section. Interactive Factors The interactive factors referred to the nature and dynamics of the therapist-couple relationship. The research identified two significant interactive factors that contributed to alliance. These factors related to 1) Goodness of Fit: Style of Therapist and 2) Therapist Skills. Goodness of fit: Style of the therapist. This theme is supported by the self-report, observational and interview data. It describes the general fit a client feels with their therapist. This seemed to be influenced by the style of the therapist, or the overall theory and interventions used to achieve the goals of the couple. Analysis of the sessions and interview data revealed that Chad and Leah formed alliances with their couples differently. Chad used theory and specific theoretical tools to align with the couples, while Leah relied more on building an emotional connection. Thus, the couples rated and discussed their alliances differently. This section provides a discussion of the differences in alliance building that were observed and discovered via the self-report, observational and interview data. 105 To assess the overall mean ratings for each therapist, the scores from each individual partner of the couples were combined to represent a total mean score. Table 4.5 describes the mean alliance scores of the couples according to therapist, as well as their overall mean alliance ratings. Table 4.5. Client Mean SOFTA-s Ratings for Therapist Total Alliance Ratings by Couple Therapist Rater Chad 1 n = 34 Leah n = 47 Couple 1 Couple 2 Total Alliance Ratings Overall M 62.11 63.50 SD 5.87 3.52 Couple 3 54.60 6.36 Couple 4 66.10 8.74 Couple 5 70.83 3.85 1 n= total number of alliance ratings over time for both couples M 62.76 SD 4.88 65.46 8.96 When all scores from the couples of each therapist are combined, Chad‘s mean alliance rating is 62.76 (SD = 4.88) while Leah‘s is 65.46 (SD = 8.96). The mean alliance ratings by couple for each therapist indicates that Couple 1 and Couple 2 have similar mean alliance ratings, with Couple 2 (m = 63.50; SD = 3.52) having slightly higher mean scores. Conversely, Leah received a greater range of scores from each couple, with Couple 5 having the highest mean rating (m = 70.83; SD = 3.85) and Couple 3 having the lowest mean score (m = 54.60; SD = 6.36). Next, the therapist ratings for each couple were analyzed to compare their scores to those of the clients. Table 4.6 illustrates the mean alliance ratings reported by the therapists for each couple. 106 Table 4.6. Therapist Mean SOFTA-s Ratings for Couples Total Alliance Ratings by Couple Therapist Chad 1 n = 18 Leah n = 24 Couple M 54.60 62.87 55.00 59.30 65.89 Couple 1 Couple 2 Couple 3 Couple 4 Couple 5 1 n=Total SOFTA-s Self-Report ratings SD 3.80 1.80 3.08 5.39 7.34 The results from the self-report data demonstrate that for five out of 6 couples, the therapists rated the alliance lower than the clients in terms of mean scores. The largest discrepancies between the couple and therapist ratings occurred for Couple 1, Couple 4 and Couple 5. Additionally, the therapist ratings showed a different perspective of the alliance from that of the clients. For instance, Chad‘s mean rating for Couple 1 implies that they are significantly less aligned than Couple 2 regardless of the reports from the couples that show relatively similar alliance ratings. This more offers more of an explanation for Couple 1‘s treatment drop-out, which is not visibly discernable when only assessing the ratings from the couples. Chad‘s alliance ratings were much lower than the couples, and indicated that he detected lower levels of alliance from the couple across treatment. Moreover, the mean alliance ratings that Leah reported for Couple 4 and Couple 5 also exemplify the difference in alliance between these couples that offers a better interpretation for Couple 4‘s abrupt drop-out in treatment. These results emphasize the importance of obtaining ratings from both clients and therapist perspectives to gain a clearer understanding of their alliance. 107 Therapist 1: Chad. Chad identifies his primary theory as Bowenian, and discussed how he uses concepts from this theory as well as Emotion Focused Therapy (EFT) to build the alliance with his couples: Well, a lot of, to use the word of emotional focused therapy, a lot of deepening of the emotion. Also, a lot of psychoeducation, I guess. I went into detail with each of them individually in the details of their genograms so that I got to know what their background was like. (Chad, T1, Interview #1) Chad believed the use of EFT to help the couples achieve their goals was complementary to building an alliance with them. His therapeutic relationship with them served as a model for their relationship with their partners. One aspect of that relationship that he felt was important was validation: The ability to take a stance of respecting other people‘s experiences without having to agree with them helps build the alliance with them, but also helps them build their relationship between each other. (Chad, T1, Interview #1) Thus, Chad‘s way of building alliance was informed through theory, and his alliance promoting behaviors were influenced by theoretical interventions. As such, he believed that alliance can be built concurrently with the interventions he uses with couples: Because whether we are in individual or conjoint, we are talking about the person, and that gives me more information about the person. It also allows me to open up a little bit more so that we are able to build that alliance. So, then when we get to the point of actually doing interventions, it is a lot easier. (Chad, T1, Interview #2) As he challenges, intervenes, or focuses on the dynamics of the couple, he also works to manage the alliance. He believed that in order to increase the alliance and effectively change the dynamics of both couples, it was necessary to elicit more emotional responses during sessions: I find it very similar to doing parenting with children where you want to make sure you have a lot of emotional weight and then do the punishing, but make sure you come back with the emotional weight. So, with them, I make sure to rebuild the alliance as I cut them down to stop. (Chad, T2, Interview #1) 108 Since his alliance promoting behaviors were interlaced with his interventions, it appeared that he established a relationship with the couples by demonstrating his competency. Due to this, his clients discussed their relationship with him differently from how Leah‘s clients expressed their feelings of the bond they had with her. When asked to describe how Chad formed a relationship with them, they would discuss how he asks questions with a purpose, gets to the root of things, is focused, stays on task, comes prepared to session, and builds upon each session. He‘s easy going, but he‘s also very focused. He knows what he is looking for and asks questions with a purpose. And, uh, he knows how to get to the root of the things hiding under the surface. And, that‘s what we are here for to get to the things behind the surface, how to get through them, and how to connect together. (Bob, T1, Interview #1) His clients indicated that his focus on theory and remaining on task showed them that he was committed to helping them achieve their goals: It means that he‘s paying attention, he‘s serious about getting at the problem and not just letting us talk. That he‘s the guide to the therapy. The therapist who understands these things in more detail doesn‘t let you go off on tangents or talk about how you feel in a broad, more loosey sense. I think he strikes a decent balance in letting us run on a little bit, but then also focusing it on the issues. (Brad, C2, Interview #1) This seemed to fit for most of the spouses of his couples. Bob, the male partner of C1, stated that he really appreciated how prepared Chad was for each session. His preparedness helped to focus therapy, guide them where they needed to go, and keep them on track. It also reduced the stress for the couple. It‘s very stress reducing because it takes pressure off of me, and I hope it takes pressure off my wife trying to figure out where do we from here. He is the conductor of the train. We are important passengers, and he wants to treat us equally and fairly, but the train is going where he wants it to go. And, he makes sure it‘s on track. That‘s important. (Bob, C1, Interview #1) From his understanding, it appears that Chad‘s commitment to staying on task and being prepared established a hierarchical position in the therapeutic relationship. Bob valued Chad‘s role as the conductor of their therapeutic process because he believed it helped keep them on 109 track, even when there was a seven week lapse in treatment. Bob‘s wife, Susan, however did not necessarily like Chad‘s role as the conductor or his determination to stay on task. She would often get frustrated in sessions when she was not allowed to discuss historical information related to the issues the couple was experiencing. Chad and Susan would often get into negative interaction cycles when attempting to ascertain what information was pertinent to the issue at hand: I know where he‘s going, because I have a medical background. But sometimes I don‘t appreciate his in your face ―no, don‘t say that, focus on this.‖ Being a female and also [having] medical experience, and also one of my jobs years ago was that I worked in a clinic for battered girls. And, so, history matters to me, and I don‘t think that‘s his style. Everybody has their own style, and to me, history matters. Like, today, I was saying ―but, you need to know the history to understand now.‖ And, he said, no you don‘t. (Susan, C1, Interview #1) Susan‘s need to provide a historical perspective often caused her to lose focus of the session by discussing these details. This influenced Chad to push her more to stay on task. In the interview, she discussed how his style makes her feel: His in your face style, I don‘t appreciate. But, it‘s not that I don‘t understand where he‘s going or why he‘s doing it, I get it. It frustrates me a lot, and I am one to say, no, I am going to say what I want, damnit. Listen to me! So, I think that frustrates him as well. (Susan, C1, Interview #1) Despite her frustration with Chad, she seemed to understand that it was not a personal attack against her, but rather his way of working with them as a couple. While her alliance scores may have been lower for the sessions in which this struggle was most obvious, she believed her overall alliance with Chad was positive. She discussed why she keeps returning to therapy even when she does not feel validated: I also think that things improve over time because he‘s a stranger to me right now. And, that we have issues that we need to learn to get past and deal with. (Susan, C1, Interview #1) 110 Susan‘s belief that relationships take time to develop allowed her to be patient with the relationship she was building with Chad. Additionally, her determination to work on the couple issues induced her to remain in treatment, even when she was frustrated. This demonstrates that Chad‘s ability to provide effective treatment was an important element in the relationship he had with this couple. Couple 2 also respected Chad‘s focus on theory and goals because it increased their confidence in his ability to help them. They were aware of his direction, which appeared to alleviate the stress on the couple by guiding them when they were unsure of what to do. While they observed his attention to theory, they also felt that he attended to the couples agenda. Couple 2 appreciated that he would begin sessions by asking them if they had any pressing or preferred issues to discuss before continuing their focus on the goals. In the first interview, Anna discussed how she liked that Chad was tailoring the therapy to them: Just that it‘s not his agenda. It‘s not like a class. You know, it‘s not like this is the teacher‘s lesson for the day. It‘s what do we need. (Anna, C2, Interview #1) Chad‘s competency was the defining ―technique‖ that Couple 2 found influential of their relationship with him. Their alliance ratings were reflective of how effective he was in helping them achieve their goals: I think sometimes the ratings on the scores reflect how I feel more of a particular session with, and not necessarily my relationship. But, sometimes it‘s both, like sometimes you kinda feel more energized if you had a good session or conversation than if it‘s not as good, or something. (Anna, C2, Interview #1) In the interviews, Brad and Anna struggled to articulate the emotional connection they had with Chad. Brad seemed to acknowledge there was some level of bond, but Anna struggled to identify how she was emotionally connected to him. I don‘t feel a strong relationship with him. I mean, I am comfortable talking, it doesn‘t affect what I say in sessions. I don‘t necessarily feel a bond. I feel that it is very much 111 like a quote unquote doctor-patient relationship. Maybe that‘s because to me the kind of clues, the kind of things I would say characterize a relationship is more like things like smiling or maybe making jokes, or being friendly. I guess I think of it as the therapist has a job, and I‘m the client, and if the they‘re doing their job but I don‘t really feel like a strong bond, and I think that has been reflected in my scores. (Anna, C2, Interview #1) She also did not seem to think that the emotional connection was necessary to the therapeutic relationship: I don‘t think I need to feel bonded to the person. I think I need to feel respected. I think he does a good job balancing his turn to talk and my turn to talk, and managing that type of thing. And, as long as I feel respected, helpful, and listened to, then it‘s fine. I feel like it‘s more like a professional patient-provider type of thing, but not very personal. (Anna, C2, Interview #1) The way the clients described their alliance with him resembled the task and goals dimension of Bordin‘s definition of alliance. They appreciated his competency, focus on goals, and commitment to staying on task in sessions. While they felt comfortable with him and liked him, they did not necessarily need to feel a bond with him. Chad discussed in an interview how he was adapting what he typically did to build the alliance with this couple. He stated that building the alliance with this couple was different from Couple 1 because he felt that sharing pieces of himself would help them to be more open. Additionally, he felt like he had to prove himself with them. This seemed to be mostly due to the interpersonal nature of the clients and expectations of the couple that the therapist be professional and knowledgeable. Thus, Chad was using more of his personal self to increase his connection with them: I‘m using more self-of-therapist stuff, or giving them more of a window into who I am, so that way it would be easier to build the relationship by giving them little [pieces]. It is more of a necessity in order to get them to be more open. So, I guess it feels like they are waiting for me to open up more so they can open up. This couple feels more like they want me to prove myself, I guess. Not necessarily as a therapist, but that I am going to hold their information in an appropriate manner, I guess. So, opening up and telling them about my children makes them feel more comfortable about telling me about their children. (Chad, T2, Interview #2) 112 In terms of the goodness of fit between Chad and his couples, the mean ratings of the couples fall into the moderate range of alliance. This indicates that overall, his style fit with the clients. While the two female partners mentioned ways in which they may not agree with his style or feel necessarily bonded with him, their focus was more on the overall progress of the therapy. Chad‘s ability to demonstrate competency was the defining way he established alliance with the couples even when the interpersonal connection was not as evident. Therapist 2: Leah. Leah identified her primary theory as Bowenian/Transgenerational in which she viewed the dynamics of the couple relationships through a lens that focused on generational influences. While Leah believed the overall direction of her work with clients was guided by theory, she strongly viewed the alliance she has with her clients as the main influence of positive outcomes: Having a strong relationship with them, knowing them, helps me feel more competent with clients. You can go through the motions of theory, but if we didn‘t have a solid relationship, what happens if I offend them? Or what if he never felt validated? Would I ever get them to open up? You know, with some clients, a strong alliance is important. They have to feel safe in order to go where they need to go in therapy. So, you will go through the motions, and they will end and everything, but I‘m not sure if you will get positive outcomes. You have to know them in order to have a good alliance and use your theory. (Leah, T2, Interview #2) Thus, her alliance building strategy was to focus on getting to know each partner individually and their relationship together as a couple during the initial stages of therapy and alliance formation. While her thoughts and techniques follow theory, she believed it is important to establish a feeling of safety and trust with couples. In order to do this, she needed to learn about them as people rather than clients: ―Because I really try to, at least in the first couple sessions, really get to know the client, or couple‖ (Leah, T2, Interview #1). She believed that due to the nature of couples therapy, it is essential to show empathy and validation in the beginning stages of therapy so that both partners feel accepted: 113 Early on, I think this is more important when you‘re building that alliance is validating both of the clients and the relationship. They will know if I am validating one person more than the other. So, making sure I‘m giving them a lot of validation and empathy, and if that‘s missing, I think that can develop a poor alliance. I think what happens is that one person will kind of feel like they are not getting favored or that one person is getting favoritism over another. (Leah, T2, Interview #1) Her clients noticed her efforts to connect with both partners through validation and empathy. Five out of the six clients stated that they felt understood, validated, and supported by her in sessions even if she was not necessarily siding with them. She showed her empathy and support by attending to each partner in the moment, as well as through facial expressions and mannerisms that indicated she allowed herself to feel their emotions: And she has eyebrows, and she uses them. You know, she talks with her body language which is really comforting because you‘re not left with monotone, you know, like from someone who is just blank. You can tell she feels the emotion. (Erin, Couple 4, Interview #1) Leah‘s clients also discussed her competency and intuitiveness that enabled her to know how to help them. Four of her clients discussed how she made sure to attend to important details in the midst of heated discussions. This was evidence of her listening skills, and focus on validating each partner. When she did address the important details, she not only acknowledged them, she also probed to uncover the meaning behind them: I guess the other thing is that when she hears something about the other person that feels or thinks something, she always goes back…she doesn‘t miss it. You know, she‘ll say, how does that make your feel or whatever. (Christina, Couple 5, Interview #1) Four of her clients also discussed her ability to move underneath the surface to identify the underlying emotions and meanings of their behaviors and interactions. This increased the couples‘ awareness about certain issues, but also exemplified her capability to help them: You know she is very good about asking us questions and asking us how we are both feeling, and with a lot of the stuff we have been going through, I think we are getting more in depth. We touch on so many different things. She‘s opened our eyes up on quite a few things. That makes me feel that she is capable. That she is qualified to give help 114 answering questions. It makes me feel comfortable to talk longer with her. (Joe, Couple 5, Interview #1) When asked how Leah built the relationship with them, her clients would repeatedly say it is ―just her,‖ which suggests that her genuine use of self influenced their feelings of connection. They struggled to separate specific skills or behaviors from her personality and behavior: Just her, just her. She just is very easy going. We can laugh, it‘s not all business. We can talk about the sun if we want to, and she will talk about it with us. So, it‘s been nice and she has definitely become important because it‘s nice to have someone else to give a different point of view on things. (Christina, Couple 5, Interview #2) She is not monotone, but she is very calm, and when it is her turn to talk and allow us to listen to what her thoughts are, she is slow, but not that we are incompetent. She is pausing when necessary and I think that is a really big skill. She allows us to understand, and if we don‘t, she knows it, and it makes us feel like she understands us and really gets us. (Erin, Couple 4, Interview #1) Additionally, observational and interviews revealed a specific alliance promoting behavior Leah used with all of the couples that increased their connection and feelings of safety with her. She would always thank the couples for participating and commend them for their ability to be vulnerable and honest during difficult sessions with her: Today, she commended us on our honesty. She‘s commending us in a way, encouraging us. And she seems genuinely appreciative of us participating. Those things stick out, and makes it easier. It makes me feel comfortable and confident in what she‘s doing. She listens to me and I feel heard. I think she has a genuine interest in wanting to help us and our relationship and that makes me feel heard and makes me open up more. I ask how long we can keep coming because I am sure we will need counseling later, but it helps. (Joe, Couple 5, Interview #1) The clients felt encouraged and appreciated by Leah due to this. Leah believed it was part of her role as a therapist to portray her understanding that therapy is difficult and that it is not comfortable to discuss couple matters with a stranger. Her purpose was to acknowledge their discomfort, and be transparent in order to prove to them that they can trust her to handle their 115 information and their emotions. The clients, however, felt that they should be thanking her for the help she is giving them. This behavior helped the clients to feel comfortable with her, and gave them hope to continue in the therapeutic process. Overall, Leah‘s clients believed that her competency and ability to increase awareness of their issues helped build their alliance with her. The greatest influence of their alliance, however, was her ―self.‖ Therefore, it appeared that the primary tools she utilized to build alliance with her couples were her personality and ability to interpersonally connect with her clients. This finding was exemplified by this quote from Erin: And from this side of the table, I can see that you just want to nail how the relationship is built. But, from this side, asking me that, it feels like I don‘t think you can teach it so much as it just has to click. So, I feel like, I don‘t know if I can pinpoint it exactly because if it is the wrong therapist, it is the wrong therapist and they won‘t say the words that make me feel good, and it just wouldn‘t work. She fits. It‘s just her. (Erin, Couple 4, Interview #2) Therapist skills. There were a variety of general skills that clients in the study indicated as important to the development of the therapeutic relationship. These skills have been identified and confirmed in previous research about individual alliance building, so they will not be discussed at length here. In summary, clients of both therapists mentioned neutrality, diplomacy, active listening, remembering details, and providing a non-judgmental stance as skills they believed were influential to the therapeutic relationship they were forming. Additionally, they discussed how the ability for the therapist to display empathy and patience played an important role in how connected they feel on a personal level with the therapist. These skills helped to dismantle the robotic image of a therapist and allowed the clients to feel comfortable when discussing difficult matters. Therapist competency. The most important skill that the couples identified was the competence of the therapist. The clients reported that the therapists‘ ability to demonstrate their 116 knowledge and utilize theory to facilitate change positively influenced the therapeutic relationship. Three (37.5% of clients in sample) clients discussed that the competency of the therapist increased their connection. When the clients were sure that the therapists were competent and knew how to help them, they felt more connected: The increase in confidence helps build the connection. You have to have confidence in someone that they can handle it and will manage it appropriately in order to express those kinds of emotions. I would just say increased confidence allows you to do that as well. (Anna, C2, Interview #2) These clients believed that therapist competency was most influential of the connection they felt with their therapist. They believed that the therapist‘s ability to help them achieve the couple goals was more important than having an individual emotional connection: I would probably want more of an emotional connection with the therapist in individual therapy. But, I guess in couples therapy, I am more looking for them to help us rather than to feel connected to the therapist. (Bob, C1, Interview #1) The personality of a therapist was also influential of the emotional connection. Adam, the male partner of Couple 3, and Brad, the male partner of Couple 2, reported that the mere fact that the therapist was useful and knew how to work with couples strengthened their relationship with them. While feeling connected to their therapist was important for most of the clients in the study, the ability of the therapist to effectively work with the couple to achieve their goals was an important element of alliance for all the clients. There appeared to be two main indicators of therapist competency based on the reports of the clients: Professionalism and Meaningful Feedback. Professionalism of the therapist. Four out of the 10 participants (40%) discussed an expectation of professionalism in which the therapist, therapeutic context, and therapeutic relationship contained elements of a traditional ―doctor-patient‖ relationship: 117 She came across very professional. I try to be somewhat stately myself, and I appreciate that from people. I expect that they are professional and have good stature and work with me as a professional by maintaining appropriate boundaries and standards of confidentiality. I guess I want them to treat this as a professional job that they are providing for me. (Mark, C2, Interview #1) Professionalism was not always easily established by the therapists in the study due to the research site being a training clinic for master and doctoral level students. While most of the clients in the study were comfortable working with relatively novel therapists, it seemed that the therapists had to overcome the ―student‖ identity that often underestimated their abilities. Anna discussed the challenge for the therapists in this type of setting: I think that one of the challenges here is that it‘s a campus and a student training center. So, establishing yourself as a professional therapist rather than as a student is important. (Anna, C2, Interview #1) Thus, establishing professionalism appeared to be a necessary task for the therapists in the current study that may not be for other therapists working in a professional practice or agency. The participants appreciated it when they sensed an air of professionalism from their therapist in terms of adherence to the expected standards of therapeutic practice. The nature of the clinic diminished the ability of the therapists to maintain typical practices of confidentiality due to the mandatory video recording of sessions and group supervision with other student therapists. This was not a strong concern for the clients, but it did pose a challenge to establishment of professionalism and competency. Other elements that influenced how the clients perceived the professionalism of their therapists was their appearance and mannerisms. If the clients thought the therapist portrayed professionalism through their appearance and mannerisms, clients assumed they were competent: Her style is very professional, which I can appreciate. She doesn‘t use slang words, which I have encountered before. Not so much that I am paying for therapy, but that they try to conform to me or come down to my level. I like someone who can maintain their level as much as they can in the context of therapy because it demonstrates 118 professionalism. If they look and act professional, I assume they are competent. (Erin, C4, Interview #1) Conversely, if they appeared more as a student than a professional, they had to prove their competency. Anna discussed how her initial impression of Chad decreased her confidence in his competency due to the setting and his overall appearance. This is a student setting, and he‘s pursuing his degree, so he‘s still a student, and his clothes and his body language made me feel, I don‘t wanna say a little more reserved, but less confident. I didn‘t see him as much as a professional, but more as a student. (Anna, C2, Interview #1) Anna discussed her assumptions and expectations of therapists that were based on past experience in therapy with older therapists who had decades of experience. Since Chad was closer in age to her and did not match her past experiences with therapists, she doubted his competence. Chad was able to overcome the negative assumptions when he began demonstrating his competency through his knowledge: Once he would talk it was clear he was familiar with the subject matter, and was educated in his field, which gave me more confidence. But, there were some initial, subconscious things when you first meet someone that affected my perception. (Anna, T1C2, Interview #1) Meaningful feedback. Clients reported that therapist feedback and interventions were also indicators of competency. Both partners of Couple 3 and Brad (Couple 2) stated that their therapist offered advice and ideas that were practical for the lives of the couple. They thought it was useful that the therapists were able to address specific issues by offering homework and other ways to connect the therapy sessions with their daily life. He‘s got the homework. The date nights, the readings, the emails, and the survey we just did. Things like that that make it more interactive, I suppose. And, you can talk and talk and talk, but how do you implement it or what does it mean? So, it‘s useful to have that connection as far as a practical level goes. (Brad, C2, Interview #2) 119 Also, the clients felt that when feedback was meaningful, it meant that the therapists were actively attending to the couples in sessions and able to evaluate and produce appropriate observations in the moment. This ability denoted their level of competency to discern significant feelings, tones, and expressions of the partners. Anna appreciated Chad‘s ability make observations that reflect his ability to synthesize and analyze the information he learned from them: I think even in the very beginning when he made some good observations. If you make an observation and make an astute analysis and synthesis of those observations, then you are listening, and you are paying attention, having to be engaged so I thought that was good. I remember moments like that where I thought, huh, that‘s an interesting point I haven‘t thought of. So, I appreciated the insight of that on a professional level. (Anna, C2, Interview #2) Anna stated that her alliance improved when Chad revealed insights that were enlightening and offered a new perspective on their relationship. This was a main component of the alliance for her since she did not have a view of the therapeutic relationship as an emotional bond or connection, but rather a professional relationship that is needed to achieve their goals. Erin also valued Leah‘s ability to give helpful feedback based on her interpretations of their relationship dynamics. She mentioned one particular session in which Leah had the couple do a sculpture activity to illustrate their relationship dynamics. She did sculptures last week. Fantastic. That was really impressive. That was something I have never seen before and I think that played a very big role in how we viewed each other, and it showed that she had a lot of knowledge. Like, that was a whole other sensory that she touched on that I didn‘t even know existed. (Erin, C4, Interview #1) Erin explained that this activity was a unique way for Leah to demonstrate the dynamics in their relationship that was safe and constructive. Thus, both Erin and Adam were able to listen to the feedback without becoming defensive or threatened. Both partners discussed how this 120 intervention opened their eyes to the underlying issues in their relationship that control their interactions with each other. Overall, the clients indicated that general therapeutic skills contributed to the development of their alliance. Clients identified the therapist competency as the most significant skill they appreciated about their therapist. When therapists demonstrated their competency by displaying professionalism and providing meaningful feedback, their alliance ratings increased. Summary of Key Components of Alliance with Couples The key components that were found to be influential to the therapeutic alliance with the couples in this study were related to a variety of client and therapist factors. The most significant client factors were the level of relationship distress and interpersonal style. These factors were generally stable characteristics and aspects of the individuals that impacted the way they interacted in relationships. As such, these dynamics influenced how the clients were able to form the alliance on the individual and couple level. The therapist factors that were most influential to alliance ratings were their personality and therapeutic skills. These factors created the important feelings of comfort and trust in the therapist‘s abilities to help the clients resolve their relationship distress, which influenced how the alliance formed and progressed throughout therapy. Research Question 2: How do Therapists Manage the Multiple Systems of Alliance in Couples Therapy? Data from the SOFTA self-report scales, observational measures, and interview transcripts were analyzed to identify how the therapists managed the multiple systems of alliance with their couples. The triangulation of the data revealed key components related to each dimension of alliance that provide a clearer picture of the systemic nature of therapeutic alliance 121 in couples therapy. A key component is identified as any client, therapist, therapeutic, or extratherapeutic factor that influences alliance. Results also identified specific alliance promoting behaviors that the therapists used to establish alliance within each system. This category has three subthemes that encompass the multiple systems of alliance in couples therapy: 1) Alliance with the Individual Partners, 2) Within-Couple alliance and 3) Collective Couple alliance. Within these themes, subthemes will were identified that further explain the components for each system of alliance, and the explicit therapist behaviors that helped manage the multiple systems. Alliance with Individual Partners This category contains the key components and behaviors that contribute to the therapeutic alliance therapists develop with the individual partners of the couple. The results are organized into the following two themes: 1) Key Components of Alliance with Individual Partners and 2) Behaviors that Promote Alliance with Individual Partners. The first theme refers to the dynamics that contributed to the relationship that each individual had with their therapist. The second theme describes the alliance promoting behaviors that were influential of building and managing those alliances throughout the therapeutic process. Key Components of Alliance with the Individual Partners The key components for this dimension of alliance were related to the following themes: 1) Engagement Level of Partners, 2) Split Alliances and 3) Synergy Effect. Engagement level of partners. In couples therapy, it is common to find that one partner was encouraged, mandated, or pushed to participate in treatment. Thus, it was postulated that the engagement level of the individual partners played a pivotal role in the alliances the individual partners had with the therapist. Analysis of the self-report and observational measures provided 122 information about the engagement level of the couples and individual partners. Table 4.7 depicts the mean scores and standard deviations for all participants on the Engagement dimension of the SOFTA self-report and observational measures. Table 4.7. SOFTA Self-Report and Observational Mean Scores for Engagement Participants Bob Couple 1 Susan 1 n = 10 Couple 2 n=8 Couple 3 n=5 Couple 4 n = 10 Chad Brad Anna Chad Mark Heather Leah Adam Erin Leah Joe Christina n=9 Leah 1 n = Total number of sessions Couple 5 2 3 ENGAGEMENT SOFTA Self-Report SOFTA Observational Mean SD Mean SD 16.50 2.06 1.50 0.52 2 2 1.69 0.83 15.50 1.80 15.50 0.84 2.90 0.78 15.12 16.00 17.62 13.40 13.00 14.60 3 0.64 2.32 1.06 0.54 1.87 2.30 1.23 3.00 3.00 2.87 1.00 1.00 2.00 2.20 0.00 0.00 0.35 1.22 1.22 0.70 0.91 3 1.11 3.00 0.00 1.25 1.41 1.50 0.97 3.00 2.75 3.00 3.00 0.00 0.70 0.00 0.00 15.44 18.66 15.70 17.33 18.33 17.22 Scores based on 8 sessions because she did not attend the 2 nd rd and 3 sessions th Scores based on 9 sessions because they did not fill out self-report for 10 session To interpret the significance of the mean self-report and observational scores, the combination of all four perspectives were considered to infer the engagement level of the clients. Results demonstrated that Couple 5 had the highest level of engagement when taking all scores into account. Joe‘s mean score was 17.33 (SD = 1.41) and Christina‘s mean score was 18.33 (SD = 1.50). Leah perceived the engagement level with this couple similarly to Joe as evidenced by her mean score of 17.22 (SD = 0.97). Finally, their mean observational ratings were also high 123 with Joe receiving a mean score of 2.75 (SD = 0.70) and Christina attaining the highest mean score possible (m = 3.00, SD = 0.00). Overall, results suggested that Couple 5 were equally and highly engaged in the therapeutic process. Likewise, while the mean engagement scores for Couple 2 were among the lowest of the sample with Brad scoring a mean of 15.12 (SD = 0.64) and Anna achieving a mean rating of 16.00 (SD = 2.32), they had the second lowest difference between their scores (0.88 of a point). Chad‘s mean score of 17.62 (SD = 1.06) was noticeably higher than either partner‘s score, which suggested that he perceived the couple to have high levels of engagement. Couple 2 also received the highest mean observational scores of all the couples with both partners receiving a mean of 3.00 (SD = 0.00). Despite their scores being in the moderate range of engagement, their similarity in scores and high observational scores indicate that both partners were equally engaged in therapy. Self-report scores for Couple 1 were slightly higher than Couple 2, however they had lower therapist and observational ratings, which suggested that they were not more engaged than Couple 2. Bob had a mean score of 16.50 (SD = 2.06) on the self-report score, which was 1 point higher than Susan‘s mean score of 15.50 (SD = 1.69). Chad‘s score (m = 15.50; SD = 0.84) suggested that he perceived the engagement level of this couple similarly to Susan. Couple 3 had the lowest engagement scores in the study. Mark‘s mean self-report score (m = 13.40, SD = 0.54) is slightly greater than his partner Heather‘s score (m = 13.0, SD = 1.87). This is only a difference of 0.40 which suggested that they had the lowest variance in their selfreport scores of all couples in the study. Leah rated the engagement level of this couple higher than either partner with a mean score of 14.60 (SD = 2.30). Mean observational scores for Mark 124 and Heather are identical at 1.0 (SD = 1.22). Their scores suggested that while they were not strongly engaged, they had similar levels of engagement. The engagement scores for Couple 4 showed conflicting dynamics. While they scored the second highest mean engagement scores with Adam‘s mean rating of 15.44 (SD = 1.23) and Erin‘s mean score of 18.66 (SD = 1.11), they had considerable variance in their mean self-report engagement scores with a difference of 3.22 points. Leah‘s mean score of 15.70 (SD = 1.25) was closer to Adam‘s score. Their observational scores were also high with a mean of 2.20 (SD = .91) for Adam and 3.00 (SD = 0.00) for Erin. Their scores suggest that while they appeared to be similarly engaged, Adam was significantly less engaged than Erin in the overall therapeutic process. Therapist self-report and observational scores. Therapist self-report and observational ratings were analyzed to determine whether the therapists‘ perception of the engagement levels of the couples influenced the amount of engagement promoting behaviors in sessions. There were two notable findings related to how the therapists rated the engagement level of the clients. For 3 out of the 5 couples, the therapists rated the engagement level similarly to the partner who reported the lowest alliance. For those cases, the therapist received their highest observational ratings, as well. This occurred in Couple 1, Couple 4 and Couple 5. Chad‘s mean score for the engagement level of Couple 1 was the same score as Susan (m = 15.50; SD = .84), who scored 1 point lower than her partner. His mean observational rating was 2.90 (SD = 0.78), which was slightly higher than his score for Couple 2. Leah also received the highest mean observational score of 3.00 (SD = 0.00) for Couple 4 and rated the engagement level (m = 15.70; SD = 1.25) similarly to Adam whose score was 3.22 points lower than Erin‘s score. Finally, her mean score of 17.22 (SD = .97) was similar to Joe‘s lower mean score of engagement, and she received the 125 mean score 3.00 (SD = 0.00) on the observational measure. These results suggest that when the therapist perceives that a member of the couple has lower engagement levels, they initiate more engagement promoting behaviors. For Couples 2 and 3, the therapists rated the alliance higher than either member of the couple, and received their lowest observational ratings. Chad‘s mean score for Couple 2 was 17.62 (SD = 1.06), which was 1.62 points higher than Anna‘s rating. He also received a mean rating of 2.87 (SD = .35) on the observational measure. For Couple 3, Leah‘s mean score of 14.60 (SD = 2.30) was 1.20 points higher than the Mark‘s rating, and she received a mean rating of 2.00 (SD = .70) on the observational measure. Thus, when the therapists perceived the engagement level as higher than it was for the couples, they utilized less engagement promoting behaviors. Gender and engagement level. There were two notable findings of the engagement level of the participants related to the gender of the participants. For three out of the five couples the female partners had higher scores than their partners on this dimension. Only the male partners of Couple 1 and Couple 3 had higher mean ratings of engagement. Likewise, for four out of the five couples, the partner who initiated treatment had higher engagement scores than their partner. The female partner was the initiator of therapy for all couples except Couple 3. Split alliances. Variance between the individual partners‘ alliances with the therapist indicated that there was a presence of a split alliance. While split alliances can occur when there are differences in total alliance scores or on any individual scales of alliance measures, Friedlander et al. (2006a) posit that split alliances are most evident in the individual scores on the client-therapist relational dimension of Emotional Connection. Therefore, the interview data, couple dynamics, and client scores on the Emotional Connection dimension of the SOFTA were 126 evaluated to investigate what dynamics were responsible for creating a split alliance. After triangulating these results with the interview and observational data, analysis revealed that split alliances were the result of both client-therapist and within-couple relational aspects that are present in conjoint therapeutic systems. The key components that contributed to a split alliance were: 1) Emotional Connection to the Therapist and 2) Couple Dynamics. Emotional connection to the therapist. Individual partner scores on the Emotional Connection (EC) dimension of the SOFTA self-report and observational measures were analyzed to ascertain whether or not a split alliance was present for any of the couples in the study. Table 4.8 illustrates the scores for each partner and therapist on the EC dimension. Scores are presented for the SOFTA self-report and observational measures for each session, as well as the mean total score. It was important to detail the scores for each session because the mean overall score may not accurately show that a split alliance had occurred during treatment. 