nu...“ an. n--~-.~u~.u..!~.y.~ ”tawny-- . .r V IlllilllllllllllllllllllllllllllllllllllllIllHHllllHllllll 31293 0183 354 This is to certify that the dissertation entitled THERAPIST AND CLIENT ADULT ATTACHMENT: A LONGITUDINAL EXAMINATION OF THE THERAPEUTIC WORKING ALLIANCE presented by Eric Martin Sauer has been accepted towards fulfillment of the requirements for Ph .D . degree in Counseling Psychology 959% Major professor Date 8-9-99 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINE return on or before date due. MAY BE RECAU£D with earlier due date if requested. DATE DUE DATE DUE DATE DUE ma Wane-p.14 THERAPIST AND CLIENT ADULT ATTACHMENT: A LONGITUDINAL EXAMINATION OF THE THERAPEUTIC WORKING ALLIANCE BY Eric Martin Sauer A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling and Educational Psychology and Special Education 1999 ABSTRACT THERAPIST AND CLIENT ADULT ATTACHMENT: A LONGITUDINAL EXAMINATION OF THE THERAPEUTIC WORKING ALLIANCE BY Eric Martin Sauer The current study examined (a) the relation of therapist and client adult attachment to measures of the working alliance and (b) the level of correspondence between therapist-client working alliance ratings over time. Participants were 28 counseling dyads drawn from university counseling centers and community agencies who were asked to complete brief survey measures following the I“) 4u‘and 7fl‘counseling sessions. Contrary to prediction, results indicated that client-therapist working alliance ratings were significantly related following the lSt and 4th therapy session but not related following the 7U‘session. In addition, results indicated that therapist and client adult attachment styles were differentially related to working alliance ratings across the three time points. Unexpectedly, therapists with insecure adult attachment styles received higher early-session client working alliance ratings; however, as expected, secure therapists received moderately higher working alliance ratings later in therapy. Also anticipated, clients with secure attachment styles received higher later-session therapist working alliance ratings. Therapists’ adult attachment styles did not moderate the relation of client and therapist working alliance ratings over time. Finally, therapist and client attachment styles did not predict the 7th correspondence of working alliance ratings following the counseling session. Copyright by ERIC MARTIN SAUER 1999 To Suzanne and Megan ACKNOWLEDGMENTS The completion of this project was a challenging and rewarding endeavor and represents the culmination of one of my major dreams in life. I would like to express my eternal gratitude to some of the people who have been most influential in the completion of this journey. I want to thank Fred Lopez, Ph.D., who for the past six years, has been my advisor, dissertation chair, coach, and mentor. I thank him not only for his support, guidance, and encouragement over the past several years but also for all of time and energy that he devoted to the completion of this project. I would also like to thank my committee members, Ken Rice, Ph.D., Gershen Kaufman, Ph.D., and Robbie Steward, Ph.D., whose guidance, support, knowledge, and feedback, were instrumental. Each of them brought diverse perspectives to this project but all of them shared an interest in my development as a counseling psychologist. I would like to thank Jim Wyssmann, Kim Thomas, Karyn Boatwright and Judy Ferris who supported me throughout this process. I cherish our friendships and will fondly remember our time together at MSU. Thank you to Nikki Brown who was instrumental in the data collection process. I appreciate your hard work and dedication to this project. vi I would like to thank my wife, Suzanne, the love of my life and my best friend. Thank you for providing me with never-ending patience, love, and encouragement. Over the years you have always been there for me. Thanks for supporting my continued education. I deeply appreciate all of the sacrifices that you made along the way. To my daughter, Megan, who has brought my life more joy and meaning than I could have ever imagined. You taught me how to walk slower through life (and to school) and helped me to keep my priorities straight. To Lisa, who has been a wonderful and loving sister. I enjoyed growing up with you. It seems to me that you were always a successful student. Thank you for inspiring me to eventually become a dedicated student. Finally, thank you to my parents, Norm and Eileen, for their support, encouragement and guidance, and for filling my life with love, laughter and joy. I thank them for teaching me the value of education (long before I was ready to accept it), how important it is to spend time with your family, and how to be a parent. vii TABLE OF CONTENTS LIST OF TABLES ............................................. x LIST OF FIGURES .......................................... xii CHAPTER I: INTRODUCTION .................................... 1 Working Alliance ........................................... 2 Working Alliance: Client and Therapist Factors ............. 3 Attachment Theory .......................................... 4 Adult Attachment .......................................... 5 Summary .................................................... 7 CHAPTER II: REVIEW OF THE LITERATURE ....................... 9 Working Alliance ........................................... 9 Working Alliance: Counseling Outcomes ................ 10 Working Alliance Ratings: Multiple Perspectives ...... 11 Convergence of Alliance Ratings ...................... 12 Working Alliance: Therapist Factors .................. 13 Working Alliance: Client Factors ..................... 15 Working Alliance: Therapist and Client Factors ....... 16 Summary .............................................. 16 Attachment Theory ......................................... 18 Contemporary Attachment Theory: Extensions to Adult Functioning and Adjustment ................................ 20 Measurement of Adult Attachment ...................... 22 Attachment and Romantic Relationships: Empirical Findings ............................................. 23 Attachment and Affect Regulation ..................... 26 Summary .............................................. 27 Extensions of Attachment Theory to the Therapeutic Relationship .............................................. 28 Client and Therapist Attachment and the Working Alliance: Empirical Findings ......................... 30 General Summary ........................................... 34 CHAPTER III: METHOD ....................................... 38 Participants .............................................. 38 Procedures ................................................ 40 Instruments ............................................... 42 Client Personal Demographic Questionnaire ............ 42 Therapist Personal Demographic Questionnaire ......... 42 Relationship Questionnaire ........................... 43 Adult Attachment Inventory ........................... 44 Working Alliance Inventory ........................... 45 Self-Understanding Scale ............................. 46 Hypotheses ................................................ 47 Analyses .................................................. 49 CHAPTER IV: RESULTS ....................................... 51 Preliminary Descriptive Analysis .......................... 51 Client Background Variables .......................... 51 viii Therapist Background Variables ....................... 53 Frequencies of Adult Attachment Styles ............... 53 Client and Therapist Working Alliance Ratings ........ 54 Results of Primary Analysis ............................... 58 Relations of Therapist and Client Ratings of the Working Alliance Over Time ........................... 58 Relations of Therapists’ and Clients’ Adult Attachment to the Working Alliance Over Time ......... 60 Relations Among Dimensional Measures of Adult Attachment and Client Working Alliance Ratings..60 Relations Among the Categorical Measure of Adult Attachment and Working Alliance Ratings...65 The Impact of Within Dyad Adult Attachment Styles on Working Alliance Ratings Over Time....67 Therapist Adult Attachment Orientation as a Moderator of Client and Therapist Working Alliance Ratings Over Time .................................... 7O Therapist and Client Adult Attachment Orientations and Correspondence of Working Alliance Ratings ....... 73 Adult Attachment Orientations and Self-Understanding Scores ............................................... 77 Working Alliance Ratings Self-Understanding Scores...77 Post-Hoc Analyses ......................................... 78 Correlations Among Working Alliance Subscale Scores..81 Adult Attachment Orientations and Working Alliance Subscale Scores ...................................... 82 Client Adult Attachment Orientations ............ 82 Therapist Adult Attachment Orientations ......... 82 Summary ................................................... 90 CHAPTER V: DISCUSSION ..................................... 93 Overview of Findings ...................................... 93 Development of the Working Alliance: Therapist and Client Perspectives .............................................. 94 Relations of Client and Therapist Adult Attachment Orientations to the Development of the Working Alliance...99 Contributions of Study and Implications for Counseling Psychology ............................................... 109 Limitations and Directions for Future Research ........... 111 Summary and Conclusions .................................. 115 APPENDIX A: CLIENT INSTRUMENTS ........................... 118 APPENDIX B: THERAPIST INSTRUMENTS ........................ 125 REFERENCES ............................................... 133 ix LIST OF TABLES Table 1: Client Demographic Data .......................... 39 Table 2: Therapist Demographic Data ....................... 41 Table 3: Correlations Among Client Background Variables and Outcome Measures ...................................... 55 Table 4: Correlations Among Therapist Background Variables and Outcome Measures ............................ 56 Table 5: Frequency of Client and Therapist Self-Reported Adult Attachment Styles on the Relationship Questionnaire (RQ) ...................................................... 57 Table 6: Client and Therapist Working Alliance Scores Over Time ...................................................... 56 Table 7: Intercorrelations Among Therapist and Client Working Alliance Ratings Over Time ........................ 59 Table 8: Intercorrelations Among Therapist and Client Adult Attachment Inventory (AAI) Scores and Working Alliance Inventory (WAI) Scores Over Time ................. 62 Table 9: Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 1 .......................... 74 Table 10: Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 2 .......................... 75 Table 11: Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 3 .......................... 76 Table 12: Correlations Among Client and Therapist Adult Attachment Orientations and Self-Understanding Scores Over Time ...................................................... 79 Table 13: Correlations Among Client and Therapist Working Alliance Inventory (WAI) Ratings and Self-Understanding Scores Over Time .......................................... 8O Table 14: Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 1 ........ 83 Table 15: Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 2 ........ 84 Table 16: Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 3 ........ 85 Table 17: Relations Among Therapist and Client Adult Attachment Inventory (AAI) scores and Working Alliance (WAI) Subscale Scores at Time 1 ........................... 86 Table 18: Relations Among Therapist and Client Adult Attachment Inventory (AAI) scores and Working Alliance (WAI) Subscale Scores at Time 2 ........................... 87 Table 19: Relations Among Therapist and Client Adult Attachment Inventory (AAI) scores and Working Alliance (WAI) Subscale Scores at Time 3 ........................... 88 xi LIST OF FIGURES Figure l: The Impact of Dyad Adult Attachment Styles on Client Working Alliance Inventory (WAI) Ratings Over Time ...................................................... 71 Figure 2: The Impact of Dyad Adult Attachment Styles on Therapist Working Alliance Inventory (WAI) Ratings Over Time ...................................................... 72 xii INTRODUCTION Counseling psychology has long been interested in factors that directly affect human development and change, and especially those that influence counseling processes and outcomes. In fact, according to Horvath and Greenberg (1994), psychotherapy research, which is entering its fifth active decade, clearly supports the efficacy of psychotherapy: Clients receiving psychotherapy demonstrate significantly more improvements or are better off than those in those in control conditions. Another consistent finding that emerges form this line of inquiry is that, with few exceptions, different therapies produce comparable therapeutic gains. With the establishment of positive therapeutic effects across diverse treatments, empirical attention has shifted to the investigation of nonspecific, or generic factors. Nonspecific factors are defined as those therapeutic factors that are not uniquely associated with specific treatment interventions but appear to be predictive of outcome (Horvath, 1994). One nonspecific factor that is common to all forms of treatment is the therapeutic relationship between the therapist and client (Horvath & Greenberg, 1994). The therapeutic relationship or “working alliance” has been consistently predictive of successful therapy outcome and may be the single most important process variable that has 1 been empirically examined in contemporary psychotherapy research (for a review see Sexton & Whiston, 1994). However, despite the general agreement among counseling psychologists regarding the critical importance of the therapeutic relationship, it has received relatively little empirical attention (Gelso & Carter, 1985). Working Alliance Originally defined by Bordin (1979), the working alliance consists of three interrelated components — tasks, goals, and bonds - that combine to create the overall perceived quality of the therapeutic relationship. Bordin hypothesized that it was this "positive collaboration" that allowed the therapist and client to work together to overcome the common foe of the client's pain. Indeed, early working alliance scores, especially from the client's perspective, are related to counseling outcome (Horvath & Symonds, 1991; Tichenor & Hill, 1989). Although there is a positive relationship between alliance scores and counseling outcome, one troublesome finding is the observed discorrespondence between early (3rd or 4U‘session) therapist and client working alliance ratings (Horvath & Symonds, 1991; Mallinckrodt, 1991; Tichenor & Hill, 1989). However, more recently, some investigators have reported low levels of correspondence between early client and therapist working alliance ratings appeared to converge over time (Kivlighan & Shaughnessy, 1995; Mallinckrodt, 1993). Despite these preliminary findings, precious little is known about factors that may be contributing to the initial discorrespondence between therapist and client ratings of the working alliance or how these factors may interact to impact the development of the working alliance over time. However, an emerging line of inquiry is beginning to examine the relations between client and therapist characteristics and working alliance ratings. Working alliance: Client and therapist factors Bordin (1979) proposed that the strength of the working alliance was determined by the personal characteristics of the client and therapist. Recently, there is some limited empirical support to suggest that certain client factors (i.e., expressed hostility, quality of interpersonal relationships, relationship expectations) and therapist factors (i.e., training level, clinical intentions, technical activity, relationship expectations) may impact the development of the working alliance (Al—Darmaki & Kivlighan, 1993; Dykeman & LaFleur, 1996; Kivlighan & Schmitz, 1992; Kokotovi & Tracey, 1990; Mallinckrodt & Nelson, 1991). Despite these limited empirical findings suggesting a relationship between therapist and client factors and the formation of the working alliance, these factors continue to be understudied and thus not well understood. Moreover, the contribution of client and therapist factors to correspondence of working alliance ratings from multiple perspectives has been ignored. In short, the overall working alliance appears to be an important variable linking process to outcome. However, there is empirical uncertainty as to which therapist and client factors contribute to the working alliance or how these factors may interact with one another to affect the development and correspondence of working alliance ratings over time. Clearly, there is a need for theory—driven research capable of addressing these empirical gaps. One existing theory —— attachment theory -- may provide a conceptual framework for understanding how client and therapist characteristics influence the development and correspondence of working alliance ratings. Attachment Theory Bowlby (1988) outlined how attachment theory could be used as a framework to guide the development of the therapeutic relationship. Like early primary caregivers, therapists must provide clients with a "safe haven" or "secure base" from which to explore the world. In fact, Bowlby (1988) suggested that, "the therapeutic alliance appears as a secure base " (p. 151). More recently, Lopez (1995) suggested attachment theory could serve as a "metaperspective" in contemporary counseling psychology. Such a perspective, he argued, "should speak to important process and outcome issues in counseling and thus deepen our understanding of how client and counselor characteristics may interact to facilitate therapeutic change" (p. 396). Adult Attachment Hazan and Shaver (1987) initially extended attachment theory to the conceptualization of intimate adult relationships. More recently, Bartholomew and Horowitz (1991) have proposed four prototypic forms of adult attachment -- secure, preoccupied, dismissive and fearful -— with two underlying dimensions, internal model of self (positive or negative) and internal model of other (positive or negative). Subsequently, attachment theory has been fruitfully applied to the examination of adult romantic relationships. Taken together, these findings demonstrate that adult attachment styles are related to individuals’ experiences in close relationships. Specifically, secure adults relative to their insecure peers, experience more positive, and well— adjusted, intimate relationships (Collins & Reed, 1990; Feeney & Noller, 1990; Lopez et al., 1997; Pistole, 1993; Sharfe & Bartholomew, 1995; Simpson, 1990; Simpson, Rholes, & Nelligan, 1992). Moreover, this line of inquiry suggests that several theory-relevant process variables (e.g., trust, self-disclosure, affect regulation, etc.) may speak to how closeness and distance is negotiated within the context of therapeutic relationships. Attachment Theory and the Therapeutic Relationship Pistole (1989) argued that attachment theory provided a useful framework for conceptualizing, directing, and evoking therapeutic change. Furthermore, she reported parallels between attachment and therapy; for instance the provision of a safe base, consistent responding, sensitivity, affect regulation and comforting. Hence, there