127 Table 4.8. SOFTA Self-Report and Observational Scores for Emotional Connection Sessions Couples Measure C1 Bob Susan Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Softa-S Softa-O Chad C2 Brad Anna Chad C3 Mark Heather Leah C4 Adam Erin Leah C5 Joe Christina Leah 1 11 +1 14 +1 12 +2 16 +2 14 +2 17 +3 16 +1 13 +1 9 +2 6 +1 13 +1 13 +3 16 +3 15 +3 10 +3 2 15 +1 17 0 15 +3 14 +3 13 +3 16 +3 15 +1 10 +1 12 +1 13 +3 13 +3 13 +2 17 +2 16 +2 12 +3 3 14 +1 ----14 +1 16 +1 14 +3 16 +3 13 +2 11 +1 13 +2 14 +3 18 +3 13 +3 18 +2 17 +3 14 +3 4 5 16 15 +3 +1 --- 17 --- +1 15 15 +3 +2 15 15 +1 +1 13 13 +1 +1 16 16 +1 +2 13 16 +2 +2 11 8 +1 +1 15 14 +3 +2 15 15 +3 +3 16 16 +3 +3 15 15 +3 +1 18 19 +2 +3 19 20 +3 +3 13 16 +3 +3 6 7 15 17 +1 0 13 15 -3 -3 16 18 +3 0 16 18 +1 +1 14 17 +1 +1 16 16 +1 +1 --- ----- ----- ----- ----- ----- --12 15 +3 +2 16 17 +3 +2 16 17 +3 +3 18 19 +3 +3 20 20 +3 +3 18 17 +3 +3 8 15 0 13 -1 18 +2 17 +2 15 +2 17 +2 ------------18 +3 18 +3 15 +3 20 +3 18 +3 18 +3 9 18 0 16 3 19 +3 ------------------------18 + 18 +3 17 +3 20 +3 20 +3 17 +3 10 16 0 16 0 17 -1 --------------------------+1 --+1 17 +3 ------------- Total EC Score M SD 15.20 1.87 0.80 0.91 15.12 1.64 -0.25 2.05 15.90 2.13 1.80 1.39 16.00 1.19 1.75 0.88 14.20 1.35 1.75 0.88 16.25 0.46 2.00 0.92 14.00 1.52 1.60 0.54 10.00 1.81 1.00 0.00 14.60 2.30 2.00 0.70 14.00 3.60 2.50 0.84 16.11 1.96 2.50 .84 15.20 1.54 2.70 0.67 18.22 1.39 2.66 1.26 18.33 1.93 2.88 0.33 15.22 3.03 3.00 0.00 A couple was considered to have a split alliance when their self-report, observational scores and total mean scores differed by more than 2 points. The scores for the EC dimension demonstrated that Couples 1, 3, and 4 had the presence of a split alliance in their overall mean score and/or during certain points during treatment. For Couple 1, their overall mean self-report 128 scores did not indicate the presence of a split alliance. Looking at the session scores, however, showed that a split alliance was present at varying times during treatment. Susan‘s score of 14 is 3 points higher than Bob‘s score of 11 after the first session. Moreover, the decline in Susan‘s self-report and observational ratings for sessions 6, 7 and 8, as well as Susan‘s negative mean observational scores (m = -0.25; SD = 2.05) indicated that they varied in their emotional connection to the therapist at these specific times during treatment. For the three sessions, Susan‘s observational scores are -3,-3, and -1, while Bob‘s are +1, 0, and 0 respectively. Couple 3 had the lowest emotional connection ratings in the study. Mark reported a mean score of 14.00 (SD = 1.52), while the mean score for Heather was 10.00 (SD = 1.81). In addition to having the greatest variance out of all couples in the study, the individual session scores revealed that they experienced a split alliance at the onset of therapy that prevailed throughout treatment. Heather‘s initial score on this dimension was 13, which was 3 points lower than Mark‘s score of 16. The difference in their scores increased after Session 2 when Heather‘s rating decreased by 3 points to 10 and Mark‘s only lowered by 1 point to 15. Finally, the last session demonstrated that Heather‘s EC rating plummeted to 8 while Mark‘s score rose to 16. The divergence in their scores during this session suggested that Heather‘s connection to Leah decreased when she felt Mark‘s alliance increase. Finally, Couple 4‘s scores on the EC dimension varied the most at the onset of treatment (6 points). It is important to note that Adam did not respond to 2 statements for this dimension on the self-report after the first session, which explains his low score of 6. He explained to the researcher that it was ―too early to tell‖ for some of the statements on this dimension. He wrote ―N/A‖ to the statements of ―The therapist has become an important person in my life,‖ and ―The therapist lacks the knowledge and skills to help me.‖ The difference in the couples‘ scores 129 fluctuated between 1 and 4 points throughout the early and mid-stages of treatment. Their scores became more similar towards the later stages of therapy (Sessions 7-9). Their mean scores for this dimension differed by 2.11 points with Adam‘s score of 14.00 (SD = 0.84) and Erin‘s score of 16.11 (SD = 0.84). The mean observational ratings for this couple were high and depicted them as having the same level of connection (m = 2.75; SD = 0.84). The scores for Couple 4 also indicated a split alliance in which Adam felt less connected to Leah than Erin. The difference in their scores did decrease over time, however, which indicated that Adam‘s emotional connection to Leah was increasing throughout treatment. Therapist emotional connection self-report and observational scores. Therapist selfreport and observational ratings were analyzed to determine whether the therapists‘ perception of the couples‘ emotional connection influenced the amount of emotional connection promoting behaviors in sessions. For 3 out of the 5 couples, the therapists mean emotional connection score was similar to the partner of the couple who had the highest emotional connection rating. This occurred for Couple 1, Couple 2 and Couple 3. The higher the therapist‘s score compared to the couple, the lower the observational mean rating for the therapist. This suggests that they engage in less emotional connection promoting behaviors when they believed to be on the same level as the most connected partner of the couple. Conversely, when Leah scored lower than the highest partner of the couple (Couple 4) or lower than both members (Couple 5), her observational ratings increased. Thus, when she perceived the emotional connection to be lower, she engaged in more emotional connection promoting behaviors. Gender and emotional connection to therapist. There were gender differences in the scores on the EC dimension. The majority of the male partners (3 out of 5; 60%) rated their emotional connection higher than their female partners. While Bob‘s mean score of 15.20 (SD = 130 1.87) was only .08 of a point difference from Susan‘s (m = 15.12; SD = 2.64), Brad‘s score (m = 16.00; SD = 1.19) was 1.80 points higher than Anna‘s (m = 14.20; SD = 1.35) score and Mark‘s score (m = 14.00; SD = 1.52) was 4 points higher than Heather‘s mean rating (m = 10.00; SD = 1.81). Couple dynamics. The couple dynamics appeared to contribute to the presence of a split alliance for Couple 1, Couple 3 and Couple 4. For Couple 1, the split alliance that was present during Sessions 6, 7 and 8 appeared to be related to high levels of conflict that was observed in those sessions. Additionally, Susan‘s relationship with the therapist began to emulate the couple relationship in which she was often combative with Chad when he attempted to stop their interactional cycle. There was a negative interaction cycle between Chad and Susan, which could have contributed to her decreased emotional connection with him during those sessions. Her scores did increase by 3 points in Session 9, which indicates that Chad was able to increase the emotional connection during treatment. The split alliance in Couple 3 may be due to the couple dynamics and Heather‘s belief that the relationship distress was solely due to Mark‘s struggle with depression. She had difficulty accepting responsibility for her role in the relationship distress. Thus, when Leah focused on couple goals and issues, Heather became frustrated because she did not think that Leah was addressing what she thought was important. When Leah challenged her during Session 2, Heather interpreted it as a personal attack. In her interview, she indicated that a rupture occurred in Session 2: There was one part that made me feel really uncomfortable. I was bringing up some issues my partner has, and she said, to me, she said, she said, oh, you had these issues when you met, right? And, um, I don‘t know, I thought that comment was inappropriate. It made me upset. (Heather, C3, Interview #1) 131 She also disclosed that the comment impacted her relationship with Leah, but she did not feel comfortable addressing this with Mark in the room. Her discomfort with addressing her relationship with Leah seemed to be influenced by her low feelings of safety with her partner. Leah sensed that the alliance was unbalanced and discussed this dynamic in terms of the couple dynamics: [Mark] identifies himself solely with the couple. His identity is enmeshed within the couple, whereas [Heather] is not. So, it is very hard to join. I think at times, [she] feels left out. I want to say my alliance is stronger with [him], but that‘s because he‘s the weaker partner. He has less power in the relationship, so I have to make sure that I really get the stronger partner I guess the fact that they are so conflictual, the moment I begin to align with one partner, I have the other partner attacking them. I feel torn. (Leah, T2, Interview #1) Leah‘s ability to simultaneously join with both partners and build equal individual partner alliances was challenged due to the differing perspectives of the presenting problem and Heather‘s belief that she was not part of the couple problem or solution. As such, Heather‘s individual alliance with Leah was negatively impacted when Leah attempted to address her role in the relationship. Thus, a split alliance occurred in which Heather‘s ratings of alliance were significantly lower than Mark‘s alliance scores. Finally, Adam‘s lower scores at the onset of Couple 4‘s treatment can be explained by his initial hesitancy to participate in the research study due to his concern that they would not connect with the therapist. Both partners stated that their initial goal was to ensure they were comfortable with the therapist. It seemed that they were not automatically trustful of the therapeutic process, which could explain his low initial score. Additionally, the couple‘s relationship distress was due to the perpetual lack of trust between them despite being together for 6 years. This lack of trust in their intimate relationship could affect how they are in all relationships, even the therapeutic one with Leah. Even though Erin‘s ratings were relatively 132 high in terms of overall alliance and the EC dimension, she was unable to articulate why she was able to trust Leah: I am not sure why I trusted her right away, but I don‘t trust therapists really quickly. It had to be about her first impression, I am sure. I have been open to therapy every time I went, so it has never been a forced thing. It was just that they weren‘t right. I mean I am also talking about how I trust her, but here I am sitting with a guy I have been with for 6 years and I don‘t trust him. (Erin, C4, Interview #1) The lack of trust in the couple relationship may have influenced Erin to immediately trust Leah. In that sense, it appeared that Erin turned to Leah for support instead of her partner. These findings suggest that couple dynamics have varying influences on how the partners connect to their therapist. Synergy effect. This theme was identified through the interview data and was actually coined by one of the participants (Anna, C2). The synergy effect refers to the propensity that a partner‘s individual alliance with the therapist is based on their perception of the relationship between their partner and therapist: I thought he did a good job making my partner talk about emotions, and seeing that my partner had confidence in him, also made me feel good. I think that affected me in a very positive way. It gave me confidence that he was able to do that with him because I knew that it meant that my partner felt very comfortable with him, and so that helped. And you know, it‘s like, well he likes him, so I like him. It‘s a synergy effect. (Anna, C2, Interview #2) Three (37.5%) clients reported that if they perceived the therapist as having a positive influence or relationship with their partner, they had more positive feelings toward the therapist. The therapist‘s ability to connect, engage, or elicit responses from their partner was an important contribution to their positive feelings about the therapist. This finding was also supported by the quantitative alliance scores. The self-report scores increased when partners witnessed this dynamic. For instance, Anna‘s ratings increased by 8 points when her partner disclosed vulnerable feelings and exhibited intense emotional reactions. The synergy effect appeared to be 133 most important for the female clients who felt that they were unable to elicit emotional responses from their partner. Thus, in a sense, they needed the therapist in order to know how their partner is feeling: But some of it is also just how well she responds to my partner. I think he is doing a lot better opening up, so that has also made me like her more because if she can get him to open up. She can knock down those walls, so it gives me better insight into her, too. It makes me feel more connected to her. I feel good and that makes her an important person in our relationship because if she can get him to open up. (Erin, C4, Interview #1) The synergy effect was also present when the therapist challenged partners. Bob‘s emotional connection with Chad increased when he perceived Chad to challenge his partner in ways he thought were necessary: It gave me even higher regard for him because he is able to gently, diplomatically, push where he felt he needed to in order to get through a defense and get to a direct line to what she is feeling. I am unable to do that, so it was nice to see him do it and get where we need to go. (Bob, C1, Interview #1) Observational data revealed that the synergy effect could also have opposite results for conflictual couples. This occurred for three couples in the study. After a gradual increase in scores during the initial phase of treatment for Couple 1, Bob‘s scores started to fluctuate after the fifth session. This instability reflected Susan‘s insecure alliance throughout the later sessions, which was caused by her conflictual relationship with Chad. Additionally, for Couple 3, the more Leah addressed and supported Mark‘s experience, the more Heather would disengage and become frustrated. This also occurred with Couple 4 when Leah supported Adam‘s feelings that contrasted with Erin‘s perception. During Session 10, Erin disclosed her feelings for regret about the couple relationship, which offended Adam. After the disclosure, she indicated that she felt vulnerable and feared Adam was angry because he stated that he did not want to respond after Leah inquired about his feelings. Leah supported Erin‘s vulnerability, but when she attempted to validate Adam‘s disconnection, Erin was instantly angry and stated, ―He is obviously showing he 134 doesn‘t care, and then you give him the benefit of the doubt about not wanting to respond after hearing that? You know, fuck that, I‘m leaving. I‘m done.‖ Despite Erin‘s high ratings of emotional connection to Leah, this moment was deleterious to her alliance as the couple never returned to therapy. Analysis of the data indicated that individual alliances with the therapist are partly based on how the client perceives the therapist responded to their partner. This has a synergy effect on the alliance in which connection increased when the client thought that the therapist responded to their partner in a way that supported their needs. However, if the client perceived the therapist to support or validate a feeling or experience for their partner that was not similar to how they feel, their emotional connection scores decreased. Behaviors that promote individual alliances. This theme describes the behaviors that therapists displayed to promote the alliances with the individual partners. The behaviors were organized into three subthemes: 1) Promote emotional connection, 2) Promote engagement and 3) Implement individual sessions. Promote emotional connection. Behaviors that promoted emotional connection were related to the self-of-the-therapist and therapist responsiveness to the needs and feelings of the clients. As discussed previously, the self-of-the-therapist played an important role in the alliance. Additionally, the emotional connection increased when the clients had a sense that the therapist knew them better: I would say maybe a little more connected in the sense of you simply got to know each other better. I shouldn‘t say gotten to know each other because I don‘t know any more about him now than I did at the beginning, but he knows us better and it was easier to get to a deeper level of talking about your emotions, or a deeper level of therapy more quickly because you have gone through all of the background things. So, I felt more confident, I felt more satisfied. (Erin, C4, Interview #2) 135 Responding to the needs and feelings of the clients was a critical component to promoting emotional connection with the couples. Both therapists believed this was necessary in their role as a therapist. Leah discussed the importance of attending to the needs of both individuals and giving a voice to the clients in building the alliance. If this is not done, she believed there would be a poor alliance: But, part of it for me is, I haven‘t attended to their needs. I haven‘t really addressed what they want from therapy or what they are expecting from this. Or I haven‘t really attended to making sure they have a voice in this therapy as well, then the alliance will not be good. (Leah, T2, Interview #1) Ensuring that both couples have a voice and are validated was foundational to the therapists‘ work with the couples. Chad believed that that if a therapist balanced and attended to both partners at all times, trust will increase, and the alliance will continue to be reinforced: I think it is a bit easier. It definitely, I mean, it seems a little more difficult. But, as long as you are giving equal weights to what each are saying, it reinforces to the individuals that you are trusting what is going on. There are reinforcers [in couple therapy] in that they can look at how I treat their spouse or partner, and can say, well, he does that for me, and therefore the trust is greater. (Chad, T2, Interview #1) He believed that validation was an important theoretical tool, but also a prerequisite for healthy relationships, including the therapeutic one. Both therapists believed that all clients want to be heard and understood, which is what typically prompts their participation in therapy. The clients also discussed their appreciation of the responsiveness of the therapists. Four clients (50% of clients in the study) stated that when the therapist specifically focused on their emotional responses, they felt an increase in connection: My connection increased because he was focusing more on my issues and talking more in depth about emotional things. So, I think that kind of lend itself to increased connection. There is a little more focus on the emotions and that increased the connection. (Brad, C2, Interview #2) Additionally, emotional connection increased when the therapist expressed their personal emotions or reactions to the experiences of the clients: 136 She will share things like, what you are saying hurts me, too. What you are saying makes me feel like that would hurt. So she does express emotions also, so that‘s nice and definitely plays a factor because if you can trust somebody, you can open up more to them and you can show them different sides of you. Like, I went to one session and cried. And it was ok because the more you trust someone, the more you can open up to them. When you have a connection with them, especially the emotional connection, then it is easier. (Christina, C4, Interview #2) Sometimes the focus on emotions did not increase the connection between the therapist and client. When Chad attempted to focus on Susan‘s (Couple 1) underlying emotions in Session 6, her score on the emotional connection dimension decreased by 4 points. Susan discussed her difficulty following Chad‘s focus on her emotions in sessions during the interview: He goes with ―well, how does that make you feel?‖ And, I will say frustrated because and he will interrupt and say not because how does that make you feel? He won‘t let me explain. But, I may just have a different point of view than he does. He‘s trying to teach me how to follow his format, and I am struggling. I just am a take charge kind of person because I was a single mom with three kids. So, it is hard when he is telling me I can‘t take charge. (Susan, C1, Interview #1) Finally, therapist responsiveness was present even when specific needs were not acknowledged or discussed by the clients in the study. For instance, Leah worked hard to facilitate authentic discussions with the partners in Couple 4 about their fears in the relationship. When it became apparent that Adam was unable to articulate his emotions, Leah shifted her focus. She slowed down and supported him by helping him to talk about his feelings: Yea, it‘s so hard to pull from him. I was just thinking that I have got to get him to this place where his level of authenticity I don‘t feel is very present yet. And, I‘m not sure how to address it. Because I don‘t know if he has ever really been present with anyone, so to have this real experience that I am asking for him is kind of a lot… I try to use my words for them…since he has not that insightful into himself and his feelings, and he doesn‘t have that level of insight. I try to use words for him to help him and also to communicate to him that I get him. I understand. (Leah, T2, Interview #2) Promote engagement. The therapists wanted to ensure that both partners were equally engaged at all times. They believed that it was their responsibility to engage clients, and to continuously address both partners regardless of the issue at hand: 137 I get both partners‘ perspective. So, when I bring up any issues, I make sure if one person responds, I want to make sure if the other agrees, is that how they see it? I try to make sure I have both partners contributing and engaged. I have both partners have active voices in the session. So, if one person is putting the other down, I don‘t just go with that because I want to know if the other is agreeing or seeing it that way. So, I want to make sure both partners are involved in sessions at all times. (Leah, T2, Interview #1) While partners noticed the continuous effort of the therapists to engage both partners at all times, they also recognized when their partners dominated the session. There were several clients who were naturally more talkative than their partners in the study. Even though the client understood this about their partner, the imbalance in engagement decreased connection: Maybe at times I feel like my partner is doing most of the talking. I am not mad, but maybe I feel a little less connected because she is the one talking and getting all the attention. But, she is more talkative. (Joe, C5, Interview #2) They also noticed if their partner was not engaged, or if they withdrew from the session. In those moments, they expect the therapist to address the process and re-engage their partner. I expect the therapist to address that. Because if he‘s withdrawing and then he says how does that make you feel? As long as he is then re-engaged. If this is clearly a topic that he withdraws from, which is part of the problem because we don‘t address this effectively at home, so it‘s your job as the therapist then to re-engage or address this or involve in some way. (Anna, C2, Interview #1) Besides ensuring that both partners were involved in the session, the therapists also conducted activities during the sessions, gave directives to the couple, and assigned homework or ways to incorporate their sessions into their daily lives. The clients appreciated the practicality of these interventions because it forced them to be accountable, as well as enabled them to see how the information they were learning inside the therapy room relates to their life. These tasks showed the clients that the therapists understood their relationship and knew how to help. I really try to check in with both of them in the session. So, if I ask one a question to one, I make sure to ask it to the other and check in with both. That way, I increase the presence of the one who is not as dominant. And, that is also a way to align with them by pulling them in. So, I think I do align with them separately in the session. (Leah, T2, Interview #1) 138 Implement individual sessions. The use of individual sessions enabled the therapists to understand the individual components of the couple relationship. Both therapists tended to do this within the first few sessions, and the structure depended on the therapeutic context. For instance, Chad‘s standard practice was to utilize the second and third appointments as individual sessions for each individual. He was unable to do so with Couple 1 because Susan refused to attend an individual session because she believed Bob needed those sessions more. This dynamic was difficult for Chad to control since Bob would arrive to sessions without Susan. Chad used the individual sessions to assess individual history and background, which allowed him to focus completely on the partner. He hoped this demonstrated that he wanted to know the partners as individuals: Generally, whenever I see a couple, I bring them both in at first, and then I have separate sessions for the second and third sessions so that I can get to know each of them individually. And, that way they know that I understand that there are individual components to the couple relationship. And, I build the alliance there, and then bring them back as a couple so that I can help them know that yes, we are working on this as an issue of the couple.(Chad, T1, Interview #1) Leah did not have a specific time frame for the individual sessions, and tended to devote complete sessions to each individual or expand the time frame of one session and see the individuals separately during that time frame. She often used them when she believed it was necessary, which incongruously happened during the third session for Couple 3, Couple 4, and Couple 5. The reasons for choosing this time for each couple was based on different reasons. She used individual sessions when she was unable to connect to Couple 3 due to their harsh interactions, while her purpose for Couple 4 was to assess the level of trust for each partner. She decided in the moment during the third session to split the partners of Couple 5 in order to evaluate domestic violence and commitment level after a two week gap in treatment. Thus, 139 Leah‘s decision to use individual session was based on couple dynamics and assessment of safety and commitment. The therapists believed that it helped them to align with the collective couple alliance by getting to know the individual partners. A clear boundary was set, however, in which the therapists ensured the partners that individual sessions were designed to address the individual issues and not couple issues. Enforcing these boundaries did not allow either member of the couple to use that time to criticize or complain about their partner. Chad believed that setting these boundaries clearly states that the role of the couple therapist is to treat the couple, not each individual. Therefore, couple sessions focused on the couple dynamics that impeded the relationship, and individual sessions addressed any partner issues that were negatively affecting the relationship. He thought that it builds the individual alliance when the partners realize that he treats each of them equally: And, I continually reiterate that in sessions about whether we are talking about couple issues or individual issues, we don‘t talk about couple issues individually, and we don‘t talk about individual issues as a couple. So, this way, they know I won‘t discuss the other partner outside of the couple context, and then they realize I won‘t do it with their partner, either. I think it builds trust. (Chad, T1, Interview #1) The therapists also believed that it was important to utilize individual sessions when they did not feel strongly aligned with a particular partner during couple sessions. This occurred for Couple 1 and Couple 3. Chad‘s primary reason to have Bob leave the room during Session 6 was to stop the harmful interaction between the partners, but it also allowed him to join with Susan, who was the more difficult partner. Likewise, Leah used her individual time with Heather to focus on connecting since it was difficult for her to do so in the couple sessions due to her critical and attacking behavior. 140 The clients indicated that they appreciated the one-on-one time with their therapist. Five out of the eight (62.5%) clients in the study indicated that the alone time allowed them to feel closer to the therapist because it allowed the therapist to get to know them: I felt more connected to her when we did our individual sessions. It was nice because I got to tell her exactly how I felt from my perspective and that is always a good thing. I think that gave her insight into who I was a person, too. I felt more comfortable after that. (Christina, C5, Interview #2) Thus, when the individual sessions were used carefully and methodically, the alliance ratings for both partners increased. This occurred for Couple 4 and Couple 5. However, when the individual sessions were not utilized correctly, negative consequences ensued for the alliance. This occurred in Couple 2 when Susan did not attend two sessions. Chad was unable to prevent this from occurring as they would not notify him prior to the session. Susan also refused to attend any individual sessions in order to balance out the alliance. The effect of these individual sessions created an imbalance in the alliance in which the alliances of Chad and Bob began to emulate one another. Furthermore, the imbalance exacerbated the already conflictual couple dynamics and created an unstable alliance for all participants for the rest of treatment. Within-Couple Alliance This category contains the key components and behaviors that contribute to the alliance that exists between the couple. The results are organized into the following two themes: 1) Key Components of Within Couple System Alliance and 2) Behaviors that Promote Within-Couple Alliance. The first theme refers to the dynamics that contributed to the relationship between partners. The second theme describes the alliance promoting behaviors that were influential of building and managing that alliance throughout the therapeutic process. Key components of within-couple alliance. In order to identify the key components of the within-couple alliance, the two within-system relational aspects of alliances were analyzed. 141 The self-report, observational and interview data for the Safety and Shared Sense of Purpose (SSP) dimensions of the SOFTA identified the within- couple alliance for the participants in the study. This section organizes the results into two categories: 1) Feelings of Safety and 2) Shared Sense of Purpose. Feelings of safety. A sense of safety is client-therapist relational dimension that is interrelated with the level of engagement and connection the clients have with their therapist, but is also influenced by the within-system dynamics of the couple (Friedlander et al., 2006). The nature of couples therapy provides natural challenges to the level of safety due to the added presence of a partner: Well, for starters, you are not alone. Your partner is in here with you and there is a degree of safety that you don‘t have to think about when you are alone. There is more of a vulnerability in couples than individual just due to your partner being in the room with you. (Christina, C5, Interview #1) It was theorized that this dimension had a reciprocal process for alliance in that the ratings of safety impact the within-couple alliance, and the couple dynamics influence the feeling of safety in the therapeutic context. Analysis of the self-report and observational measures provided information about the couples and individual partners feelings of safety. Table 4.9 depicts the mean scores and standard deviations for all participants on the Safety dimension of the SOFTA self-report and observational measures. 