is a theoretical basis for extending attachment to the study of counseling process. Attachment theory (Bowlby, 1988) has recently been extended into the domain of counseling processes, and appears to offer a useful lens for understanding the relations between client and therapist characteristics and the development of the therapeutic relationship. In fact, a limited number of studies have indicated a relation between client and therapist attachment orientations and the quality of the therapeutic relationship. Specifically, the evidence indicates that client attachment orientations are related to more positive therapeutic relationships (i.e., higher working alliance ratings) (Mallinckrodt, 1991; Mallinckrodt, Coble & Gantt, 1995; Satterfield & Lyddon, 1995). One study reported that therapist adult attachment orientations were positively related to working alliance scores (Dunkle and Friedlander, 1996). In short, the aforementioned literature indicates some limited evidence that client attachment may be significantly related to working alliance ratings. Despite these findings, therapist attachment has been largely ignored although the results of one study suggest that variation in the therapistsl attachment security may also affect the formation of the therapeutic relationship. Clearly, there is a need to examine the contributions of both therapist and client adult attachment to the development of the therapeutic working alliance over several counseling sessions. Also warranted is an examination of the relation between adult attachment orientations to more session specific impacts (i.e., changes in self-understanding). Thus, the following research questions will be examined in this study. First, do client and therapist adult attachment styles/orientations make unique contributions to working alliances? Second, do client and therapist adult attachment orientations make unique contributions to the correspondence of alliance ratings over time? Third, do therapist adult attachment orientations moderate the correspondence between client and therapist ratings of working alliances? Fourth, how are client and therapist adult attachment orientations and working alliance scores related to client self- understanding scores over time? Summary The following conclusions can be drawn: (a) the overall working alliance is an important relationship variable linking process and outcome; (b) there is empirical uncertainty regarding the lack of correspondence between client and therapist working alliance ratings, or how correspondence changes over time; (c) client characteristics (i.e., attachment memories, attachment security, attachment styles) may be related to the working alliance; (d) therapist characteristics (i.e., technical activity, training level, attachment history, etc.) may be related to the working alliance but this link has been understudied and thus not well understood; (e) it remains unclear what client and therapist factors contribute to positive working alliance ratings, or how these factors interact and (f) there is a need for theory capable addressing these empirical gaps. REVIEW OF THE LITERATURE This chapter will present an organized review of the literature relevant to the current study. First, it will briefly review the theoretical underpinnings of the working alliance, as well as studies that have demonstrated its relation to psychotherapy outcome. It will then examine research that has explored the discorrespondence of alliance ratings across multiple perspectives - client, therapist and observer, as well as more recent studies that have demonstrated a relation between therapist and client factors and the development of the working alliance. Next, it will briefly review attachment theory and its subsequent application to the conceptualization of adult romantic relationships. It will then consider several empirical investigations that have examined adult attachment and romantic relationships with a specific focus on those studies that examined therapy-relevant processes. Finally, it will present findings from recent clinical studies that have specifically examined the relationship between client or therapist attachment and working alliance ratings. Working Alliance Investigations over the past twenty years have established that therapy "in general" has positive effects. However, there have been few reliable differences in the efficacy among treatments. This may suggest that the positive effects of treatment may be process factors common to a variety of interventions (Horvath & Greenberg, 1989). One pantheoretical concept called the working alliance emphasizes the importance of the therapist-client collaboration. "The positive relation between a good alliance and successful therapy outcome is well documented across a variety of different treatments" (Horvath & Luborsky, 1993, p. 569). Although originally grounded in psychoanalytic theory, Bordin (1979) described the working alliance as the change- inducing relationship between the therapist and client. He believed that it was this "positive collaboration" that allowed the therapist and client to work together to overcome the client's pain. In fact, he stated that, "The effectiveness of a therapy is a function in part, if not entirely, of the strength of the working alliance" (p. 253). The three components of the working alliance (i.e., tasks, bonds, and goals) combine to create the overall perceived quality of the therapeutic relationship. Tasks are defined as the in—counseling behaviors that are generally expected of the therapist and client; goals refer to a therapist's and client's mutual acceptance and valuing of the direction (outcomes) of treatment; bonds are defined as the feelings of closeness or positive attachments between the therapist and client. Working Alliance: Counseling Outcomes More recently, investigators have developed several instruments to measure Bordin’s (1979) tripartite model of 10 the working alliance(for reviews see Horvath and Symonds, 1991; Tichenor & Hill, 1989). In fact, empirical research examining the relations of working alliance ratings to counseling outcomes has produced consistent findings across instruments and therapies (for a meta-analytic review see Horvath & Symonds, 1991). For instance, a reliable but moderate relation has been established between working alliances and therapy outcomes (ES = .26) (Horvath & Symonds, 1991). Likewise, early working alliance scores, especially from the client's perspective, are predictive of outcome (Horvath and Symonds, 1991; Tichenor & Hill, 1989). However, despite the established connection between working alliance ratings and counseling outcome, some have suggested that alliance ratings from multiple perspectives may be unrelated (Tichenor & Hill, 1989). Working Alliance Ratings: Multiple Perspectives A recent line inquiry has explored the working alliance from multiple perspectives - client, therapist, and observer. One troublesome finding in this literature is the observed lack of correspondence between client and therapist ratings of the working alliance. In fact, investigators have reported a low level of correspondence between early client and therapist working alliance ratings (Horvath & Symonds, 1991; Mallinckrodt, 1993; Tichenor & Hill, 1989), especially for novice therapists (Mallinckrodt, & Nelson, 1991). For instance, Tichenor and Hill (1989) compared several measures of the working alliance within a small clinical 11 sample (8 dyads). Results indicated that client and therapist and ratings of the working alliance were not significantly related to each other (r = .09). Moreover, they suggested that working alliances rated from different perspectives (i.e., client, therapist, observer) were not interchangeable. Elsewhere, in a meta-analytic review, Horvath and Symonds(1991) reported alliance ratings across different perspectives —- client, therapist and observer -— were differentially predictive of treatment outcome. Results indicated that the quality of the working alliance, as rated by the client, was more predictive of therapy outcome than were either therapist or observer ratings. Convergence of Alliance Ratings More recently, investigators have begun to explore the discorrespondence between client and therapist working alliance ratings. In fact, there is some limited empirical evidence indicating that client and therapist ratings may converge over time. For instance, Kivlighan and Shaughnessy (1995) reported that, in a sample of university counseling center dyads, early (4th session) nonsignificant relations between client and therapist working alliance ratings became significant over time. Likewise, Mallinckrodt (1993) reported that nonsignificant correlations between early client and therapist working alliance ratings (r = .07) became significantly related (r = .59) at termination. 12 Despite this emerging evidence suggesting the convergence of client and therapist working alliance ratings over time, there remains empirical uncertainty regarding which factors may be contributing to the lack of correspondence between early (3mior 4U‘session) client and therapist working alliance ratings or how these factors may impact correspondence over time. A related line of inquiry has suggested that therapist and client factors may impact development of the working alliance. Working Alliance: Therapist Factors Bordin (1979) proposed that the strength of the working alliance was determined by personal or pretherapy characteristics of therapists and clients. More recently, studies have examined the relation between therapist characteristics and the development of the working alliance. For instance, Mallinckrodt and Nelson (1991) examined the relationship between therapist training level and the three components of the working alliance. Specifically, they hypothesized that therapist training level (i.e., whether therapists were novices, advanced trainees or experienced counselors) would be differentially related to the three components of the working alliance following the end of the of the third counseling session. Using a clinical sample of 50 therapist-client dyads, they found a positive relation between the goal and task dimensions of the working alliance and training level. In other words, among more experienced therapists, these alliance components were more highly l3 rated. However, the bond dimension was not significantly related to training level. These results were consistent across both therapist and client ratings and suggested that the more technical tasks of therapy may be more difficult for novice therapists to execute. Elsewhere, Kivlighan (1990), using an analogue design, explored the impact of technical activity on the formation of client-rated working alliances. More specifically, this study explored whether 19 general therapist intentions were related to working alliance ratings. Participants were 42 student volunteer clients and 42 counselor trainees. Results indicated that, after controlling for client interpersonal attitudes, therapists’ intentions were related working alliance scores. Specifically, negatively related to client- rated working alliance scores following the 2m counseling session were therapists intentions to engage in assessing (i.e., gathering information, clarifying), exploring (i.e., probing thoughts, feelings, behaviors) and supporting (i.e., offering encouragement) activities. More recently, Kivlighan and Schmitz (1992) examined the relationship between therapist personal characteristics —- technical activity -- and the development of the working alliance over four simulated counseling sessions. Counseling dyads consisted of therapists in training and recruited undergraduate students volunteering to serve as "clients." As hypothesized, therapists were more challenging and here- and-now focused in counseling dyads that improved over time. 14 Moreover, these therapists increased the use of these two activities across the four “therapy” sessions. Working Alliance: Client Factors In a related line of inquiry, researchers have examined the relation between client characteristics and the development of the working alliance. For example, in a naturalistic study of counseling dyads, Kokotovic and Tracey (1990) examined the relation between client characteristics (i.e., expressed hostility, quality of interpersonal relationships) and clients' and therapists' ratings of the working alliance following the first counseling session. Results indicated that clients, who were rated by their therapists as having poor interpersonal and family relationships, as well as those who expressed hostility, tended to have poorer working alliances with their therapists. More recently, Dykeman and LaFleur (1996) explored the relation between working alliance scores and adjective descriptors of clients in university counseling centers. Results indicated that several therapists’ adjective descriptors of clients were related to alliance scores. Specifically, positive adjective correlates (e.g., good— natured, affectionate, cooperative, friendly, honest, etc.) were related to positive working alliances; whereas negative adjective correlates (e.g., infantile, sulky, cynical, defensive, evasive, etc.) were related to weaker alliances. 15 Working Alliance: Therapist and Client Factors Only one study was located that examined the concurrent impact of client and therapist factors and the working alliance. In this study, Al-Darmaki and Kivlighan (1993) examined the relation between congruence of client-therapist relationship expectations and working alliance ratings. For this study, “relationship expectations” were defined as the client’s and therapist's expectation that the client would “self-disclose in the context of an egalitarian relationship” (p.382). Congruence was operationalized as the absolute value of the difference between therapist and client relationship expectation ratings. Thus, lower discrepancy scores represented greater congruence. Using a sample of 25 client-therapist dyads, results indicated that, following the 3K’therapy session, congruence of therapist— client relationship expectations predicted working alliance formation. Specifically, congruence significantly predicted client and therapist working alliance subscale scores. Interestingly, the relative congruence of relationship expectations was unrelated to the working alliance components. Thus, as these investigators concluded, “what matters is whether the client’s and counselor’s expectations for relationship match and not whose expectations are higher or lower” (p. 383). Summary To summarize, the overall working alliance has been established as an important variable linking process to 16 outcome in counseling. However, there is empirical and theoretical uncertainty regarding the discorrespondence between alliance ratings from multiple perspectives (i.e., therapist and client). It also remains unclear how the working alliance develops over time. Some emerging evidence has suggested that client and therapist characteristics may be conjointly related to the formation of the working alliance; however, the vast majority of the aforementioned studies relied on working alliance ratings from a single- time—point and thus cannot speak to the “development” or “formation” of the working alliance. Furthermore, with the exception of one study, these studies also measured working alliance ratings from a single perspective (i.e., therapist or client). Thus, it remains unclear what client and therapist factors contribute to working alliance ratings, or how these factors may interact to impact the correspondence and development of the working alliance. Another factor adding to the aforementioned uncertainty is that past studies have lacked a unifying theory or underlying conceptual framework in their attempts to identify client and therapist factors that may contribute to the working alliance. Clearly, there is a need for theory- driven research capable of addressing these empirical gaps. One existing theory -- attachment theory - which has served as a useful lens for understanding the development of human relationships across the life span, may provide the necessary scaffolding for examining how client and therapist l7 characteristics influence the formation of the working alliance. Attachment Theory The following section will provide a brief overview of attachment theory, with an emphasis on its contemporary extensions to adult functioning and adjustment. It will then review the empirical literature demonstrating relations between adult attachment variables and individuals’ experiences in close relationships. This review will specifically highlight those studies that examined features of close relationships that are therapy-relevant (i.e., self—disclosure, trust, collaborative problem solving, affect regulation, etc.). Attachment theory (Bowlby, 1977) can serve "as a way of conceptualizing the propensity of human beings to make strong affectional bonds to particular others" (p. 201). According to Bowlby (1969/1982), attachment theory is considered ethological in nature in that it conceptualizes an infant's tie to his/her primary caregiver as distinctive and preprogrammed. This early motivational and behavioral repertoire has the goal of keeping the child in close proximity with the mother-figure. Proximity seeking behaviors (e.g., smiling, sucking, rooting, grasping, crying, clinging, etc.) have evolutionary significance in that, by serving to provide security and protection from environmental dangers (i.e., predators), they increase the fitness of the species. Additionally, the attachment 18 relationship provides the child with a psychological and relational "safe haven" or "secure base" from which to explore the world. Ainsworth, Blehar, Waters, and Wall (1978) developed an empirical procedure called the "Strange Situation" which allowed for the standardized observation of infants' behavioral responses to separations and reunions with their primary caregivers (most often their mothers). In this procedure, experimenters provided differential conditions from which to observe the parent-child interaction. Observations of infants in the strange situation have led to the identification of distinctive patterns of attachment behaviors. These patterns or attachment styles were originally clustered into three primary groups -— secure, anxious/ambivalent, and avoidant. "Secure" infants have confidence that parents will be available should the child encounter a frightening/dangerous situation. Additionally, secure children use their parents as a "secure base" or "safe haven" from which to explore the world. "Secure" infants cry less, are easier to soothe when upset, and freely explore their environments in the presence of their mothers. Next, "anxious/ambivalent" infants are uncertain about parental responsiveness. They demonstrate increased anxiety, clinging behaviors and restricted exploration. Finally, "avoidant" or "not-yet attached" infants demonstrate an undifferentiated attachment pattern towards their mothers. These infants are compulsively self-reliant l9 and appear to be emotionally disconnected from parental figures. In short, attachment theory (Bowlby 1969/1982; Ainsworth et al., 1978) underscores the importance of early childhood experiences with primary caregivers in guiding early human development. Through these early relationship experiences, children develop "internal working models" of themselves (and others) that may be positive or negative. These early caregiver experiences provide a relatively stable pattern of responding in later adult relationships. In fact, according to Bowlby (1977), "there is a strong