142 Table 4.9. SOFTA Self-Report and Observational Mean Scores for Safety Participants Bob Couple 1 Susan 1 n 10 Couple 2 n=8 Couple 3 n=5 Couple 4 n = 10 Couple 5 n=9 1 Chad SAFETY SOFTA Self-Report SOFTA Observational Mean SD Mean SD 15.00 2.10 1.00 1.15 2 2 2.13 1.12 12.37 0.87 10.90 1.52 1.70 1.05 Brad Anna Chad Mark Heather Leah Adam 15.12 16.37 14.37 14.60 11.20 12.80 3 14.00 1.64 1.30 1.68 3.20 1.77 1.78 1.22 1.50 1.50 0.50 0.60 0.60 1.60 2.00 0.53 0.53 1.06 0.89 0.89 1.14 1.33 Erin 19.11 14.20 16.00 17.33 16.33 3 1.26 2.00 1.88 2.29 2.23 1.22 2.39 2.40 1.77 2.33 3.00 1.07 1.92 0.86 0.00 Leah Joe Christina Leah n=total number of sessions 2 Scores are based on a total of 8 sessions because she did not attend the 2 3 nd rd and 3 sessions th Scores are based on a total of 9 sessions because they did not fill out the 10 session self-report Results revealed that Couple 5 and Couple 2 had the highest ratings of safety. Joe‘s mean score was 16.00 (SD = 2.23) and Christina‘s was 17.33 (SD = 1.22). They also had the highest mean rating for safety by the therapist with Leah‘s mean score of 16.33 (SD = 2.39). Couple 2 has the next highest shared mean ratings out of the couples. They also have the least difference between their scores with Brad‘s score (m = 15.12, SD = 1.64) being 1.25 point lower than Anna‘s mean score (m = 16.37, SD = 1.30). Their scores on the safety dimension suggest that Couple 5 and Couple 2 had positive within-system alliance due to their high levels of safety. Couple 1, Couple 3 and Couple 4 had variances in their scores for safety. Couple 1 had a difference of 2.63 points between Bob‘s mean score of 15.00 (SD = 2.10) and Susan‘s mean 143 score of 12.37 (SD = 2.13). Bob also achieved a slightly higher observational rating of 1.00 (SD= 1.15) compared to Susan‘s rating of 0.87 (SD = 1.12). These scores indicate that Bob felt moderately safe in sessions, but Susan had significantly less feelings of safety in the therapeutic context. The scores for Couple 3 varied and were the lowest in the sample. Mark had a mean score of 14.60 (SD = 3.20), while Heather scored 3.40 points lower with her mean of 11.20 (SD = 1.77). Thus, the results suggested that this couple had different levels of safety, and that Mark had significantly higher levels of safety than Heather. Their observational ratings were also the lowest of the sample with both partners achieving a mean rating of .60 (SD = 0.89), indicating that this couple did not portray behaviors that signified they felt safe during sessions. Finally, Couple 4 had the greatest difference in their mean self-report scores (5.11 points), yet their observational mean scores were identical at 2.00 (SD = 1.33 for Adam; SD = 1.88 for Erin). Erin‘s mean self-report score was high (m = 19.11; SD = 1.26) while Adam‘s score was in the moderate range (m = 14.00; SD = 1.22). These scores suggest that Erin felt extremely safe in sessions, while Adam felt a significantly lower sense of safety within the therapeutic process. Due to the variances in the Safety scores for the partners of these three couples, results suggest that Couples 1, 3 and 4 did not have positive within-system alliance. Also, in terms of gender, the female partners in three out of the five couples had higher scores than their partners on this dimension. Again, the male partners of Couple 1 and Couple 3 had higher mean ratings of safety than their female partners. Therapist self-report and observational ratings for Safety. Therapist self-report and observational ratings were analyzed to discern whether the therapists‘ perception of the couples‘ feelings of safety influenced the amount of safety promoting behaviors in sessions. For Couple 1, 144 Chad‘s self-report mean score of 10.90 (SD = 1.52) was significantly lower than the couples‘ ratings. His observational mean rating was 1.70 (SD = 1.05), which suggests that he was employing safety related alliance behaviors during their sessions. For Couple 2, Chad‘s mean score for safety was 14.37 (SD = 1.68), which was slightly lower than the lowest rating of the couple. His observational mean rating was 0.50 (SD = 1.06). The results suggested that Chad‘s perception of the level of safety did influence how much he promoted safety with the couples. When he perceived that there were low levels of safety such as in the case of Couple 1, he exhibited more safety promoting behaviors during sessions. When safety was not as much of a concern for Couple 2, he exhibited less safety promoting behaviors. Leah‘s ratings of safety for Couples 3 and 4 were associated with the amount of safety promoting behaviors she utilized in sessions. Her mean rating of 12.80 (SD = 1.78) for Couple 3 fell in between their scores (m = 14.20, SD = 1.78), suggesting that she sensed the discomfort for both partners in the sessions. The observational mean score of 1.60 (SD = 1.14) indicated that she was exhibiting various safety promoting behaviors in sessions. For Couple 4, her mean rating (m = 14.20; SD = 2.29) was similar to Adam‘s rating of safety which was 5.11 points lower than Erin‘s mean score. This suggests that she perceived the difference in feelings of safety between the partners and rated couple‘s level of safety based on the partner who exhibited the lowest amount of safety. Leah‘s observational mean score for this couple was 2.40 (SD = 1.07), which suggests that she focused on safety promoting behaviors frequently during sessions. Interestingly, Leah‘s mean self-report and observational scores for Couple 5 were the highest out of all 3 couples. Her mean score on the self-report (m = 16.33; SD = 2.39) fell in between the couple‘s ratings which suggests that she perceived the safety level of the couple rather than a 145 particular individual. Her mean score of 3.00 on the observational measure indicated that she was continuously implementing safety promoting behaviors in sessions. The findings from these three couples suggest that for Couples 3 and 4, Leah implemented safety promoting behaviors in sessions when she perceived their safety low. She utilized more promoting behaviors for Couple 4, which could be due to her perception of their fragmented feelings of safety. For Couple 5, she exhibited the highest amount of safety promoting behaviors, even though she perceived this couple as having the highest level of safety out of the three couples. It is also reasonable to speculate that their scores were due to her strong efforts to promote safety. Shared sense of purpose. A shared sense of purpose is a within-system relational dimension of alliance that refers to the commitment level of the partners, and degree to which they have shared goals and work collaboratively together in therapy. This aspect of alliance is influenced by the level of cohesiveness or distress the couple is experiencing. As such, the more cohesive a couple is, the stronger their shared sense of purpose. Conversely, the more distressed a couple is, the more likely they are at odds with one another and have a weaker shared sense of purpose. Analysis of the self-report and observational measures provided information about the couples and individual partners shared sense of purpose. Table 4.10 depicts the mean scores and standard deviations for all participants on the Shared Sense of Purpose (SSP) dimension of the SOFTA self-report and observational measures. 146 Table 4.10. SOFTA Self-Report and Observational Mean Scores for Shared Sense of Purpose Participants Bob Couple 1 Susan 1 n = 10 Couple 2 n=8 Couple 3 n=5 Couple 4 n = 10 Couple 5 n=9 1 2 3 Chad Brad Anna Chad Mark Heather Leah Adam Erin Leah Joe Christina Leah SHARED SENSE OF PURPOSE SOFTA Self-Report SOFTA Observational Mean SD Mean SD 2 16.90 1.44 2.58 -0.12 2 2 2.58 2 1.03 3 1.06 0.74 1.55 1.13 1.34 1.78 2.99 1.83 1.87 1.87 1.37 -0.20 -0.20 2.60 1.11 3 1.01 1.11 2.52 1.56 0.86 2.45 1.93 2.40 1.77 1.77 2.11 0.96 1.98 1.98 1.16 17.25 12.40 3.01 -0.12 2.48 16.62 17.62 16.12 14.60 13.40 13.80 17.22 19.44 15.30 19.33 17.44 17.33 1.12 1.12 1.30 1.64 1.64 1.11 2.52 n=total number of sessions Scores are based on a total of 8 sessions because Susan did not attend the 2 nd rd and 3 sessions th Scores are based on a total of 9 sessions because they did not fill out the 10 session self-report Results of the self-report and observational analysis for the SSP dimension showed that Couple 2 and Couple 5 had the highest scores. Joe had a mean score of 19.33 (SD = 0.86), which was 1.89 points higher than Christina‘s mean score of 17.44 (SD = 2.45). Their mean observational rating was 1.77 (SD = 1.98). For Couple 2, Brad‘s mean score was 16.52 (SD = 1.06), which was 1.10 points lower than Anna‘s mean score of 17.62 (SD = 0.74). They achieved a mean observational rating of 1.87 (SD = 1.12). These results support the hypothesis that Couple 5 and 2 had strong within-couple alliances. Couple 4‘s SSP scores were the third highest in the study with Adam‘s mean score of 17.22 (S = 2.99) and Erin‘s mean score of 19.44 (S = 1.01). Their scores differed by 2.22 points. 147 Their mean observational rating was 1.11 (SD = 2.52). Couple 1 had the least difference (0.35 of a point) between their self-report scores. Susan‘s score of 17.25 (SD = 3.01) was slightly higher than Bob‘s mean score (m = 16.90; SD = 1.44). However, they achieved negative mean observational ratings (m = -0.12; SD = 2.28). Finally, Couple 3 scored the lowest on this dimension. Mark‘s mean score (m = 14.60; SD = 1.13) was 1.20 points higher than Heather‘s mean score of 13.40 (SD = 1.34). Their mean observational rating was -0.20 (SD = 1.64). These results support the postulation that Couples 1, 3 and 4 would have weaker within-couple alliances. Therapist self-report and observational ratings for shared sense of purpose. Therapist self-report and observational ratings were analyzed to discern whether the therapists‘ perception of the couples‘ feelings of shared sense of purpose influenced the amount of SSP promoting behaviors in sessions. For Couple 5, Leah‘s rating of the SSP (m = 17.33; SD = 1.93) was similar to Christina‘s rating, who had the lowest score of the couple. Leah achieved a mean observational rating of 2.11 (SD = 1.16), which indicates that she was consistently displaying behaviors that promoted SSP frequently in sessions. Chad‘s mean rating for Couple 2 (m = 16.12; SD = 1.55) was similar to Brad‘s score, who had the lowest score of the couple. He achieved a mean observational of 1.37 (SD = 1.33), which suggests that he was moderately exhibiting SSP promoting behaviors. Leah‘s mean rating of 15.30 (SD = 1.56) was 1.92 points lower than Adam‘s score, who was the partner who scored the lowest for Couple 4. Leah put forth considerable effort to promote their SSP throughout treatment as evidenced by her mean observational rating of 2.40 (S = 0.96) for this dimension. A notable finding that is not displayed in the table is that her efforts to promote SSP for this couple were also supported by the changes in Adam‘s ratings of 148 SSP over the course of treatment. His scores increased from 10 after the initial session to 20 after the ninth session. Erin‘s ratings of SSP were consistently high throughout treatment. Her initial score for SSP was 19 and was 20 after the ninth session. Despite the large gap in scores at the beginning of treatment, they achieved the least variance between their scores on this dimension compared to their ratings on the other three parts of the SOFTA scale. Finally, for the couples who scored the lowest on this dimension, the therapist observational ratings were the highest. Chad‘s mean rating (m = 12.40; SD = 2.48) was significantly lower (4.50 points) than the lowest rating of Couple 1. He received a mean observational score of 1.25 (SD = 1.83), which suggests that he was attending the SSP of the couple throughout treatment. Likewise, Leah achieved a mean observational rating of 2.60 (SD = 1.11), which indicates that she was exhibiting consistent SSP promoting behaviors with Couple 3. Gender and shared sense of purpose. In terms of gender differences on the SSP dimension, the results of this study showed that 3 out of 5 (60%) female partners had higher scores. The female partners in Couple 1, Couple 2 and Couple 4 scored higher than their male counterparts. The male partners in Couple 3 and Couple 5 scored higher than their female partners on the SSP dimension. Behaviors that promote within-couple alliance. The observational and interview data provided details about the specific behaviors that strengthened the within-couple alliance via the client, therapist and independent observer perspectives. The behaviors that were identified related to two categories: 1) Promote Safety and 2) Promote Shared Sense of Purpose. Promote safety. There were some distinct client and therapist behaviors that the interview and observational data found to be influential of safety. The degree to which the partners were 149 able to feel safe was based in their overall connection with the therapist. If they believed they fit with the therapist, that he/she was neutral, and that the therapist can help them, they instantly had a feeling of safety. But, with her, it was like an instant ok, she‘s going to be neutral, she‘s going to be normal, and she‘s going to hear both of us, she‘s going to open both of our eyes, so I guess it‘s more knowing that it‘s going to work with her for us to feel safe, and we do feel safe. (Adam, C4, Interview #1) The therapist behaviors that promoted safety were related to the following three categories: 1) Slowing the Process: Moving at the Client‘s Pace, 2) Creating a Safe Place and 3) Effectively Managing Conflict. Slowing the process: Moving at the client’s pace. Trust was essential for all clients in the study. They needed to trust the therapist in order to feel comfortable to be vulnerable in sessions. It was clear that some clients had a natural ability to be open to the therapeutic process, which appeared to help them feel safe. However, half (n = 4) indicated that they were not initially comfortable, and it took time for them to be able to share their feelings. The therapists were able to build their trust by not pressuring them to disclose anything until they were ready. Through reassurance and slowing the pace of therapy, the therapist increased the overall comfort level of the clients. Instead of pushing these partners, Leah repeatedly assured them that she was accepting of their pace and was not going to make them proceed at a speed that was too fast for them: She has made me feel more comfortable during those heated moments and that it is ok to talk about it. She doesn‘t make me feel pressure, you know. She makes me feel more comfortable by not pressuring me to talk when I don‘t want to. She has made that clear— that she won‘t make me respond when I am not ready or can‘t. (Joe, Couple 5, Interview #1) Moving at a comfortable pace was important to build trust and increase the comfort level of the clients. Additionally, the therapists had to learn how to elicit the desired responses from the 150 partners. For three of the male clients in the study, it was not probable that they would disclose their thoughts or feelings on their own without any prompting from the therapists. The therapists adapted their styles to match the clients style in order to engage the client into uncomfortable discussions: I feel like, um, I‘m not really a talkative person, so that first and second session were great because if I don‘t have someone asking me questions, I won‘t talk. So, she would ask the questions, and she knows how to get it out of me, so I feel a lot more comfortable opening up. She is going at my pace and not forcing me to go faster or anything. (Adam, C4, Interview #1) Provide a safe place. When the clients were asked about how their therapist promoted safety, the majority stated that (n=6; 75%) their presence made them feel safe because it changed the dynamics that typically induced conflict for the couple when they were alone: Just the very fact that he‘s asking the questions so I am not just bringing stuff up out of the blue. And [my wife] realizes that maybe I am being honest about things and maybe I am not so open when we are alone, although she tries to probe and find things out. I think she realizes that here is the better venue to do that as opposed to at home getting worked up about something small probably. The therapy sessions have helped in that regard. (Bob, C1, Interview #1) It‘s that when she asks a question, there is no defense because he doesn‘t know her agenda, where when you are in a relationship, he has always made an assumption that this is what I am going to say, or whatever, it‘s great in general because as humans, we all do it. But to be able to look outside of the box and see outside of our relationship. That is why it is really successful because we are allowed to be without feeling judged. (Erin, C4, Interview #1) Additionally, the presence of the therapist held them accountable for their behavior, which helped control the dynamics that were too intense for them to handle alone: ―Again, that whole having to answer to her thing. It forces you to do good behavior and start practicing good habits to hopefully change the bad ones,‖ (Erin, C4, Interview #1). In order to establish this safe place, the therapists employed several rules that set the foundation for their work with couples. First, they educated the couples about what behavior is 151 acceptable in therapy, and what behavior they would not support at the onset of treatment. They informed the couple that they would not tolerate any critical or harmful interactions that they felt would not be productive for therapy or the couple. They also stated that their role was not to play ―judge‖ in which they would be confirming who was right or wrong, or policing the couple‘s behavior. They clearly stated their role as their therapist was to provide guidance and to collaboratively work with them to establish their goals. Moreover, they clearly defined the difference between couple and individual therapy. They delineated the differences between the two contexts and emphasized that the ―client‖ is the couple rather than one partner. Thus, a boundary was immediately formed that would not allow either partner to make it appear as if the relationship distress was the result of one partner‘s behavior. Finally, the therapist encouraged that the couple establish their own rule to refrain from discussing issues that arouse conflict outside of the session: She just made us promise that what we talk about in the room stays in the room, if it‘s super-heated, she will stay with us a little longer to try to help us work it out, or if something to where we can just drop it to the next session and pick it up again where we left off, we will. She made that promise and we have all stuck to it. (Adam, C4, Interview #1) These rules formed the structure that enabled the therapists and overall therapeutic process to function as a safe place for the couple. Effectively managing conflict. The therapist‘s ability to interrupt, mediate and restructure negative interaction cycles also influenced feelings of safety for the clients. Intense moments occurred frequently in sessions and where characterized as times when emotional issues are raised and partners are in vulnerable places. These moments did not involve high conflict, but rather discussions that were hard for the partners to discuss. There were two key therapist behaviors that appeared to help the clients feel safe during these moments. First, the therapists 152 actively reframed any statements that appeared blaming, defensive or harmful without silencing the partners. Leah repeatedly did this by portraying understanding and acceptance of both partners‘ positions: It was nice because she gave us both a chance to speak and say how we feel, and then she was not judgmental at all. She lets you know that she understands how you feel. It‘s nice because you don‘t think she is judging or against you. It‘s nice. (Joe, C5, Interview #2) Second, they validated and explained why they had to interrupt the partner they were silencing. Thus, instead of simply shutting them down, they would make direct validating statements to them before they gave them negative feedback: She told me to shut up once. She was like, [Christina], hold on. Let me talk to him, just hold on and then I will get back to you. I hear that you are upset and angry, and I get it. I do, but I want to stop this part right here because it is getting negative, and I don‘t want to lose him in this process. (Christina, C5, Interview #2) By doing this, the therapists earned the clients‘ trust that enabled them to be more directive and immediate when it was necessary to interrupt direct conflict that was harmful to the relationship. When the therapist was not able to soften harmful interactions, they would stop the session or separate the partners. This occurred twice in the study. Chad did this in Session 6 with Couple 1 when he was unable to soften Susan‘s harsh words against Bob. In order to protect Bob and attend to Susan‘s needs of releasing the anger, Chad asked Bob to leave the room. In the interview, Bob discussed what it was like for him to leave the room: Yea, well last session was dramatic compared to all the other ones. She was quite emotional about certain things and she had a right to be. It got to a point where he had to get me out of the room to work with her to get down to her feelings. (Bob, T1, Interview #1) Leah also interrupted Couple 5 during escalating conflict to protect both partners. Separating the couple prevented further harm to the couple, but also reaffirmed that both 153 therapists were not there to witness or facilitate hurtful behavior. Ceasing the destructive patterns in the moment was an important part to build safety for the couples. Promote shared sense of purpose. While all the couples entered therapy with some degree of conflict, 3 out of the 5 (60%) couples presented with couple dynamics that weakened their sense of shared purpose. Observational and interview data revealed that this fragmented within-couple system was the single most challenging element of therapy for the therapists. The therapist behaviors to promote SSP were related to the following categories: 1) Encouraging Commitment: Establishing Shared Needs and Goals, 2) Emphasizing Common Feelings and Validating Perspectives and 3) Uniting the Couple. Encouraging commitment: Establishing shared needs and goals. One of the first SSP promoting techniques the therapists did, particularly Leah, was to assess the commitment level of the couples. She inquired about their commitment to therapy, as well, and explained that the commitment to the relationship and therapy are critical to being successful. She also did this to reframe the relationship for the couple since they tend to have an overall negative perspective of it: I usually ask couples this, how committed are you to the relationship? I also try to assess for if leaving the relationship is an option because if they tell me no, I go back on that and say you both say you want to stick through this, and I will work with you, you know, so that we can move you forward. (Leah, T2, Interview #2) The continuous focus on the couple goals, needs and commitment was especially needed for Couples 1, 3 and 4. This was due to the differing perspectives about what the couple issues were. While Couple 1 appeared united in their goal to stay together, they were not unified in what needed to change in order to achieve that goal. The couple sessions often shifted into discussions about Bob‘s role in the distress, and Chad struggled to reframe this as a couple issue. Similarly, Couple 3 struggled to establish common goals due to their dissimilar views on what influenced 154 their couple distress. At one point, Leah asked the partners what they ultimately wanted, and Heather stated she wanted Mark to stop being depressed, while he responded that he wanted to make her happy. Mark‘s identity was enmeshed with the couple, while Heather was unable to accept that she had a role in the relationship distress. Finally, Couple 4 expressed their desire to stay together, but their low levels of trust disabled them from having hope that they could overcome their challenges. There were intense moments in treatment when they questioned their desire and ability to stay in the relationship. For these couples, the consistent focus on goals was necessary, as well as continuous praise about their decision to come to therapy, ability to face their issues instead of ignore or run away from the problems, as well as solution-focused interventions that emphasized the times they were not distressed and identify the positive aspects of their relationship. This feedback increased their belief in each other and also served to create a new narrative for the couple that was positive and hopeful: She noticed today and said, well, guys, I think you are headed in the right direction, what can I do to help you in your problem area. She wasn‘t stumped, but it pointed out that, look at all the great things you guys have to offer each other. It was then pointed out to me, that not only do we have a therapist to answer to each week, and that we made a commitment to come here and to participate, but that we are doing things on our own and maybe can fix all this. (Erin, C4, Interview #2) Emphasizing common feelings and validating perspectives. This behavior occurred most during arguments and moments of gridlock between the couples. The therapist listened intently for common themes in their complaints, feelings, and needs. They would emphasize these whenever possible. Typically, couples had common feelings even if they were related to different complaints. When the partners had diverse needs, the therapists had to normalize the individual desires in order to promote understanding and acceptance. 155 The therapists also had to act as a translator for the couples to get them to hear each other‘s perspective. There were three ways the therapists did this. With couples who were more conflictual, the therapists would move from one partner to the other directing them to talk about their perspective and then translating the perspective to the other partner. It was necessary for them to have the couple talk through the therapist due to the tendency for the conversation to escalate. Other times, they would have the partners repeat what they heard from each other and ask how they interpreted the message. This addressed any distortions or misinterpretations they had about the partner‘s perspective. The therapist educated the partners about how the assumptions they hold about the words, tone or looks from their partners prohibit them from hearing each other. Leah did this with Couple 5 when they were unable to hear each other when they were in an argument: She kind of brought us into, you know [Christina] and I were explaining how we didn‘t feel like we were hearing each other, and she put in so many words of, and helped me see in terms of [Christina‘s] perspective, and [Christina] said that she pointed out things that made her see things from my perspective. And that really helped. She brings things together to help us see why we are feeling this way, and that has been very helpful. (Joe, Couple 5, Interview #1) The most powerful way the therapists were able to get the partners to hear each other was by eliciting the emotions behind their perspectives. Chad discussed how he did this through deepening the emotion and modeling how to validate each other: Well, a lot of, to use the word of emotional focused therapy, a lot of deepening of the emotion. Also, a lot of validation, emphasizing that I respect each of their opinions despite whether or not they are true or not. Basically, I respect what they bring to the table. (Chad, T2, Interview #1) By overtly validating each other‘s perspectives, Chad demonstrated that it is possible to respect each other even when they do not agree. This moved the couple out of the gridlock that caused them to fight to be right and into a softer place that allowed them to have understanding and 156 compassion for one another. Promoting validation and softening the emotional responses enabled the partners to hear each other. When sessions facilitated these types of interactions, alliance scores would increase. The change in scores related to the emotional connection to the therapist safety, and the shared sense of purpose dimensions of the alliance. Uniting the couple. The last defining technique the therapists did to promote shared sense of purpose was to help the couple feel united as a team against the issue they were facing. This was essential at the beginning stages of therapy, specifically for Couple 1 and Couple 3. These two couples did not function as a team from the start due to the differing beliefs about the presenting problem. Both female partners believed the need for therapy and reason for their relationship distress was due to their male partner‘s psychological distress. The therapists focused on encouraging the female partners to view the distress in the context of the relationship and identify their roles in the couple distress. They consistently reframed individual symptoms and attempted to shift discussions about the male partner to conversations about couple goals or the dynamics of the relationship. As discussed previously, these maneuvers challenged the female partners, and ultimately the therapeutic process, due to their inability to be united in their perspectives of the problem. Three out of the five couples presented with a common complaint or goal for therapy. They would appear to be on the same page during the initial session when they explained the reasons why they decided to begin therapy. Over time, however, this unity became fragmented as they disclosed their perspectives and feelings about the relationship issues. With these couples, the therapists externalized the issues to help unite the couple. They continuously inquired about how the ―couple‖ was going to handle these challenges, or how they were going to prevent these issues from destroying their relationship to encourage the partners to feel like a team fighting the 157 issues. Thus, the need for the therapists to promote unification and develop a sense of team was influential of positive shared sense of purpose. It appeared that the couples felt like a team for varying reasons. The partners in Couple 4 and Couple 5 felt more united when they were able to hear their partner‘s perspectives and feelings: A lot of it has to go back to listening. Having that third party there, man, because if we didn‘t have that third party, we were always taking what each said personally and as an insult. Instead of having someone there saying maybe they are saying this, or look at it this way helps. We could not hear each other, and so, we didn‘t know what we were feeling. There‘s no other better way to say it as of right now, you know what I mean? (Adam, C4, Interview #1) On the other hand, Couple 2 appeared to feel united the most when Chad introduced homework and activities or facilitated productive conversations. This created ways for the couple to connect by giving them specific homework or tasks to do together and time to have important discussions: I think to me, I feel like we are more connected, more like a team, if it‘s a productive conversation, if I feel like it was engaging, and maybe, too, if I learned something. Where if I feel like I have learned something as a result of his analysis and synthesis, then I think we are working well together. (Anna, C2, Interview #2) The focus on tasks and engagement made sense for this couple, who were rebuilding their relationship after a two year separation, because they struggled more on the emotional level of the relationship. Collective Couple Alliance This theme refers to the therapeutic alliance therapists develop with the couple as a whole. The collective couple alliance was influenced by the aforementioned client-therapist and within-couple relational dynamics. Thus, it was hypothesized that the couples who had strong individual partner alliances and within-couple alliances would have a strong collective couple 158 alliance. Conversely, it was presumed that the couples who did not have strong individual partner alliance and within-couple alliances would have a weaker collective couple alliance. Analysis of the self-report, observational and interview data identified the overall level of collective couple alliance for the participants in the study. In order to assess the couple alliance, mean scores for the couples as reported by the clients and therapists, along with the mean scores for each partners were analyzed to infer the overall couple alliance for the participants in the study. Table 4.11 illustrates the ratings according to client, therapist and independent-observer. Table 4.11. Client SOFTA Self-report and Observational Mean Scores for Collective Couple Alliance Clients’ Mean Alliance Scores Observational Mean Scores Couple Mean Alliance Scores Couple Bob Susan M 63.60 Couple 2 1 60.25 1 n = 10 Brad 62.88 Couple 2 Anna 64.13 n=8 Mark 59.00 Couple 3 Heather 50.20 n=5 3 Adam Couple 60.44 4 3 Erin 72.89 n = 10 Joe 70.77 Couple 5 Christina 70.89 n=9 1 n = number of sessions 2 3 Therapist Mean Alliance Scores for Couple M SD 54.60 3.80 SD 5.17 6.49 M .84 .62 SD 1.40 1.94 M 62.11 SD 5.87 1.81 2.83 1.50 1.50 .53 .53 63.50 3.52 62.87 1.80 2.30 3.08 .75 .65 1.25 1.26 54.60 6.36 55.00 3.08 7.00 1.97 1.54 66.10 8.74 59.30 5.39 4.73 2.07 1.72 4.55 3.29 1.77 2.33 1.92 .86 70.83 3.85 65.89 7.34 Score is based on 8 sessions because she did not attend two sessions Score is based on 9 sessions because couple left session before completing paperwork The couples who reported the highest mean collective alliances were Couple 5 (m = 70.83; SD = 3.85) and Couple 4 (m = 66.10; SD = 3.85). The lowest mean collective couple 159 alliance was reported by Couple 3 (m = 54.60; SD = 6.36). The therapist ratings of their couples depict Couple 5 (m = 65.89, SD = 7.34) and Couple 2 (m = 62.87; SD = 1.80) as having the highest mean collective couple alliances. The couple with the lowest mean alliance rating as reported by the therapist is Couple 1 (m = 54.60; SD = 3.80). Next, the individual partners‘ mean SOFTA self-report and observational alliance scores were analyzed in order to evaluate how strongly aligned they were as a couple. It was hypothesized that the couples whose partners had the least variance in their individual ratings would have stronger collective couple alliances with their therapist. Analysis of the self-report ratings indicated that Couple 5 (difference of .12 point) and Couple 2 (difference of 1.25 point) had the least differences between their individual ratings. This supported the perspectives of the therapists that indicated these two couples had the strongest collective couple alliances. The couples with the greatest differences between their self-report ratings were Couple 4 (difference of 12.45 points) and Couple 3 (difference of 8.80 points). The observational scores identified Couple 5 as having the greatest difference between their alliance ratings (.56 of a point) followed by Couple 1 (.22 of a point). Couple 2 had identical mean ratings (m = 1.50, SD = .53) and Couples 3 and 4 both differed in their alliance ratings by .10 of a point. These results need to be carefully considered, however, due to the manner in which they were averaged. The SOFTA-o scores are not easily comparable to the self-report scores due to the fact that they do not produce an overall total score for alliance, but rather individual ratings for each dimension of alliance. Thus, four scores for each session per individual were inputted to calculate the mean. The results suggest that Couple 5 and Couple 2 had the strongest collective couple alliances since they had the least variance between their individual mean alliance ratings and the highest mean total client and therapist self-report ratings. Conversely, Couple 3 had the lowest 160 collective couple alliance due to their low self-report scores and the significant discrepancy between their individual scores. While Couple 4 had high mean self-report scores, their individual scores had the greatest variance of all couples, which suggests they did not have a strong collective couple alliance. Moreover, the variances between the mean scores for the individual partners suggest that Couples 1, 3 and 4 had a split alliance, which further supports that they did not have a strong sense of collective couple alliance. The data supports the hypothesis that Couple 5 and Couple 2 had strong collective alliances as suspected by their individual and within-system alliances with the therapist. Additionally, the postulation that Couple 1, Couple 3 and Couple 4 did not have a strong collective alliance based on the presence of a split alliance and negative couple dynamics was supported. This indicates that the individual alliances and within-system alliance contribute to the overall collective couple alliance. The dynamics that appear to influence strong collective alliance are the engagement level of the individual partners, how emotionally connected each partner is to the therapist, and the within-couple dynamics that enable a feelings of safety and strong sense of shared purpose. The elements that are harmful to the collective couple alliance are the present of a split alliance and conflictual couple dynamics that prohibit the couple from feeling safe and working together towards their goals. Repairing the Alliances Both therapists stated that it is essential for therapists to remain vigilant to signs that suggest the alliance is becoming unbalanced and in danger of causing a split alliance, or that a rupture has occurred. The therapists mentioned overt and covert signs that indicate they need to balance the alliance. 