causal relationship between an individual's experiences with his parents and his capacity to make affectional bonds" (p. 206). Contemporary Attachment Theory: Extensions to Adult Functioning and Adjustment In a seminal investigation, Hazan and Shaver (1987) proposed that attachment theory provided a useful framework for the conceptualization of adult romantic love. They posited that the three—category model infant attachment styles developed by Ainsworth and her colleagues — secure, avoidant, and anxious/ambivalent - could be translated into concepts that appropriately described important variations in adult romantic relationships. Specifically, the authors stated “that romantic love is an attachment process (a process of becoming attached), experienced somewhat differently by different people because of variations in 20 their attachment histories” (p. 511). In their investigation, the authors reported that the three identified adult attachment styles had similar frequencies to the three observed infant attachment styles - that is roughly 60% classifying themselves as secure and the remainder split roughly evenly between the avoidant and anxious/ambivalent types. More recently, Bartholomew and Horowitz (1991) argued that adult attachment was better conceptualized into four, rather than three, attachment styles. They posited that there were two primary (underlying) dimensions of adult attachment, internal model of self (positive or negative) and internal model of other (positive or negative). These four dimensions create four prototypic attachment styles. "Secure" individuals have a positive View of self (lovability) and an expectation that others will be nurturing and responsive to their needs. These adults seek out relationships and feel comfortable with intimacy. "Preoccupied" individuals have negative self-views (unlovability) and a positive view of others. These adults gain self-acceptance through the acceptance/respect of others. "Fearful" adults feel unworthy (unlovable) and have an expectation that others will be harmful and/or rejecting. These individuals avoid intimacy to reduce the likelihood of rejeCtion. Finally, "dismissing" individuals have a positive View of self but a negative view of others. These individuals maintain interpersonal distance and deny or 21 disavow having intimacy needs. Measurement of Adult Attachment Hazan and Shaver (1987) reported that attachment styles differed in predictable ways in their experience of romantic love. In addition, they introduced a single-item, self— report instrument to measure the three proposed adult attachment styles — secure, avoidant, and anxious/ambivalent. More recently, Bartholomew and Horowitz (1991) proposed another self-report measure of adult attachment. Using their 4—group model - secure, preoccupied, dismissing and fearful, the authors developed a single item measure that utilized the four prototypical descriptions of each attachment style. Alternatively, others (e.g., Collins and Reed, 1990; Simpson, 1990, etc.) have developed continuously-scaled measures of adult attachment. For example, Simpson (1990) proposed a continuously scaled self-report measure of adult attachment. By decomposing the sentences found on the Hazan and Shaver (1987) instrument, Simpson created a 13—item, dimensional measure of adult attachment that provided two factor—analytically-derived subscales related to attachment. The avoidance/security subscale assessed respondents' comfort in close relationships. The anxiety subscale measured the level of tension or worry that individuals typically reported in close relationships. Elsewhere, Collins and Reed (1990) translated Hazan and Shaver’s (1987) categorical measure of adult attachment into 22 an 18-item, continuously-scaled instrument. This measure yields three factors - depend, close and anxious — that capture individuals’ comfort in romantic relationships (e.g., feeling like they can depend on others, fear of abandonment, etc.). Adult Attachment and Romantic Relationships: Empirical Findings Using the three-group, four—group, or dimensional measures of adult attachment, subsequent studies have demonstrated that, in general, adult attachment styles and adult attachment orientations are associated with diverse experiences within close relationships. For example, Simpson (1990) explored the impact of attachment styles on romantic relationships. In a longitudinal study of 144 dating couples, results indicated that adult attachment styles were associated with “qualitatively” different experiences in romantic relationships. That is, those with secure attachment tended to be in “relationships characterized by higher levels of interdependence, trust, commitment, and satisfaction” (p. 977). Likewise, Hazan and Shaver (1987) reported that secure adults characterized their romantic relationships in more positive terms (e.g., happy, trusting, friendly, etc.). In contrast, avoidant and anxious/ambivalent adults reported more negative characterizations of their romantic relationships (e.g., emotional extremes, jealousy, etc.). In a similar study, Feeney and Noller (1990) evaluated 23 the utility of attachment styles as a predictor of adult romantic relationships. As anticipated, results indicated that attachment styles exerted a powerful influence on close relationships. That is, secure respondents reported more trusting attitudes towards others. Conversely, avoidant respondents endorsed items indicating mistrust and interpersonal distance. Finally, anxious-ambivalent subjects expressed a desire for dependence in close relationships. The authors suggested that attachment patters appear to reflect individuals’ general views about the costs and benefits of close relationships. Mikulincer and Nachshon (1991) examined the relation of attachment styles on self-disclosure. They reported that “both variables have been found to play a central role in the formation of close relationships” (p. 322). In a study of 352 undergraduate Israeli students, all of whom were involved in heterosexual relationships, results indicated that secure and ambivalent respondents demonstrated more self-disclosure and they also disclosed more personal information than their avoidant peers did. Moreover, secure respondents demonstrated a pattern called “responsive self— disclosure” which also indicated that they were more responsive the self-disclosures of others and more likely to promote or reinforce their partner’s disclosures. Pistole (1993) examined the relations between adult attachment style dimensions and self-disclosure and trust. Using a sample of 98 undergraduate students who were asked 24 to fill out surveys with their “most important ‘romantic’ love relationship in mind,” results indicated that securely attached respondents, compared to their avoidantly attached peers, reported increased levels self-disclosure and more comfort with self-disclosure. Moreover, the findings also indicated that securely attached respondents were more likely to trust their partners than were their insecurely attached peers. The author noted that these findings were consistent with previous studies that have suggested that secure attachment is associated with increased functioning or “competence” in close relationships. Collins and Reed (1990), using a sample of 71 undergraduate students, explored the relations between attachment styles and relationship quality in dating couples. In general, results indicated that attachment dimensions were related to the quality of romantic relationships. For example, respondents with more secure attachment had more positive views about self and others, and reported higher levels of communication, trust and satisfaction within their romantic relationships, than did their insecurely attached peers. Sharfe and Bartholomew (1995) examined the relation among attachment styles and conflict resolution in young couples. Results indicated that secure attachment was positively related to the use of constructive accommodation strategies to resolve relational conflict, whereas insecure attachment (fearful) was related to the use of destructive 25 interpersonal patterns (i.e., distancing or neglecting partners) during conflict. This finding was corroborated by Lopez et al. (1997) who reported that, within their romantic relationships, secure adults reported a stronger orientation towards collaborative problem solving than did their fearful counterparts. Horowitz, Rosenberg and Bartholomew (1993) examined links between attachment styles and interpersonal functioning. In their study of close, nonintimate relationships, dismissive respondents had the most difficulty providing accurate descriptions of significant others. These investigators suggested that, “the interpersonal problems associated with dismissive attachment styles prevent the person from knowing other people well, so that the person’s internal representations of others is relatively unclear” (p.558). Attachment and Affect Regulation In a related line of inquiry, attachment theory has been employed to examine how adult experience, manage or regulate their affect. In general, findings indicate that, attachment styles are predictive of the ways people experience affect and cope with stressful events (for reviews see Fuendeling, 1998; Kobak & Sceery, 1988; Mikulincer & Florian, 1998). For example, Kobak and Sceery (1988) examined the relation between adult attachment and affect regulation. Results indicated that secure adults reported less distress and higher levels of social support 26 and were judged by their peers as more ego—resilient, less anxious and less hostile, than their insecure counterparts. In a related study, Simpson (1990) reported secure respondents characterized their relationships as having higher occurrences of positive affect (e.g., excited, happy, passionate, etc.) and lower occurrences of negative affect (e.g., angry, jealous, hostile, etc.) than did insecurely attached respondents. Recently, Fuendeling (1998) provided a comprehensive review of the literature on adult attachment and affect regulation. As expected, the evidence “does suggest consistent, or stylistic, ways of regulating affect that are particular to each attachment style” (p. 291). In fact, several findings indicate that individuals with secure attachment orientations demonstrate more constructive and interpersonally oriented methods (e.g., support seeking, self-disclosure, etc.) to manage or cope with distressful affect. Summary Taken together, investigations using the three-group, four-group or dimensional measures of adult attachment indicate that secure adults, as compared to their insecure peers, experience more positive, and well—adjusted, intimate relationships. In addition to enhanced relationship functioning, securely attached adults may be better equipped to accurately evaluate important relational characteristics (i.e., working models) of themselves and others. Finally, 27 there appear to be stylistic differences indicating that attachment styles are related to how individuals experience or regulate affect. Generally speaking, this line of inquiry suggests that several therapy-relevant processes variables (e.g., trust, self-disclosure, collaborative problem solving, relationship satisfaction, affect regulation, etc.) may speak to how closeness and distance is negotiated within the context of the therapeutic relationship. In fact, some investigators have begun to explore the relation between attachment and the working alliance. Extensions of Attachment Theory to the Therapeutic Relationship The therapeutic relationship shares many common features with romantic relationships - felt security, trust, self-disclosure, affect regulation, collaborative problem solving, etc. In fact, Bowlby (1988) outlined how attachment theory could be used as a framework to guide and understand the therapeutic relationship. In general, he posited that therapists provided clients with a "secure base" form which to explore painful thoughts and emotions. This position, he suggested was similar to that of early primary caregivers who provide a child with a secure base or “safe haven” from which to explore the world. Of particular importance was his postulation that the therapeutic alliance eventually emerges as the secure base. Likewise, Dolan, Arnkoff and Glass (1993) have suggested that therapists should assess clients' 28 attachment style and alter their interpersonal stance and interventions to enhance the therapeutic working alliance. Elsewhere, Pistole (1989) theorized that attachment theory provided a useful framework for understanding, directing and evoking therapeutic change. She reported that, “Since adults often seek counseling for relationship difficulties, connecting attachment with emotionally important adult relationships makes attachment theory directly relevant to the counseling profession” (p. 190). She also noted the parallels between attachment relationships and psychotherapy relationships (i.e., emotional availability, soothing, safety, security, exploration, etc.). Like Bowlby, she argued that the critical feature of effective therapy was the provision of a safe base from which clients can explore their worlds. Others have similarly argued that attachment theory could be applied to the process of psychotherapy (Biringen, 1993; Paterson & Moran, 1988; Rutter, 1995). By outlining the clinical utility of several of the theory’s key concepts, investigators have suggested that attachment theory “has the potential to provide valuable insights regarding the process and techniques of psychotherapy” (Paterson & Moran, 1988, p. 611). Thus, there is a theoretical foundation for inquiring about the role of adult attachment in the counseling process. 29 Client and therapist attachment and the working alliance: Empirical findings Recently, attachment theory has been used to examine the therapeutic working alliance. In fact, a limited number of studies have examined the relation between attachment and the working alliance. In general, it appears that client and therapist attachment dynamics may be related to the quality of the therapeutic working alliance. For example, in his study of 102 client—counselor dyads, Mallinckrodt (1991) examined how client attachment memories of parents affect clients’ and therapists’ BNL session working alliance ratings. Results indicated that clients' bonds with fathers were significantly related to counselor-rated working alliances. Interestingly, there was only a moderate relation (r = .32) between clients' and therapists' evaluations of the working alliance. This moderate association suggested that therapists and clients may use different criteria in their respective assessments of the working alliance. One notable limitation of this study was the reliance on a single-time-point assessment of the working alliance. In another study of client-counselor dyads, Mallinckrodt, Coble and Gantt (1995) examined the impact of female clients’ memories of attachment bonds with parents and working formation. Drawing clients from community and university settings, results indicated that recalled parental bonds accounted for 23% of the variance in clients' 30 working alliance ratings. Interestingly, recalled bonds with fathers were generally a more robust predictor of the working alliance than were recalled bonds with mother. A notable limitation of this study was that working alliance ratings were completed at various points during counseling. Elsewhere, Dolan (1992) examined the hypothesis that clients' attachment styles would be related to Working Alliance Inventory (WAI) ratings. Contrary to prediction, following the third session, attachment styles -- secure, anxious-ambivalent and avoidant -— were not significantly related to the quality of the therapeutic alliance. However, dimensional measures of attachment were significantly associated with alliance ratings. More specifically, correlational findings indicated that higher attachment security was related to higher working alliance ratings -- bonds, goals and total WAI scores —- from the therapists' perspective, whereas higher avoidance scores were related to poorer total WAI ratings and less correspondence regarding the goals of therapy. Satterfield and Lyddon (1995) explored the relation among three client attachment dimensions and the quality of client rated working alliance ratings following the 3rd counseling session. The attachment dimensions were defined as: (a) depend, or the level that individuals trust or depend on others, (b) anxiety, or the level individuals fear rejection or abandonment by other, and (c) close, or the level individuals are comfortable with closeness and 31 intimacy. Dyads consisted of 60 clients and 38 graduate student counselors. Results indicated a significant positive relation between the depend dimension of the attachment measure and client WAI ratings. In particular, clients who reported lower depend scores were more likely evaluate the working alliance in negative terms early in counseling. Unfortunately, this study relied on counselors in training, single-time—point ratings of the working alliance, and reported a high attrition rate (37.5%) due to incomplete protocols and premature terminations. Only two studies were located that examined the relations between therapist attachment and the quality of the therapeutic relationship. Although limited in number, this emergent literature suggests that therapists’ attachment may also impact the development of the working alliance. In one study, Dozier, Cue, and Barnett (1994) examined the relation between clinicians' adult attachment orientations and the nature and style of intervention strategies. These clinical dyads included clinical case managers and adults with serious psychopathological disorders. Case managers in this study provided a wide range of services to clients including helping them with financial matters, as well as other interpersonal and intrapersonal issues. Results indicated that case managers’ attachment orientations were related to their clinical intervention strategies. Specifically, secure case managers were more 32 likely to respond to clients' underlying needs, whereas insecure case manages only attended to the most obvious need. In addition, case managers with preoccupied orientations intervened more intensely than did their dismissing peers. One disturbing implication of this study is that insecure case managers seemed to be responding to their own internal attachment needs rather than clients' intrapersonal/interpersonal issues. It should be noted that the majority of the case managers in this study had limited training and experience and many of the interventions that they provided were not psychotherapy. In one particularly relevant study, Dunkle and Friedlander (1996) attempted to “shed light on our understanding of the contribution of selected therapist variables to the working alliance” (p. 459). In this study, the relations between therapists' degree of comfort with attachment and clients' WAI ratings were examined. Results indicated that, early in treatment (between the Budennd SU‘ session), therapists who reported greater comfort with intimacy were more likely to receive favorable bond ratings on the working alliance. Taken together, the aforementioned research indicates that “therapists, like clients, bring to the therapeutic relationship a personal history that affects their interactions” (Dunkle & Friedlander, 1996, p.459). Despite these preliminary findings, the impact of client and therapist characteristics on the development of the working 33 alliance remains understudied and thus not well understood. Moreover, some noteworthy