161 Alliance imbalances. One recognizable sign that indicates an unbalance is when they notice they are focusing on one partner more than the other. This tends to happen with the one who is more vocal in sessions, appears to be more engaged, or is comfortable expressing emotions: You just have to balance it and it is very easy to side especially if that one talks more, gives you much more to work with, is more open, more in touch with their emotions. It‘s very easy to get more aligned with one more person. I just kinda know it. I feel it. I just know if I‘ve been engaged with one person more than the other. That‘s how it looks. It looks like I am more engaged with one partner than the other. (Leah, T2, Interview #1) The clients who were more vocal in sessions tended to hold more emotional power in the relationship. This typically was the female partner in the current study. With the exception of Couple 2 and 3, these partners were the ones who initiated therapy and had higher engagement and safety scores throughout treatment. Another observable sign that there is a potential to unbalance the alliance is when there are individual client and couple dynamics. These dynamics can interfere with the ability to properly balance the alliance and create a strong collective couple alliance. A common occurrence is for one partner to be more aligned than the other due their power position in the couple. The therapists stated that it is more difficult to join with the partner who is more critical, attacking or defensive in their interactions with their partner. This dynamic was present for the female partners in Couple 1, Couple 3 and Couple 4. The therapists struggled to validate and join with the stronger partner due to their harsh interactions. Leah discussed how this dynamic was occurring for Couple 4: Cause she‘s more verbal. So, it is definitely easier to pull from her. I also empathize with him and so at times, I feel aligned with her, but at times I feel very protective of him because he is the one that is less wordy and less insightful. So, I feel like I have to help him more…so sometimes I may not always validate her as much because I am trying to emphasize with him so much. (Leah, T2, Interview #1) 162 The main overt sign that the therapists discussed that alerts them to attending to the alliance is their feelings and self-of-the-therapist issues. The therapists discussed how they were able to sense when the alliance was not balanced. They both stated that the therapist-client interactions feel intense, and they feel challenged. These feelings increased their anxiety level, and impacted their ability to maneuver with the couple. Leah described how this affected her in sessions: I feel like I screwed up. It is like I don‘t know what I‘m doing and I just feel lost. Yea, it‘s like I feel lost. Because at any moment, they can drop out. And, if you lose an alignment with your clients, then that‘s it. (Leah, T2, Interview #2) Therapeutic ruptures. While protecting the weaker partner can be a therapeutic intervention that serves to challenge the overall dynamics of the couple, this can cause a rupture with the stronger partner and increase the likelihood that the couple will end therapy. When this happened for the three couples, it had deleterious effects on the alliance for these couples. The three female partners in these couples all indicated that they experienced a rupture at some point in their treatment. The female partners in Couple 1 and Couple 3 openly discussed in the interviews about the times when they were frustrated with the therapist. They both ended treatment unexpectedly. Likewise, the female partner in Couple 4 abruptly walked out of the last session after displaying anger with the therapist for supporting her partner. The meaning for their rupture or termination of therapy was more related to the weak shared sense of purpose that they had with their partner. The ruptures occurred during times when the therapist was challenging the overall dynamic of the couple, or forcing the clients to discuss and face the couple issues. All of the ruptures occurred during moments of conflict, and did not appear to matter how strongly aligned the client was to the therapist. Leah discussed her 163 feelings of helplessness to prevent the ruptures that result from the couple relationship more than the client-therapist relationship: I don‘t think it has to do with you. Even if you process and metacommunicate about intensity before it occurs, the therapist can only do so much. Honestly. Some of this is them having to work through whatever it is and even if they want to deal with some of the things in the relationship (Leah, T2, Interview #2) Leah compared these couples with those who are able to stay engaged in the process with her when she raises the intensity. She believed the difference was in their commitment level to the relationship: Because if I challenge them, I can raise the intensity, and they may be just as scared, but they are not going anywhere. I think it would take a lot. Whereas with the other couples who do have one foot out the door, when you raise the intensity, that is an easy way out. (Leah, T2, Interview #2) While some ruptures are easily recognized, others are not. Clients discussed times when they felt a little disconnected from the therapist during treatment. The disconnection did not turn into a therapeutic rupture, but did affect how connected they were with the therapist. There were a variety of ways the clients felt connected, and some of the influences of these feelings were unrelated to the therapist. For instance, Joe discussed how there are times he feels disconnected during session when his partner is talking more than him: Maybe at times I feel like my partner is doing most of the talking. I am not mad, but maybe I feel a little less connected because she is the one talking and getting all the attention. But, she is more talkative. (Joe, C5, Interview #2) Joe was not a talkative person during sessions, which made this finding even more important due to the potential of Leah not recognizing it based on their personalities. Ruptures and feelings of disconnect can occur when the style of the therapist did not match the client, as in the case of Susan. She indicated that she often felt frustrated and helpless in sessions when she was unable to discuss what she wanted. Moreover, negative feelings may arise if the therapist focuses on an issue that the client does not feel is important. This occurred 164 twice in the study with Couple 3 and Couple 4. When Leah brought up a topic that Couple 4 stated was in the past, Erin inquired about what her reasons were to bring up the topic: The female brought up something that occurred in the previous session about something I had touched on that were past issues. From my perspective, It probably was quite uncomfortable whether they would admit it to me or not. She basically asked why are we talking about this if we are already past it, and asked me what I was going to do to help them if I am bringing it up. (Leah, T2, Interview #2) Leah appreciated that Erin was able to address this with her so that they can have an open dialogue about it. Clients can also feel disconnected due to behaviors that occur outside of sessions. Erin indicated that there were times they felt less connected when she went to tell Leah something after session and noticed a difference in how the therapist treated her after a session: Our time‘s up, so it was probably inappropriate for me to tell her, so it was brief. I think she did the right thing, but I felt a little disconnect there. While I feel so connected in the room, I felt disconnected in the hall. But then I realized what it was and it is ok. I took responsibility for it. (Erin, T4, Interview #2) This disconnection had a natural repairing in which the client was able to recognize the meaning behind the therapist‘s behavior. However, the clients discussed that when they feel a disconnection or rupture with the therapist, the feel frustrated, vulnerable, and their overall feelings of alliance decreases. Therapists reported that they used immediacy to prevent ruptures. They believed it was important to check in and assess what is happening when they see signs. In order to do this, they had to work in the moment: In order to prevent a rupture, you have to work in the moment. If their needs are not attended to, a person may feel silenced, or may feel like they‘re not getting what they want. Will they come back? Probably. Now, if it keeps happening, no, they aren‘t coming back. Because the message that you are sending is, I see something is going on, but I have my own agenda, and I‘m going to follow my own agenda. (Leah, T2, Interview #2) 165 Another way the therapists repaired ruptures was to give the clients space to change the direction of therapy for their benefit. Chad did this with Couple 1 when the sessions were becoming more and more focused on Susan challenging him than working on the couple issues. He refocused therapy on the goals and implemented a format would prevent the negative interactions they were having. Summary of the Multiple Systems of Alliance in Couples Therapy The results for the second research question identified the key components for each system of alliance in couples therapy: alliances with individual partners, within-couple alliance and collective couple alliance. Results suggested that engagement level and emotional connection were important components of the individual alliances with partners. When either alliance is not strong, the propensity for split alliances and imbalanced alliances was increased. Another notable finding regarding this level of alliance was the process of the synergy effect. This effect was common with the couples in the sample in which partners rated their individual alliances with the therapist based on how they perceived their partner‘s alliance was with the therapist. This had positive and negative effects on alliance depending on the level of distress the couple was experiencing. The second alliance system that was explored was the within-couple alliance. This alliance was the most powerful systems of alliance due to its influences on the individual and collective couple alliances. If the within-couple alliance was intact, the partners rated the alliance more favorably. If the within-couple alliance was not intact, the partners‘ alliances experienced negative patterns, such as split or imbalanced alliances. Finally, the collective couple alliance resulted from the culmination of the other two alliances. Thus, it was not possible to have a collective couple alliance without intact alliances on the other two levels. The section concluded with findings that identified how the therapists repaired imbalanced and 166 split alliances. The most important strategy involved directly addressing indications of disconnect or ruptures in the moment with the couples. Research Question 3: How does Therapeutic Alliance Evolve over Time in Couples Therapy? Data from the SOFTA self-report scales, observational measures, and interview transcripts were analyzed to identify how therapeutic alliance evolved over time for the couples in the study. Triangulation of the data revealed specific dynamics that influenced the development of alliance for the couples in this study. The results are presented according to each couple. The analysis provides an overview of the key components and behaviors that were influential of the alliance patterns. These key components and systems reflect the already identified elements presented in the first two research questions. Within-Case Analysis of Alliance Formation and Management This theme provides the significant findings of the study organized as alliance evolved over therapy for each couple. Session-by-session SOFTA self-report scores are presented for the couples over the course of therapy to provide a process oriented view of alliance as it developed with each member of the therapeutic system: therapist, male partner and female partner. SOFTA observational scores are not included in the figures due to the manner in which the scores are configured. The observational ratings simply provide a score for each subscale of the SOFTA and not a total score. Thus, the observational scores will be discussed when they are necessary to substantiate a finding. Couple 1 The alliance pattern for Couple 1 resembled the often challenging and fluctuating dynamics of the couple and therapist relationship. Figure 4.1 depicts the alliance scores for Couple 1. 167 1 Figure 4.1 . Couple 1: SOFTA-S Scores over Time Couple 1 SOFTA Self-Report Scores 80 70 60 50 40 30 20 10 0 Bob Susan Therapist S1 55 67 46 S2 61 62 53 S3 62 S4 68 56 S5 63 61 58 55 S6 57 46 51 S7 70 65 56 S8 70 59 59 S9 63 64 57 S 10 67 58 55 1 For interpretation of the references to color in this and all other figures, the reader is referred to the electronic version of this dissertation The formation and management of alliance for Couple 1 was impacted by several dynamics that evolved throughout treatment. These dynamics are discussed in this section and are organized according to the following themes: 1) Extratherapeutic Factors, 2) Client Factors, 3) Individual Alliance with Partners and 4) Within-Couple Alliance. Extratherapeutic factors. Extratherapeutic factors presented challenges to the therapeutic process throughout the course of therapy. The couple was experiencing a significant amount of stress related to health, work and family issues. Susan had health issues that interfered with her ability to work and limited some aspects of her functioning. She experienced chronic pain, which also created significant emotional stress. The work and family stress impeded their level of engagement through inconsistency of sessions, significant gaps in treatment, and overall 168 ability for the couple to focus and commit to therapy. Their schedules were busy and contributed to extensive breaks in therapy. For example, a gap of 7 weeks occurred between the fourth and fifth sessions. Chad felt that these issues were impeding therapy by slowing down progress: It‘s been difficult because time wise, by now, I would hope to be approaching termination and getting over their issues. But, therapy session wise, it is almost like the fourth session. So, it feels like we are just starting as opposed to where I think we should be ending. And it‘s because of all the cancelled sessions. (Chad, T1, Interview #1) Not only did these dynamics reduce the overall effectiveness of therapy, they prohibited the ability for the therapist and couple to properly develop a therapeutic alliance. Client factors. Both individual and couple dynamics significantly contributed to the alliance for this couple. These aspects were most influential of alliance during the early phase of treatment because they set the tone for therapy and impacted how the alliance was initially formed. Individual dynamics. The individual dynamics were mostly related to Susan. Her personality was a challenging factor throughout therapy. Her dominant personality tended to control the content of sessions and negatively react when Chad attempted to direct sessions. The couples‘ initial alliance ratings appeared to be related to this dynamic. Susan‘s rated the alliance high with a score 67, while Bob‘s score was 12 points lower. Susan was the initiator of therapy, and she controlled the session with complaints about Bob. It was apparent that she believed the reasons for their couple distress were mostly related to Bob‘s individual issues. Chad‘s attempts to engage Bob throughout session were often disabled by Susan‘s tendency to interrupt and talk over her partner. The alliance shift at Session 2 was due to Chad‘s strategy to gain control of the session by engaging the couple in an activity that explored family dynamics. The activity mandated that only one partner talk at a time. He had Bob participate first, and this took the entire session. While this session employed several emotional connection promoting behaviors, 169 and softened the couple‘s interaction by eliciting emotional responses from both partners, Susan‘s score decreased by 5 points. Thus, when she was not as engaged in session by partaking in the actual content, her alliance ratings decreased. Couple dynamics. The couple dynamics that contributed to the alliance formation related to their perspectives about the presenting problem and belief about how to resolve the couple issues. It was clear that the presenting problem was believed to be due to Bob‘s personal issues. Susan‘s description of the couple issues clearly indicated that she believed her partner was the presenting problem. This dynamic did not emerge in the self-report data for the shared sense of purpose dimension of alliance, however. Their initial ratings for this aspect of alliance were significantly high. However, the belief that Bob held most of the responsibility for the couple issues ultimately had deleterious effects on alliance outcome. As treatment unfolded, the couple dynamic and perspective of presenting problem affected all systems of alliance. Individual alliances with partners. Chad‘s individual alliances with the partners of this couple were challenged by Susan‘s perspective of the couple issues. Figure 1 depicts how Chad‘s alliance ratings over the course of therapy began to emulate Bob‘s ratings. This dynamic had negative consequences for the partners‘ individual and couple alliances by creating an imbalance in alliance with Chad and split alliance between the partners. Imbalance in alliance. It was clear in the observational and interview data that Chad had a stronger alliance with Bob. This imbalance emerged as a result of Susan‘s absence from the second and third sessions. This allowed Bob to have two individual meetings with Chad. While this one-on-one time positively affected Bob‘s alliance scores, it increasingly made it difficult for Chad to develop an individual alliance with Susan: I kind of have felt that he becomes easier to build the alliance with. The alliance is stronger because I am feeling what he is feeling in their relationship. I am kind of 170 cautious about that because I don‘t want that to imbalance the relationship. So, whenever he does that, I try to focus more on what she‘s feeling, and try to help her feel less sided against. (Chad, T2, Interview #1) Chad‘s attempts to balance the alliance by having individual sessions with her were thwarted due to her belief that Bob needed individual therapy to solve the couple issues. Thus, Bob‘s scores gradually grew over the first four sessions when he had the individual sessions with Chad. While his scores began to fluctuate later, his alliance was relatively stable and strong with Chad. Even after Susan rejoined therapy, the alliance did not become balanced. Thus, the alliance imbalance had long-term effects on the overall alliance for Couple 1. Split alliance. Susan‘s scores frequently shifted over time after Session 5. It appeared that her alliance was increasing and decreasing from session to session and depicted an unstable alliance. This dynamic was related to her relationship with Chad that increasingly became conflictual between Session 6 and Session 8. As mentioned previously, a split alliance was present during this point in therapy. The split alliance appeared to be due to the continuing imbalance of alliance that exacerbated Susan‘s individual dynamics with Chad. The content of Session 6 consisted of strong emotional reactions to Chad‘s efforts to soften her tone with Bob and connect with her through eliciting softer emotional reactions. He eventually split the couple during the session to prevent further harm to the relationship and to address Susan‘s need to be heard. Both partners rated the alliance lower during this session, but their scores increase again after Session 7, which indicated that Chad‘s efforts to provide safety and validation had a positive effect on alliance, even if it was brief. Within-couple alliance. When the couple sessions resumed at Session 5, the ratings became more unstable and begin to mirror each other‘s scores. This finding suggests that the couple dynamics were negatively affecting the alliance after the fifth session. This appeared to 171 be a reciprocal process throughout the later sessions. The alliance imbalance created instability in the overall couple alliance by initiating a split alliance between Bob and Susan. This provided an unsafe place for the couple to address their issues. As they addressed the issues, the alliance was challenged. As the alliance was challenged, the couple dynamics became more conflictual. Despite this, the couple maintained a sense of shared purpose in their commitment to working on the goals, which is what prevented them from ceasing treatment when the conflictual dynamics of the alliance were the most intense after the fifth session. The couple did not return for therapy after the 10th session when Chad gave them resources to use at home that would allow them to continue working on their relationship despite their busy schedules that were prohibiting consistent therapy sessions. Couple 2 Couple 2‘s alliance scores over time revealed a relatively stable pattern. It appeared that the couple was aligned with Chad, and that it had a gradual progression throughout the eight sessions. The scores for Couple 2 over the course of 8 sessions are illustrated in Figure 4.2. Figure 4.2. Couple 2: SOFTA-s Scores over Time SOFTA-S Scores Couple 2 80 70 60 50 40 30 20 10 0 Brad Anna Therapist S1 65 64 63 S2 63 59 66 S3 62 59 62 S4 61 61 64 172 S5 61 64 60 S6 63 65 64 S7 62 73 68 S8 66 68 68 There were various factors that influenced the alliance pattern of Couple 2. These factors related to 1) Extratherapeutic Factors, 2) Client Factors, 3) Value on Progress and 4) Synergy Effect. Extratherapeutic factors. Soon after the onset of treatment, the couple discovered that Brad needed to relocate out of state for his job and that they would be ending therapy prematurely. Not only did this place the couple under a significant amount of stress to prepare for the move, but it forced the therapeutic process to excel in order to accomplish as much as possible. This had important implications for the alliance development: I‘ve spent less time getting to know them. And, since they are cognitive, it is part of them, as well. There‘s less emotion exploration, I guess. So, yea, I mean I had a couple who came in the same time that they did, and we‘re just barely started on some of the stuff we started on them a couple weeks ago. (Chad, T2, Interview #1) The impending move placed pressure on the couple and Chad to initiate discussions that they were not ready to have. Since emotional discussions was difficult for them already, pushing them to do this because of the accelerated format of therapy reduced feelings of comfort and safety: It‘s tough, and I think if we had more time it would be better. Last week he was pressing a little bit, not a lot, but I think if we had more time maybe it would have been a more gradual thing. So, yea, I was a little more apprehensive because that is something you don‘t like to talk about or deal with. (Brad, C2, Interview #2) Despite the increased pace of therapy, the couple maintained alliance throughout treatment. They never cancelled, completed homework assignments between sessions, and appeared actively involved in the tasks and goals of therapy. Client factors. The alliance pattern over the course of the sessions does not show much fluctuation. The alliance for this couple appeared stable with no significant indications of strong emotional connection or weak alliance. This appeared to be due to the personalities of the couple. 173 Interpersonal dynamics. Their interpersonal dynamics were typically void of emotion. They struggled when Chad would attempt to elicit emotional reactions from them, as well as when they were asked to discuss the bond they had with him during the interviews. Thus, they did not display many observable signs of emotional connection towards Chad. Anna repeatedly stated that she did not expect to feel connected to Chad, or any therapist, as their role for her was to be a professional provider. Yet, it seemed that she did enjoy seeing more personal aspects of Chad in terms of personality and facial expressions. Brad also disclosed that he felt more reserved in sessions due to his discomfort sharing information with other men. This was related to his experiences in his family-of-origin. Thus, this couple did not display many observable signs of emotional connection towards Chad during sessions. Couple dynamics. The most prominent feature of their alliance pattern is the similarity between the three perspectives throughout treatment. The scores for all three are comparable in the actual rating of alliance, and have the same general pattern. Thus, the scores have a parallel progression in which they oscillate together over the course of therapy. This suggests that they had congruent perceptions of the alliance and were in tune with the dynamics of the relationship in order to sense the changes. The congruency between perspectives seemed to be enabled by the couple cohesiveness. The couple dynamics did not hinder Chad‘s ability to form balanced alliances with both partners and the couple as a whole. Thus, he was able to monitor the alliance and have accurate assessments of their alliances. Focus on progress. Two particular pivotal moments of alliance occurred that explain the alliance fluctuations during treatment. The couple‘s scores decreased after Session 1 and did not 174 rise again until Session 4. The decrease in scores during the initial sessions appeared to be due to the slower progress that is typical of the beginning phase of therapy: I remember my wife commented about it. She said, that session was just ok. And it may have been him just figuring out how to approach it, or whatever. I think the second session was beyond the introduction but still feeling out where were going, and talking about other things that were kind of obliquely related, but not as useful as other sessions. (Brad, C2, Interview #1) It became clear over time that the alliance for this couple was strongly related to the productivity and effectiveness of each session. The focus on tasks and goals contributed to their ratings of alliance more than the emotional connection they felt with Chad. Thus, their alliance improved when they perceived to make progress and learn new strategies. Interestingly, their scores began to rise again after the fourth session in which Chad revealed personal information about his experiences as a father. His self-disclosure appeared to positively influence their alliance. Synergy effect. Another pivotal moment occurred at Session 7. While both of their scores increased, Anna‘s rating jumped 8 points. This upsurge in scores was due to the ―synergy effect‖ that Anna discussed in her interview. Her alliance increased due to Chad‘s ability to elicit information and emotional responses from her partner that she was unable to do. This suggests that Brad‘s alliance with Chad contributed to her experience of the alliance. Couple 2‘s alliance with Chad was unique in that it was forced to immediately form and progress quickly due to an imposed deadline. This diminished the amount of time Chad had to connect and learn about the couple before he had to actively intervene to help them achieve their goals. At times, this did not seem to affect the couple‘s alliance, especially since they rated the alliance based on the progress and effectiveness of the sessions. However, there were times when the increased pace created feelings of discomfort by forcing the couple to address issues they were not ready to openly discuss. Ironically, when they were pushed, and emotions were 175 displayed, their alliance scores increased. Their alliance again increased when Chad facilitated discussions focused on Brad‘s emotional responses. The increase in scores could suggest that the most important emotional connection for this couple was the one they had with each other rather than the therapist. Couple 3 The alliance pattern for Couple 3 signified differing alliance scores for all perspectives and explains their premature drop-out of therapy. Figure 4.3 illustrates the alliance ratings over time. Figure 4.3. Couple 3: SOFTA-s Scores over Time SOFTA-S Scores Couple 3 80 70 60 50 40 30 20 10 0 Mark Heather Therapist S1 62 60 52 S2 60 49 53 S3 56 49 55 S4 56 49 60 S5 61 44 55 The overall alliance progression portrayed the conflictual dynamics of the couple. Mark‘s scores are generally the highest throughout treatment, while Heather‘s are the lowest (with the exception of the first session). The following elements influenced the poor alliance outcome for this couple: 1) Client Factors, 2) Individual Alliances with Therapist, 3) Within-Couple Alliance and 4) Incongruent Perspectives. 176 Client factors. There were two client factors that influenced the alliance outcome for this couple related to individual and couple dynamics. These factors played a role in how the alliance was formed and managed. Interpersonal skills. Heather had a disability that impaired some motor and speech functioning. While she had high cognitive functioning, her responses were often void of an emotional connection. Thus, her interpersonal skills were not as conducive to forming a personal connection with others. While she indicated that she liked Leah in the interview, her connection with Leah was not easily discernable in the sessions. Conversely, Mark was open and comfortable with sharing his feelings with others. He described himself as an ―engineered socialite‖ in which it was relatively easy for him to connect with others. The differences in personality and interpersonal skills contributed to the dynamic in which Mark had a higher individual alliance with Leah. Couple distress. What was observable in the sessions was the high degree of conflict between the partners. Thus, the greatest contribution to the poor alliance with this couple was the conflictual nature of the relationship. It appeared that Mark was much more committed to the relationship than Heather. This was evidenced by his motivation to initiate therapy and active involvement in the sessions. From the onset of treatment, Leah struggled to ignite a sense of shared purpose with the couple in order to set mutual goals for therapy. When she inquired about what they each wanted to achieve, Mark stated he wanted to make Heather happy, while Heather remarked that he wanted him to fix his depression. As such, any focus on couple issues created distress during sessions and affected the alliance. The combination of the relationship dynamics and Heather‘s belief that Leah was not focusing on what she believed to be the problem appeared to contribute to their treatment drop-out after the fifth session. 177 Individual alliances with therapist. The partners individual alliances with Leah differed significantly, as evidenced by the variation in their alliance scores. While the interpersonal and couple dynamics contributed to their individual alliances, there were also two other occurred to impact this dimension of alliance, as well as the overall therapeutic relationship: 1) Therapeutic Rupture and 2) Split Alliance. Therapeutic rupture. Heather‘s decrease of 11 points between Session 1 and Session 2 indicated that a rupture occurred. Analysis of the interview data revealed the reason for the rupture was due to Leah‘s focus on her role in the relationship distress. Heather discussed that she was becoming increasingly frustrated that her partner was not being truthful in sessions and that she feared that Leah was not getting accurate information and would not understand her side: In a lot of the sessions he‘s not being honest, and I think that screws everything up. He makes himself out to be better than me, and I think this influences how she sees me. It‘s hard with the therapist because she doesn‘t realize it. I mean, I tell her, but, I don‘t know. it mainly affects my relationship with him. (Heather, C3, Interview #1) The rupture and frustration about how the sessions were moving appeared to be mostly related to her relationship with Mark since she repeatedly stated that she still trusted Leah. While she stated that Mark‘s dishonesty mainly affected her relationship with him, it appeared to continually weaken her alliance since her ratings never improved over the course of treatment. Split alliance. This couple had a split alliance throughout treatment. While a rupture occurred in Session 2, the couple issues contributed more to the split alliance than the rupture. Due to the amount of distress in the relationship, Mark began to depend on Leah for support. Since he was the partner who was more invested in the relationship, his alliance was based on how the relationship was changing: 178 She‘s sorta tied to our relationship. I mean, emotionally to me. My behavior with her, is unfortunately, related to my relationship with my partner. Today, I felt like I wanted to use her as a crutch in a way because I felt disconnected from my partner. She was good at putting me back on track and redirecting. When I feel less connected with my partner, I rely more on the therapist. (Mark, C3, Interview #1) Leah was aware of this dynamic and discussed how she struggled to build a stronger alliance with Heather, but was unable to due this within the couple work because Heather was not as committed to the relationship: I guess because their interactions are so harsh and so mean, and they interact from places of hurt, so, it‘s very difficult to attend to the couple issues together. So, I met with them individually more out of hope. I wanted to address their commitment to the couple relationship and to therapy. (Leah, T2, Interview #1) Heather did not identify with the relationship and did not have the commitment or motivation to work on the issues. Thus, she would disconnect when they discussed couple issues. Synergy effect. This couple experienced a synergy effect that was deleterious to the alliance. She disconnected when Leah focused on Mark‘s experiences. When she perceived Mark‘s alliance to improve, or that there was more focus on his experience, her alliance decreased. This happened during the fifth session when Leah processed the recent loss of Mark‘s father. Heather did not engage in the conversation, and after the session, she became angry while completing the paperwork. She indicated that she was angry because she did not think the session was useful and that they were not making any changes. Shortly after the session, she called Leah and firmly told her they would not be returning to therapy. Within-couple alliance. The couple did not have an intact alliance between each other due to their differing perspectives on the problem for the couple distress that influenced the various dynamics that occurred during their treatment. Leah was unable to promote this aspect of alliance due to the challenging individual and couple dynamics. This system of alliance impacted 179 the other two in which the couple had split individual alliances and poor collective alliance. The instability on all levels of alliance ultimately explains their dropout after the fifth session. Incongruent perspectives. Leah‘s scores did not shift in relation to the couple. Her scores generally increased over time until the last session, while the couple‘s scores generally decreased over time. This suggests that there was a disconnect between the therapist and couple‘s experiences of alliance. Despite Leah‘s awareness of Heather‘s disengagement in therapy, the self-report data suggested that she was not aware of how much Heather‘s alliance was decreasing. It seemed that Leah related the changes in Heather‘s alliance more to her disconnect from the relationship than to the therapeutic alliance with her. The combination of all these dynamics explain the negative alliance pattern for Couple 3. The couple presented with a weak within-couple alliance that impacted all aspects of alliance. Their interpersonal and couple dynamics contributed to a split alliance that was evident throughout most of treatment. Heather‘s perspective of the presenting problem challenged Leah‘s ability to promote the within-couple alliance, and any attempt to do so would cause Heather to further disconnect from therapy and Leah, which ultimately influenced her to end therapy. Couple 4 The scores for Couple 4 depicted a parallel progression of alliance for all members. Thus, the alliance for all participants varied together throughout treatment. Figure 4.4 illustrates this pattern. 180 Figure 4.4. Couple 4: SOFTA-s Scores over Time 80 70 60 50 40 30 20 10 0 Adam Erin Therapist S1 43 68 52 S2 61 65 60 S3 62 74 57 S4 63 75 62 S5 64 74 60 S6 59 68 51 S7 60 76 56 S8 65 78 62 S9 67 78 68 S10 56 The alliance pattern revealed that Erin‘s initial alliance rating was significantly higher than Adam‘s score after the first session. This variation decreases during Session 2, when Adam‘s scores increased by 18 points. However, the gap in their scores remerged as treatment progressed from the third session until the end of treatment. Their alliance pattern can be explained by 1) Congruent perspectives, 2) Within-couple alliance and 3) Individual alliances with therapist. Congruent perspectives. Leah‘s ratings changed in relation to the couple ratings which indicates that she was properly monitoring the alliance. However, her scores were generally lower than both partner scores, which suggests that she did not believe their alliance was as strong as they were reporting. This indicates that Leah may have been rating the alliance based on her sense that they were not on the same level in terms of their connection with one another, commitment and overall relationship goals. These relationship dynamics were the defining 181 influence of alliance for this couple as they suggested there was a split alliance between the partners. This appeared to be the contributing factor for their divergent scores. Within-couple alliance. This couple was not necessarily highly conflictual during sessions. However, they did not have a strong within-couple alliance due to lack of trust and doubts about the relationship. There were signs of hope throughout treatment, and moments when the couple appeared to be improving, but the doubt about the relationship was a strong influence of their alliance and treatment outcome. As sessions progresses, it seemed that Erin was having increased doubts about the relationship which influenced her high individual ratings. The more she disconnected from the relationship, the more she aligned with Leah for support. Individual alliances with therapist. The individual alliances with the therapist did not initially present a threat to the overall alliance for this couple. While it appeared that Erin was more comfortable in sessions, Adam seemed to have a strong alliance with Leah. The self-report scores, however, indicated there was a difference between how connected they actually were to Leah. Split alliance. Erin‘s scores were significantly higher than Adam‘s ratings for all sessions. This suggests that the couple had a split alliance from the onset of treatment that persisted over the course of therapy. This split alliance appeared to be influenced by the lack of trust in the relationship. As such, Erin‘s alliance and trust with Leah was established instantly as a way for her to get her relationship needs met. Her alliance ratings would increase when Leah would challenge or elicit an emotional response from her partner: Somehow she took care of my partner, so I was allowed to just take care of me. And, in our relationship when we argue, I can‘t do that. I have to take care of him in a fight, and the moment I take care of me, he‘s out of the door. So, that was helpful, I think that was a definite reason I felt more alliance with her because she challenged him so I didn‘t have to. (Erin, C4, Interview #2) 182 Thus, Erin increasingly relied on Leah to change the relationship instead of using Leah as a guide to help the couple work together towards the goals. When Leah gave meaningful feedback, or identified important dynamics through in-session activities, Erin‘s alliance would increase. Adam‘s alliance increased as well, which also suggested he began to depend on Leah as the couple relationship continuously disconnected. While they both portrayed strong alliance behaviors in session, and reported feeling connected to Leah during interviews, their alliance with each other was decreasing. Towards the end of treatment, Erin began disclosing her regrets and concerns about the relationship. She even disclosed in the second interview that she felt safe to do this because she felt as if Leah was pushing her to do so when they had an individual session: I just think we have this other relationship that happens that she is having with me and not [partner]. I‘m not sure if she thinks that our relationship is something we should pursue. I think she sees something else. I think she is giving me kinda clues. Not that she‘s saying ―end the relationship,‖ but what if? She touches on all of that, so it could be just me, but she gives me the tools to figure it out. I think she just gives me the space to explore it. Educating me, I guess. (Erin, C4, Interview #2) What is important to note about the split alliance is that it seemed to be more related to Erin‘s doubts about the relationship, which caused her to strongly align with Leah when she perceived that she was giving her insight into this process. Thus, the split alliance was not due to Adam having a weaker alliance with Leah as it is believed that his ratings represent a strong alliance with Leah. Synergy effect. There was a positive and negative synergy effect for this couple. Erin‘s scores increased during times when Leah was able to elicit emotional responses from Adam during sessions. Her alliance also improved when she observed Leah to attend to Adam during intense discussions because it allowed her to focus on her own feelings: 183 When we had an argument 2 weeks ago during session, she took care of my partner, so I was allowed to just take care of me. And, in our relationship when we argue, I can‘t do that. I have to take care of him in a fight, and the moment I take care of me, he‘s out of the door. So, that was helpful, I think that was a definite reason I felt more alliance with her because it was safe. (Erin, C4, Interview #2) The strong alliance she had with Leah also contributed to the rupture that occurred in Session 10 when she abruptly left the session because she felt that Leah was supporting Adam‘s experience at the expense of hers. Thus, this case illustrated that having a strong alliance does not prevent ruptures or treatment drop-outs. Rather, it may cause a therapeutic triangle that weakens the within-couple alliance. Despite the couples‘ high scores and ratings of alliance, they did not return to therapy. The alliance pattern for this couple signified the importance of assessing alliance even when it appears to be strong. Leah‘s lower scores indicated that she sensed that the couple did not have as strong of an alliance due to the couple dynamics. However, the increased individual alliances indicated that the partners were leaning on Leah for support, which had negative consequences for their alliance. Couple 5 Alliance ratings for Couple 5 demonstrated that they had the highest levels of alliance in the study. Figure 4.5 illustrates their high ratings of alliance. 184 Figure 4.5. Couple 5: SOFTA-s Scores over Time SOFTA-S Scores Couple 5 80 70 60 50 40 30 20 10 0 Joe Therapist S1 70 52 S2 66 58 S3 69 65 S4 63 62 S5 74 68 S6 69 72 S7 74 70 S8 76 73 S9 76 73 Leah‘s scores were again lower than the participants except for Session 8 when she rated the alliance in between Christina and Joe. The couple‘s scores suggested that the alliance was shifting frequently, but a decrease in scores did not generally last longer than one session. All ratings achieved greater stability between Session 7 and Session 9. Overall, their alliance appeared to be strong with fluctuating patterns that continued to increase over time. The notable components for this couple were: 1) Strong Collective Couple Alliance and 2) Positive withinsystem couple system. Strong collective couple alliance. The couple reported high alliance scores from the start of treatment, which suggests that they both felt strongly connected to Leah during the first session. Interestingly, their scores decreased between Session 1 and Session 2 and depict a more fluctuating pattern until the fifth session. While Christina‘s scores became more stable after Session 5, Joe‘s scores fluctuated throughout treatment, with the most notable change occurring between the fourth and fifth sessions when his score increased by 11 points. This significant increase occurred when Leah focused on his negative involvement in the court system regarding 185 his children. Thus, the upsurge in ratings appeared to relate to the emotional connection he felt with Leah. Overall, they both appeared to feel equally connected to Leah. Intact within-couple system. The defining characteristic of the alliance for Couple 5 was their high levels of cohesiveness and shared sense of purpose. They were not in therapy because of significant couple distress, but rather to enhance their relationship and learn skills to resolve typical couple issues they periodically struggled with. Thus, they were focused on one another, rather than Leah, to work on their goals. They both felt equally connected to her, but not at the expense of their relationship with each other. They did have some minor conflict arise throughout treatment, but were responsive to Leah‘s suggestions and were able to recover from these negative interactions. Summary of Couple Alliance Ratings over Time The preceding sections presented session-by-session scores to explain how alliance evolved over the course of the first four sessions. Alliance patterns indicated that most clients (n = 8; 80%) had high initial scores of alliance followed by a sharp decrease in scores at Session 2 that eventually stabilized over the sessions. Conversely, the therapists rated the alliance lower at the onset, and while their scores persistently increased, they consistently rated the alliance lower than the couples. The results for the couple ratings over time suggested that there were important elements of alliance that emerged during particular phases of treatment related to the couple dynamics. It appeared that the degree of cohesiveness between the partners contributed to the degree of congruency between the perspectives of alliance. When a couple had a strong within-couple system, the couple and therapist tended to have similar alliance ratings that were higher and more stable over time. Additionally, conflictual dynamics within the couple, or between the therapist 186 and one particular partner, challenged the ability to form and maintain stable alliances. Results suggested that harmful aspects of alliance can occur at any point during treatment, which tended to contribute to alliance instability. In this study, positive and negative alliance fluctuations occurred during the initial sessions of therapy (Sessions 1-4), as well as during later sessions, regardless of the strength of the alliance. 187 CHAPTER V DISCUSSION The purpose of the current study was to explore therapeutic alliance in couples therapy. The objectives of the research were to: (1) Identify therapeutic components that contribute to alliances with couples, (2) Understand how alliance is managed between the multiple systems in conjoint treatment, and (3) Explore how therapeutic alliance evolves over time in couples therapy. The ultimate goal was to further develop, improve, and disseminate effective ways of building strong alliances with couples. The predominant finding of this study revealed that therapeutic alliances in couples therapy encompass an array of key components that are present in each system of alliance. While each system of alliance is influenced by the others, the systems that emerged as being most influential to alliance patterns for the couples in this study were the individual partner alliance and the within-couple alliance and their interaction with each other. These alliance systems generate the collective couple alliance and had powerful effects on the overall alliance formation and maintenance process. The remainder of the chapter recapitulates and interprets this study‘s findings as they relate to the previous empirical research. The chapter is organized according to the research questions. Following the discussion, thoughts regarding the clinical implications of the research, strengths and limitations of the study, and suggestions for future research are provided. The chapter concludes with a summary of the current research study. Research Question 1: Key Therapeutic Components of Therapeutic Alliance in Couples Therapy The first research question addressed the key components that influence alliance in couples therapy. In this study, the components that impacted the formation of alliance consisted of therapeutic factors, client factors, and therapist factors. These factors represented the 188 significant elements, qualities, and dynamics that influenced how alliances formed, progressed and fluctuated throughout the therapeutic process. Thus, they were instrumental to understanding the nature of the therapeutic alliances with the couples in this study. Therapeutic Factor: Nature of the Training Clinic The context of where the therapy sessions occurred appeared to have mixed influences on the alliance for the participants in this study. Clients were aware of the training nature of the clinic in which all treating therapists were graduate students with varying levels of experiences. As such, the clients participating in services at this clinic were informed, and consented to, the recording of all sessions, having their cases be supervised by licensed supervisors, and forgoing some aspects of confidentiality for purposes of training and supervision. Regardless of their consent and knowledge of these elements, these aspects influenced how some clients felt and behaved in sessions. There were two ways the training nature of the clinic impacted alliance. First, some clients revealed that they were more guarded in sessions despite feeling comfortable with the therapist. This has significant implications for therapeutic alliance outcomes. Feelings of safety are foundational to the client-therapist relational dimensions of engagement and emotional connection, as well as the within-couple relational dimension of shared sense of purpose (Friedlander et al., 2006a). In this study, the effects of the context appeared to impact the alliance by influencing increased feelings of reservation and caution. A second manner in which the therapeutic context challenged the alliance with the couples in the study regarded how they perceived their therapist. Some clients indicated that the student status of the therapist invoked feelings of uncertainty about the level of skill and ability to effectively help the couple achieve their goals. This doubt appeared to only affect the initial 189 alliance formation for a few of the clients. The self-report scores and clients indicated that their trust improved over time as the therapist demonstrated their knowledge and skills. For the majority of the participants, the student status did not create doubt in the therapist‘s ability to treat them. With some clients, there were specific points of connection such as also being a student, that helped to create a stronger alliance. This finding supports existing literature that posits clients and therapists have higher ratings of alliances when they have similar life experiences and characteristics (Ackerman & Hilsenroth, 2003; Crits-Christoph et al., 2006). In conclusion, the effects of the decreased confidentiality and training status of the therapist had more short-term effects on the initial formation of the alliance. These effects disappeared as the therapist demonstrated their competency and as the process of therapy evolved. This demonstrates the importance of therapists considering how therapeutic elements that limit the anonymity and confidentiality of clients, such as insurance billing, can impact alliance over the course of treatment, as well as the importance of establishing confidence in therapist skills and abilities as early as possible in treatment. Client Factors Consideration of the role in which client factors play in the therapeutic alliance is pivotal to the interpretation of any and all research related to therapeutic processes. Research on therapeutic outcome has demonstrated that the person of the client and the factors in their life are more influential of outcome than any other element (Bohart & Tallman, 2010). In this study, any characteristics or dynamics that were specific to the individual partners or couple as a whole were influential. Client expectations of therapy/therapist. Research indicates that one of the most prominent predictors of treatment dropout relates to the client expectations about therapy 190 (Bohart& Tallman, 2010). Clients will prematurely end treatment if they do not believe that therapy is meeting their expectations (Callahan, Aubuchon-Endsley, Borja, & Swift, 2009). Likewise, the degree to which therapists fulfill the clients‘ expectations to be listened to and validated, has been shown to have significant positive effects on therapeutic alliance (Bohart& Tallman, 2010). The most important expectation that couples in this study had for their therapist and general therapeutic process revolved around feelings of safety. Clients expected that therapy would be a safe place where the therapist would provide a neutral perspective and manage the intense interactions between the partners. This expectation was especially important for the clients due to the nature of couples therapy. Interestingly, the men in the study articulated this need for neutrality and safety more than their female partners. Women in the study expressed their expectation that the issues would be resolved. These findings address a dynamic that was evident throughout the study related to the couples. Across cases, the female partners seemed to be the partner with the most ―therapeutic‖ power in the relationship. They initiated treatment and were also the initiator of drop-out (when it occurred). Even in the case in which the male partner initiated therapy, the female‘s perception of a rupture ceased treatment. These dynamics can be explained by the literature of men and women in therapy that postulate men typically have power over tangible resources while women hold more emotional power and control over the intimate aspects of the relationship (Symonds & Horvath, 2004). According to that perspective, women would have more control over therapeutic dynamics, including initiating and ending treatment. This also explains why the men desired a safe context to discuss the issues in which they perceived their partner to hold all of the power 191 It was also observed that female partners had specific expectations of the therapist and therapy. They looked to the therapist to ―change‖ their partner. When they heard the therapist validate the partner‘s response instead of condemning or reprimanding them, their alliance ratings would decrease. There were two instances when this type of interaction caused a rupture, and ultimately led to premature drop-out. These findings emphasize the need to directly, and repeatedly, discuss the nature of couples therapy and the role of the therapist in that context. Additionally, therapists need to address the expectations of the couple in order to assess the motives and goals for each partner. These types of interactions are inherent to relational therapy, and any misguided expectation about the role of the therapist can influence negative alliance patterns. Relationship distress. It is common, if not expected, that couples will present in therapy with one or both partners experiencing distress. Research indicates that the pre-existing relationship of the couple has an influence on the therapeutic alliance (Symonds & Horvath, 2004). Explicitly, incongruity in both relationship satisfaction and distress challenge the manner in which therapists build alliances with the couple. Couples who are more distressed will necessitate more effort and time in forming a therapeutic alliance. Conversely, couples who are less distressed at the onset of treatment may develop a therapeutic alliance more easily (Glebova et al., 2011). Results of this study confirm these findings. The two couples who presented with similar, lower levels of distress had stronger, more stable alliances over time. There were fluctuations at times, however there was congruency between the therapist, male partner, and female partner ratings. For the remaining couples, there were significant ways the conflict affected couple alliances in the study. These findings point to the difficulty in appropriately assessing and monitoring the multiple alliances with couples who 192 are conflictual. In the study, the therapists struggled with gauging the alliance levels when partners were conflictual. They tended to rate it similarly to who they perceived to be the most distressed partner. At first, this would seem to make sense because it would imply that they were viewing the alliance as low if a partner was distressed. However, the most distressed partner typically had the highest level of alliance because they were using the therapeutic relationship as support. This finding is supported by research that found that the more troubled the relationship, the more quickly the most distressed partner will align with the therapist (Knerr, 2010). Furthermore, when both couples feel desperate, as was the case with Couple 4, it would appear that the alliances were strong within all systems. However, the stronger the individual alliances became with the therapist, the weaker the within-couple alliance became. The current study supported previous research that demonstrates that building strong and healthy alliances with conflictual couples is a perpetual challenge for therapists due to the varying ways conflict can impact the alliance patterns. This process may be isomorphic to the couple process in which couples who experience high levels of distress form a relationship with the therapist that has the same level of distrust and conflict as they have with each other. As such, the therapist would be aware of the struggle to form alliances yet feel helpless to do so. On the other hand, if the distress is not as obvious, the alliances may form quickly and strongly, and obscure the instability that this bond is creating for the couple alliance. Thus, it seems important that therapists immediately and continuously address the perspectives and levels of couple distress in order to keep the alliance strong. Interpersonal characteristics and past experiences. All clients have individual differences in terms of how they promote and experience interpersonal interactions. Research posits that the client interpersonal relations are more predictive of the therapeutic alliance than 193 even psychological distress (Saunders, 2001). Thus, this client factor is important to consider when understanding the key components of the alliance with couples. For the couples in the current study, the interpersonal characteristics of the clients contributed to the strength and nature of the alliance they formed with the therapists. There was a wide range of interpersonal qualities that were evident in the individual clients. The couples who had the more positive alliances consisted of partners who were talkative, outgoing, and trusting. They engaged more in sessions, had higher levels of safety and a general openness to therapy and the therapist. They were able to connect with their emotions and articulate their feelings, which increased the feeling of a connection with the therapist. The interpersonal quality that challenged the formation of alliances most clearly was a dominant personality that tended to control the overall goals and structure of therapy, as well as the focus and tasks of each session. Importantly, the presence of deficient interpersonal skills or dominant personality does not immediately indicate they will have weaker alliances. Saunders (2001) found that clients who experienced interpersonal difficulties such as being close and emotionally open with others were more likely to have lower ratings on the bond component of the alliance. Thus, while they appear to be less emotionally connected, they could still have strong alliances related to the goals and tasks aspects of alliance. Client gender. Results from the current study indicated that the gender of the client plays a role in alliance with heterosexual couples. Analysis of total alliance scores for all clients in the study showed that women had slightly higher mean ratings of alliance than men. Additionally, for all but one couple in the sample, therapy was initiated by the female partner. This finding can be explained by existing literature that posits women are more likely than men to initiate therapy, and are more comfortable talking to others about their problems (Berger, 1979; Rait, 2000). 194 Furthermore, research by Garfield (2004) suggests that there may be a disadvantage and power differential for men due to the nature of therapy that typically focuses on emotions and relationships. Since men do not typically excel in those types of discussions (Garfield, 2004), their alliance may appear lower than their female partners. This does not necessarily mean that their alliances are lower and poses the question of how alliance is assessed. Studies of alliance have concluded that clinicians and researchers tend to gauge the level of alliance by how strong they appear on the bonds dimension (Friedlander et al, 2006a; Knobloch-Fedders et al, 2007). As discussed previously, some clients may feel more connected to their therapist based on the engagement level of the alliance in which they are actively involved the therapeutic process. The results of this study showed that the individual scores for some men depicted strong alliances. As such, the men in this study did not necessarily conform to the gender stereotype present in alliance research that posits they have lower alliances. Moreover, even the specific self-report ratings can be subjective. A score of 70 out of 80 can signify strong alliances to one client, just as a score of 40 out of 80 can indicate a strong alliance for another. Ratings can be individually based and related to that client‘s interpersonal characteristics. Thus, comparing scores may not always be the best way to gauge the strength of alliances and an idiographic approach may be more explanatory than a nomothetic method. Empirical research has consistently reported that there are differences in how alliance ratings for men and women influence treatment outcomes. Multiple studies have concluded that men‘s higher alliance scores are a stronger predictor of positive therapy outcomes than higher ratings of women (Anker, Owen, Duncan & Sparks, 2010; Knobloch-Fedders et al., 2007; Symonds & Horvath, 2004). These findings suggest that resolution of relationship distress depends on how strong the male partner is aligned with the therapist. Since the current study did 195 not assess treatment outcome, results cannot be interpreted to further this conclusion. However, this study did find that female partners in the most conflictual couples had a strong influence on couple retention. For the couples in this sample, the female partner‘s lower alliance scores and/or experience of a rupture influenced treatment drop-out. Thus, when the female partners had lower overall alliance scores or experienced a rupture, the couple dropped out of treatment. The findings of the current study suggest that for the couples in this particular sample, lower female alliance ratings were a significant predictor of poor alliance outcomes. This result was surprising due to the literature that supports the notion that men have greater influence on therapeutic alliance and treatment outcome in couples therapy (Bourgeois et al., 1990; KnoblochFedders et al., 2004; Mamodhoussen, et al., 2005). The results of the current study coincide with the research of Quinn and colleagues (1997) who found that the outcome of therapy is more positive when the women‘s alliance is higher than the men‘s alliance. Additionally, the current finding can be explained by research that found evidence that suggest men‘s alliance ratings may be more predictive of outcomes in short-term therapy, but women‘s scores emerge as the more critical predictor of alliance in longer termed therapy (Anker et al., 2010). Another finding of the current study related to gender of clients revealed that for some couples, the female partners‘ ratings of the alliance were based on the relationship they had with the therapist, as well as how they perceived their partner‘s relationship was with the therapist. As such, if they perceived their partner to have a positive relationship, they rated their alliance strong. This was true in the opposite sense as well whereas when women perceived their partner to have a stronger alliance with the therapist than they believed to have with the therapist, they rated the alliance lower, experienced ruptures and initiated treatment dropout. The degree of conflict and expectations and goals of therapy contributed to this process for the couples in this 196 study. When the level of distress was lower and the couple had the same expectations and goals for therapy, women‘s alliance ratings increased when they perceive their partner to have a positive relationship with the therapist. When the conflict was higher and the couple differed in their expectations and goals for therapy, they rated the alliance lower if they believe their partner had a stronger alliance with the therapist. Lastly, the current study confirms prior research by Gellhaus Thomas et al., (2005) by discovering that there were differential effects of therapist and client alliance behaviors for male and female partners. Men‘s alliance ratings appeared to fluctuate according to the presence of negative and attacking behavior on the part of their partner. They also increased slightly when they perceived the therapist to address the partner and interrupt their behavior. However, this increase was very slight, as the men in the study often rated the alliance lower when they perceived their partner to have a negative interaction with the therapist. This finding supports research by Gellhaus Thomas et al. (2005) that found that a consistent harmful predictor of therapeutic alliance for men was negative statements made by their partner. These findings support Friedlander‘s et al., (2006b) finding that an important aspect of CFT alliance is the development of safety within the therapeutic context in order to prohibit couples from displaying attacking and/or critical behavior that may damage alliances. Therapist Factors Research has demonstrated the importance of including therapist variables into their consideration of the factors that influence therapeutic processes. Common factors research implicates therapists as one of the contributing components to effective (and ineffective) therapeutic outcomes (Blow et al., 2007). Indeed, therapists are important variables in the therapeutic relationship. 197 Intrapersonal dimension. It has generally been concluded that that the client – therapist match is an important predictor of successful therapeutic relationships (Berzins, 1977; Beutler, 1991; Reis & Brown, 1999). Demographic similarity, such as ethnicity and gender have specifically been shown to contribute to successful treatment (Berzins, 1977; Beutler, Clarkin, Crago, & Bergan, 1991; Flaskerud, 1990; Nelson &Neufeldt, 1996). Additionally, research has demonstrated that clients prefer therapists who are warm, nurturing, and affirming (Bischoff & McBride, 1996; Horvath &Bedi, 2002; Wampold, 2010). This aspect of therapists has been credited for retaining clients for longer periods of time, and regarded as the dimension that clients respond to immediately and initially (Horvath & Greenberg, 1994). Generally, clients have better alliance ratings when they regard their therapists as experienced, trustworthy, and comforting with their therapist (Horvath &Bedi, 2002). The current study also found that clients appreciated specific mannerisms and aspects of the therapists‘ personalities that represented the overall interpersonal style of the therapist. The most significant aspect that clients appreciated was the ability of the therapist to show parts of their personality. The clients wanted to see the features of the therapist that were more human and relatable. They stated that seeing the therapist‘s ―real‖ personality offset the traditional doctor-patient‖ dynamic that was often intimidating to clients. This quality helped the clients to feel comfortable and adjust to the uncomfortable elements of the therapeutic process. These qualities appeared to be responsible for creating the initial connection with clients in the current study. As such, they appeared to be most influential of the initial bond the clients felt with their therapist. While the therapists‘ interpersonal characteristics persisted over the course of treatment, they were not found to be as influential of alliance later in therapy. Thus, it seems that there is an important window of opportunity for therapists to demonstrate their 198 warmth and personality. Once it is perceived by the clients, it remains as a stable characteristic over the course of therapy unless, of course, there is a therapeutic mistake or rupture that is not resolved. Some research is related to this finding. Knobloch-Fedders et al. (2007) found that alliance patterns form in the initial session and remain consistent throughout therapy. Accordingly, it seems that first impressions matter in therapy. For clinicians, this means that the initial interactions including the first phone call and session, are essential in creating a therapeutic alliance with clients. There were conflicting findings related to the importance of therapist-client matching in terms of race and ethnicity. Research in alliance has not identified a significant link between ethnic similarity and the therapeutic relationship. What is suggested is that while clients may prefer a therapist of similar race/ethnicity, it does not mean they have a more positive therapeutic relationship or outcome (Ricker, Nystul & Waldo, 1999). The results from the current study support this idea and confirmed previous evidence in the clinical literature that claims clients put more weight on similar attitudes, values, and personality than on ethnicity (Atkinson, Furlong, Poston, 1986). The most significant finding in terms of clinical implications regarding this aspect is that clients appreciated it when the therapist openly addressed this dynamic in sessions with them. Therefore, it seems important for therapists to attend to, and even directly address, how race influences the way clients feel in the room. As one participant stated, it is ―there‖ and the therapist‘s ability to discuss these matters could increase the comfort level for clients. In terms of therapist gender, findings of the current study provided some support for the previous research that posits both male and female clients prefer women therapists (Dailey, 2004; Johnson, 2005; Scher, 2005). In this study, three men and one woman stated that they believed they would feel more comfortable with a female therapist. This finding appeared to be 199 influenced by other dynamics. For instance, the discomfort one of the male clients was feeling with his male therapist appeared to be related to the lack of male presence in his family-oforigin. He discussed in the interview how he was more comfortable in relationships with women than men because he was raised by his mother. Additionally, there was a difference in the amount of warmth expressed by the two therapists. The female therapist displayed more warmth in sessions than the male therapist, which could explain these results. Finally, the clients‘ preferences for a female therapist could be related to gendered stereotypes that posit women are more nurturing and accepted in the role as a therapist, as well as the common belief that it is not acceptable for men to display vulnerability with other men. Regardless of why the clients believed that they would feel more comfortable with a female therapist, the quantitative alliance scores did not show that their alliance was impacted by the gender of their therapist. Thus, it seems important for therapists to remain cognizant of the impact of past experiences, gendered beliefs and stereotypes on clients and directly address these matters at treatment onset. Results indicated that the demographics of the therapists contributed to the overall aspects of the ―self‖ of the therapist. They are undeniable characteristics of the therapist, and did play a role in how the couples formed an alliance with their therapists. The most significant demographic was gender due to its contribution to an inevitable imbalance in therapy that initially hindered feelings of safety for a small number of male clients in the study. Over time, the gender imbalance did not have a significant impact on alliance ratings. This suggests that the gender imbalance is a significant contributor to initial alliance ratings. However, if therapists openly addressed this dynamic, and increased feelings of safety by ensuring that they equally validated and elicited both partners‘ experiences, clients no longer felt an imbalance. The 200 transparency of the therapist helped to establish a dialogue that increased feelings of safety and confidence in the therapist. Interactive Factors The interactive factors relate to the idea of therapist-client complementarity and collaboration (Horvath &Bedi, 2002). Complementarity refers to the idea that harmonious interactions are evident in positive relationships, rather than negativity or hostility. Higher alliance ratings are associated with friendly and autonomy-enhancing relationships rather than competitive, hostile, or controlling interactions (Henry &Strupp, 1994). In this study, participants identified their relationship with the therapist and therapeutic skills as important interactive factors in their alliances. Goodness-of-fit: Style of therapist. The interactive factor of how well a therapist fits with their client has been well documented in the clinical literature (e.g., Duncan & Moynihan 1994; Herman, 1988; Hubble, Duncan & Miller, 1999). The goodness-of-fit is important in terms of theory, personality and client expectations (Hubble et al., 1999). Most research on this interactional relationship is related to individual therapy, but it is reasonable to assume that couples have the same expectations of therapy as do individuals. However, the goodness-of-fit between a therapist and couple is more complicated to determine due to the interrelated aspects of both partners who each have their own expectations, agenda and motives for therapy. To the extent that the therapists matched these for both partners determined how well the couple fit with the therapist. It seemed as if this fit was sometimes achieved by adapting to the couples to meet their needs and personalities. The therapists adapted their techniques, pace and focus in order to meet the needs of the couple. The times when this was not effective occurred with couples who had weak within-couple alliance in which they had differing beliefs about the cause of the 201 distress, as well as lower commitment levels to the relationship . As such, the results suggest that the context of couples therapy interferes with some of the more natural ways a client-therapist match typically occurs. This again supports the suggestion to directly address the nature of couples therapy, as well as the client expectations early and often in therapy. A discussion about the particular working model and theoretical orientation of the therapists is warranted as well. This study supported previous research that stresses the importance of therapists matching their clients. This was found to be a complicated task with the most conflictual couples in the study. It was not clear if there were ways to avoid this. When clients have their own perspectives about therapy, research suggests that therapist attempt to converge with those views (Horvath, 1994). In couples therapy this may be harmful, as aligning with one has the potential to damage the alliance with the other. The results indicated that the when the goodness-of-fit was right, it was a natural occurrence that did not have specific, discernable techniques to identify in order to understand how to build positive alliances. As the clients stated in the interviews, ―it just fits.‖ Therapist skills. The importance of therapeutic skills resembles the idea of collaboration in the therapeutic alliance research. Collaboration refers to the idea that the therapist and client are mutually working together to accomplish the tasks and goals of therapy. Collaboration is thought to be the foundation for strong therapeutic alliances (Bordin, 1979). Indeed, research suggests that collaboration and cooperation are conducive to stronger alliances, and in turn, better outcomes (Brossart et al., 1998; Kowalik et al., 1997). How clients perceive a therapist to be collaboratively working with them to achieve the couple goals is through the skills that therapists demonstrate in sessions. The current study discovered that the general skills that the clients perceived as helpful to their relationship related to therapist competency. There were two 202 ways the therapists portrayed their competency for participants in the study. Clients perceived the therapist as more competent when they displayed professionalism and incorporated meaningful feedback into sessions in the form of interventions, observations and specific activities. Professionalism. Results indicate that professionalism was an important quality of the therapists that influenced the perception of therapist competency. Clients had more confidence in their therapist‘s effectiveness when they perceived the therapist as professional. Importantly, professionalism was found to be subjective and based on the clients‘ expectations and preferences for therapist mannerisms, such the language they use and their overall appearance. Thus, a therapist appeared professional for varying reasons related to the particular client. While some clients preferred professional attire, others appreciated how therapists established appropriate boundaries that defined the therapeutic relationship. As discussed previously, this particular therapeutic setting provided challenges to establishing professionalism. Overall, the therapists were able to establish professionalism by demonstrating their knowledge and training. Meaningful feedback. Research has not clearly identified how important interventions and specific skills are to therapeutic alliance. However, some research suggests that successful outcomes are more correlated with the alliance dimensions of tasks and goals rather than the aspect of bonds (Friedlander et al., 2006a; Horvath, 1994; Kaufman, 2000). Data from the current research supports this finding. In the interviews, clients indicated that when they perceived specific sessions or therapist feedback and interventions as meaningful, they rated the alliance more positively. Indeed, observations confirmed that sessions in which there were specific activities, interventions, or insight-oriented feedback contributed to higher alliance scores for clients. Thus, when the couples perceived the amount and type of feedback as 203 meaningful, it ultimately strengthened their alliance. How this is connected to progress or outcome is not known, as this research did not focus on measuring those variables. Although this was not surprising, this finding did demonstrate one of the principal differences between therapeutic alliances with couples and individual clients. Since conjoint therapy is not one-on-one interaction, couples do not necessarily need to feel as connected to the therapist. They are more concerned about their partner and how the therapist is helping the couple resolve the issues. While a bond is important, and they desire to be validated and understood, they do not need to necessarily feel personally connected in order to stay engaged and have strong alliances. This finding is in line with other research that emphasizes the importance of the tasks and goals aspect of alliance (Symonds & Horvath, 2007). There were some clients in the study who indicated that the bond was significant to their relationship with the therapist. Importantly, these clients were partners in the most conflictual relationships. This again supports the presumption that overly strong bonds could signify a split alliance and weak within-couple alliance. Thus, therapists should keep in mind that elevated scores on the engagement level, or tasks and goals dimensions does not mean the couples are not connected. Similarly, high ratings on the emotional connection or bonds scale does not indicate a strong, healthy alliance. Findings suggest that there needs to be a balance among all these dimensions to achieve a stable, strong alliance. Research Question 2: Managing the Multiple Systems of Alliance in Couples Therapy The findings of the current study revealed key components and specific alliance promoting behaviors related to the three alliance systems in couples therapy: individual alliances with partners, within-couple alliance, and collective couple alliance. The most prominent alliance 204 systems were found to be the individual alliances with partners and within-couple alliance. As such, the discussion will pertain to these two important alliance systems in couples therapy. Individual Alliances with Partners This system was found to be one of the most significant, and most destructive, systems of alliance for the couples in the study. Results suggested that when this alliance was not properly formed-usually due to couple distress-it had deleterious effects on the within-couple alliance and collective couple alliance. Conversely, if too strong of an alliance was formed on this dimension with conflictual couples, a split alliance tended to occur. This ensued most strongly in the couples in which one partner believed the couple distress was due to their partner‘s psychological distress. Notably, this manifestation transpired when the alliance appeared to be strengthened at the expense of the female partner‘s alliance. When the couples were less distressed, had similar perspectives on the presenting problem, and displayed positive within-system alliances, the individual alliance did not affect the couple relationship. Thus, while the individual system is an important alliance to develop in terms of engagement and emotional connection, therapists need to be aware of how this impacts alliance when the level of distress is high and the couple has a weak within-couple alliance. These results are similar to those found in two other studies. A study of the between-and within-system alliances effect on distress in couples therapy found that when the male partner‘s alliance with the therapist increased, his female partner‘s distress increased. Conversely, when the within-couple alliance strengthened, his partner‘s distress decreased (Anderson & Johnson, 2010; Knobloch-Fedders et al., 2007). While these findings occurred specifically in the context of a partner‘s psychological distress, it is reasonable to suggest that this could be a similar pattern for the partners in this study. Interestingly, the male partners were considered to have the 205 psychological distress in the couples of the current study. They were also the partners who formed stronger alliances with the therapist. Engagement and emotional connection. The therapist-client relational dimensions of Engagement and Emotional Connection were found to be important aspects of alliance to attend to when building the individual alliances with the partners. How the therapists addressed these elements depended on their ratings of the couples on these individual subscales. For the partners in the current study, women tended to have higher engagement scores than their male partners. This supports past research that posits women both initiate and participate in therapy at higher levels than men (Rait, 2000). There were not any distinctive differences between the engagement levels of the partners at the beginning of therapy, however the women did seem to have more power in terms of initiating and ending treatment. For the three couples who were most conflictual, the female partners decided when treatment ended. This result conflicts with previous research that suggests men are the initiators of canceling sessions or ending therapy (Berg & Rosenblum, 1977). Thus, the findings of the current study imply that therapists cannot take either partner‘s initial level of engagement for granted in couple therapy. Particularly for distressed couples, even the initiator and most engaged partner is at risk for ceasing treatment prematurely. These recent findings may also be explained by Symonds and Horvath (2004) research that suggested more positive outcomes are likely when men have a stronger alliance than women at the onset of therapy and they both continue together with a positive alliance. These findings suggest that a balanced therapeutic alliance necessitates that therapists strongly engage the male partner immediately and continue to stay engaged with the female partner throughout therapy (Garfield, 2004). 