gaps have emerged in this literature that need further investigation. First, past studies have examined either client or therapist attachment styles/orientations and thus cannot speak to the conjoint contribution or interaction of these variables on working alliance ratings. Next, because past studies relied on single-time-point ratings of the working alliance, they cannot speak to the development or formation of the alliance. Finally, the impact of therapist and/or client attachment orientations on the correspondence of working alliance ratings over time has yet to be examined. General Summary The therapeutic working alliance has been established as a critical component of successful psychotherapy. Despite this connection, there appears to be discorrespondence between working alliance ratings across multiple perspectives — client, therapist, observer. However, there is some limited empirical support suggesting that therapist or client factors may contribute to working alliance ratings but these connections remain understudied and thus not well understood. Attachment theory offers a useful lens for exploring the relation between client and therapist factors and the development of the working alliance. In fact, several of the therapy-relevant processes (e.g., trust, self-disclosure, affect regulation, collaborative problem solving, 34 relationship satisfaction, etc.) that have been associated with adult attachment characteristics in close relationships, may also speak to how closeness and distance are negotiated within the context of therapeutic relationships. In fact, although few in number, recent empirical findings lend support the theoretical proposition that personal characteristics of the client and therapist impact the development of the working alliance. However, despite this preliminary linkage, several noteworthy gaps remain. For example, past studies have-generally a) relied on single-time-point methodology, b) not concurrently examined client and therapist attachment, c) not explored the impact of attachment on correspondence of working alliance ratings, and d) not inquired about the relation of attachment to more session specific impacts. Thus, it remains unclear how client and therapist attachment may impact the correspondence and the development of the working alliance over several counseling sessions. The purpose of this investigation was to examine the relations of therapist and client adult attachment to measures of the working alliance across seven counseling sessions. Specifically, it explored how therapist and client adult attachment styles/orientations affected the correspondence and development of the working alliance. The following hypotheses were derived for this investigation: 1. Therapist and client ratings of the working alliance 35 would be differentially related across three time points. More specifically, (a) therapist and client ratings of the working alliance would not be significantly related at Time 1 (after the first session), or at Time 2 (after the fourth session), but that they would be significantly related at Time 3 (following the seventh session). 2. Therapists' and clients’ adult attachment security would be positively related to ratings of the working alliance across all three time points. 3. Therapists' adult attachment orientations would moderate the relation of therapist and client ratings of the working alliance at each time point. More specifically, among therapists who self-report a secure attachment orientation, therapist and client working alliance ratings would be more strongly and positively related. 4. Therapist and client adult attachment orientations should significantly predict correspondence of client and therapist ratings of the working alliance at time 3. More specifically, therapist and client adult attachment security would be positively related to working alliance correspondence. To explore how therapist and client attachment orientations and working alliance scores were related to more session-specific measures (i.e., assessments of client self-understanding), the following research questions were 36 proposed for this study: 1. Are client and therapist adult attachment orientations related to client self-understanding scores over time? 2. Are client- and therapist-rated working alliance scores related to client self-understanding scores over time? 37 METHOD This chapter will outline methodological procedures employed in the current study. First, it will present a description of the clients and therapists that participated in this study. Next, it will detail the procedures employed in this study including recruiting practices, informed consent and data collection strategies. It will then describe the five self-report measures that were used in this study and, where appropriate, it will provide psychometric information on the reliability/validity of the instruments. Finally, statistical analyses that were conducted on the obtained data will be described. Participants Twenty—eight clients (8 men and 20 women), aged 20-56, receiving counseling services in university counseling centers (3 = 21) community counseling centers (n = 6), and other professional settings (3 = 1) were participants for the study (see Table 1). The racial/ethnic background was Caucasian (93%) and Asian—American (7%). The frequencies of clients’ current marital statuses were as follows: single (68%), married (18%) and divorced (14%). A slight majority of the sample (54%) reported that they had received previous counseling. 38 Table 1 Client Demographic Data N 28 Age (M years) 29.44 (10.32)a Education (M)b 2.29 (1.67) PREVC Yes 15 (54%) No 13 (46%) Ngte; PREVC = Previous counseling experience. a Standard deviations in parentheses. b 1 = High school diploma; 2 = Associate’s degree; 3 = Bachelor’s degree; 4 = Master’s degree; 5 = Doctorate; 6 = Other. 39 Twenty therapists (5 men and 15 women) aged 23-52 also participated in the study (see Table 2). The racial/ethnic background of the therapists was Caucasian (80%), African- American (15%) and Asian-American (5%). Seven (35%) of the therapists had little to no therapy/counseling experience (0-1 years), seven (35%) had some experience (2-4 years), and 6 (30%) had a moderate level of experience (5 or more years). Procedures Therapists and clients were recruited from graduate- level clinical training courses, university counseling centers and community counseling centers. Therapists from the various settings were asked to participate in a study that aimed “to learn more about characteristics that contributed to the development of therapeutic relationships.” Therapists were informed that if they chose to participate they would be asked to complete a brief survey packet on three different occasions. Therapists who agreed to participate in the study signed and returned an informed consent form (see Appendix B). Participating therapists were informed to notify the researcher when they were assigned a new adult client whom they planned to see for at least seven sessions of therapy. At that time, 40 Table 2 Therapist Demographic Data N 20 Age (M years) 31.60 (9.10)b Education (M)a 1.85 (.81) GRAD Yes 15 (75%) No 5 (25%) PTO Psychodynamic 7 (35%) Eclectic 3 (30%) Cognitive-Behavioral 4 (22%) Systems 2 (20%) Other 1 (5%) Note. GRAD = Currently enrolled in a graduate program; PTO = Primary theoretical orientation. 3 1 = Bachelor’s degree; 2 = Master’s degree; 3 = Doctorate. b Standard deviations in parentheses. 41 therapists were then given all of the survey packet materials and, using the same protocol, were instructed to recruit their client during the first session of therapy. Clients who agreed to participate signed and returned an informed consent form (see Appendix A). As an incentive, all respondents who finished the study received a small monetary gift ($10.00 video rental gift certificate). To ensure confidentiality and anonymity, surveys were coded and no identifying information was requested from clients. Clients were also informed that therapists would not be seeing their survey responses. Therapist and client survey packets were administered following the lst, 3rd and 7th counseling sessions. Packets contained the measures provided below. Instruments Client Personal Demographic Questionnaire (CPDQ) This questionnaire gathered background information on clients’ sex, age, race/ethnicity, current marital status and highest level of education. Respondents were also asked to indicate if they have ever received counseling before their current counseling experience. Therapist Personal Demographic Questionnaire (TPDQ) This questionnaire gathered background information on therapists’ sex, age, race/ethnicity, current marital status and highest level of education. Respondents were also asked to indicate the number of years of counseling/therapy experience, their primary theoretical orientation as well as status of graduate school (name of program, year in school, 42 etc.) if applicable. Relationship Questionnaire (RQ) This self-classification instrument provides four short paragraphs describing each of the four prototypical attachment styles -- secure, dismissive, preoccupied and fearful (Bartholomew & Horowitz, 1991). Respondents are asked to indicate which one of the four paragraphs best describes their feelings about close relationships. A sample item is, “It is relatively easy for me to become emotionally close to others. I am comfortable depending on others and having others depend on me. I don't worry about being alone or having others not accept me” (Bartholomew & Horowitz, 1991, p. 244). Scharfe and Bartholomew (1994) reported that 56% of men and 63% of women self-reported the same RQ classification over an eight-month interval. Elsewhere, Kirkpatrick and Hazan (1994) reported that 70% percent of their adult respondents reported the same RQ attachment style over a 4-year period. Bartholomew and Horowitz (1991) demonstrated that the four attachment styles related in theoretically-expected directions to measures of self- concept and sociability in a college sample. These findings were consistent across self— and friend-reports. Elsewhere, Horowitz, Rosenberg and Bartholomew (1993) reported that different attachment styles corresponded to different types of interpersonal problems using the RQ in a college sample. More recently, Brennan, Clark and Shaver (1998) reported convergent validity of this measure by demonstrating that 43 the 4 self-reported attachment categories responded in theoretically-expected directions on concurrent measures of anxiety, avoidance, touch, sexual preferences and postcoital emotions. Moreover, these authors reported that findings suggest that, “Anxiety is similar to Bartholomew’s self- model dimension, and Avoidance is similar to her other-model dimension” (p. 64). Adult Attachment Inventory (AAI) This 13-item self-report instrument uses a 7-point scale ranging from “strongly disagree” (1) to “strongly agree” (7) to measure adult attachment orientations. The AAI provides two factor-analytically-derived subscales related to attachment (Simpson, 1990; Simpson, Rholes, & Nelligan, 1992). The avoidance/security subscale indicates respondents' comfort in close relationships. A sample item from this scale is, “I’m not very comfortable having others depend on me” (Simpson, 1990, p. 973). The anxiety subscale indicates the level of tension or worry that individuals typically reported in close relationships. A sample item from this subscale is, “I often want to merge completely with others, and this desire sometimes scares them away” (Simpson, 1990, p. 973). Higher scores on these two subscales respectively indicated greater levels of avoidance and anxiety within close relationships. In a sample of 144 dating couples, Simpson (1990) demonstrated that the attachment security subscales related in theoretically- expected directions to measures of interdependence, 44 commitment, trust, relationship satisfaction and emotionality. Simpson et al. (1992) reported a Cronbach alpha of .81 for the avoidance/security subscale for both men and women in a college sample. Cronbach alphas for the anxiety subscale were .58 and .61 for men and women respectively in a college population. Lopez et al. (1997) reported Cronbach alphas of .83 (avoidance/security) and .70 (anxiety) within a mixed-sex college population. In the present study, the obtained Cronbach’s alpha coefficients for clients were .83 and .70 for the avoidance/security and anxiety subscales, respectively; for therapists, the obtained Cronbach’s alpha coefficients were .85 and .54 for the avoidance/security and anxiety subscales, respectively. For the current study, scores on these two subscales were also aggregated to form a composite insecurity score, which essentially represented a secure versus fearful adult attachment dimension. Obtained alpha coefficients for this composite score were .80 and .84 for clients and therapists, respectively. Working Alliance Inventory (WAI) This 36-item self-report instrument uses a 7-point rating scale ranging from "never" (1) to "always" (7) to measure the quality of the working alliance (Horvath & Greenberg, 1989). Parallel forms are available for the client (WAI-C) and therapist (WAI-T). A sample of an item that appeared on both versions was, “I feel uncomfortable with ” (Horvath & Greenberg, 1989, p. 226). Horvath 45 (1994) reported that “a number of separate investigations provide support of the WAI’s validity” (p.115). For example, according to Tichenor and Hill (1989), the WAI correlated positively with other alliance measures (i.e., California, Psychotherapy Alliance Scales, The Penn Helping Alliance Ratings Scale, The Vanderbilt Therapeutic Alliance Scale). Horvath (1994) reported that the WAI is related in theoretically—expected directions to other global measures of the therapeutic relationship, including Rogerian dimensions (i.e., empathy, unconditional positive regard, empathy). Horvath and Greenberg (1989) reported Cronbach alphas of .93 and .87 for the overall client version and therapist versions, respectively. Kokotovic and Tracey (1990) reported alphas of .91, .88, and .93, respectively, for task, bond and goal scales on the therapist form, and .90, .88, and .91, respectively, for task, bond, and goal on the client form. In the present study, only WAI total scores were used in testing the study’s main hypothesis, although the WAI subscales were examined in several post hoc analyses. Obtained Cronbach’s alpha coefficients for the total WAI-C (client) were .92, .81, and .81 for Time 1, Time 2 and Time 3, respectively. Alphas for the total WAI-T (therapist) were .89, .91, and .86 for Time 1, Time 2, and (Time 3, respectively. Self-Understanding Scale (SUS) This 3-item self-report instrument, taken from the 16- item Session Impacts Scale (SIS), uses a 5-point scale 46 ranging from “not at all” (1) to “very much” (5) to assess the level of understanding (i.e., realization about self, realization about others, clearer awareness) experienced by clients in therapy sessions (Elliot & Wexler, 1994). A sample item was, “As a result of this session, I now have insight about another person or have understood something new about someone else or people in general" (Elliot & Wexler, 1994, p. 173). Stiles et al. (1994) reported a reliability estimate of .78 for the Understanding Scale and also suggested validity by demonstrating that the scale was positively associated with subscales of the Session Evaluation Questionnaire (i.e., Depth, Positivity, and Good Therapist). Likewise, Hill, Diemer and Heaton (1997) reported an internal consistency (alpha) of .75 for a sample of volunteer adult clients. Obtained alphas in the current study were .81, .37, and .15 for Time 1, Time 2, and Time 3, respectively, indicating that, beyond the initial assessment, this measure did not evidence acceptable levels of reliability. Hypotheses First, it was hypothesized that therapist and client ratings of the working alliance would be differentially related across three time points. More specifically, (a) therapist and client ratings of the working alliance would not be significantly related at Time 1 (after the first session), or at Time 2 (after the fourth session), but that they would be significantly related at Time 3 (following the 47 seventh session). Secondly, it was hypothesized that therapists' and clients’ adult attachment security would be positively related to ratings of the working alliance across all three time points. Additionally, it was hypothesized that therapists' adult attachment orientation would moderate the relation of therapist and client ratings of the working alliance at each time point. More specifically, among therapists who self-report a secure attachment orientation, therapist and client working alliance ratings would be more strongly and positively related. Third, it was hypothesized that therapist and client adult attachment orientations would significantly predict correspondence of client and therapist ratings of the working alliance at time 3. More specifically, therapist and client adult attachment security would be positively related to working alliance correspondence. Finally, the following research questions were also explored in this investigation: Are client and therapist adult attachment orientations related to client self- understanding scores over time? Are client- and therapist- rated working alliance scores related to client self— understanding scores across the three time points? 48 Analyses First, descriptive statistics were conducted on the client and therapist background variables. Next, correlations among client and therapist background variables and outcome measures, working alliance total scores over time and frequencies of adult attachment styles were calculated. Intercorrelations were then computed to explore the relations among therapist and client working alliance ratings across the three time points (hypotheses 1). Several analyses were then conducted to examine the prediction that therapist and client adult attachment security would be positively related to working alliance ratings across three time points (hypothesis 2). First, intercorrelations were computed among therapist and client adult attachment orientations and client and therapist working alliance ratings across the three time points. Next, a series of T-tests were conducted to examine the mean differences on working alliance ratings across two therapist and client self-reported attachment style groupings - secure and insecure (i.e., dismissive, preoccupied and fearful). Finally, two 3 x 2 (Time X Pair) ANOVAs with repeated measures were conducted to examine the impact of dyad attachment security on client— and therapist-rated working alliance scores over time. To examine the third hypothesis a series of three multiple regressions were conducted to examine if therapist 49 attachment security moderated the relation of therapist and client working alliance ratings over time. To examine this hypothesis, therapist adult attachment orientations - anxiety and avoidance — were again combined to create a composite “insecurity” score. Three multiple regressions were then conducted to predict the criterion variable -— client working alliance ratings -— at each time point. Next, the predictor variables (i.e., therapist attachment insecurity, therapist working alliance scores) were centered, and the therapist attachment insecurity X therapist working alliance interaction was created by multiplying the two centered predictors. These variables were then entered into a hierarchical regression equation. Finally, regression analysis was conducted to examine if therapist and client adult attachment orientations predicted the correspondence in their respective ratings of the working alliance (hypothesis 4). For this analysis, the criterion variable called “correspondence” scores were computed by obtaining the absolute value of the difference between therapist and client working alliance scores at Time 3. The predictor variables -- therapist attachment security, Client attachment insecurity -- were then entered simultaneously into a regression equation. 