206 Promoting engagement was important during the initial phase of therapy due to the common dynamic in which one partner was pushed to attend therapy. Therapists tended to rate the engagement level similarly to the partner who scored the lowest. When this occurred, the observational ratings of the therapist were the highest. Thus, they exhibited more engagement promoting behaviors when they perceived the engagement level was low. When they rated the engagement level higher than the couple, their observational scores were lower, signifying that they displayed less engagement behaviors for that particular session. Interestingly, the couples were also more conflictual in those sessions. Perhaps this means that the therapist‘s rating reflected more of their personal feelings of engagement due to the increased effort put forth on their part to facilitate the session. However, they may have been attending to other aspects of the alliance, such as safety. Important engagement promoting behaviors according to therapists and clients related to ensuring both partners had equal chances to talk, facilitating specific dialogues in sessions and initiating in-session and out of session activities. The emotional connection the clients had with the therapists played a significant role in the couple alliance. If one partner was more connected to the therapist, a split alliance occurred. This was due to client factors, such as personality and couple dynamics, and the occurrence of a rupture. In the sample, three couples experienced a split alliance either at some point during treatment or throughout treatment. The couple dynamics influenced the split alliance in all three of these couples. Interestingly, the male partners tended to rate their emotional connection higher than the female partners. This was surprising as past research addressing males in therapy suggests that they identify more strongly with the alliance aspects of tasks and goals than the dimension of bonds (Garfield, 2004; Symonds & Horvath, 2004). 207 The therapists in the study tended to rate the emotional connection according to the partner who had the higher scores on this dimension. This may have contributed to the occurrence of the split alliances. When partner scores differ on this dimension, and the therapist rated the emotional connection similar to the one who was more connected, it would make sense that a split occurred. This appeared to be related to the amount of emotional connection promoting behaviors that were used during sessions. It was observed that the higher the therapist‘s score compared to the couple, the lower the observational mean rating for the therapist. This suggests that they engaged in less emotional connection promoting behaviors when they believed to be on the same level as the most connected partner of the couple. Conversely, if the therapists scored lower than the highest partner of the couple or lower than both members, their observational ratings increased. Thus, when they perceived the emotional connection to be lower, they engaged in more emotional connection promoting behaviors. It is important to understand why this dynamic occurred, especially since the scores on all other alliance dimensions indicated that they rated these dimensions similar to the partner who had lower scores, or considerably lower than both partners. This seemed to indicate that the therapists rated the other dimensions based on the overall couple, rather than one specific partner. Thus, if only one partner was engaged, they believed that the couple was not highly engaged. For emotional connection, it appeared that they were not incorporating the other partner, or simply overestimating the emotional connection of that partner. This could mean that the emotional connection dimension is the one that is most noticed and sensed by therapists since it signifies the relationship they have with their clients. Perhaps it is more difficult to accurately assess this aspect of alliance when there are two partners in the room. More research needs to explore this finding as it could be unique to this particular sample. 208 Significant ways in which the therapists promoted emotional connection were to focus on the emotional components of the individual, couple, and relationship between the therapist and client. Clients indicated that they felt closer and safer when the therapists promoted these elements in the sessions. However, for a small number of clients, focusing on their emotional reactions was not conducive to building an alliance. This was related to their personality and misfit with a style of therapy that focused on emotional responses, and implies that therapists need to carefully consider the therapeutic tools and theories they use with clients. While attending to the emotional components of a client has been recognized as an effective method of treatment (Johnson & Talitman, 1997), if it does not match the client‘s perspective of the problem or overall personality, this form of treatment will not be successful. Furthermore, it can weaken the alliance and increase the likelihood of client dropout. This finding is in line with common factors research which suggests theories are not one-size-fits all in nature and therapists need to adapt their methods to meet the needs of their clients (Blow et al., 2007; Sprenkle & Blow 2004; 2007). One specific method of building individual alliances with the partners was to implement individual sessions. There were mixed results about how effective this strategy was to manage the individual alliances. The individual sessions served to increase alliances for the individual partners when they occurred. Clients indicated that they appreciated the individual focus and believed it helped them to connect to the therapist. However, they had detrimental effects in the case when the couples had low within-system alliances. While those sessions were beneficial to the individual alliances, they weakened the within-couple alliance. This has important implications for therapists. Many clinicians who practice couples therapy do not condone the use of individual sessions when working with a couple due to the risk 209 of alliance imbalance or the potential for other issues to arise in those sessions that would challenge the couples work. As such, therapists need to ascertain a specific strategy that identifies when and how to use the sessions with couples. The use of these sessions have the potential to contribute to strong, positive couple alliances, but also have the same potential to harm the connection if the relationship is weak. The within-couple alliance can be more difficult to accurately assess because it is not directly related to the therapist. It may be easier to discern how well an individual is connected to the therapist than to their partner, especially if there are multiple motives and perspectives occurring that are not always obvious. Thus, it is essential that therapists employ proper, continuous assessment of all levels of alliance in order to understand how this strategy may impact the overall couple alliance. Synergy effect. One of the most significant findings related to the individual alliances was the synergy effect that occurred for participants. This dynamic related to the finding that some partners rated their alliance according to how they perceived their partner‘s alliance was with therapist. This finding denotes that individual alliances with the therapist are partly based on how the client perceives the therapist responded to their partner. This has a synergy effect on the alliance in which connection increased when the client thought that the therapist responded to their partner in a way that supported their needs. However, if the client perceived the therapist to support or validate a feeling or experience for their partner that was not similar to how they felt, their emotional connection scores decreased. This discovery has been revealed in other studies of alliances with couples (Friedlander et al., 2006a; Knobloch-Fedders et al., 2007; Symonds & Horvath, 2004) and highlights the significance of how individual alliances impact the couple. As such, it verifies the importance for therapists to attune to how they are connecting with partners in a dyadic relationship. 210 Within-Couple Alliance The within-couple alliance was the most salient to alliance progress in this study. The strength and weaknesses of the alliance depended on this system. As previously discussed, the relationship dynamics of couples is one of the most significant obstacles to the alliance. It is within this dimension, then, that the challenges existed for the therapists and couples of the current study. Research has found that the within-system alliance appears to be particularly significant when the treatment goal is to resolve relational distress (Friedlander et al., 2006a). Indeed, for both partners in the couple, the within-system alliance is more strongly associated with improvement in the relationship than the between system alliance. Thus, this system is of utmost important to couples therapists due to the purpose and goal for initiating therapy is to relieve relationship distress. The dimensions of safety and shared sense of purpose were found to influence this alliance in that couples with lower scores on both scales had weaker within-couple alliances. Safety. The feelings of safety appeared to have a circular process in that the ratings of safety impacted the within-couple alliance, and the couple dynamics influenced the feeling of safety in the therapeutic context. A notable finding regarding this dimension of the within-system alliance was that it contributed to the alliance between the partners and therapist at times more than to the within-couple alliance. The degree to which the partners were able to feel safe was based in their overall connection with the therapist. If they believed they fit with the therapist, that he/she was neutral, and that the therapist can help them, they instantly had a feeling of safety. This again could was explained by the couple dynamic and within-system alliance. The cases in which this dynamic was most obvious were the ones who had the most conflict. This 211 suggests that the clients depend on the therapist more when they have lower feelings of safety in their relationship. Shared sense of purpose. It was believed that the shared sense of purpose dimension would be the ―heart‖ of the within-couple alliance. Indeed, there were distinct differences between the couples who appeared to have a high sense of purpose from those who did not. Most of the characteristics resembled those already discussed in the current literature (for review see Friedlander et al., 2006a). The important finding in the current study regarded the tendency for the couples with strong within-couple alliances to focus on the evolution of the couple relationship and goals rather than a personal connection with the therapist. Whether or not they felt a strong connection to their therapist was not as important as the progress that was occurring in their relationship. For these couples, the presence of the therapist in the room was not the main focus. While they appreciated his or her facilitation of discussions, the couples would focus more on one another in the moment. Furthermore, this aspect was evident from the start with these couples and did not depend on the therapist‘s efforts to promote a feeling of shared purpose. These couples had higher and more stable alliances throughout treatment. For the couples who did not enter therapy with a strong sense of purpose, there were larger variations between the partners‘ alliance ratings at the onset of treatment. How their alliance evolved depended on the active promotion of the couple relationship. Despite the efforts of the therapists to develop couple goals, promote unity, and facilitate similar perspectives, their alliances had increased fluctuations during therapy. Additionally, there was the presence of a split alliance, as well as therapeutic ruptures. Thus, it appeared that the most significant indicator of how alliance developed was the degree of shared sense of purpose the couples had before walking into the therapy room. 212 These findings are supported by the literature on the within-couple alliances that posits the within-couple alliance significantly impacts alliance development and maintenance in couples therapy (Anderson & Johnson, 2010). Beck et al. (2006) posited that the within-system alliance is most salient to clients, who referred to it more frequently and in more detail than the between-system alliance in their interviews. Similarly, the couples in this study referred to the between-system alliance more frequently when the within-couple alliance was not as strong. This increased focus appeared to be related to the need for support from the therapist or their dissatisfaction with the therapist‘s way of addressing the couple issues. Collective Couple Alliance Results confirmed that in order to have a collective couple alliance, the partners had to have stable individual alliances with the therapist and an intact alliance within the couple. Thus, it was not possible to have a strong collective couple alliance without achieving the other two forms of alliance. These findings echoed past research that posits the collective alliance is difficult to acquire in conjoint treatment due to the levels of conflict and tendency for one partner to not believe they are part of the problem or solution (Friedlander et al., 2006). As evidenced in this study, the therapists needed to develop specific ways to form a relationship with the couple who appeared to hold the most power in the relationship. Most times, this was difficult due to the personal and couple dynamics, and the therapists found themselves more aligned with the weaker partner, which served to further the imbalance. Monitoring and Repairing Alliances The findings from the current study emphasize the importance of monitoring and repairing ruptures throughout treatment. It was apparent that the intensity of a rupture existed on a continuum in which minor ones were characterized by feelings of disconnect and severe ones 213 created feelings of anger, hurt and vulnerability. While the extreme forms are more apparent, the less intense feelings of disconnect could have an insidious process that will result in a rupture that is not repairable. It appeared that some ruptures improved on their own when they were minor and a result of misinterpretations and the therapist was able to address and clarify their behavior. Conversely, some ruptures, even when directly addressed in the moment, were not repaired due to the dynamics of the couple and particular partner characteristics. These findings indicate that the manifestation of ruptures is probable in couples therapy and the level of intensity and ability for them to be repaired was due more to client factors, such as the interpersonal skills and couple conflict. The findings from this study elucidated the complexity of therapeutic alliance with couples. Deciphering which system to attend to when and why could impact the pattern of alliance over the course of therapy. The results of this study suggest that the within-couple alliance is the prominent alliance to attend to for couples in therapy. Since most couples enter counseling with some level of distress, assessment of their within-couple alliance is necessary at the onset of therapy. If it is clear that one partner does not view the relationship distress as a couple issue but rather a manifestation of their partner‘s dysfunction, therapists need to strongly align with that partner in order to begin establishing a unified view of the goals for therapy. Once that partner is engaged and connected to the therapeutic process, the therapist needs to continually work to strengthen the within-couple alliance. While each alliance system reinforces the other, if the within-system alliance is not intact, the other alliances are at risk for ruptures and imbalances. Furthermore, assessment of the couple dynamics is crucial when deciding which alliance promoting behaviors to use. As was demonstrated, some strategies, such as scheduling individual sessions with each partner, could have unforeseen detrimental effects on the within- 214 couple alliance. Understanding each of these alliances and how they form independently as well as interdependently with the other systems is a complicated task when faced with couples in high conflict. The key is immediate and continuous assessment of all alliances. Research Question 3: Couple Alliance Patterns over Time The third research question explored how the alliance formed and progressed over the course of therapy for the couples in the study. The findings revealed notable patterns of alliance for the couples. These patterns will be discussed and interpreted as they relate to previous research on alliance patterns in couples therapy. Notable Patterns in Therapist and Observational Ratings There were notable patterns in the therapist and observational ratings. First, the therapists‘ ratings of the alliance were typically disparate to those of the clients. How dissimilar their scores were appeared to depend on their perception of the couple‘s alliance. If the therapist rated the alliance as high, their scores more closely resembled those of the clients. Conversely, if they rated the alliance as poor, the therapist scores were significantly lower than the ratings of the partners. This suggests that couples who had poorer alliances had a stronger, albeit negative, influence on the therapist‘s alliance ratings. Another finding was that the therapists tended to have lower ratings of alliance in terms of total and subscale scores. This is a common finding in alliance research in which clients rate alliance higher than the therapist (Beck et al., 2006; Fitzpatrick, Iwakabe, & Stalikas, 2005; Hersoug, Hoglend, Monsent & Havik, 2001; Hilsenroth, Peters, & Ackerman, 2004). Researchers have speculated that this may be the result of the difference between therapist and client reference points used in their evaluation of alliance. Clients may base their ratings on how they compare to other therapeutic experiences, as well as other relationships they have outside of therapy. Therapists may rate the alliance based on 215 comparisons of other clinical experiences with clients. (Tyron & Kane, 1993; Hovarth & Symonds, 1991). This would make sense for the couples in this study as they all had prior therapeutic experiences. Observational data further obscured the ability to distinguish which perspective was more accurate. The dimensions that are based on affect related behaviors, such as emotional connection and safety, often had greater variance between observational and self-report scores. Typically, the clients rated alliance higher on the shared sense of purpose dimension than the therapist did. Importantly, the therapist and observational ratings provided more accurate explanation for the couple dynamics, within-couple alliance, split alliances, and treatment dropout that occurred for some couples. Past research has posited that this could arise due to the tendency for self-reports perceptions of client and therapists to reflect the cumulative impact of therapy rather than the individual session (Anderson & Johnson, 2010; Knobloch-Fedders et al., 2004). Moreover, clients‘ perception of the shared sense of purpose are related to the dynamics that occur outside of session and their ratings would reflect their overall sense of the withincouple relationship. The results of this study do not completely support this hypothesis. While it may be true that the ratings of the therapists were based on their global perceptions of alliance (which would suggest lower scores), the client ratings did not reflect their global perception of the relationship since dynamics of several sessions would indicate that their shared sense of purpose was low. If they rated this dimension based on their relationship outside of therapy, it would imply that they are less conflictual at home than they are in therapy. These conflicting findings highlight the complexity that multiple perspectives of alliance creates related to the varying levels of objectivity and subjectivity that are inherent to the process. 216 Alliance Patterns Research suggests that alliance studies need to assess the client perceptions of the alliance because they are superior to predicting outcome (Beck et al., 2006; Horvath et al., 1993; Horvath, 1995; Kaufman, 2000; Knobloch-Fedders et al., 2004). While this study did not assess how the alliance ratings related to outcome, the findings do suggest that clients‘ ratings were predictive of treatment retention and drop-out, while therapist and observational perspectives were influential to understanding the reason for the negative alliance patterns. Thus, this study focused on the ratings of the clients in order to determine alliance patterns. The therapist and observational scores were utilized to explain the ratings. Initial alliance. Some research depicts that therapeutic alliance from the clients‘ perspective is established at treatment outset with very little change over the early sessions (Glebova et al., 2010; Knobloch-Fedders, 2007). The results of the current study did not fully support this finding. Generally, the scores fluctuated in the early sessions. It was not until the fifth session that the scores became more stable and reflective of their average ratings. This pattern is more in line with the research that posits therapeutic alliance changes over time during the initial stages of therapy (Friedlander et al., 2006). The fluctuating pattern of the alliances in the early stage supports the postulation that alliance is not fully established until the period between the third and fifth sessions (Horvath & Symonds, 1991). Alliance fluctuations. Alliance fluctuations ensued during two specific times in therapy for the couples. There was a sharp or gradual decrease in scores between the first and third sessions for most clients. This alliance pattern has been discussed by Gelso and Carter (1994) who postulated that the alliance course in treatment follows an immediate weakening after the 217 initial development. In successful therapies, the alliance will increase to the earlier, high levels. This was found to be true for most of the couples in this study. The reason for this decline in scores could be explained by the pace of the beginning phase of therapy. Clients indicated that these sessions were slower, involved discussion of historical background, and did not significantly contribute to their goals. While clients understood the importance of assessment to the therapeutic process, their ratings of alliance still decreased. Importantly, two participants who rated the initial session relatively lower, actually had increased scores during the second and third sessions. These partners, both male, were also the members of the couple who were ―influenced‖ to attend therapy by their female partners who initiated therapy. Thus, their scores can be related to the lower level of engagement they felt during the intake session. These findings make sense in light of the general stages of treatment. The initial stage is when clients are most mobilized and hopeful about their work in therapy (Gelso & Carter, 1994). High initial session scores can represent the excitement and energy clients feel when they first embark on the progress towards a goal. This is similar to what the literature refers to as the hope, or placebo, factor that has been found to contribute to treatment outcomes (Blow & Sprenkle, 2001). In terms of alliance, feelings from the initial sessions may be more related to their hope for change than their actual connection to the therapist or therapeutic process. As such, the first session can appear exciting, fresh and provoke feelings of connection that are not necessarily related to a long-term bond. It would make sense for the alliance ratings to decrease over the next few sessions as that initial ―high‖ wears off and the real work begins. Yet another explanation relates to the propensity that the therapist mainly focuses on creating a connection with the couple during the first session. They may display more warmth, 218 humor, and overall connection promoting behaviors that would create feelings of connection. These behaviors may not be as evident in the subsequent sessions as the therapist begins to focus on the couple goals and issues, which often initiate feelings of discomfort for clients. As to how this early phase alliance pattern influenced the alliance ratings over the course of treatment, the results were mixed. A second phase of alliance ensued between the sixth and tenth session in which there were noticeable fluctuations for the couples. For the strongly aligned couples, their scores typically fluctuated to either a higher or lower score, followed by a return to previous levels of alliance. The scores for the more distressed couples followed the same pattern, but did not achieve a stable alliance towards the tenth session. Thus, the overall alliance patterns for the couples in this study appear to reflect the alliance patterns found by Horvath et al. (1994). For most couples, there was an initial phase of development for the alliance within the first five therapy sessions. If the alliance was not stable or strong for one or both partners, the couple ceased treatment, which was the case for one couple in this study. The next phase of alliance which is often a more critical phase of alliance, transpired between the fifth and tenth sessions (Horvath, 1994). This phase of alliance was characterized by increased focus on the couple dynamics in which the therapists challenged the negative patterns inherent to the relationship. This typically weakened the alliance for the couples. In the case of two least distressed couples, the alliance was able to recover. For the other two cases, who had higher levels of distress, retention was challenged and both did not return to therapy after the tenth session. Results from this study indicate that alliance in couples therapy is co-constructed by the therapist and clients, and is influenced by the level of couple distress, perspective of presenting problem, and ability of the therapist to strengthen the within-couple alliance. The alliance ratings 219 for the initial and middle phases of treatment appeared to be influenced by different factors. Alliance formation appeared to be due to the level of couple distress and shared sense of purpose that the couple presented with, as well as the interpersonal characteristics of the individual partners and therapists that contributed to their ability to establish a personal connection. The couples who had the most unstable alliance patterns experienced ruptures and split alliances during the initial phase of treatment. Since research posits that the presence of a split alliance during the first initial sessions of therapy is typically more related to history of distressed familyof-origin relationships that to the therapeutic process (Knobloch-Fedders et al., 2004), it appeared that the reason for these negative alliance dynamics during the first few session related more to the couple dynamics than the therapist. As such, the client factors that the couples brought into therapy were significant influences of initial alliance formation. The degree to which the therapist and couple were able to establish common goals and a unified perspective of the couple distress determined how the alliance shifted over time. Thus, therapist skills such competency and behaviors that promoted safety and a shared sense of purpose with the couple, as well as the stability of the within-couple alliance contributed to the alliance fluctuations during the middle phase of treatment. Clinical Implications This study has important implications for couple therapists. First, the study identified salient factors related to therapeutic, client and therapist aspects that influenced the way in which alliances are formed and maintained. In order to appropriately and effectively evaluate and manage alliance requires careful consideration of how each factor may emerge and impact the alliance. The most prominent way of managing these dynamics was found to be to openly and directly addressing them with the couple. 220 The most significant factor related to client factors was found to be the level of relationship distress that the couples presented with when entering therapy. This finding supported past research that found the level of distress prior to therapy potentially affects the couple‘s ability to build alliances within the first few sessions (Knobloch-Fedders et al, 2004). This suggests that therapists should assess the level of relationship distress at the onset of treatment. One way that was shown to be useful in the current study is using the Revised Dyadic Adjustment Scale (RDAS; Busby et al., 1995) prior to the beginning of the first session. This scale is easily accessible and therapists can score each member of the couple with relative ease. The results from this assessment can help therapists assess for, and address any discrepancies in the partners‘ reports of relationship satisfaction. If the couple is highly distressed, therapists need to openly discuss how the level of distress could impact the various systems of alliance. By addressing this at the onset of treatment, the clients may be prepared for any negative feelings that may arise during the course of therapy and understand that they may be more due to the relational distress than the therapist or therapeutic process. Finally, instilling hope for the most distressed couples is an important first goal for therapists to achieve in order to form a positive initial alliance with the couple. Therapists should plan to focus on strengths and identify positive aspects of the couple relationship before focusing on the weaknesses or destructive interaction patterns that are significant contributors to the relationship distress in order to begin building the within couple alliance as soon as possible. Another factor that is related to both clients and therapists is that of gender. The current study, backed by past research, found that there may be gender related dynamics that impact alliance formation and treatment retention. Women tended to be the initiators of therapy in this study. Clinical assumptions would suggest they are the strongest aligned partner at the onset of 221 therapy. Thus, clinicians may focus most of their attention on engaging and forming an alliance with the man. While male partners tended to rate the alliance lower and appear to be less engaged or committed to the relationship and therapeutic process, that was not necessarily the case. This needs to be carefully examined as their alliance may be just as strongly intact as their partner‘s alliance. Low ratings may not be an indicator of weak alliance, and their reserved manners in the therapy room may be more related to the couple dynamics and their general discomfort with sharing personal information. This may be an important clinical issue, but not necessarily an indication of alliance. Overly focusing on the male partner may cause an alliance imbalance which could result in a split alliance in which the female partner has lower levels of alliance. If the female partner holds more of the relational power, as was true of the females in the current study, the couple is at risk for dropping out of treatment. The more effective approach may be to promote the alliance immediately by establishing goals for treatment that engage the male client while promoting alliances with both partners simultaneously. In terms of therapist gender, this research found that some clients have gender related preferences for therapists. In the study, a few clients indicated that they felt more comfortable with a female therapist even if their alliance ratings did not support this assertion. This is also supported by past research and emphasizes the need for therapists to recognize how their gender impacts clients in the room, particularly with couples. There is often a gender imbalance (with heterosexual couples) that cannot be prevented or ignored. Therapists need to consider how these dynamics, as well as any other demographic variances (such as age, race/ethnicity, culture, religion, etc) influence the alliance they have with the couple. Therapists should openly acknowledge and facilitate a discussion about these demographics to increase feelings of safety for the clients. 222 The existence of multiple alliances poses several clinically relevant implications. The results of the study suggest that the most salient system to overall alliance with couples is the within-couple relationship. As such, therapists need to remain cognizant of how this dimension is impacting the alliance throughout treatment. More importantly, therapists need to carefully consider which alliance is important to focus on at specific times during treatment. The current findings suggest that the within-system alliance needs to be promoted continuously from treatment onset. However, this is not always possible without forming strong alliances with each partner, particularly for highly distressed couples. Most therapists tend to focus on their relationship with the partner who appears to be the least engaged, or weaker partner, in the therapy and/or couple relationship. This decision did not have successful results for the couples in the current study. The partner who was responsible for both initiating and ending therapy was more often the partner with the higher alliance scores at treatment onset. Thus, therapists need to engage the most influential partner of the couple in order to promote couple retention. This necessitates therapists to resist the inclination to focus on the weaker partner in order to protect them, and to monitor their relationship with this partner as they are typically the ones who present the least challenges to forming alliances. The results suggest that therapists need to carefully consider how they choose to promote the different systems of alliance. As indicated in this study, strategies for building individual alliances consisted of equally engaging both partners, focusing on emotional connection, and implementing individual sessions as appropriate. Therapists need to be careful when they implement individual sessions, however. While this was mutually beneficial for the partners in most couples, it was detrimental for the couples who experienced high levels of distress. 223 Therefore, it seems imperative that therapists should monitor all three alliances, and any focus on a particular alliance should be done within the context of the within-couple alliance. It was important for the therapists to accurately assess how strong the within-couple alliance was in order to implement any strategy to promote alliance. Signs that signified positive within-couple alliances included the tendency for partners to focus on each other instead of the therapist, emphasis on the evolution of the couple relationship and goals rather than a personal connection with the therapist, and little variation between the partners‘ ratings of alliance. Conversely, couples with weaker shared sense of purpose began treatment with increased differences between the partner scores, experienced increased fluctuations during therapy, and were more likely to experience a split alliance and therapeutic ruptures. Successful ways to promote the within-couple alliance appeared to be related the levels of safety and shared sense of purpose for the couples in this study. Therapists need to actively work to manage conflict, cease attacking or critical behavior, and promote unified perspectives of the presenting problem throughout treatment. Perhaps the most important implication that emerged from the current study is the need for therapists to monitor all systems of alliance consistently from the outset of treatment. Alliance monitoring should be considered part of the theoretical framework and treatment plan for working with couples. It is just as important to establish a balanced, strong alliance with the couple as it is to change their maladaptive relationship patterns. Therapists should use the SOFTA self-report and observational measures as one way to assess therapeutic alliance with their couples. The self-report is a resourceful instrument that requires minimal time to complete and score. As such, this tool can be easily incorporated into treatment and should be given frequently throughout the course of therapy. It is recommended that this measure be completed 224 by the therapist and both partners of the couple after each session. This will allow the therapist to assess how alliance is changing from session to session, as well as the congruity between the alliance ratings. Although recording therapy sessions may not feasible in settings other than training clinics, the SOFTA observational measure can also be utilized by therapists as a way to learn how to recognize in-session alliance behaviors. This has two important consequences. First, it would help the therapist move away from assessing alliance by how they feel in the room which can sometimes be misleading due to the likelihood for increased feelings of anxiety and tension that is often experienced during couple sessions. Also, by learning how to recognize behavioral signs of alliance, the therapist can begin to identify any signs that the alliance is shifting and/or weakening during the session. Lastly, therapists in the current study had positive results when they openly addressed the therapeutic relationship in the moment. The therapists addressed behaviors that suggested the therapeutic relationship was changing, such as a particular partner‘s disconnection during session, the overall atmosphere of the session that suggested tension and discomfort, and overt tones or attitudes that portrayed anger or frustration. Clients appreciated when these dynamics were addressed by the therapists. For any alliance monitoring techniques to be successful, therapists need to be comfortable and willing to address the therapeutic relationship with their couples even if this feedback could be negative. Promoting a collaborative relationship can serve as a strong model for both the couple and the therapeutic relationship. Implications for Therapist Training and Supervision The most important implication for training and supervision concerns the need to incorporate educational resources that teach student therapists how to overtly assess and process 225 therapeutic alliance with couples. As mentioned previously, the SOFTA self-report and observational measures are important resources that can help supervisors teach therapists how to recognize alliance indicators during sessions. Supervisors should mandate therapists to know the significant factors that contribute to alliance formation and management in couples therapy. Therapists should be aware of the client factors, such as relationship distress and interpersonal skills that challenge the ability to build and sustain alliances in therapy. In addition, teaching therapists about the importance of expressing warmth and portraying aspects of their personalities with clients may increase the likelihood that they will form strong connections with both partners of the couple. Supervision should also include ways to help therapists become more comfortable with openly discussing the therapeutic alliance and receiving feedback from their clients. Incorporating role plays into supervision can increase therapist comfort and teach them how to discuss the therapeutic alliance openly with clients. For instance, this study identified the importance for therapists to address the gender imbalance that was present in the room. Teaching therapists how to raise this issue with clients is an important way that supervisors can help therapists to strengthen their alliance with couples. Additionally, supervisors should assess selfof-the-therapist issues that may prohibit the therapist from inquiring about their role in the alliance, or any therapeutic process. This can help therapists develop the clinical skills that are necessary to develop strong therapeutic alliances with couples. Teaching therapists that they are only one part of the alliance puzzle can also help to alleviate fears of receiving critical feedback from clients. Finally, raising awareness of the potential therapeutic, client and interactive factors may decrease their reactivity to negative feedback. 226 Limitations of the Current Study The current study has several limitations. First, the data was collected at a University training clinic with graduate level therapists. As such, results may not be representative of other community settings or clinicians with more experience. Clients who seek services at the training clinic may not be typical of other clients. The reduced rates offered for treatment may influence the type of clients that seek services, as well as treatment retention. Additionally, the nature of the clinic posed distinct challenges that were found to have an influence on the research findings. Confidentiality was limited due to the need for supervision and monitoring of the therapists‘ cases. Thus, results may not represent typical therapeutic alliance outcomes. Another limitation of the current study is the small sample size. For the purpose of this research, the small sample size of five couples and two therapists was sufficient to achieve the goal of exploring how therapeutic alliance evolves over time in couples therapy. Furthermore, the current sample consisted of heterosexual couples and was mostly comprised of Caucasian participants. As such, it is not known whether the results can be applied to same-sex couples or other clients from different cultural backgrounds. Finally, this research study was unable to thoroughly answer the third research question that inquired about how therapeutic alliance evolves over time due to the limitation of the data analysis method. While thematic content analysis provided detailed descriptions of the factors and systems of alliance influential to alliance development, it was not able to provide a complete analysis of how alliance evolved over the course of treatment for the couples in this study. Strengths of the Current Study The most significant strength of this study was the incorporation of the multiple perspectives of alliance that exist in the context of couples therapy. Gathering data from each 227 member of the therapeutic system and independent observer enabled this study to examine the multiple alliances in the context of the couple rather than individual data. Equally as important was this study‘s inclusion of multiple sources of data to obtain a more comprehensive description of the evolvement of alliance over time. This also strengthened the results through data triangulation. Moreover, this study collected data all throughout the therapeutic process in order to understand how alliance forms, grows and sustains in couples therapy. This addressed the deficiency in the research literature regarding the need for continuous, longitudinal assessment of therapeutic alliance in conjoint treatment. Finally, this research included the actual voices of the couples and therapists involved in the therapeutic relationship through multiple interviews that occurred at specific treatment intervals. This allowed for deeper understanding of the contextual and personal experiences of the alliance as it evolved over time. Through these methods, this study has provided an extensive description of the multifaceted process of alliance formation and maintenance in couples therapy. Suggestions for Future Research This study identified several suggestions for future research. First, future research should employ methods similar to those of the current study to allow further understanding of the nature and process of therapeutic alliance with couples. As such, researchers are urged to incorporate therapist, client and observational perspectives into their methodology to obtain a complete picture of the alliance. In particular, future research should explore client and therapist factors that are influential to alliance formation and growth. Family-of-origin issues, interpersonal skills and personality characteristics were found in the current study to be influential of how couples formed the alliance with their therapist. More research is needed to assess the role of these individual dynamics in alliance with couples. Research also needs to focus more specifically on 228 therapist and client perceptions of alliance as these were also significant contributions to the findings of this study. For example, the therapist‘s perception of emotional connection was not always congruent with both partners‘ assessment and appeared to be related to this aspect of alliance being the easiest dimension to ―sense‖ by participants. This could have a negative impact on alliance as this dimension was found to be important for individual partner alliance and the development of a split alliance. Future research should identify if this phenomenon occurs outside of the current sample. An important implication that emerged from this study was the need for therapists to implement methods of receiving feedback from the couples about the therapeutic alliance. Future research needs to explore what types of feedback are most influential of positive therapeutic alliance outcomes. The SOFTA self-report and observational measure was utilized in this study to assess the varying perspectives of alliance at each session. While this measure was successful in chronicling the alliance over time, it was not used to provide feedback directly to the therapists. Thus, it is not known whether or not the information provided by these measures would positively or negatively impact alliance for the couples in the study. One method that was helpful in the current study was for therapist to directly inquire about the therapeutic relationship in the moment with their couples. However, this often places pressure on the clients to provide positive feedback. Other methods may elicit more honest feedback that can serve to positively influence the alliance. Research needs to identify which methods of feedback are useful for therapists to accurately assess the alliance, as well as to determine what they need to refine, change or focus on to strengthen alliances. Additionally, future research should incorporate a larger sample size of couples in community practices rather than training facilities to increase the likelihood that the results can 229 be generalized to all couples. It is important to recognize that any research that aims to analyze therapeutic processes, such as alliance, may be impacted by the reduced confidentiality inherent to the research methods inherent to these types of studies. Also, research needs to recruit samples that include same-sex couples, as well as clients and therapists from other cultural backgrounds to ascertain how similar or different the process may be. Furthermore, research needs to extend alliance assessment past the middle phase of therapy to understand how alliance evolves after this critical time period. Alliance patterns appear to be particularly vulnerable during the middle phase of therapy, and more research is needed to understand what methods are effective in sustaining strong alliances past this challenging time frame. Finally, while quantitative studies of the alliance are necessary and important to the understanding of couple alliances, research needs to continue incorporating qualitative methods, such as interviews, in order to achieve a deeper, more meaningful understanding of the therapeutic alliance with couples. Conclusion The purpose of the current study was to adopt a process-oriented model of measuring therapeutic alliance with couples by exploring it over the course of treatment. By incorporating multiple sources and perspectives of data, this study obtained a comprehensive understanding of the formation and evolvement of therapeutic alliance with five couples. Results identified the key components that contributed to how alliance progressed throughout therapy. The significant client factors that influenced the nature of alliance were found to be level of couple distress, gender of client and strength of the within-couple system. The most significant therapist factors appeared to be personal characteristics such as personality and gender, and their overall therapeutic style. All three alliance systems were evaluated to identify how each contributed to 230 the overall alliance patterns of the couples. The research identified that the within-couple alliance was most influential of alliance for this study‘s sample of couples. Whether the within-couple system was intact or weak impacted the formation and management of alliances. The alliance configurations for the couples depicted a pattern that did not become stable until the fourth or fifth session, followed by fluctuations during later sessions due to the challenging dynamics of the therapy sessions. How the couples recovered from weaker alliances was related to the strength of their within-couple alliance. The findings of the research both support and further current empirical literature on therapeutic alliances in conjoint treatment. The nature and impact of these alliances are instrumental to the treatment of couples. Therapists who are able to successfully form and sustain alliances with couples are better able to assist in the resolution of their relationship distress. As such, this study hoped to identify the important factors and systems that contribute to the therapeutic relationship in order to further develop and disseminate effective methods of strengthening therapeutic alliances with couples. 231 APPENDICES 232 APPENDIX A Script for Client Recruitment To be given to any couple calling to set up an intake session with a Ph.D level therapist. The only couples who are ineligible for study are ones who present with domestic violence or are court mandated. We are currently conducting a research study that involves couples. The study is entitled, ―An In-depth Process Study of Therapeutic Alliance Development and Management in Couples Therapy,‖ and is being conducted by a doctoral student in our program. The purpose of this study is to identify what helps therapists develop and maintain therapeutic alliances with couples. Therapeutic alliance has been identified as a common factor of successful therapies and is instrumental in treatment outcome for couples. However, research on therapeutic alliance with couples has not provided a clear answer as to how therapeutic alliance is developed and maintained in couples therapy. If you would like to participate in this research, I will have the researcher, Sara Timmons, contact you to provide more information regarding the research study. If you decide to participate, Sara will set up a time to meet with you prior to your first session. I would also like to emphasize to you that your participation in this research study is completely voluntary and does not preclude your participation in therapy. If you chose to not participate, you are still able to participate in therapeutic services here. Would you like Sara to contact you to provide more information about this research study? 233 APPENDIX B THERAPIST CONSENT FOR PARTICIPATION IN RESEARCH STUDY Tina Timm, Ph.D, Assistant Professor, Michigan State University Sara Timmons, M.S., Doctoral Candidate, Michigan State University Michigan State University 1. Sara Timmons, a doctoral student in the Department of Family and Child Ecology at Michigan State University has requested my participation in a research study at this institution. The title of this study is: ―An In-Depth Process Study of Therapeutic Alliance Development and Management in Couples Therapy.‖ 2. You have been informed that the purpose of this research is to identify what helps therapists develop and maintain therapeutic alliances with couples. Therapeutic alliance has been identified as a common factor of successful therapies and is instrumental in treatment outcome for couples. However, research on therapeutic alliance with couples has not provided a clear answer as to how therapeutic alliance is developed and maintained in conjoint treatment. 3. Your participation includes providing therapy to couples as is regularly practiced at The Family and Child Clinic (FCC). This includes the routine videotaping of all therapy sessions. In addition to this, you will also complete the following assessments: a. You will complete the System for Observing Family Therapy Alliances self-report th (SOFTAs) at the end of each session until the end of treatment (or 12 session, whichever comes first). It takes approximately 5 minutes to complete. th th b. You will participate in an individual interview with Sara Timmons after the 4 , 8 , th and 12 session of treatment. These interviews will take approximately 30-45 minutes to complete. c. You will participate in brief interviews if needed during treatment with Sara Timmons. These interviews will take approximately 10-15 minutes to complete. 4. The only costs associated with this study will be your time. 5. You have been informed that there are possible risks to you if you agree to participate in the study. You may experience some level of psychological discomfort due to being asked to reflect on your therapeutic alliance with your clients. 6. You have been informed that the investigator has the right to terminate your participation in the study when, in the investigator‘s judgment, it is in your interest to do so, or under certain circumstances such as the inability to keep scheduled appointments, non-cooperation with treatment, or other administrative reasons. 234 7. You have been informed that the results of the research study may be published but that your name or identity will not be revealed and that your record will remain confidential. You have been informed that the research team will destroy videotapes at the conclusion of the study. You have been informed that only the research team will have access to your information. You also have been informed that the Michigan State University Institutional Review Board (the Board that is responsible for protecting the welfare of human subjects recruited to participate in research), may review your study records. There is no identifying information associated with the storage of these records. 8. The possible benefits of your participation in the research study are a contribution to the field of couple and family therapy. You have been informed that participation in this study may not benefit you directly. 9. You have been informed that the alternative to participation in this study is nonparticipation. 10. You have been informed that your participation is voluntary and that refusal to participate will involve no penalty to you or loss of any benefits to which you are otherwise entitled. You have been informed that you may withdraw from the research study at any time without penalty or loss of benefits to which you are otherwise entitled. 11. If you have concerns or questions about this study, please contact the Principal Investigator: Tina Timm, Ph.D., School of Social Work, 220 Baker Hall, Michigan State University, East Lansing, MI 48824. Dr. Timm can be reached by phone (517) 432-7112 or e-mail: timmt@msu.edu. 12. If you have questions or concerns regarding your rights as a survey participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University‘s Human Research Protection Program at (517) 355-2180, Fax: (517) 432-4503, e-mail irb@msu.edu, or regular mail at 207 Olds Hall, Michigan State University, East Lansing, MI 48824. 13. I have read the above statement and have been able to ask questions and express concerns, which have been satisfactorily responded to by the investigator. I have been informed of the purpose of the study as well as the potential benefits and risks that are involved. I hereby give my informed and free consent to be a participant in this study. __________________ Date __________________________________ Consent Signature of Subject This form is only valid if the IRB‘s current stamp of approval is shown below. 235 14. I certify that I have explained to the above individual(s) the nature and purpose and the potential benefits and possible risks associated with participation in this research study, have answered any questions that have been raised, and have witnessed the above signature. 15. These elements of informed consent conform to the assurance given by Michigan State University to protect the rights of human subjects. 16. I have provided the subject/client with a copy of this signed consent document. _________________ Date _________________________________ Signature of Investigator 236 APPENDIX C CONSENT FOR PARTICIPATION IN RESEARCH STUDY Tina Timm, Ph.D., Assistant Professor, Michigan State University Sara Timmons, M.S., Doctoral Candidate, Michigan State University Michigan State University 1. Sara Timmons, a doctoral student in the Department of Family and Child Ecology at Michigan State University has requested my participation in a research study at this institution. The title of this study is: ―An In-Depth Process Study of Therapeutic Alliance Development and Management in Couples Therapy.‖ 2. You have been informed that the purpose of this research is to identify what helps therapists develop and maintain therapeutic alliances with couples. Therapeutic alliance has been identified as a common factor of successful therapies and is instrumental in treatment outcome for couples. However, research on therapeutic alliance with couples has not provided a clear answer as to how therapeutic alliance is developed and maintained in conjoint treatment. 3. Your participation will be comprised of participation in therapy as is regularly practiced at The Family and Child Clinic (FCC) which includes the routine videotaping of all therapy sessions. You will also complete the following assessments: a. You will fill out the Family Data Form, an intake/demographics form, at the outset of treatment as mandated by standard clinical procedures. This form takes approximately 10 minutes to complete. b. You will fill out The Dyadic Adjustment Scale (DAS) at the initial intake session th and final session (or 12 session, whichever comes first). It takes approximately 10 minutes to complete. c. You will fill out the Outcome questionnaire at the initial intake session and after th the 12 or final session, whichever comes first. It takes approximately 10 minutes to complete. d. You will complete the System for Observing Family Therapy Alliances selfreport (SOFTAs) after the first session, as well as at the end of each subsequent th session until the end of treatment (or 12 session, whichever comes first). It takes approximately 5 minutes to complete. th e. You will participate in an individual interview with Sara Timmons after the 4 , th th 8 , and 12 session of treatment. These interviews will take approximately 30-45 minutes to complete. f. You will participate in brief interviews at other times during treatment with Sara Timmons. These interviews will take approximately 10-15 minutes. 237 4. The only costs associated with this study will be your time. You will be given $25 per interview for a total compensation amount of $75 per person. 5. You have been informed that there are possible risks to me if you agree to participate in the study. The risks are very similar to the risks of standard therapy at any therapeutic setting. One risk may include the possible loss of confidentiality due to your participation in face-to-face interviews with the researcher and taping of therapy sessions. Additionally, you may experience some level of psychological discomfort due to being asked to reflect on your therapeutic alliance with my therapist. You have been informed that every effort is made to ensure confidentiality in relation to your therapy (as in all therapy). All therapeutic records are kept in a locked file cabinet at FCC. All research materials will be kept separately in a locked file cabinet at FCC. The information from the study will not be revealed to any other person or agency without your written permission. This is different from your clinical rights. With regard to your confidentiality as it pertains to therapy, the clinical information from your sessions will not be revealed to any other person or agency without your written permission. You are aware that (as mandated by law) there is certain information that the therapist is obligated to disclose to outside sources. You have been informed that the mandatory reporting situations are: A) If you threaten bodily harm to another person or to yourself. B) If you reveal information related to child abuse or neglect. C) If a court of law issues a subpoena, the therapist is required to reveal information specifically described in the subpoena. You have been informed that the investigator has the right to terminate your participation in the study when, in the investigator‘s judgment, it is in your interest to do so, or under certain circumstances such as the inability to keep scheduled appointments, non-cooperation with treatment, or other administrative reasons. 6. You have been informed that the results of the research study may be published but that your name or identity will not be revealed and that your record will remain confidential. You have been informed that the research team will destroy videotapes at the conclusion of the study. Your therapy case file will be stored and destroyed in line with regular procedures at FCC. You have been informed that only the research team will have access to your information. You are also informed that the Michigan State University Institutional Review Board (the Board that is responsible for protecting the welfare of human subjects recruited to participate in research), may review your study records. There is no identifying information associated with the storage of these records. 238 7. The possible benefits of your participation in the research study are a contribution to the treatment provided to couples who participate in couple‘s therapy. You have been informed that participation in this study may not benefit you directly. 8. You have been informed that the alternative to participation in this study is nonparticipation. Nonparticipation in the study does not preclude your participation in therapy. 9. You have been informed that your participation is voluntary and that refusal to participate will involve no penalty to you or loss of any benefits to which you are otherwise entitled. You have been informed that you may withdraw from the research study at any time without penalty or loss of benefits to which you are otherwise entitled. 10. You have been informed that if you choose to discontinue your participation in therapy, that does not discontinue your participation in the research study unless you wish to do so. The researcher will still contact you to do a final interview. 11. If you have concerns or questions about this study, please contact the Principal Investigator: Tina Timm, Ph.D., School of Social Work, 220 Baker Hall, Michigan State University, East Lansing, MI 48824. Dr. Timm can be reached by phone (517) 432-7112 or e-mail: timmt@msu.edu. 12. If you have questions or concerns regarding your rights as a survey participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University‘s Human Research Protection Program at (517) 355-2180, Fax: (517) 432-4503, e-mail irb@msu.edu, or regular mail at 202 Olds Hall, Michigan State University, East Lansing, MI 48824. 13. I have read the above statement and have been able to ask questions and express concerns, which have been satisfactorily responded to by the investigator. I have been informed of the purpose of the study as well as the potential benefits and risks that are involved. I hereby give my informed and free consent to be a participant in this study. __________________ __________________________________ Date Consent Signature of Subject This form is only valid if the IRB‘s current stamp of approval is shown below. 14. I certify that I have explained to the above individual(s) the nature and purpose and the potential benefits and possible risks associated with participation in this research study, 239 have answered any questions that have been raised, and have witnessed the above signature. 15. These elements of informed consent conform to the assurance given by Michigan State University to protect the rights of human subjects. 16. I have provided the subject/client with a copy of this signed consent document. _________________ Date ____________________________________ Signature of Investigator 240 APPENDIX D Confirmation of Compensation for Participation in Research Study By signing below, I hereby confirm receipt of one (1) gift card for the amount of $25 as compensation for my contribution to the Sara Timmons‘ doctoral research on Therapeutic Alliance. __________________________________________ Name 241 ________________ Date APPENDIX E Revised- Dyadic Adjustment Scale Most persons have disagreements in their romantic relationships. Please indicate below the approximate extent of agreement or disagreement between you and your partner for each item in the following list. Circle the number in the box that best describes your situation. Always Agree 1. Religious Matters 2. Demonstrations of Affection 3. Making Major Decisions 4. Sex Relations 5. Conventionality (correct or proper behavior) 6. Career Decisions Almost Occasionally Frequently Always Disagree Disagree Agree Almost Always Disagree Always Disagree 5 4 3 2 1 0 5 4 3 2 1 0 5 4 3 2 1 0 5 4 3 2 1 0 5 4 3 2 1 0 5 4 3 2 1 0 More often Occasionally than Not All the Time 7. How often do you discuss or have you considered divorce, separation or terminating your relationship? 8. How often do you and your partner quarrel? 9. Do you ever regret that you married (or live together)? 10. How often do you and your mate ―get on each other‘s nerves‖? Most of the Time 0 1 2 0 1 0 0 Rarely Never 3 4 5 2 3 4 5 1 2 3 4 5 1 2 3 4 5 242 Every Day Occasionally Rarely Nev er 4 11. Do you and your mate engage in outside interests together? Almost Every Day 3 2 1 0 Never 12. Have a stimulating exchange of Ideas 13. Work together on a project 14. Calmly discuss something Less than Once a Month Once or Twice a Month 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 243 Once or Twice a Week Once a Day More Often APPENDIX F SOFTA-S (CLIENT VERSION) Evaluate the following phrases and indicate your level of agreement by circling the appropriate number: 1. What happens in therapy can solve our problems. 2. The therapist understands me. 3. The therapy sessions help me open up (share my feelings, try new things…) 4. All my family members who come for therapy want the best for our family and to resolve our problems. 5. It is hard for me to discuss with the therapist what we should work on in therapy. 6. The therapist is doing everything possible to help me. 7. I feel comfortable and relaxed in the therapy sessions. 8. All of us who come for therapy sessions value the time and effort we all put in. 9. The therapist and I work together as a team. 10. The therapist has become an important person in my life. 11. There are some topics I am afraid to discuss in therapy. 12. Some members of the family don‘t agree with others about the goals of the therapy. 13. I understand what is being done in therapy. 14. The therapist lacks the knowledge and skills to help me. 15. At times I feel on the defensive in therapy. 16. Each of us in the family helps the others get what they want out of therapy. Not at all 1 A little 2 1 1 3 4 Very much 5 2 2 3 3 4 4 5 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 244 Moderately A lot APPENDIX G SOFTA-S (Therapist Version) Evaluate the following phrases and indicate your level of agreement by circling the appropriate number: 1. What happens in therapy can solve this family‘s problems. 2. I understand this family. 3. The therapy sessions are helping family members open up (share my feelings, try new things…) 4. All of the family members who are coming for therapy want the best for the family and to resolve their problems. 5. It is hard for me and the family to discuss what we should work on in therapy. 6. I am doing everything possible to help this family. 7. Family members feel comfortable and relaxed in the therapy sessions. 8. All of those who come for therapy sessions value the time and effort the others put in. 9. The family and I are working together as a team. 10. I have become an important person in this family‘s life. 11. There are some topics that the family members are afraid to discuss in therapy. 12. Some members of the family don‘t agree with others about the goals of the therapy. 13. What this family and I are doing in therapy makes sense to me. 14. I lack the knowledge and skills to help this family. 15. At times some family members feel on the defensive in therapy. 16. Each person in the family helps the others get what they want out of therapy. Not at all 1 A little 2 Moderately 3 A lot 4 Very much 5 1 1 2 2 3 3 4 4 5 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 245 APPENDIX H SOFTA-O (Client Version) Please read the definition of each of the following four constructs. Then, on the coding pages, identify the family members to be rated in the top row. As you observe the session, mark each behavior that occurs in the appropriate column. At the conclusion of the session, use these marks to make a judgment about each family member‘s alliance on Engagement, Emotional Connection, and Safety. Rate the entire family structure on Shared Sense of Purpose. Use the guidelines in the training manual to go from check marks to ratings. Note that items in italics reflect a lack of engagement, poor emotional connection, a lack of shared sense of purpose, or a lack of safety. Use the following ordinal scale: -3 = Extremely problematic -2 = Moderately problematic -1 = Somewhat problematic 0 = Unremarkable or neutral +1= Somewhat strong +2= Moderately strong +3= Extremely strong Engagement in the Therapeutic Process: Safety Within the Therapeutic System: The client viewing treatment as meaningful; a sense of being involved in therapy and working together with the therapist, that therapeutic goals and tasks in therapy can be discussed and negotiated with the therapist, that taking the process seriously is important, that change is possible. The client viewing therapy as a place to take risks, be open, flexible; a sense of comfort and an expectation that new experiences and learning will take place, that good can come from being in therapy, that conflict within the family can be handled without harm, that one need not be defensive. Emotional Connection to the Therapist: Shared Sense of Purpose Within the Family: The client viewing the therapist as an important person in her/his life, almost like a family member; a sense that the relationship is based on affiliation, trust, caring, and concern; that the therapist genuinely cares and ―is there‖ for the client, that he/she is on the same wavelength with the therapist (e.g., similar life perspectives, values), that the therapist‘s wisdom and expertise are valuable. Family members seeing themselves as working collaboratively to improve family relations and achieve common family goals; a sense of solidarity in relation to the therapy (―we‘re in this together‖); that they value their time with each other in therapy; essentially, a felt unity within the family in relation to the therapy. 246 Engagement in the Therapeutic Process: Client indicates agreement with the therapist goals. Client describes or discusses a plan for improving the situation. Client introduces a problem for discussion. Client agrees to do homework assignments. Client indicates having done homework or seeing it as useful. Client expresses optimism or indicates that positive change has taken place. Client complies with therapist‘s request for an enactment. Client leans forward. Client mentions the treatment, the therapeutic process, or a specific session. Client expresses feeling “stuck,” questions the value of therapy, or state that therapy is not/has not been helpful. Client shows indifference about the tasks or process of therapy (e.g., paying lip service, “I don’t know,” tuning out). Rate Engagement in the Process for each family member: 247 _____ Family Member _____ ______ ___ _____ _____ ______ ___ _____ _____ _____ _____ ______ ______ ___ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ -3 -2 -1 0 +1 +2 +3 Emotional Connection to the Therapist: Family Member Client shares a lighthearted moment or joke with the therapist. Client verbalizes trust in the therapist. Client expresses interest in the therapist‘s personal life. Client indicates feeling understood or accepted by the therapist. Client expresses physical affection or caring for the therapist. Client mirrors the therapist‘s posture. Client avoids eye contact with the therapist. Client refuses or is reluctant to respond to the therapist. Client has hostile or sarcastic interactions with the therapist. Client comments on the therapist’s incompetence or inadequacy. Rate Emotional Connection for each family member: 248 _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ _____ _____ ______ ______ ___ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ -3 -2 -1 0 +1 +2 +3 Safety within the Therapeutic System: Client implies or states that therapy is a safe place. Client varies his/her emotional tone during session. Client shows vulnerability (e.g., discusses painful feelings, cries). Client has an open upper body posture. Client reveals a secret or something that other family members didn‘t know. Client encourages another family member to ―open up‖ or to tell the truth. Client directly asks other family members for feedback about herself/himself as a person. Client expresses anxiety nonverbally (e.g., taps or shakes). Client protects self in nonverbal manner (e.g., crosses arms over chest, doesn’t take off jacket or put down purse, sits far away from group, etc.). Client refuses or is reluctant to respond when directly addressed by another family member. Client responds defensively to another family member. Client makes an uneasy/anxious reference to the camera, observation, supervisor, or research procedures. Rate Safety Within the Therapeutic System for each family member: 249 _____ _____ Family Member _____ ______ _____ ______ ___ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ -3 -2 -1 0 +1 +2 +3 Shared Sense of Purpose Within the Family: Family members offer to compromise. Family members share a joke or a lighthearted moment with each other. Family members ask each other for their perspective. Family members validate each other‘s point of view. Family members mirror each other‘s body posture. Family members avoid eye contact with each other. Family members blame each other. Family members devalue each other’s opinions or perspectives. Family members try to align with the therapist against each other. Client makes hostile or sarcastic comments to family members. Family members disagree with each other about the value, purpose, goals, or tasks of therapy or about who should be included in the sessions. Rate Shared Sense of Purpose Within the Family for each family member: 250 _____ Family Member _____ ______ ___ _____ _____ ______ ___ _____ _____ _____ _____ _____ _____ _____ _____ _____ _____ ______ ______ ______ ______ ______ ___ ___ ___ ___ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ _____ _____ ______ ___ -3 -2 -1 0 +1 +2 +3 APPENDIX I SOFTA-O (Therapist Version) Please read the definition of each of the following four constructs. As you observe the session, mark each behavior as it occurs. At the conclusion of the session, use these marks to make a judgment about the therapist‘s contribution to Engagement, Emotional Connection, Safety, and the family‘s Shared Sense of Purpose. Use the guidelines in the training manual to go from check marks to ratings. Note that items in italics reflect negative contributions to the engagement, emotional connection, a shared sense of purpose, or safety. Use the following ordinal scale: -3 = Extremely problematic -2 = Moderately problematic -1 = Somewhat problematic 0 = Unremarkable or neutral +1= Somewhat strong +2= Moderately strong +3= Extremely strong Engagement in the Therapeutic Process: Safety Within the Therapeutic System: The client viewing treatment as meaningful; a sense of being involved in therapy and working together with the therapist, that therapeutic goals and tasks in therapy can be discussed and negotiated with the therapist, that taking the process seriously is important, that change is possible. The client viewing therapy as a place to take risks, be open, flexible; a sense of comfort and an expectation that new experiences and learning will take place, that good can come from being in therapy, that conflict within the family can be handled without harm, that one need not be defensive. Emotional Connection to the Therapist: Shared Sense of Purpose Within the Family: The client viewing the therapist as an important person in her/his life, almost like a family member; a sense that the relationship is based on affiliation, trust, caring, and concern; that the therapist genuinely cares and ―is there‖ for the client, that he/she is on the same wavelength with the therapist (e.g., similar life perspectives, values), that the therapist‘s wisdom and expertise are valuable. Family members seeing themselves as working collaboratively to improve family relations and achieve common family goals; a sense of solidarity in relation to the therapy (―we‘re in this together‖); that they value their time with each other in therapy; essentially, a felt unity within the family in relation to the therapy. 