50 RESULTS This chapter will include several subsections. First, it will review the relevant descriptive statistics that were collected on this sample. It will then systematically report the results of the statistical analyses for each of the main hypotheses and the research questions under investigation. Finally, it will present the findings from a limited number of post hoc analyses that were conducted to further examine the results of the initial analyses. Preliminary Descriptive Analyses Intercorrelations were computed to determine if therapist or client background variables (i.e., sex, age, level of education, previous counseling experience) were related to any of the criterion variables (i.e., adult attachment orientations, working alliance ratings, and self- understanding scores). Client Background Variables For clients, results indicated that none of the background variables were consistently related to the outcome variables over time (see Table 3). However, a few significant relationships were observed. Specifically, age was related to self-understanding scores at Time 1 (r = .63, p < .01), with older clients reporting higher self- understanding scores; education level was related to working alliance scores at Time 3 (r = —.54, p < .05), with less educated clients reporting higher alliance scores; and 51 Table 3 Correlations Among Client Background Variables and Outcome Measures AV ANX WAIl WAIZ WAI3 3051 susz SUS3 1. Sexa -.13 .20 .13 .02 .24 -.13 -.29 —.13 2. Age .05 .05 .14 .42 -.25 .63** .26 —.18 3. 300b .12 -.06 .13 .02 -.54* .33 .31 -.11 4. PREC -.36 -.26 -.13 -.04 .24 -.47* -.24 .47 Note. AV = Avoidance subscale score of the Adult Attachment Inventory (AAI); ANX = Anxiety subscale score of the AAI; WAIl = Working Alliance Inventory (WAI) total score at Time 1 (n = 28); WAIZ = WAI total score at Time 2 (n = 22); WAI3 = WAI total score at Time 3 (n = 17); SUSl = Self— Understanding score at Time 1 (n = 28); SUSZ = Self— Understanding score at Time 2 (n = 22); SUS3 = Self- Understanding score at Time 3 (n = 17); PRE = Previous counseling. a l = Male; 2 = Female. b 1 = High School Diploma; 2 = Associates Degree; 3 = Bachelor’s Degree; 4 = Master’s Degree; 5 = Doctorate. C 1 = Yes; 2 = No. * E < .05. ** E < .01. 52 previous counseling was related to session impacts scores at Time 2 (r = -.47, p < .05), with clients who had received previous counseling reporting higher self-understanding scores. Since none of the client background variables were systematically related to the outcome variables, they were not used as covariates in subsequent analyses. A series of t-tests were also conducted to examine whether clients who completed all seven counseling sessions (n = 17) and those who dropped out prematurely (n = 11) significantly differed from one another with regard to their adult attachment orientations or to their initial ratings of the working alliance. Results indicated no significant differences between these two groups on these measures. Therapist Background Variables For therapists, results indicated that none of the background variables (i.e., sex, age, highest level of education, years of therapy experience) were consistently related to the outcome variables over time (see Table 4). Since none of the therapist background variables were systematically related to the outcome variables, they were not Used as covariates in subsequent analyses. Frequencies of Adult Attachment Styles Table 5 shows the obtained NS and frequencies of therapist self-reported adult attachment styles: secure (n = 11; 55%), dismissive (n = 2; 10%), preoccupied (n = 2; 10%), and fearful (n = 5; 25%). For clients, the obtained NS and 53 frequencies were secure (3 = 6; 21%), dismissive (n = 5; 18%), preoccupied (n = 6; 21%), and fearful (n = 11; 40%). Compared to other investigations using the Relationship Questionnaire with non—clinical samples (Horowitz, Rosenberg, & Bartholomew, 1993; Lopez et al., 1997; Pistole, 1995), the proportion of therapist self-reporting a secure attachment style (55%) was higher than previously reported, whereas the proportion of clients self—reporting a secure attachment style (21%) was lower than previously reported. However, the higher proportion of insecure clients found in this study is consistent with another study (Dolan, 1992) that found that insecure attachment styles were over represented in a clinical sample. A Chi—square analysis revealed that, within the current sample, therapists were more likely to self-report a secure adult attachment style than were clients, X2 (1, N = 48) = 5.75, p < .05. Client and Therapist Working Alliance Ratings Table 6 displays the means and standard deviations for client and therapist working alliance ratings across the three time points. These findings indicate that the average working alliance ratings by clients and by therapists increased across the three time points demonstrating that both clients and therapists viewed their relationship as improving over time. In general, these working alliance scores were consistent with those observed elsewhere (Mallinckrodt, 1993). 54 Table 4 Correlations Among Therapist Background Variables and Outcome Measures AV ANX WAIl WAI2 WAI3 1. Sexa .29 .32 .10 .06 .10 2. Age .23 -.39 .22 .09 .24 3. EDUCb .17 —.14 -.01 .04 .33 4. YRSC .34 .13 .01 .08 .42 Note. AV = Avoidance subscale score of the Adult Attachment Inventory; ANX = Anxiety subscale score of the Adult Attachment Inventory; WAIl = Working Alliance Inventory Working Alliance Inventory total score at Time 1; WAI2 total score at Time 2; WAI3 Working Alliance Inventory total score at Time 3; MSTAT Marital status; EDUC = Highest degree obtained; YRS Years of counseling/therapy experience. a 1 = Male; 2 = Female. b 1 = Bachelor’s degree; 2 = Master’s degree; 3 = Doctorate. C 1 = 0—1 years; 2 = 2-4 years; 3 = 5 or more years. * p < .05. ** p < .01. 55 Table 5 Frequency of Client and Therapist Self-Reported Adult Attachment Styles on the Relationship Questionnaire (RQ) § 2 3 _F_ TOT THERAPIST 11(55%) 2(10%) 2(10%) 5(25%) 20 CLIENT 6(21%) 5(18%) 6(21%) 11(40%) 28 TOTAL 17(35%) 7(15%) 8(17%) 16(33%) 48 Note. S = Secure; D = Dismissive; P = Preoccupied; F = Fearful. 56 Table 6 Client and Therapist Working Alliance Scores Over Time M CWAIl 199.18 CWA12 217.50 CWAI3 221.88 29.20 15.88 13.90 TWAIl TWAI2 TWAI3 IZ 193.93 201.91 210.18 .77 .97 .87 Note. CWAIl total score at Time 1; Inventory total score at Time 2; Working Alliance Inventory total score at Time 3; TWAIl Client-rated Working Alliance Inventory CWAI2 CWAI3 Client-rated Therapist-rated Working Alliance Inventory total score at Time 1; TWAI2 = total score at Time 2; TWAIB Alliance total score at Time 3. 57 Therapist-rated Working Therapist-rated Working Alliance Inventory Client-rated Working Alliance Results of Primary Analyses The primary purpose of this investigation was to examine the relations among client and therapist adult attachment styles/orientations to measures of the therapeutic alliance over time. The following section will present results of the analyses of the main hypotheses and research questions under investigation. Relations of Therapist and Client Ratings of the Working Alliance Over Time The first hypothesis predicted that therapist and client ratings of the working alliance would be differentially related across the three time points. More specifically, therapist and client ratings of the working alliance would not be significantly related at Time 1 (after the first session), or at Time 2 (after the fourth session), but would be significantly related at Time 3 (following the seventh session). To examine this hypothesis, intercorrelations among therapist and client working alliance ratings across the three time points were computed (see Table 7). Contrary to prediction, results indicated that therapist and client working alliance ratings were significantly related at Time 1 (r = .42, p < .05), and at Time 2 (r = .62, p < .01), but not at Time 3 (r = .10). This finding indicated that therapists and clients shared similar perceptions of the working alliance early in therapy (i.e., following the 1St 4th and sessions); however, by the 7U‘session, therapist and 58 Table 7 Intercorrelations Among Therapist and Client Working Alliance Ratings Over Time TWAIl TWAI2 TWAI3 CWAIl .42* .26 -.14 CWAIZ .36 .62** .19 CWAI3 .45 .OO .10 Note; CWAIl = Client-rated Working Alliance Inventory total score at Time 1; CWAI2 = Client-rated Working Alliance Inventory total score at Time 2; CWAIB = Client-rated Working Alliance Inventory total score at Time 3; TWAIl = Therapist-rated Working Alliance Inventory total score at Time 1; TWAIZ = Therapist-rated Working Alliance Inventory total score at Time 2; TWAI3 = Therapist—rated Working Alliance total score at Time 3. * P < .05. ** P < .01. 59 client perceptions of the working alliance were virtually unrelated. Relations of Therapists’ and Clients’ Adult Attachment to the Working Alliance Over Time The second hypothesis predicted that therapists' and clients’ adult attachment security would be positively related to working alliance ratings across all three time points. To examine this hypothesis, several different analyses were conducted. First, using the dimensional measure of adult attachment (Adult Attachment Inventory), intercorrelations were computed to explore the relations between therapist and client adult attachment orientations and working alliance ratings across three time points. Next, using the categorical measure of adult attachment (Relationship Questionnaire), the relations among therapist and client adult attachment styles and working alliance ratings over time were explored. Finally, the relations between within-dyad adult attachment styles and working alliance ratings over time was examined via an analysis of variance with repeated measures. Relations Among Dimensional Measures of Adult Attachment and Client Workinnglliance Ratings To first examine if therapist and client adult attachment security was positively related to working alliance scores over time (hypothesis 2), intercorrelations were computed among therapist and client adult attachment orientations and client and therapist working alliance 6O ratings. For these analyses, attachment orientations were represented by the two dimensional subscales — avoidance and anxiety. In addition, because therapist avoidance and anxiety scores were significantly intercorrelated (r = .51, p < .01), scores on both indexes were aggregated to form a composite insecurity score, which essentially represented a secure versus fearful adult attachment dimension. That is, low scores on the composite scale were in the secure direction (i.e., low avoidance, low anxiety), whereas high scores were in the fearful direction (i.e., high avoidance, high anxiety). Therapist adult attachment orientations and client working alliance ratings over time. Results indicated that therapist avoidance was not related to client working alliance scores at Time 1 (r = .33), Time 2 (r = -.07), or Time 3 (r = -.16). However, therapist anxiety was significantly related to client-rated working alliance scores at Time 1 (r = .40, p < .05) but not at Time 2 (r = -.28) or Time 3 (r = —.24) (see Table 8). A similar pattern was observed using the composite therapist insecurity score. These results indicated that therapist composite insecurity was related to client working alliance ratings at Time 1 (r = .40, p < .05), but not at Time 2 (r = -.16) or Time 3 (r = -.21) (see Table 8). The above findings suggested that there was a possible “reversal” in the relationship between therapist adult attachment orientations and client-rated working alliance 61 Table 8 Intercorrelations Among Thergiist and Client Adult Attachment Inventory (AAI) Scores and Working Alliance Inventory (WAI) Scores Over Time Scale 1 2 3 4 5 6 7 8 9 10 11 12 1. CA .31 .86* -.14 -.04 -.O7 .16 .23 .20 .00 -.10 -.14 2. CX .75” -.01 -.22 .32 .00 .08 .03 .22 -.22 .23 3. CC -.ll -.l6 .12 .ll .20 .16 .19 -.20 .02 4. Cl .50" .24 .33 .40* .40" .42" .26 -.14 5. C2 .36 -.07 -.28 -.16 .36 .62" .19 6. C3 -.16 -.24 -.2] .45 .00 .10 7. TV .51" .94" -.04 .13 -.O6 8. TX .77" .05 .02 .l 1 9. TC -.01 .10 .00 10. T1 .34 .22 11. T2 .56“ 12. T3 Ngtg CA = Client avoidance subscale score on the AAI; CX = Client anxiety subscale score on the AAI; CC = Client composite insecurity score on the AAI; C1 = Client WAI total score at Time 1; C2 = Client WAI total score at Time 2; C3 = Client WAI total scores at Time 3; TA = Therapist avoidance subscale score on the AAI; TX = Therapist anxiety subscale score on the AAI; TC = Therapist composite insecurity score on the AAI; T1 = Therapist WAI total score at Time 1; T2 = Therapist WAI total score at Time 2; T3 = Therapist WAI total score at Time 3. * p<.05. ”p<.OI.Table8 62 scores over time. That is, following the 1St counseling session, therapists with higher anxiety and composite insecurity scores received significantly higher client-rated working alliance scores. Conversely, following the 7th session, therapists with higher anxiety and composite insecurity scores received moderately lower client—rated working alliance scores. This emergent trend was in the expected direction and indicated that, as sessions progressed, securely attached therapists formed more positive client-rated working alliances than did their insecurely attached peers. Therapist adult attachment orientations and therapist working alliance ratings over time. Using the same procedure as described above, results indicated that, contrary to prediction, therapist adult attachment orientations were not significantly related to therapist-rated working alliance scores over time (see Table 8). That is, therapist avoidance was not significantly related to therapist working alliance ratings at Time 1 (g = —.04), Time 2 (r = .13), or Time 3 (E = -.06). Likewise, therapist anxiety was not related to therapist working alliance ratings at Time 1 (g = .05), Time 2 (r = .02), or Time 3 (g = .11). Finally, therapist composite insecurity was not related to therapist working alliance ratings at Time 1 (E = -.12), Time 2 (r = .02), or Time 3 (r = .11). 63 Client adult attachment orientations and client working alliance ratings over time. Results indicated (see Table 8) that client anxiety was not significantly related to client working alliance ratings at Time 1 (g = -.01), Time 2 (g = -.22) or Time 3 (g = .32). Similarly, client avoidance was not significantly related to client working alliance ratings at Time 1 (g = -.14), Time 2 (g = -.04) or Time 3 (E = -.07). Finally, because client avoidance and anxiety scores were intercorrelated (5 = .31), scores on both indexes were aggregated to form a composite insecurity score, which essentially represented a secure versus fearful adult attachment dimension. Using this “insecurity” composite score, results indicated that client composite insecurity was not significantly related to working alliance scores at Time 1 (E = -.11), Time 2 (g = — .16), or Time 3 (E = .12). Taken together, these results indicated that, contrary to prediction, client adult attachment orientations did not appear to be significantly related to client-rated working alliance scores across the three time points. Client adult attachment orientations and therapist— rated working alliance scores over time. Results indicated that client adult attachment orientations were not related to therapist-rated working alliance scores across the three time points (see Table 8). Specifically, client avoidance was not significantly related to therapist working alliance scores at Time 1 (E = .00), 64 Time 2 (r = -.10) or at Time 3 (r = -.14). Likewise, client anxiety was not significantly related to therapist working alliance scores at Time 1 (g = .22), Time 2 (g = -.22) or at Time 3 (E = .23). Finally, client composite insecurity (as described above) was not significantly related to therapist working alliance scores at Time 1 (r = .12), Time 2 (g = - .20), or at Time 3 (r = .02). Taken together, these results indicated that, contrary to prediction, client attachment security did not appear to be significantly related to therapist-rated working alliance scores across the three time points. Relations Amonggthe Categorical Measure of Adult Attachment and Working Alliance Ratingg Examining the relation of client and therapist self- reported adult attachment and client- and therapist-rated working alliance scores provided an alternative test of this hypothesis. Due to the small Ns and low frequencies of some of the insecure categories, for both therapists and clients, those participants with insecure attachment styles (i.e., preoccupied, dismissive and fearful) were respectively aggregated into a common insecure group, thus permitting a secure versus insecure comparison within each participant group. Therapist adult attachment styles and working alliance scores over time. A series of T-tests were conducted to examine the mean differences on client-rated working alliance scores between 65 secure and insecure therapists (i.e., dismissive, preoccupied and fearful) across the three time points. Contrary to prediction, results indicated a significant difference for therapist adult attachment style at Time 1 (E = 2.11, p < .05, with insecure therapists having significantly higher client-rated working alliance scores than did secure therapists. However, no significant differences were observed at Time 2 or at Time 3. Results also indicated that there were no significant differences between therapist adult attachment styles and therapist-rated working alliance scores at Time 1, Time 2, or Time 3. In short, insecure therapists had higher client- rated working alliances following the first counseling session; however, no other significant differences were observed between therapist attachment style and client- or therapist-rated working alliance scores at each time point. Client adult attachment styles and working alliance scores over time. Using the same procedure as described above, a series of T-tests were again conducted to examine the mean differences on working alliance ratings between secure and insecure clients (i.e., dismissive, preoccupied and fearful) across the three time points. Contrary to prediction, results indicated no significant differences on client-rated working alliance scores between secure and insecure clients at Time 1, Time 2, or Time 3. However, at Time 3, secure clients reported moderately higher working alliance scores 66 than did insecure clients, this difference was in the expected direction and approached significance, t = —2.08, p = .055. Results also indicated that there were no significant differences between client adult attachment styles and therapist-rated working alliance scores at Time 1 or Time 2. However, at Time 3, as predicted, results indicated that secure clients received significantly higher therapist-rated working alliance scores, t = —2.72, p < .05, than did insecure clients. In short, the results indicating that insecure therapists had more positive client-rated working alliances at Time 1 was unexpected. However, the finding, at Time 3, that secure clients had significantly higher therapist-rated