251 Therapist’s Contributions to Engagement in the Therapeutic Process: Therapist explains how therapy works. *Therapist ask clients what they want to talk about in session. *Therapist encourages clients to articulate their goals for therapy. Therapist asks clients whether they are willing to do a specific in-session task (e.g., enactment). *Therapist asks clients whether they are willing to follow a specific suggestion or do a specific homework assignment. *Therapist expresses optimism or note that a positive change has taken place or can take place. Therapist pulls in quiet clients (e.g., by deliberately leaning forward, calling them by name, addressing them specifically). Therapist asks if the clients have any questions. Therapist praises client motivation for engagement or change. Therapist defines therapeutic goals or imposes tasks or procedures without asking the clients for their collaboration. Therapist argues with the clients about the nature, purpose, or value of therapy. Therapist shames or criticizes how clients did (or did not do) a prior homework assignment. Rate therapist contribution to Engagement: _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ -3 252 -2 -1 0 +1 +2 +3 Therapist’s Contributions to Emotional Connection: *Therapist shares a lighthearted moment or joke with the clients. *Therapist expresses interest in the clients apart from the therapeutic discussion at hand. *Therapist expresses caring or touches clients affectionately yet appropriately (e.g., handshake, pat on head). Therapist discloses his or her personal reactions or feelings toward the clients or the situation. Therapist remarks on or describes how his or her values or experiences are similar to clients‘. Therapist (verbally or nonverbally) expresses empathy for the clients‘ struggle (e.g., ―I know this is hard,‖ ―I feel your pain,‖ crying with the client). Therapist reassures or normalizes a client‘s emotional vulnerability (e.g., crying, hurt feelings). *Therapist has hostile, sarcastic, or critical interactions with clients. Therapist does not respond to clients’ expressions of personal interest or caring for him or her. Rate therapist contribution to Emotional Connection: _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ -3 253 -2 -1 0 +1 +2 +3 Therapist’s Contributions to the Safety Within the Therapeutic System: *Therapist acknowledges that therapy involves taking risks or discussing private matters. Therapist provides structure and guidelines for safety and confidentiality. *Therapist invites discussion about intimidating elements in the therapeutic context (e.g., recording equipment, reports to third parties, treatment team observation, one-way mirror, research, etc.). Therapist helps clients to talk truthfully and not defensively with each other. Therapist attempts to contain, control, or manage overt hostility between clients. Therapist actively protects on family member from another (e.g., from blame, hostility, or emotional intrusiveness). Therapist changes the topic to something pleasurable or non-anxiety arousing (e.g., small talk about the weather, room décor, TV shows, etc.) when there seems to be tension or anxiety. Therapist asks one client (or a subgroup of clients) to leave the room in order to see one client alone for a portion of the session. Therapist allows family conflict to escalate to verbal abuse, threats, or intimidation. Therapist does not attend to overt expression of client vulnerability (e.g., crying, defensiveness). Rate therapist contribution to Safety within the Therapeutic System: _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ -3 254 -2 -1 0 +1 +2 +3 Therapist Contributions to a Shared Sense of Purpose Within the Family: *Therapist encourages clients to compromise with each other. *Therapist encourages clients to ask each other for their perspective. *Therapist praises clients for respecting each other‘s point of view. Therapist emphasizes commonalities among clients‘ perspectives on the problem or solution. Therapist draws attention to clients‘ shared values, experiences, needs, or feelings. Therapist encourages clients to show caring, concern, or support for each other. *Therapist encourages clients to ask each other or feedback. *Therapist fails to intervene when family members argue with each other about the goals, value, or need for therapy. *Therapist fails to address one client’s stated concerns by only discussing another client’s concerns. Rate therapist contribution to a Shared Sense of Purpose Within the Family: _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ -3 255 -2 -1 0 +1 +2 +3 APPENDIX J THERAPIST INTERVIEW I am going to ask you some questions regarding your understanding and monitoring of the therapeutic alliance with your couple. Some questions will relate to your general knowledge of the alliance, while others will focus specifically on the four dimensions that have been empirically validated as important elements of therapeutic build and maintain alliance with couples in reference to specific skills, strategies, and experiences. First, I would like to start by getting some general information about your understanding of the therapeutic alliance and how you attend to it in sessions with your couple. 1. What is your understanding of the therapeutic alliance? (Probes: How do you define the alliance? How does working with couples differ from individuals?) Now, I am going to ask you questions specific to your relationship with the couple you are currently seeing. 2. How do you feel about the relationship that you and your clients have established? (Probes: How do you know that there is a positive/negative alliance between you and the couple?) 3. What specific therapeutic elements and techniques have you used to build therapeutic alliance with your couple? (Probes: What strategies do you use to promote alliance? What are the specific skills/behaviors you employ to build an alliance? If there is poor alliance or rupture, what do you do to repair it?) 4. How do you build the relationship with both individuals of the couple as well as the couple as a whole? (Probes: How do you know when there is a poor alliance or rupture between you and the couple, or an individual in the couple? How do you notice? How does treatment feel different?) 5. How do you think your relationship with your clients‘ changes from session to session? (Probes: How do you monitor the alliance during and in between sessions? Do you ever openly discuss the therapeutic relationship with the couple? Are you aware of the alliance most of the time or only at certain points? What has caused changes in the alliance?) 256 APPENDIX K CLIENT INTERVIEW I am going to ask you some questions about your overall relationship with your therapist. These questions will focus on your engagement in the therapeutic process, feeling of safety within the therapeutic process, emotional connection to the therapist, and shared sense of purpose within your family. 1. How do you feel about the relationship that you and your therapist have established? (Probes: How was your relationship established? What did your therapist do to promote the relationship? What behaviors helped build the relationship?) 2. What type of therapist behavior/skills/characteristics do you feel were instrumental in developing the relationship between you and your therapist? (Probes: What did your therapist do to promote the relationship? What behaviors helped build the relationship? 3. Can you tell me about any times that either of you feel more or less connected with your therapist as individuals? As a couple? (Probes: Were there ever times in which you did not feel united? If so, how did your therapist help to promote this?) 4. How do you think your relationship with your therapist changes from session to session? (Probes: What causes changes in your relationship? Are you aware of the relationship at all times or only at certain points?) 257 REFERENCES 258 REFERENCES Ackerman, S.J., & Hilsenroth, M.J. (2003). A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clinical Psychology Review, 23, 1-33. Alexander, J. E, Holtzworth-Munroe, A., & Jameson, P. (1994). The process and outcome of marital and family therapy: Research review and evaluation. In A. Bergin & S. Garfield (Eds.), Handbook of Psychotherapy and Behavior Change. (pp. 595-612). New York: Wiley. Allgood, S. M., & Crane, D. R. (1991). Predicting marital therapy dropouts. Journal of Marital and Family Therapy, 17, 73-79. Anderson, S. R. & Johnson, L. N. (2010). A dyadic analysis of the between- and within-system alliances on distress. Family Process, 49(2), 220-220-35. Angus, L. & Kagan, F. (2007). Empathic relational bonds and personal agency in psychotherapy: Implications for psychotherapy supervision, practice, and research. Psychotherapy: Theory, Research, Practice, Training, 44, 371-377. Anker, M. G., Owen, J., Duncan, B. L., & Sparks, J. A. (2010). The alliance in couple therapy: Partner influence, early change, and alliance patterns in a naturalistic sample. Journal of Consulting and Clinical Psychology, 78(5), 635-645. doi:10.1037/a0020051 APA Presidential Task Force on Evidence-Based Practice. (2006). Evidence-based practice in psychology. American Psychologist, 61, 271–285. Bachelor, A. & Salame, R. (2000).Participants‘ perceptions of dimensions of the therapeutic alliance over the course of therapy. Journal of Psychotherapy Practice and Research, 9, 39–53. Baldwin, S.A., Wampold, B.E., & Imel, Z.E. (2007). Untangling the alliance-outcome correlation: Exploring the relative importance of therapist and patient variability in the alliance. Journal of Consulting and Clinical Psychology, 75, 842-852. Barber, J. P., Connolly, M. B., Crits-Christoph, P., Gladis, L., & Siqueland, L. (2000). Alliance predicts patients‘ outcome beyond in-treatment change in symptoms. Journal of Consulting and Clinical Psychology, 68, 1027-1032. Barber, J.P., Luborsky, L., Crits-Christoph, P., Thase, M.E., Weiss, R., Frank, L., et al. (1999). Therapeutic alliance as a predictor of outcome in treatment of cocaine dependence. Psychotherapy research, 9, 54-73. 259 Beck, M., Friedlander, M.L., & Escudero, V. (2006). Three perspectives of clients‘ experiences of the therapeutic alliance: A discovery-oriented investigation. Journal of Marital and Family Therapy, 32, 355-368. Benner, P. (1985). Quality of life: A phenomenological perspective on explanation, prediction, and understanding in nursing science. Advances in Nursing Science, 8(1), 1-14. Berg, B., & Rosenblum, N. (1977). Fathers in family therapy: A survey of family therapists. Journal of Marriage and Family Counseling, 3, 85-91. Berger, M. (1979). Men‘s new family roles: Some implications for therapists. Family Coordinator, 28, 638-646. Bischoff, M.M. & McBride, A. (1996). Client perceptions of couples and family therapy. The American Journal of Family Therapy, 24,117-128. Retrieved from http://ezproxy.msu.edu/login?url=http://search.proquest.com/docview/61565515?account id=12598 Blow, A.J. & Sprenkle, D.H. (2001). Common factors across theories of marriage and family therapy: A modified delphi study. Journal of Marital and Family Therapy, 27, 385-401. Blow, A.J., Morrison, N.C., Tamaren, K., Wright, K. Schaafsma, M., & Nadaud, A. (2009). Change processes in couple therapy: An intensive care analysis of one couple using a common factors lens. Journal of Marital and Family Therapy, 35, 3, 350-368. Blow, A.J., Sprenkle, D.H., & Davis, S.D. ( 2007). Is who delivers the treatment more important than the treatment itself? The role of the therapist in common factors. Journal of Marital and Family Therapy, 33, 298-318. Blow, A.J., Timm. T.M., & Cox, R. The role of therapist in therapeutic change: Does therapist gender matter? Journal of Feminist Family Therapy: An International Forum, 20, 66-86. Bordin, E.S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, research, and practice, 16, 252-260. Bordin, E. S. (1994). Theory and research on the therapeutic working alliance: New Directions. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research and practice (pp. 13-37). New York: John Wiley. Bourgeois, L., Sabourin, S., & Wright, J. (1990). Predictive validity of therapeutic alliance in group marital therapy. Journal of Consulting and Clinical Psychology, 58, 608–613. Bowlby, J. (1969). Attachment and Loss: Vol. 1: Attachment. NY: Basic Books. 260 Brown, P.D., & O‘Leary, D.K. (2000). Therapeutic alliance: Predicting continuance and success in group treatment for spouse abuse. Journal of Consulting and Clinical Psychology, 68, 340-345. Busby, D., Christensen, C., Crane, D., & Larson, J. (1995). A revision of the dyadic adjustment scale for use with distressed and nondistressed couples: Construct hierarchy and multidimensional scales. Journal of Marital and Family Therapy, 21, 289-308. Callahan, J. L., Aubuchon-Endsley, N., Borja, S. E., & Swift J. K. (2009). Pretreatment expectancies and premature termination in a training clinic environment. Training and Education in Professional Psychology, 3, 11–119. Castonguay, L.G., Constantino, M.J., & Holtforth, M.G. (2006). The working alliance: where are we and where should we go? Psychotherapy: Theory, Research, Practice, Training, 43, 271-279. Catherall, D. (1984). The therapeutic alliance in individual, couple, and family therapy. Unpublished doctoral dissertation, Northwestern University, Evanston, IL. Christensen, L. L., Russell, C. S., Miller, R. B., & Peterson, C. M. (1998). The process of change in couples therapy: A qualitative investigation. Journal of Marital and Family Therapy, 24, 177-88. Retrieved from http://ezproxy.msu.edu/login?url=http://search.proquest.com/docview/220979763?accoun tid=12598 Conners, G.J., Carroll, K.M., DiClemente, C.C., Longabaugh, R. & Donovan, D.M. (1997). The therapeutic alliance and it relationship to alcoholism treatment participation and outcome. Journal of Consulting & Clinical Psychology, 65, 510-514. Constantino, M. J., Schwaiger, E. M., Smith, J. Z., DeGeorge, J., McBride, C., Ravitz, P., & Zuroff, D. C. (2010). Patient interpersonal impacts and the early therapeutic alliance in interpersonal therapy for depression. Psychotherapy: Theory, Research, Practice, Training, 47, 418-424. doi:10.1037/a0021169 Creswell, J.W. (2003). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Thousand Oaks, CA: Sage. Crits-Christoph, P., Cooper, A., & Luborsky, L. (1988). The accuracy of therapists‘ interpretations and the outcome of dynamic psychotherapy. Archives of general psychiatry, 56, 490-495. Crits-Christoph, P., Connolly-Gibbons, M. B., Crits-Christoph, K., Narducci, J., Schamberger, M., & Gallop, R. (2006). Can therapists be trained to improve their alliances? A preliminary study of alliance-fostering psychotherapy. Psychotherapy Research, 16, 268– 281. 261 Crits-Christoph, P., Gibbons, M.B.C., & Hearon, B. (2006). Does the alliance cause good outcome? Recommendations for future research on the alliance. Psychotherapy: Theory, Research, Practice, Training, 43, 280-285. Dailey, P. (2004). Gender differences in client self-disclosure: The influence of gender role identification and working alliance. Dissertation Abstracts International. (UMI No. 3129012). Denscombe, M. (1998) The Good Research Guide: For Small-Scale Social Research Projects. Buckingham, Open University Press. Denzin, N. K. (1994). The art and politics of interpretation. In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 500-515), Thousand Oaks, CA: Sage. Duncan, B.L., & Miller, S.D. (2000). The Heroic Client: Doing Client Directed, OutcomeInformed Therapy. San Francisco: Jossy-Bass. Duncan, B.L. & Moynihan, D. (1994). Applying outcome research: Intentional utilization of the client frame of reference. Psychotherapy, 31, 294-301. Dunkle, J.H. & Friedlander, M.L. (1996). Contribution of therapist experience and personal characteristics to the working alliance. Journal of Counseling Psychology, 43, 456-460. Eames, V. & Roth, A. (2000). Patient attachment orientation and the early working alliance: A study of patient and therapist reports of alliance quality and ruptures. Psychotherapy Research, 10, 421-434. Elkin, I., Falconnier, L., Martinovich, Z, & Mahoney, C. (2006). Therapist effects in the National Institute of Mental Health Treatment of Depression Collaborative Research Program. Psychotherapy Research, 16, 144-160. Escudero, V., Friedlander, M.L., Varela, N., & Abascal, A. (2008). Observing the therapeutic alliance in family therapy: Associations with participants‘ perceptions and therapeutic outcomes. Journal of Family Therapy, 30, 194-214. Fenton, L. R., Cecero, J. J., Nich, C., Frankforter, T. L., & Carroll, K. M. (2001). Perspective is everything: The predictive validity of working alliance instruments. Journal of Psychotherapy Practice and Research, 10, 262–268. Fitzpatrick. M.R., Iwakabe, S., & Stalikas, A. (2005). Perspective divergence in the working alliance. Psychotherapy Research, 15, 69-79. Flicker, S. M., Turner, C. W., Waldron, H. B., Brody, J. L., & Ozechowski, T. J. (2008). Ethnic background, therapeutic alliance, and treatment retention in functional family therapy with adolescents who abuse substances. Journal of Family Psychology, 22,167-170. doi:10.1037/0893-3200.22.1.167 262 Florsheim, P., Shotorbini, S., Guest-Warnick, G., Barratt, T., & Hwang, W. (2000). Role of the working alliance in treatment of delinquent boys in community-based programs. Journal of Clinical Child Psychology, 29, 94-107. Frank, J.D. & Frank, J.B. (1991). Persuasion and Healing: A Comparative Study of Psychotherapy,3rd ed. Baltimore: Johns Hopkins University Press; 1991. Freud, S. (1940). The dynamics of transference. In J. Strachey (Ed.) The Standard Edition of the Complete Psychological Works of Sigmund Freud. (Vol. 12) (pp. 122-144). London: Hogarth Press. Friedlander, M.L., Escudero, V., & Heatherington, L. (2001). SOFTA-o for Clients. Unpublished instrument. Available from SOFTA Web site, http://www.softa-soatif.net Friedlander, M.L., Escudero, V., Haar, N., & Higham, J. (2005). Introducing the SOFTA-s (selfreport version of the System for Observing Family Therapy Alliances). Poster session presented at the annual conference of the Society for Psychotherapy Research, Montreal, Quebec, Canada. Friedlander, M.L., Escudero, V., & Heatherington, L. (2006a). Therapeutic Alliance in Couple and Family Therapy: An Empirically Informed Guide to Practice. Washington, D.C.: American Psychological Association Friedlander, M.L., Escudero, V., Horvath, A.O., Heatherington, L., Cabero, A., & Martens, M.P. (2006b). System for observing family therapy alliance: A tool for research and practice. Journal of Counseling Psychology, 53, 214-224. Friedlander, M.L., Lambert, M. Muniz de la Pena, C. (2008). A step toward disentangling the alliance/improvement cycle in family therapy. Journal of Counseling Psychology, 55, 118-124. Friedlander, M.L., Lehman, P., McKee, M., Field, N., & Cutting, M. (2001). The alliance in couples and family therapy: Initial stages in the development of an observational rating system. Paper presented at the conference of the North American Society for Psychotherapy Research, Puerto Vallarta, Mexico. Garfield, R. (2004). The therapeutic alliance in couples therapy: Clinical considerations. Family Process, 43(4), 457-465. Gaston (1990). The concept of the alliance and its role in psychotherapy: Theoretical and empirical considerations. Psychotherapy, 27, 143-153. 263 Gaston, L., Piper, W.E., Debbane, E.G., Bienvenu, J.P., & Garant, J. (1994). Alliance and technique for predicting outcome in short-and long-term analytic psychotherapy. Psychotherapy Research, 42, 121-135. Gaston, L., Thompson, L., Gallager, D., Cournoyer, L.G., & Gagnon, R. (1998). Alliance, technique, and their interactions in predicting outcome of behavioral, cognitive, and brief dynamic therapy. Psychotherapy research, 8, 190-209. Gelso, C.J. & Carter, J.A. (1994). Components of the psychotherapy relationship: Their interaction and unfolding during treatment. Journal of Counseling Psychology, 41, 296306. Gelso, C. J., & Hayes, J. A. (1998). The psychotherapy relationship: Theory, research, and practice. New York: Wiley. Gilgun, J. (1992). Definitions, methodologies, and methods in qualitative family research. In J. Gilgun, K. Daly, & G. Handel (Eds.), Qualitative Methods in Family Research (pp. 22– 40). Newbury Park, CA: Sage Publications. Glebova, T., Bartle-Haring, S., Gangamma, R ., Knerr, M., Delaney, R. O., Meyer, K., . . . . Grafsky, E. (2011). Therapeutic alliance and progress in couple therapy: Multiple perspectives. Journal of Family Therapy, 33(1), 42-42-65. doi:10.1111/j.14676427.2010.00503.x Glesne, C., & Peshkin, A. (1992). Becoming Qualitative Researchers: An Introduction. White Plains, NY: Longman. Golden, B.R., & Robbins, S.B. (1990). The working alliance within time-limited therapy. Professional Psychology: Research and Practice, 21, 476-481. Greenberg, L.S. & Pinsof, W.M. (1986). A Psychotherapeutic Process: A Research Handbook. NY: Guilford Press. Greenson, R.R. (1965). The working alliance and the transference neuroses. Psychoanalysis Quarterly, 34, 155-181. Guba, E.G. (1981). Criteria for assessing the trustworthiness of naturalistic inquiries. Educational Communication and Technology Journal , 29, 75–91. Hartley, D.E. & Strupp, H.H. (1983). The therapeutic alliance: Its relationship to outcome in brief psychotherapy. In J. Masling‘s (Ed.) Empirical studies in analytic theories. (pp. 137). Hillside, NJ: Earlbaum. Heatherington, L. & Friedlander, M.L. (1990). Couple and family therapy alliance scales: Empirical considerations. Journal of Marital and Family Therapy, 16, 299-306. 264 Helmeke, K. B. and Sprenkle, D. H. (2000). Clients‘ perceptions of pivotal moments in couples therapy: A qualitative study of change in therapy. Journal of Marital and Family Therapy, 26, 469–483. doi: 10.1111/j.1752-0606.2000.tb00317.x Henry W R. & Strupp, H. (1994) The therapeutic alliance as an interpersonal process. In A. O. Horvath & L. S. Greenberg (Eds.) The Working Alliance: Theory, Research and Practice, New York: Wiley. Henry, W.P., Strupp, H.H., Butler, S.F., Schacht, T.E., & Binder, J.L. (1993). Effects of training in time-limited dynamic psychotherapy: Changes in therapist behavior. Journal of Consulting and Clinical Psychology, 61, 434-440. Herman, S.M. (1998). The relationship between therapist-client modality similarity and psychotherapy outcome. The Journal of Psychotherapy Practice and Research, 7, 56-64. Hersoug, A. G., Høglend, P., Monsen, J. T, & Havik, O. E. (2001). Quality of working alliance in psychotherapy: Therapist variables and patient/therapist similarity as predictors. Journal of Psychotherapy Practice and Research,10, 205–216. Hilsenroth M. J., Peters E. J., & Ackerman S. J. (2004). The development of therapeutic alliance during psychological assessment: Patient and therapist perspectives across treatment. Journal of Personality Assessment, 83, 332–344. Hight, C.T. (1997). The effects of the perceived therapeutic alliance on outcome in couples therapy. (Doctoral dissertation.). Retrieved from: http://ezproxy.msu.edu/login?url=http://search.proquest.com/docview/304355817?accoun tid=12598 Holstein, James A. & Gubrium, J.F. (1995). The Active Interview. Newbury Park, CA: Sage. Holtzworth-Munroe, A., Jacobson, N.S., DeKlyen, M., & Whisman, M.A. (1989). Relationship between behavioral marital therapy outcome and process variables. Journal of Consulting and Clinical Psychology, 57, 658-662. Horvath, A. O. (1994). Empirical validation of Bordin‘s pantheoretical model of the alliance: The working alliance inventory perspective. In A. O. Horvath & L. S. Greenberg (Eds.), The working alliance: Theory, research, and practice (pp. 109–128). Oxford, England: Wiley. Horvath, A. O. (1994). Research on the alliance. In A. O. Horvath & L. S. Greenberg (Eds.), The Working Alliance: Theory, Research and Practice (pp. 259-287). New York: Wiley. Horvath, A. O. (2001). The alliance. Psychotherapy: Theory, Research, Practice, Training, 38, 365–372. 265 Horvath, A. O., & Bedi, R. P. (2002). The alliance. In J. C. Norcross (Ed.), Psychotherapy Relationships that Work: Therapist Contributions and Responsiveness to Patients (pp. 37–70). New York: Oxford University Press. Horvath, A., Gaston, L., & Luborsky, L. (1993). The therapeutic alliance and its measures. In N. Miller, L. Luborsky, J. Barber, & J. Docherty‘s (Eds.) Psychodynamic Treatment Research (pp. 247-273). NY: Basic Books. Horvath, A. O. & Greenberg, L. S. (1989). Development and validation of the Working Alliance Inventory. Journal of Counseling Psychology, 36, 223. Horvath, A.O. & Luborsky, L. (1993). The role of therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology, 38, 139-149. Horvath, A.O. & Marx, R.W. (1991). The development and decay of the working alliance during time-limited counseling. Canadian Journal of Counseling, 24, 240-259. Horvath, A. O., & Symonds, B. D. (1991). Relation between working alliance and outcome in psychotherapy: A meta-analysis. Journal of Counseling Psychology, 36, 139–149. Johnson, N.G. (2005). Women helping men: Strengths of and barriers to women therapists working with men clients. In G.R. Brooks & G.E. Good (Eds.), The New Handbook of Psychotherapy and Counseling with Men. (pp. 696-718). San Francisco: Jossey-Boss. Johnson, S.M. & Talitman, E. (1997). Predictors of success in emotionally focused marital therapy. Journal of Marital and Family Therapy, 23, 135-152. Kaufman, M. (2000). Effects of therapist self-monitoring on therapeutic alliance and subsequent therapeutic outcome. The Clinical Supervisor, 19, 41-59. Kerr, M. E. & Bowen, M. (1988). Family Evaluation: An Approach based on Bowen Theory. New York: W.W. Norton & Company. Kim, D. M., Wampold, B. E., & Bolt, D. M. (2006). Therapist effects in psychotherapy: A random effects modeling of the NIMH TDCRP data. Psychotherapy Research, 16, 161172. Kivlighan, D.M. (2007). Where is the relationship in research on the alliance? Two methods for analyzing dyadic data. Journal of Counseling Psychology, 54, 423–433. Kivlighan, D.M., Patton, M.J., Foote, D. (1998). Moderating effects of client attachment on the counselor experience-working alliance relationship. Journal of Counseling Psychology, 45, 272-278. 266 Kivlighan, D. M., & Shaughnessy, P. (2000). Patterns of working alliance development: A typology of client‘s working alliance ratings. Journal of Counseling Psychology,47, 362371. Kivlighan, D. M., Jr., & Shaughnessy, S. (1995). Analysis of the development of the working alliance using hierarchical linear modeling. Journal of Counseling Psychology, 42, 338349. Knerr, M. and Bartle-Haring, S. (2010) Differentiation, perceived stress and therapeutic alliance as key factors in the early stage of couple therapy. Journal of Family Therapy, 12, 94– 118. Knerr, M., Bartle-Haring, S., McDowell, T., Adkins, K., Delaney, R.O. & Gangamma, R. (2011). The impact of initial factors on therapeutic alliance in individual and couples therapy. Journal of Marital and Family Therapy, 37 (2),182–199. doi: 10.1111/j.17520606.2009.00176.x Knobloch-Fedders, L.M., Pinsof, W.M., & Mann, B.J. (2004). The formation of the therapeutic alliance in couple therapy. Family Process, 43, 425-442 Knobloch-Fedders, L., Pinsof, W.M., Mann, B.J. (2007). Therapeutic alliance and treatment progress in couple psychotherapy. Journal of Marital and Family Therapy, 33, 245-257. Kramer, U., Roten, Y., Beretta, V., Michel, B., & Despland, J.N. (2008). Patient‘s and therapist‘s view of early alliance building in dynamic psychotherapy: Patterns and relation to outcome. Journal of Counseling Psychology, 55, 89-95. Krippendorff, Klaus (2004). Content Analysis: An Introduction to its Methodology. Thousand Oaks, CA: Sage. Krupnik, J. L., Sotsky, S. M., Simmens, S., Moyer, J., Elkin, I., Watkins, J. & Pilkonis, P. A. (1996). The role of the therapeutic alliance in psychotherapy and pharmocotherapy outcome: findings in the National Institute of Mental Health treatment of depression Collaborative Research Program. The Journal of Consulting and Clinical Psychology, 64, 532-539. Leininger, M. M. (1985). Ethnography and ethnonursing: Models and modes of qualitative data analysis. In M. M. Leininger (Ed.), Qualitative Research Methods in Nursing (pp. 33-72). Orlando, FL: Grune & Stratton. Ligiero, D.P. & Gelso, C.J. (2002). Countertransference, attachment, and the working alliance: The therapist‘s contributions. Psychotherapy: Theory, Research, Practice, Training, 39, 3-11. Lincoln, Y.S. & Guba, E.G. (1985). Naturalistic Inquiry, Beverly Hills: Sage, 1985. 267 Lincoln, Y. S., & Guba, E., G. (2000). Paradigmatic controversies, contradictions and emerging confluences. In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of Qualitative Research nd (2 ed., pp. 163-188). Thousand Oaks, CA: Sage Publications, Inc. Luborsky, L. (1976). Helping alliances in psychotherapy. In J.L. Cheghorn‘s (Ed.) Successful psychotherapy. (pp. 92-116). NY: Brunner/Mazel. Luborsky, L. (1993). Recommendations for training therapists based on manuals for psychotherapy research. Psychotherapy, 30, 578-586. Luborsky, L., Crits-Christoph, P. Alexander, L., Margolis, M., & Cohen, M. (1983). Two helping alliance methods for predicting outcomes of psychotherapy: A counting signs vs. a global rating method. Journal of nervous and mental disease, 171, 480-491. Luborsky, L., Johnson, S. & McLellan, T. (1994). The Penn Helping Alliance Questionnaire (PAq-I): Its Composition and its Research Supports. Philadelphia: University of Pennsylvania. Luborsky, L., McLellan, T., Woody, G., O‘Brien, C. & Auerbach, A. (1985). Therapist success and its determinants. Archives of General Psychiatry, 42, 602-611. Mallinckrodt, B. (2000). Attachment, social competencies, social support and interpersonal process in psychotherapy. Psychotherapy Research, 10, 239-266. Mallinckrodt, B., Coble, H.M., Gantt, D.L. (1995). Toward differentiating client attachment from working alliance and transference. Journal of Counseling Psychology, 42, 320-322. Mamodhoussen, S., Wright, J., Tremblay, N., & Poitras-Wright, H. (2005). Impact of marital and psychological distress on therapeutic alliance in couples undergoing couple therapy. Journal of Marital and Family Therapy, 31, 159-169. Marmar, C., Gaston, L., Gallagher, D., & Thompson, L.W. (1989). Alliance and outcome in latelife depression. Journal of Nervous and Mental Disease, 177, 464-472. Marmar, C., Weiss, D.S., & Gaston, L. (1989). Toward the validation of the California Therapeutic Alliance Rating System. Psychological Assessment, 1, 46-52. th Marshall, C. & Rossman, G. (2006). Designing Qualitative Research, 4 Ed. London: Sage. Martin, D.J. (1998). Relation of the therapeutic alliance with outcome and other variables: A meta analytic review. Unpublished doctoral dissertation, Ohio University. Martin, D., Garske, J. & Davis, M. (2000). Relation of the therapeutic alliance with outcome and other variables: A meta-analytic review. Journal of Consulting and Clinical Psychology, 68, 438. 268 Marzieli, E. (1984). Three viewpoints on the therapeutic alliance scales similarities, differences, and associations with psychotherapy outcome. Journal of Nervous and Mental Disease, 172, 417-423. Marzieli, E., Marmar, C., & Krupnick, J. (1981). Therapeutic alliance scales: Development and relationship to psychotherapy outcome. American Journal of Psychiatry, 138, 361-364. Miles, M.B. & Huberman, A.M. (1994). Qualitative data analysis: An expanded source book nd (2 ed.). Thousand Oaks, CA: Sage. Minuchin, S. & Fishman, S.C. (1981). Family Therapy Techniques. Cambridge, MA: Harvard University Press. Morrow, S.L. (2005). Quality and trustworthiness in qualitative research in counseling psychology. Journal of Counseling Psychology, 52, 250-260. Nelson, M.L. & Neufeldt, S.A. (1996). Building on an empirical foundation: Strategies to enhance good practice. Journal of Counseling & Development, 74, 609-615. Orlinsky, D.E. & Howard, K.I. (1986). Process and outcome in psychotherapy. In S.L. Garfield rd & A.E. Bergin (Eds.), Handbook of Psychotherapy and Behavior Change (3 ed., pp. 311-381). NY: Wiley. rd Patton, M. Q. (2002). Qualitative evaluation and research methods (3 ed.). Thousand Oaks, CA: Sage Publications, Inc. Pinsof, W.M. (1988). The therapist-client relationship: An integrative systems perspective. Journal of Integrative and Eclectic Psychotherapy, 7, 303-313. Pinsof, W.M. (1992). Toward a scientific paradigm for family psychology: The integrative process systems perspective. Journal of Family Psychology, 5, 432-447. Pinsof, W.M. (1994). An integrative systems perspective on the therapeutic alliance: Theoretical, clinical and research implications. In A.O. Horvath & L.S. Greenberg (Eds.) The working alliance: Theory, research, and practice (pp. 173-195). New York: Wiley. Pinsof, W.M. & Catherall, D.R. (1986). The integrative psychotherapy alliance: Family, couple, and individual therapy scales. Journal of marital and family therapy, 12, 137-151. Pinsof, W.M. & Wynne, L.C. (2000). Toward process research: Closing the gap between family therapy practice and research. Journal of Marital and Family Therapy, 26, 1-8. Quinn, W.H., Dotson, D., & Jordan, K. (1997). Dimensions of therapeutic alliance and their associations with outcome in family therapy. Psychotherapy Research, 7, 429-438. 269 Rait, D.S. (1995). The therapeutic alliance in couples and family therapy: Theory in practice. In Session: Psychotherapy in Practice, 1, 59-72. Rait, D. S. (2000), The therapeutic alliance in couples and family therapy. Journal of Clinical Psychology, 56, 211–224. doi: 10.1002/(SICI)1097-4679(200002)56:2211::AIDJCLP73.0.CO;2-H Raytek, H. S., McGrady, B. S., Epstein, E. E., & Hirsch, L. S. (1999). Therapeutic alliance and the retention of couples in conjoint alcoholism treatment. Addictive Behaviors, 24, 317330. Rennie, D.L. (2004). Reflexivity and person-centered counseling. Journal of Humanistic Psychology, 44, 182-203. Ricker, M., Nystul, M., & Waldo, M. (1999). Counselors' and clients' ethnic similarity and therapeutic alliance in time-limited outcomes of counseling. Psychological Reports, 84, 674-676. Robbins, M. S., Liddle, H. A., Turner, C. W., Dakof, G. A., Alexander, J. F., & Kogan, S. M. (2006). Adolescent and parent therapeutic alliances as predictors of dropout in multidimensional family therapy. Journal of Family Psychology, 20, 108-116. doi:10.1037/0893-3200.20.1.108 Robbins, M., Turner, C., Alexander, J., & Perez, G. (2003). Alliance and dropout in family therapy for adolescents with behavior problems: Individual and systemic effects. Journal of Family Psychology, 17, 534-544. Rugel, R. P. (1997). Husband-Focused Marital Therapy: An Approach to Dealing with Marital Distress. Springfield, IL: Charles C. Thomas Safran, J.D. & Muran, J.C. (1998). The Therapeutic Alliance in Brief Psychotherapy. Washington, D.C.: American Psychological Association. Safran, J.D. & Muran, J.C. (2000). Negotiating the Therapeutic Alliance: A Relational Treatment Guide. NY: The Guilford Press. Saunders, S.M. (2001). Pretreatment correlates of the therapeutic bond. Journal of Clinical Psychology, 57, 1339-1352. Saunders, S.M., Howard, K.I., & Orlinsky, D.E. (1989). The therapeutic bond scales: Psychometric characteristics and relationship to treatment effectiveness. Psychological Assessment, 1, 323-330. Scher, M. (2005). Male therapist, male client: Reflections on critical dynamics. In G.R. Brooks & G.E. Good (Eds.), The New Handbook of Psychotherapy and Counseling with Men (Vol. 2, pp. 696-718). San Francisco: Jossey-Bass. 270 Sexton, H.L., Hembre, K., & Kvarme, G. (1996). The interaction of the alliance and therapy micorprocess: A sequential analysis. Journal of Consulting and Clinical Psychology, 64, 471-480. Sexton, T.L., Robbins, M.S., Hollimon, A.S., Mease, A.L. & Mayorga, C.C. (2003). Efficacy, effectiveness, and change mechanisms in couple and family therapy. In T.L. Sexton, G.R. Weeks, & M.S. Robbins (Eds.) Handbook of Family Therapy, (pp. 229-262). NY: Brunner-Routledge. Silverman, D. (2000). Doing Qualitative Research: A Practical Handbook. London: Sage. Skowron, E.A. & Friedlander, M.L. (1998). The differentiation of self inventory: Development and initial validation. Journal of counseling psychology, 45, 235-246. Sprenkle, D.H., & Blow, A.J. (2004). Common factors and our sacred models. Journal of Marital and Family Therapy, 30, 113-129. Sprenkle, D.H., & Blow, A.J. (2007). The role of the therapist as the bridge between common factors and therapeutic change: More complex than congruency with a worldview. Journal of Family Therapy, 29, 109-113. Stake, R.E. (1995). The Art of Case Study Research. Thousand Oaks, CA: Sage. Stiles, W.B., Glick, M.J., Osatuke, K., Hardy, G.E., Shapiro, D.A., Agnew-Davies, R., et al. (2004). Patterns of alliance development and the rupture-repair hypothesis: Are productive relationships U-shaped or V-shaped? Journal of Counseling Psychology, 1, 81-92. Strauss, A. & Corbin, J. (1990). Basics of Qualitative Research: Grounded Theory and Techniques. CA: Sage Publications. Strauss, A. & Corbin, J. (1998). Basics of Qualitative Research: Techniques, and Procedures nd for Developing Grounded Theory (2 edition). Thousand Oaks, CA: Sage Publications. Symonds, B.D. (1998). A measure of the alliance in couple therapy. Paper presented at the annual conference of the International Society for Psychotherapy Research, Snowbird, UT. Symonds, B.D. & Horvath, A.O. (2004). Optimizing thee alliance in couple therapy. Family Process, 43, 443-455. Taylor, S. J., & Bogdan, R. (1998). Introduction to qualitative research methods: A guidebook rd and resource (3 ed.). Hoboken, NJ,: John Wiley & Sons Inc. Thomas, S. E. G., Werner-Wilson, R. J., & Murphy, M. J. (2005). Influence of therapist and client behaviors on therapy alliance. Contemporary Family Therapy, 27(1), 19-35. 271 Tyron, G. S., & Kane, A. S. (1993). Relationship of working alliance to mutual and unilateral termination. Journal of Counseling Psychology, 40, 33-36. Vera, E. M., Speight, S. L., Mildner, C., & Carlson, H. (1999). Clients‘ perceptions and evaluations of similarities to and differences from their counselors. Journal of Counseling Psychology, 46, 277-283. Waldron, H.B., Turner, C.W., Barton, C., Alexander, J.F., & Cline, V.B. (1997). Therapist defensiveness and marital therapy process and outcome. American Journal of Family Therapy, 25, 233-243. Wampold, B. E. (2001). The Great Psychotherapy Debate: Models, Methods, and Findings. Mahwah, NJ: Erlbaum. Wampold, B. E. (2010). The research evidence for the common factors models: A historically situated perspective. In B. L. Duncan, S. D. Miller, B. E. Wampold, & M. A. Hubble nd (Eds.), The Heart and Soul of Change (2 ed.): Delivering what Works in Therapy (pp. 49-81). Washington, DC: American Psychological Association. Wampold, B. E., & Brown, G. S. (2005). Estimating therapist variability: A naturalistic study of outcomes in managed care. Journal of Consulting and Clinical Psychology, 73, 914-923. th Yin, R.K. (2009). Case Study Research: Design and Methods, 4 Ed. Los Angeles: Sage. 272