alliance scores was anticipated and provided partial support for the hypothesis that client and therapist attachment would be positively related to working alliance scores over time. Likewise, although not statistically significant, the finding that secure clients reported higher working alliance scores at Time 3 also provided tentative support for this hypothesis. The Impact of Within Dyad Adult Attachment Styles on Working Alliance Ratings Over Time As an alternative means of testing the second hypothesis, and to more sensitively assess the impact of client and therapist adult attachment styles on the formation of the working alliance across the major time 67 points in this study, the combined attachment styles of each of the 28 therapist-client dyads were obtained. Due to the small Ns and low frequencies of some of the insecure categories, the insecure attachment styles (i.e., preoccupied, dismissive and fearful) of clients and therapists were again aggregated into a common insecure group. Using this procedure, there were 13 secure therapist-insecure client dyads and 9 insecure therapist- insecure client dyads. The remaining 6 dyads (consisting of 3 secure therapists-secure clients dyads and 3 insecure therapist-secure clients dyads) were dropped due to low frequencies. Thus, the following analysis is a comparison of dyads containing secure therapists versus dyads with insecure therapists over the three time points. In all of these dyads, the clients were insecure. Dyad adult attachment styles and client working alliance scores. A 3 x 2 (Time X Pair) ANOVA with repeated measures was conducted to examine the impact of dyad attachment security on client working alliance ratings over time. A significant main effect was obtained for Time, F (2,11) = 5.76, p < .05. A significant Time X Pair interaction was also obtained, F (2,11) = 5.76, p < .05. The significant interaction effect was then explored in more detail via within subject contrasts which indicated a significant linear effect, F (1,12) = 8.11, p < .05, and quadratic effect, F (1,12) = 9.68, p < .01. The quadratic effect indicated that the two 68 dyads (those with insecure therapists versus those with secure therapists) had different slopes (or average change) across the three time points. A series of 3 ANOVAs indicated a significant between—subject effect at Time 1, F (1,21) = 5.23, p < .05, with secure therapist dyads receiving lower client-rated working alliance scores than did insecure therapist dyads. No significant between-subject effects were observed at Time 2, F (1,17) = .54, p = .47, or at Time 3, F (1,13) = .26, p = .62. In order to visually represent these results, the client-rated working alliance scores by time are plotted on Figure 1. This figure illustrates that dyads containing secure therapists tended to have initially lower client working alliance ratings, but over time demonstrated a significant pattern of change from low—to-high; conversely dyads with insecure therapists tended to have initially higher client working alliance scores, but over time demonstrated virtually no change. Dyad adult attachment styles and therapist working alliance scores. A 3 x 2 (Time X Pair) ANOVA with repeated measures was conducted to examine the impact of dyad attachment security on therapist working alliance scores. Significant main effects were obtained for Time, F (2,11) = 4.05, p < .05. However, the Time X Pair interaction effect was nonsignificant. A series of 3 ANOVAs indicated that secure therapist dyads did not demonstrate significant differences on therapist-rated alliance scores at Time 1, F (1,21) = 69 .02, p_= .89, or Time 2, F (1,17) = .54, p = .47. However, at Time 3, secure therapist dyads received modestly higher working alliance ratings, than did insecure therapist dyads, F (1,13) = 1.97, p = .19. In order to visually represent these results, the therapist-rated working alliance scores by Time are plotted on Figure 2. Therapist Adult Attachment Orientation as a Moderator of Client and Therapist Working Alliance Ratings Over Time The third hypothesis predicted that therapists' adult attachment orientations would moderate the relation of therapist and client ratings of the working alliance at each time point. More specifically, it was anticipated that among therapists who self—reported a secure attachment orientation, therapist and client working alliance ratings would be more strongly and positively related at each time point. To examine this hypothesis, therapist adult attachment orientations - anxiety and avoidance - were again combined to create a composite “insecurity” score. Three multiple regressions were then conducted to predict the criterion variable -- client working alliance ratings -- at each time point. Following the recommendations of Aiken and West (1991), the predictor variables (i.e., therapist attachment insecurity, therapist working alliance scores) were centered, and the therapist attachment insecurity X therapist working alliance interaction was created by multiplying the two centered predictors. These variables were then entered into a hierarchical regression equation. 70 Figure 1 The Impact of Dyad Adult Attachment Styles on Client Working Alliance Inventory (WAI) Ratings Over Time ‘5. . ... . - Secure Therapist Dyads l+ Insecure Therapist Dyads 2301 220 «L M1144 ’—.:M/. g, 210 F .E ' E 200 4- 2 E 190 - c .. .2 i 23 180 170 160 i I fl T1 T2 T3 Time 71 Figure 2 The Impact of Dyad Adult Attachment Styles on Therapist Working Alliance Inventory (WAI) Ratings Over Time i [r m. .. Secure Therapist Dyads l ( + Insecure Therapist Dyads 215 ‘T:_ . i .p E I .' ; 210 5E— ; 1 ." a g 205 J _. __5 .E I ,l : 2 200 L .5 a E E 195 -: a i e I '2 I ,_ 190 i 185 E 1 180 LWLAWWW T1 T2 T3 I Time 72 Results indicated that therapist attachment security did not moderate the relationship between client and therapist working alliance ratings over time as none of the interaction terms approached significance at any of the three time points (see Tables 9-11). Thus, contrary to prediction, these findings indicated that therapists’ adult attachment orientations did not moderate the relation of therapist and client working alliance ratings over time. Therapist and Client Adult Attachment Orientations and Correspondence of Working Alliance Ratings The fourth and final hypothesis predicted that therapist and client attachment orientations would significantly predict correspondence of client and therapist ratings of the working allianCe at Time 3. More specifically, therapist and client attachment security would be positively related to the correspondence in their respective ratings of the working alliance. To test this hypothesis, a client—therapist working alliance correspondence score was first computed by obtaining the absolute value of the difference between therapist and client working alliance scores at Time 3. Next, using the composite attachment insecurity scores (as previously described), the predictor variables -- therapist attachment insecurity and client attachment insecurity -- were entered simultaneously into a regression equation predicting Time 3 discrepancy scores. Contrary to prediction, results indicated that therapist and client attachment security did 73 Table 9 Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 1 Model Summary: R R Square Adj. R Std. Error Model 1 .59a .35 .29 24.55 Model 2 .60b .36 .28 24.85 a Predictors: (Constant) Therapist WAI scores at Time 1 (centered), Therapist AAI composite scores (centered). bPredictors: (Constant) Therapist WAI scores at Time 1 (centered), Therapist AAI composite scores (centered), interaction term at Time 1 (centered). Change Statistics: R Square F dfl df2 Sig. F Model 1 .35 6.60 2 25 .005 Model 2 .01 .39 1 24 .538 74 Table 10 Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 2 Model Summary: R R Square Adj. R Std. Error Model 1 .66a .44 .38 12.52 Model 2 .68b .47 .47 12.53 a Predictors: (Constant) Therapist WAI scores at Time 2 (centered), Therapist AAI composite scores (centered). b Predictors: (Constant) Therapist WAI scores at Time 2 (centered), Therapist AAI composite scores (centered), interaction term at Time 2 (centered). Change Statistics: R Square F dfl df2 Sig. F Model 1 .44 7.38 2 19 .004 Model 2 .03 .98 1 18 .334 75 Table 11 Regression Analysis Testing the Moderating Role of Therapist Attachment Security (AAI) on Client Working Alliance (WAI) Ratings at Time 3 Model Summary: R R Square Adj. R Std. Error Model 1 .23a .05 -.08 14.46 Model 2 .29b .08 -.13 14.78 a Predictors: (Constant) Therapist WAI Scores at Time 3 (centered), Therapist AAI composite scores (centered). bPredictors: (Constant) Therapist WAI scores at Time 3 (centered), Therapist AAI composite scores (centered), interaction term at Time 3 (centered). Change Statistics: R Square F dfl df2 Sig. F Model 1 .05 .39 2 14 .684 Model 2 .03 .41 1 13 .533 76 not significantly predict correspondence of client-therapist working alliance ratings at Time 3, F(2, 24) = .351, p = .707. Adult Attachment Orientations and Self-Understanding,Scores The first research question asked: Are client and therapist adult attachment orientations related to client self—understanding scores over time? To examine this question, intercorrelations were computed among client and therapist attachment orientations (i.e., anxiety, avoidance, composite insecurity) and self-understanding scores across the three time points (see Table 12). Results indicated that client and therapist attachment orientations were not significantly related to client self-understanding scores across the three time points. Working Alliance Ratings and Self-Understanding Scores The second research question asked: Are client- and therapist-rated working alliance scores related to client self-understanding scores across the three time points? To explore this question, intercorrelations were computed among self-understanding scores and client- and therapist-rated working alliances. Results indicated (see Table 13) that client-rated working alliance scores were significantly related to client self-understanding scores at Time 1 (E = .58, p = .001), at Time 2 (p = .60, p = .008), and moderately related at Time 3 (g = .38). In other words, clients rating the therapeutic relationship more positively also reported higher self-understanding scores at each of 77 the three time points. Therapist-rated working alliance scores were not significantly related to session impacts scores at Time 1 (E = .17), but were significantly related at Time 2 (p .55, p < .05), and marginally related at Time 3 (E = .40 p = .11). These results indicated that, over the final two time points, therapists who rated the working alliance more favorably had clients who correspondingly reported higher self-understanding scores. Post hoc analyses In order to clarify the nature of relations between client and therapist adult attachment and the development of the working alliance over time, a limited series of post hoc analyses were conducted. First, the primary analyses indicated that therapist and client working alliance ratings were differentially related over time. To further explore this relationship, additional analyses explored how client and therapist working alliance subscale scores -- bond, goals and tasks -- were related over time. Next, the primary analyses indicated that client and therapist adult attachment orientations were differentially related to working alliance scores over time. To further assess this relationship, additional analyses examined how client and therapist adult attachment orientations were related to working alliance subscale scores over time. Finally, results indicated that client and therapist attachment security did not predict the correspondence of client—therapist working alliance ratings. To further examine this finding, post—hoc 78 Table 12 Correlations Among Client and Therapist Adult Attachment Orientations and Self—Understanding Scores Over Time SUSl SUS2 SUSB 1. CAV -.1O .26 -.13 2. CAX .01 .06 .21 3. CAVAX -.O7 .21 .02 4. TAV .03 .43 .12 5. TAX -.02 .14 .10 6. TAVAX .02 .37 .12 Note. CAV = Client avoidance subscale score on the Adult Attachment Inventory; CAX = Client anxiety subscale score on the Adult Attachment Inventory; CAVAX = Client composite insecurity score on the Adult Attachment Inventory; TAV = Therapist avoidance subscale score on the Adult Attachment Inventory; TAX = Therapist anxiety subscale score on the Adult Attachment Inventory; TAVAX = Therapist composite insecurity score on the Adult Attachment Inventory; SUSl = Self—Understanding score at Time 1; SUS2 = Self— understanding score at Time 2; SUS3 = Self-Understanding score at Time 3. 79 Table 13 Correlations Among Client and Therapist Working Alliance Inventory (WAI) Ratings and Self-Understanding Scores Over 21.99 SUSl SUSZ SUSB l. CWAIl .58** .44 .19 2. CWAIZ .59** .60** .10 3. CWAI3 .14 -.02 .38 4. TWAIl .16 .33 .42 5. TWAI2 .35 .55* .04 6. TWAI3 .08 .25 .40 Note. CWAIl = Client WAI total score at Time 1; CWAI2 = Client WAI total score at Time 2; CWA13 = Client WAI total score at Time 3; TWAIl = Therapist WAI total score at Time 1; TWAI2 = Therapist WAI total score at Time 2; TWAI3 = Therapist WAI total score at Time 3; SUSl = Self- Understanding score at Time 1; SUSZ = Self-understanding score at Time 2; SUSB = Self—Understanding score at Time 3. * E < .05. ** E < .01. 80 analyses explored the discrepancy of client and therapist working alliance scores over time. The results of these post hoc analyses are presented below. Correlations Among Working Alliance Subscale Scores Intercorrelations were computed to examine how client and therapist working alliance subscale scores — bond, goals, and tasks -- were differentially related over time (see Tables 14-16). Results indicated that, client and therapist goal scores were significantly related at Time 1 (E = .58, p < .001) and Time 2 (E = .43, p'< .05), but not at Time 3. Client and therapist task scores were not related at Time 1 (E = .30), significantly related at Time 2 (p = .56, p < .01), but were not related at Time 3 (p = -.08). Interestingly, client and therapist bond scores were not significantly related at Time 1 (p = .24), at Time 2 (E = .04), or at Time 3 (p = .06). These results indicated that therapists and clients had common perceptions of the goals and tasks components of the working alliance early, but not later in therapy. These findings were generally consistent with the primary finding of this study that clients and therapists held common perceptions of the overall working alliance early in therapy. However, the low correlations among bond subscale scores appeared to indicate that, when it came to feelings of closeness or positive attachment within the counseling relationships, therapists and clients had divergent perceptions at each of the three time points. 81 Adult Attachment Orientations and Working Alliance Subscale Scores Intercorrelations were also computed to determine if there was a relation between client and therapist adult attachment orientations and therapist and client ratings of working alliance subscale scores across the three time points. Client Adult Attachment Orientations Results indicated that client avoidance and client composite insecurity (as described above) were not significantly related to any client— or therapist-rated working alliance subscale scores across the three time points (see Tables 17—19). However, client anxiety was negatively related to client-rated task scores at Time 2 (g = -.44, p < .05), which indicated that less anxious clients had higher client-rated task scores following the 4th counseling session. Therapist Adult Attachment Orientations Results indicated that therapist adult attachment orientations were not related to any therapist—rated working alliance subscale scores across the three time points. However, therapist adult attachment orientations were significantly related to several client-rated working alliance subscale scores across the three time points (see Tables 17-19). First, therapist anxiety was related to bond (p = .38, p < .05) and task (E = .38, p < .05) scores at Time 1, which indicated that therapists with higher anxiety 82 Table 14 Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 1 T-Bond T-Goal T-Task C-Bond .24 .32 .27 C-Goal .36 .58** .40* C-Task .17 .39* .30 Note. C—Bond = Client WAI bond score; C-Goal = Client WAI goal score; C-Task = Client WAI task score; T—bond = Therapist WAI bond score; T—Goal = Therapist WAI goal score; T-Task = Therapist WAI task score. * p < .05. ** p < .01. 83 Table 15 Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 2 T-Bond T-Goal T-Task C-Bond .04 .42 .32 C-Goal .52* .43* .29 C-Task .46* .65** .56** Note. C-Bond = Client WAI bond score; C-Goal = Client WAI goal score; C-Task = Client WAI task score; T-bond = Therapist WAI bond score; T-Goal = Therapist WAI goal score; T—Task = Therapist WAI task score. * E < .05. ** E < .01. 84 Table 16 Correlations Among Therapist and Client Working Alliance Inventory (WAI) Subscale Scores at Time 3 T-Bond T—Goal T-Task C-Bond .06 -.10 -.02 C-Goal .38 .06 .06 C-Task .42 .09 -.08 Note. C—Bond = Client WAI bond score; C—Goal = Client WAI goal score; C-Task = Client WAI task score; T—bond = Therapist WAI bond score; T-Goal = Therapist WAI goal score; T-Task = Therapist WAI task score. * p < .05. ** p < .01. 85 Table 17 Relations Among Therapist and Client Adult Attachment Inventory (AAI) scores and Working Alliance (WAI) Subscale Scores at Time 1 CAV CAX CVX TAV TAX TVX 1. CWAB -.20 .05 -.11 .33 .38* .39* 2. CWAG -.O4 -.O7 -.O7 .33 .36 .38* 3. CWAT -.15 -.O2 -.12 .26 .38* .34 4. TWAB -.O8 -.15 .02 -.O4 -.17 -.10 5. TWAG .Ol -.20 .09 .05 .15 .09 6. TWAT .07 -.24 .22 -.15 .09 -.08 ngg; CAV Client AAI avoidance subscale score; CAX = Client AAI anxiety subscale score; CVX = Client AAI composite insecurity score; TAV = Therapist AAI avoidance subscale score; TAX = Therapist AAI anxiety subscale score; TVX = Therapist AAI composite insecurity score; CWAB = Client WAI bond score; CWAG Client WAI goal score; CWAT = Client WAI task score; TWAB = Therapist WAI bond score; TWAG = Therapist WAI goal score; TWAT = Therapist WAI task score. * 2 < .05. ** E < .01. 86 Table 18 Relations Among Therapist and Client Adult Attachment and Working Alliance Subscale Scores at Time 2 CAV 1. CWAB .29 2. CWAG -.22 3. CWAT -.16 4. TWAB -.19 5. TWAG .05 6. TWAT -.lO CAX CVX TAV TAX TVX .17 .30 .38 .27 .37 -.24 —.29 -.32 -.54** -.44* —.44* -.37 -.19 -.35 —.27 - 15 — 22 .08 - 11 01 - 20 — 08 .18 17 19 - 24 - 21 .07 01 05 Note. CAV Client AAI avoidance subscale score; CAX = Client AAI anxiety subscale score; CVX = Client AAI composite insecurity score; TAV = Therapist AAI avoidance subscale score; TVX = Therapist Client WAI bond Client WAI task = Therapist WAI TAX = Therapist AAI anxiety subscale score; AAI composite insecurity score; CWAB = score; CWAG Client WAI goal score; CWAT = score; TWAB Therapist WAI bond score; TWAG goal score; TWAT = Therapist WAI task score. * B < .05. ** E < .01. 87 Table 19 Relations Among Therapist and Client Adult Attachment and Working Alliance Subscale Scores at Time 3 CAV CAX CVX TAV TAX TVX 1. CWAB -.08 .27 .09 -.24 -.17 -.23 2. CWAG .OO .32 .17 -.22 -.43 -.32 3. CWAT -.11 .25 .05 .12 .OO .09 4. TWAB .05 .17 .13 .04 .OO .03 5. TWAG -.O6 .17 .05 .05 .21 .12 6. TWAT -.28 .20 -.O9 -.19 .06 -.ll Note. CAV Client AAI avoidance subscale score; CAX = Client AAI anxiety subscale score; CVX = Client AAI composite insecurity score; TAV = Therapist AAI avoidance subscale score; TAX = Therapist AAI anxiety subscale score; TVX = Therapist AAI composite insecurity score; CWAB = Client WAI bond score; CWAG Client WAI goal score; CWAT = Client WAI task score; TWAB Therapist WAI bond score; TWAG = Therapist WAI goal score; TWAT = Therapist WAI task score. * E < .05. ** E < .01. 88 scores received higher bond and task ratings following the lét counseling session. Conversely, therapist anxiety was negatively related to goal scores at Time 3 (g = —.54, p < .01), which indicated that therapists with higher anxiety scores received lower goal scores following the 7U‘therapy session. Next, therapist avoidance was not related to any of the working alliance subscale scores across the three time points. Finally, therapist composite insecurity was positively related to bond (p = .39, p < .05) and goal (E = .38, p < .05) scores at Time 1, which indicated that therapists with higher composite insecurity scores received higher bond and task scores following the lfi'therapy session. Taken together, these findings indicated that, early in therapy, therapist attachment insecurity was associated with higher client-rated working alliance subscale ratings (i.e., bonds and tasks). Conversely, later in therapy, therapist attachment security was associated with higher client-rated goal scores. The Impact of Within Dyad Attachment on the Discrepancy of Therapist and Client Working Alliance Ratings Over Time For this analysis, the combined attachment styles of the 14 dyads that completed the study were obtained. Due to the small Ns and low frequencies of some of the insecure categories, the insecure attachment styles were aggregated into a common insecure group. Using this procedure, there were 7 secure therapist—insecure client dyads and 7 insecure 89 therapist—insecure client dyads. Thus, the following analysis is a comparison of dyads containing secure therapists versus those dyads with insecure therapists over the three time points. A 3 x 2 (Time X Pair) ANOVA with repeated measures was conducted to examine the discrepancy of therapist and client working alliance scores over time. A significant main effect was obtained for Time, F (1,12) = 5.14, p < .05. However, the Time X Pair interaction effect was not significant. These final post—hoc results indicated that dyad attachment security did not appear to moderate the discrepancy of therapist and client working alliance ratings over time. Summary The preliminary analyses indicated that therapists in this study were more likely to self—report a secure adult attachment style than were clients. These analyses also indicated that the average working alliance ratings by therapists and clients increased across the three time points, demonstrating that both viewed the relationship as improving over time. Results of the primary analyses indicated that client and therapist working alliance ratings were differentially related across seven counseling sessions. Contrary to prediction, therapist and client working alliance ratings were significantly related following the lfi'and 4u‘therapy sessions but were virtually uncorrelated following the 7th session. Interestingly, this pattern appeared to be related 90 to therapist and clients sharing common perceptions of the goals and tasks components of the working alliance early, but not later in therapy; bond scores, by contrast, were divergent across each of the three time points. This study also found that therapist adult attachment security was significantly related to working alliance scores over time. Unexpectedly, therapist insecurity (as measured by orientations or styles) was consistently associated with more positive client-rated working alliances early in therapy. However, later in therapy (following the 4U‘and 7u‘sessions, this trend began to reverse itself, as more secure therapists began to receive higher client-rated working alliance scores. Interestingly, the post-hoc examination of working alliance component scores indicated that, following the 1St therapy session, therapist insecurity was positively associated with client-rated bond, goals and task scores; however, following the fourth and seventh therapy sessions, therapist insecurity was negatively associated with client—rated goal scores. Similarly, this study also found that client adult attachment styles were also related to working alliance scores over time. Specifically, clients who self-reported a secure attachment style received higher therapist-rated working alliance scores following the 7u‘therapy session. Likewise, secure clients also reported moderately higher working alliance scores following the 7fl‘therapy session. Another noteworthy finding that emerged from this study was 91 that the development of the working alliance varied over time as a function of dyad adult attachment styles. That is, dyads containing insecure therapists, as compared to dyads with secure therapists, had higher client-rated working alliance scores following the lfl'therapy session. However, across the seven therapy sessions, dyads with insecure therapists demonstrated flat or relatively unchanged client- rated working alliances. Conversely, dyads with secure therapists, although initially receiving lower client—rated working alliance scores following the 1St therapy session, demonstrated a pattern of change from low-to—high across the seven counseling sessions. 92 DISCUSSION This study examined the impact of client and therapist adult attachment on the working alliance across seven counseling sessions. It was predicted that therapist and client working alliance ratings would be differentially related across seven counseling sessions. It was also proposed that therapist and client adult attachment security would be positively related to working alliance ratings across the seven counseling sessions. Furthermore, therapist adult attachment security was expected to moderate the relation of client and therapist working alliance ratings at each time point. Finally, therapist and client adult attachment orientations were anticipated to predict the correspondence of client and therapist working alliance ratings at Time 3. The following section will include a summary of the overall findings of this study. It will also offer possible explanations for some of the current findings. It will then discuss the contributions of this study and implications for counseling psychology. Finally, limitations and directions for future research will be provided. Overview of Findings Although several of the initial hypotheses for this study were not supported, this study offers some unique findings related to the development of therapeutic working alliance. Specifically, this study contributes to our 93 understanding of how client and therapist adult attachment styles/orientations are differentially related to the formation of working alliance over time. This study also sheds light on the level of correspondence of clients’ and therapists’ perceptions of the working alliance ratings over time. A discussion related to these findings is provided below. Development of the Working Alliance: Therapist and Client Perspectives Unexpectedly, results of this study indicated that therapists and clients developed a common perception of the working alliance early in therapy (i.e., following the 1St and 4U‘sessions); however, by the 7u‘session, therapist and client perceptions of the working alliance were virtually unrelated. This finding was inconsistent with previous studies that have reported a low level of correspondence between early (3“‘session) therapist and client working alliance ratings (Horvath & Symonds, 1991; Kivlighan & Shaughnessy, 1995; Mallinckrodt, 1991; Tichenor & Hill, 1989). However, the early-session correspondence between therapist and client working alliance ratings may lend empirical support to several theoretical propositions about the nature of the developing therapeutic alliance between clients and therapists in brief therapy. For example, Mann (1973) theorized that the early stage of brief therapy or Phase I, usually the first four sessions, is marked by positive feelings between the 94 therapist and client. During this “honeymoon” stage, positive transference, “symptom relief” and “magical expectations” about counseling outcome are common. Theoretically speaking, a pattern of early-session convergence may be in part related to clients and therapists having generally positive feelings (i.e., hope, optimism, etc.) about the process of therapy or therapeutic relationship at the outset of therapy. The early convergence of alliance ratings also lends indirect support to Gelso and Carter’s (1985) proposition that in briefer interventions, “it is important that the alliance be established very early, as early as the first session” (p. 165). Although these authors are speaking about the “quality” of the working alliance and not the “correspondence” of perceptions of the alliance per se, it seems reasonable to assert that developing early common perceptions of the alliance, as found in the current study, may also be of critical importance, especially in briefer forms of therapy. Curiously, this study found that therapist and client working alliance ratings were virtually unrelated following the 7U‘therapy session. Although this finding was unexpected, several different theories may lend support to the current finding. For example, the above finding is consistent with Mann’s (1973) suggestion that once brief therapy dyads move to the middle point of therapy (7”‘or 8U‘neeting), negative 95 transference, ambivalence and resistance begin to emerge. Since the clients and therapists in the current study reported divergent views of the working alliance following the 7u‘counseling session, this divergence may indicate that dyads were theoretically moving out of the “honeymoon” phase to a more difficult stage of therapy. Thus, it may be more difficult for dyads to feel a sense of collaboration during this more difficult phase of therapy. Likewise, the above finding of divergent client- therapist perceptions of the working alliance later in therapy, is consistent with one empirical study (Horvath & Marx, 1990). These authors found that counselor-rated working alliance development demonstrated a three-part cycle of development—rupture-repair over the course of brief therapy (10 sessions). In contrast, client-rated working alliance scores increased steadily over the course of therapy. The important implication is that clients and therapists had differential patterns of working alliance ratings over time. Conversely, the lack of correspondence following later therapy sessions contrasts with several empirical studies that have reported that therapist-client alliance ratings converge over time (Kivlighan & Shaughnessy, 1995; Mallinckrodt, 1993). For example, Kivlighan and Shaughnessy (1995) reported that, “Although early-session correlation coefficients were small and nonsignificant, the correlations between client and therapist working alliance ratings were 96 moderate to large during later counseling sessions” (p. 346). These authors reported that over time it appeared that clients and therapists began to develop common perceptions of the therapeutic working alliance. One possible explanation for the later stage discorrespondence among therapist and client alliance ratings found in the current study may be related to the varying importance of the therapeutic working alliance over time. For example, Gelso and Carter (1995) posited that the importance of the working alliance “waxes and wanes” during the different stages of therapy. Moreover, they suggested that during the initial stage of therapy the alliance may be in the “foreground” as therapists and clients attempt to collaborate about the difficult process that lies ahead. However, once this alliance is established, it may then recede until it is needed again. Thus, for therapists and clients in the current study, the pattern of early convergence may have been related to dyads initially focusing on collaborating about the different aspects of therapy. Correspondingly, the later divergence found following the seventh session might have been an indication that dyads had entered a “working” stage, and thus were less concerned about, or less focused on, the working alliance. In sum, the results of the current study suggest that it is during the middle phase of brief therapy that clients and therapists have the most divergent views about the therapeutic working alliance. It is during this stage that 97 it may be particularly difficult for clients and therapists to collaborate or come to a shared perception about the bonds, goals, and tasks of therapy. The results of this investigation’s post—hoc analyses, which examined the relations between therapists’ and clients’ ratings of the three working alliance components over time, shed some further light on this issue. In general, client and therapist ratings of the three working alliance components were differentially related over time. In actuality, the positive relations between early-session therapist and client overall working alliance ratings reported above appear to be largely a function of similar perceptions regarding the goals and tasks of therapy. Conversely, therapists’ and clients’ perceptions of the therapeutic bond were consistently unrelated across the seven counseling sessions. The above finding may lend empirical support to claims by other investigators (Gelso & Carter, 1985; Horvath & Greenberg, 1989) who have theorized that bond aspect of the working alliance develops more slowly than the task and goal dimensions. Likewise, it may also support Horvath and Greenberg’s (1989) proposition that the “strength of the bond component may be more critical in latter phases of treatment when the client is faced with difficult and painful choices” (p. 229). Thus, one possible explanation for the above finding may be related to time. That is, in the current study, data 98 were collected across the first seven counseling sessions. Theoretically speaking, this may not have provided sufficient time for therapists and clients to develop common perceptions of the bond dimension of the working alliance, which may take longer to develop than the other working alliance dimensions. Relations of Therapist and Client Adult Attachment to the Development of the Working Alliance The overall finding that therapists’ adult attachment orientation/styles was related to working alliance scores provided empirical support to theoretical propositions about the relationship of therapist attachment and the development of the working alliance. For example, it was Bordin (1979) who originally suggested that the therapeutic working alliance was influenced by the personal characteristics of therapists. Moreover, others have hypothesized that therapists’ own attachment styles influence the process of therapy (Dozier, Cue & Barnett, 1994; Dozier & Tyrrell, 1998; Dunkle & Friedlander, 1996). Although the general impact of therapists’ adult attachment has received little empirical attention, the above finding is related to one empirical study that examined the relation between therapist (case manager) attachment organization and intervention strategies (Dozier, Cue & Barnett, 1994). The results of the Dozier et al. study found that clinicians’ attachment orientations influenced intervention strategies. Specifically, securely attached 99 therapists (case managers) responded to clients’ underlying needs or in ways that challenged clients’ existing attachment orientations. Conversely, insecure therapists tended to respond to the most obvious needs or in ways that were consistent with or complementary to clients’ models of close relationships and thus failed to challenge clients underlying assumptions about close relationships. “For example, clients who are preoccupied present themselves as needy, expect to be treated as if they were fragile, and find that clinicians who are more insecure intervene in a manner consistent with these expectations” Dozier, Cue & Barnett, 1994, p. 798). Interestingly, insecure case managers intervened more intensely than their dismissive peers did. One disturbing implication of Dozier et al. study was that insecure case managers appeared to be responding to their own internal attachment needs rather than to their clients’ therapeutic needs. In sum, there is some emerging evidence to suggest that therapist attachment security may indeed be related to the development of the therapeutic working alliance. However, beyond simply finding an overall connection between therapist adult attachment and the working alliance, the following section will specifically examine the differential impact of therapist adult attachment orientations on working alliance ratings over time. The current study found that therapists with secure adult attachment orientations/styles received moderately 100 higher client-rated working alliance scores following the seventh counseling session. Similarly, although initially receiving lower client-rated working alliance scores than did their insecure peers, over time dyads led by secure therapists demonstrated steady increases in client working alliance scores. These findings suggest that secure therapists may be more likely to “build” or “create” positive working alliances. One possible explanation for the above finding might be related to clients developing more accurate or more “real” perceptions of their therapists over time. Such a pattern lends support to the proposition by Gelso and Carter (1994) that the “real relationship” or less distorted component of the therapeutic relationship, develops over time and is positively associated with stronger alliances. As noted earlier, Dunkle and Friedlander (1996) found that therapists who reported greater comfort with intimacy received higher client-rated working alliances early in therapy. Although this finding indicated that therapist attachment security had a positive impact on early-session client working alliances (which appears to in contrast with the results of the current study), it is important to note that these investigators assessed the working alliance at one time point - sometime following the third, fourth or fifth therapy sessions. Thus, there is uncertainty as to when therapist security became positively associated with working alliance ratings. Notwithstanding, it is reasonable 101 to conclude that sometime during the early-to-middle phase of therapy (i.e., between the 3“1and 7U‘sessions), both Dunkle and Friedlander’s results as well as those of the current study indicate at least a moderate relation between therapist adult attachment security and working alliance ratings. Unexpectedly, the results of the current study demonstrated that insecure therapists (those reporting insecure adult attachment orientations/styles), as well as insecure therapist-client dyads, received higher working alliances following the first therapy session, than did their secure counterparts. This finding is puzzling and somewhat difficult to explain. However, the finding appeared to indicate that, at least from the clients’ point of View, insecure therapists initially “looked” better than secure therapists did. One possible explanation for this finding is that clients’ early-session ratings of their therapists are distorted. That is, during the early phase of therapy, clients' perceptions of therapists may be largely based on transference reactions. According to Gelso and Cater (1985), the transference relationship is an “unreal” relationship because it involves displacing feelings, thoughts and behaviors from past relationships with significant others onto the therapist. That is, according to these authors, “transference entails a misperception or misinterpretation of the therapist, whether positive or 102 negative” (p. 170). Thus, it seems reasonable to conclude that the current sample of clients — most of whom were insecurely attached -- may have been prone to misperceive their insecure therapists. It seems plausible that insecurely attached therapists —- those whom are likely to be experiencing anxiety or discomfort in close relationships -- may have somehow triggered more positive perceptions from clients and thus appeared to “look” better or more “ideal” to clients during the initial therapy session. It may be that insecure therapists are ill-equipped to manage countertransference pulls from their clients. Therefore, they may have manifested a tendency to act in ways to accommodate client’s maladaptive interpersonal stances. Such behaviors may have created a more collaborative or less challenging interpersonal climate for insecure clients. However, according to Dozier and Tyrrell (1998), therapists need to resist the countertransference pull from their client in order to be therapeutic. These authors also reported that “clinicians relying on autonomous strategies (secure attachment styles) appear best able to provide clients with experiences that challenge working models” (p. 254). Likewise, Gelso and Carter (1994) reported that “a primary task of therapists is to monitor their own countertransference issues so they are not injurious to the alliance” (p. 299). Another possible explanation for the higher early- 103 session alliance ratings among insecure therapist is that clients were idealizing their insecure therapists. In fact, a recent study by Hatcher and Barends (1996) provides support for this alternative explanation. In their factor analytic study of 3 working alliance measures (i.e., Working Alliance Inventory, Psychotherapy Alliance Scales, Penn Helping Alliance Questionnaire), these authors identified a 6 factor model. Of particular note was a factor called “Idealized Relationship,” which was an indication of degree to which a client was able to disagree with their therapist. Results indicated “that patients who are the most reluctant to disagree with their therapist, relative to their overall alliance, actually reported less progress in therapy” (p. 1329). Based on the above study, it appears reasonable to speculate that the clients in the current study (most of whom were insecurely attached themselves) may have been prone to idealize their insecure therapists early in counseling, and that insecure therapists, more so than their secure counterparts, may have been inclined to collude with these projections. The current study’s post-hoc analyses, which examined the relations between therapist and client adult attachment orientations and therapist and client ratings of the three working alliance components over time, attempted to further clarify this issue. These findings corroborated the primary analyses of this study by indicating that, following the 1St therapy session, therapists reporting secure adult 104 attachment orientations received higher working alliance subscale scores. However, later in therapy, this trend began to reverse itself as secure therapists received higher goal scores. Interestingly, therapist adult attachment orientations were not related to bond scores following the 4u‘and 7th therapy session. This finding is unexpected and in contrast with a recent study that reported that secure therapists -- those who reported greater comfort with intimacy -- received significantly higher early-session WAI bond scores (Dunkle & Friedlander, 1996). As expected, there was a significant relationship between client adult attachment orientations and working alliance scores. That is, following the 7U‘therapy session, clients with secure adult attachment orientations received higher working alliance ratings from their therapists. The above finding is consistent with several previously mentioned studies that reported that client characteristics are related to working alliance development (Al—Darmaki & Kivlighan, 1993; Dykeman & LaFleur 1996; Kokotovic & Tracey, 1990). More generally, this finding is also consistent with the empirically established connection between adult attachment and individuals’ experiences in close relationships. As previously reported, secure adults, relative to their insecure peers, experience more positive, satisfying, and well-adjusted relationships. 105 However, it is curious that client attachment orientations were not related to working alliance scores early in therapy. Unlike insecure therapist attachment, that was associated with higher early-session working alliance ratings, insecure client attachment was not associated with working alliance scores following the lfl‘and 4fl‘therapy sessions. Interestingly, post-hoc results that demonstrated client anxiety was negatively related to client-rated task scores following 4th counseling session, indicated that clients who reported less anxious adult attachment orientations expressed more therapy-appropriate task expectations (e.g., self-disclosure, trust, affect regulation, collaborative problem solving, etc.). Contrary to prediction, therapist adult attachment orientations did not moderate the relation of therapist and client working alliance ratings. Given the clearly established link between individuals’ adult attachment styles and their experiences in close relationships, this study had anticipated that secure therapists would be more successful in promoting shared perceptions of the working alliance with their clients. Measurement issues may have been implicated in the failure to observe these relationships. Although the current study relied on paper-and-pencil measures to assess client and therapist adult attachment styles, Dozier and Tyrrell (1998) have recently argued that concepts of adult 106 attachment and internal working models are not interchangeable. More specifically, they described attachment styles as accessible to conscious awareness and self—report. Conversely, internal working models are more unconscious and must be “assessed through discourse analysis of subjects’ discussion of their early attachment relationships” (p. 224). As suggested by Bartholomew and Shaver (1998), “Self- report measures focus on conscious, potentially inaccurate summaries by a person of his or her own experiences and behaviors” (p. 29); whereas measures such as the Adult Attachment Interview (AAI) focus on the way individuals talk about their attachment experiences with primary caregivers attempting to identify communication patterns (e.g., defensiveness, preoccupation, etc.) that are “not necessarily noticed by the people who exhibit them? (p. 29). Similarly, Dozier and Tyrrell (1998) argued that therapeutic relationships have more in common with parent- child relationships then they have with adult romantic relationships. As these authors suggested, “the therapist is a prototypical example of an attachment figure in adulthood. The relationship is caretaking and nonreciprocal, and the client usually perceives the therapist as stronger and wiser” (p.226). Therefore, the implementation of a measure that assessed internal working models via a semi- structured interview (e.g., Adult Attachment Interview), as opposed the self-report paper-and-pencil measures of adult 107 attachment that were employed in the current study, may have yielded the expected relationships between therapist attachment and the development of the working alliance. Contrary to prediction, therapist and client adult attachment orientations were not found to significantly predict correspondence of therapist and client working alliance following the 7th counseling session. This finding was inconsistent with previous studies that have indicated that adult attachment is related to the capacity to describe close relationships as well as the accuracy of individuals’ perceptions of others’ internal working models (Horowitz, Bartholomew, Rosenberg, 1993; Kobak & Hazan, 1991). For example, Horowitz, Rosenberg and Bartholomew (1993) reported that insecure adults, especially those with dismissive attachment styles, had more difficulty providing clear or unambiguous descriptions of significant others. These investigators speculated that, "the interpersonal problems associated with dismissive attachment styles prevent the person from knowing other people well, so the person's internal representation and descriptions of other people are relatively unclear" (p. 558). One possible explanation for the lack of relation between attachment and working alliance correspondence may be related to the low frequencies of some of the insecure attachment styles. For example, in the current sample, the frequencies of dismissive therapists (10%) and clients (18%) were low. 108 As previously mentioned, another possible explanation for the above finding was that the paper-and—pencil measures used to assess attachment were not the best options. Assuming that the therapeutic relationships are similar to parent-child relationships, it may have been more fruitful (although far more labor intensive) to attempt to assess clients’ and therapists’ internal working models. Such models may have served as better predictors of working alliance correspondence. Measurement issues may also explain the failure to find significant relationships between therapist and client adult attachment orientations and self-understanding scores over time. That is, beyond the initial assessment, this measure did not evidence acceptable levels of reliability, which may explain why these theoretically-anticipated relationships were not demonstrated in the current study. Contributions of the Current Study and Implications for Counseling Psychology This study was unique in that it longitudinally examined the concurrent impacts of therapist and client adult attachment styles on the development of the working alliance. Specifically, it is the first study to explore how client and therapist adult attachment orientations/styles respectively and conjointly affected the development of working alliance ratings across several counseling sessions. Past studies have generally relied on single-time-point methodology and/or analogue designs and none have examined 109 the combined effects of client and therapist attachment styles. Given the general movement in counseling research towards the identification of client and therapist factors that may influence the working alliance, the current study provides some interesting information about how important therapist and client factors (i.e., adult attachment styles and orientations) may differentially impact the working alliance at different phases of therapy. For example, this study demonstrated that, early in therapy (following the 1St counseling session), therapists reporting insecure adult attachment styles/orientations received more positive working alliance ratings; conversely, later in therapy (following the 7th counseling session) secure therapists and clients began to receive higher working alliance ratings. The results of this study also found that early-session therapist and client working alliance ratings might be more correspondent than previously reported. If this finding can be replicated, it may support the theory that the working alliance is most prominent (or active) during the early stages of therapy as therapists and clients attempt to collaborate on the difficult processes that are ahead. However, once established, the current study found that therapists’ and clients’ perceptions of the working alliance became divergent, which may be an indication that, once established, the alliance may fade into the background until it is needed (to resolve difficult issues) in the future. 110 Limitations and Directions for Future Research There are several limitations of the current study that should be noted. First, due to correlational nature of the present study, cause—effect linkages are not warranted. This study also relied on survey measures, which introduced a possible mono-method (or self-report) bias. For example, respondents in this study may have responded in socially desirable ways. Another limitation of this study was its relatively small sample size. According to Heppner, Kivlighan and Wampold (1992), low statistical power, which can be in part due to the number of subjects in the study, can contribute to nonsignificant findings. The current study started with a small but reasonable clinical sample (n = 28 dyads). However, as the number of dyads decreased over time, it became increasingly more difficult to detect significant effects. One factor that contributed to this phenomenon was premature or unilateral termination. Of the 28 clients who began the study, 11 (39%) unilaterally terminated before the‘ 'Wm session -- a rate comparable to that reported by Tyron and Kane (1993). Therefore, as with other longitudinal clinical studies, client attrition was a significant problem in this study. Fourth, the generalizablity of the current findings to other populations and settings was somewhat limited. For clients, a vast majority of the current sample was female (71%), white (93%), with a mean age of 29, and over half 111 (54%) had received counseling before. For therapists, although somewhat more diverse with a 20% nonwhite representation, a vast majority were female (75%), with a mean age 31. Therefore, these results may best generalize to clients and therapists with similar demographic characteristics. In addition, some of the recruitment procedures may have also limited the external validity of these findings. For this study, therapists volunteered to participate and then recruited clients to participate. Therapists and clients who agreed to participate in this study may not adequately represent the general population of therapists and clients. Future studies should continue to longitudinally examine the relation between important client-therapist characteristics (i.e., adult attachment styles) and therapeutic processes. In particular, these studies should examine the working alliance beyond seven counseling sessions. By extending data collection, investigators could examine what happens to the correspondence of working alliance ratings during later stages of therapy. Similarly, such studies could explore whether or not clients and therapist eventually develop common perceptions of the therapeutic bond. In addition, by extending data collection, future studied could provide an empirical examination of Mann’s (1973) propositions about the “end phase” of therapy. 112 Specifically, he posited that once therapy begins to move to the later stages of brief therapy feelings of sadness, loss and grief begin to emerge as clients begin to deal with the “termination and separation” from the therapist. It is during this stage that therapists must also face the end of therapy. According to Mann, “it is absolutely incumbent on the therapist to deal directly with the reaction to termination in all of its painful aspects and affects...” (p.36). Precious little research has examined the later stages of therapy. It would be especially interesting to explore how client and therapist adult attachment styles impact their respective abilities to manage difficult termination issues. Given their increased ability to manage negative affect, it would seem that secure clients and therapists would be better able to manage this “intensely affect-laden” phase of therapy. Finally, by extending data collection beyond seven sessions, future studies could also continue to examine the development, as well as the correspondence of therapist and client working alliance ratings. Specifically, if the “rupture” in the correspondence of alliance ratings that occurred in the current study following the seventh session is replicated, it would be possible to examine if the rupture is eventually “repaired.” In fact, Mallinckrodt (1999) has recently argued that the process of building or repairing breaches in the working alliance may help clients develop critical social skills. In line with Mallinckrodt’s 113 suggestions, future studies should continue to explore these alliance ruptures as well as attempt to identify specific therapeutic techniques that can be employed “to repair the inevitable breakdowns” that occur within therapeutic relationships. Future researchers should attempt to gather considerably larger clinical samples. In fact, several of the anticipated relationships in the current study may have reached significance with a larger sample size. Factors such as client drop out and premature termination are common problems with longitudinal clinical research. Finally, as mentioned, future studies should use alternative adult attachment measures. For example, the Adult Attachment Interview (AAI), which is a semi—structured interview, may help to uncover more unconscious aspects therapist and client attachment, especially regarding their internal working models of self, others and important relationships. If therapy is indeed more like a parent- child relationship, the implementation of such a measure may yield more fruitful information regarding the impact of attachment on the process of therapy. Future studies may also consider using the recently developed Client Attachment to Therapist Scale (CATS; Mallinckrodt, Coble, & Gantt, 1995), which asks clients to respond to 36 attachment-related questions about their counselor. This instrument was specifically “designed to measure the psychotherapy relationship from the perspective 114 of attachment theory” (p. 307) and thus may be a more valid indicator of the client—therapist relationship than the working alliance measures that were used in the current study. Similarly, future studies should continue to employ measures that more clearly assess ongoing process-dynamics within therapy sessions. For example, measures that capture sessions-specific changes (e.g., Session Impacts Scale) may be a more valid indicator of ongoing relationships between clients and therapists. However, rather than relying of the self-understanding subscale, which demonstrated poor reliability in the current study, future studies should use the complete Session Impacts Scale. Summary and Conclusions Counseling psychology has long been interested in factors that directly affect human development and change, and especially interested in those that influence counseling process or outcome. Given this general focus, the findings of the current study are particularly noteworthy in that they help to identify how important therapist and client factors (i.e., adult attachment styles/orientations) may impact the develOpment of therapeutic relationships. Likewise, goal of the current study is in line with a recent argument by Lopez (1995), that attachment theory might be a useful “metaperspective” in counseling psychology. As he suggested, such a perspective should \ ‘...deepen our understanding of how client and counselor 115 characteristics may interact to facilitate therapeutic change” (p. 396). As anticipated, attachment theory appears to be a useful lens from which to explore the development of the therapeutic relationship. More specifically, the current study provides information about how therapist and client attachment may be differentially related to the working alliance, which according to Sexton and Whiston (1994) may be the most important process variable in contemporary psychotherapy research. More specifically, this study examined the impact of therapist and client adult attachment on the development of the working alliance across several counseling sessions. Most importantly, this study found that development of the working alliance varies as a function of therapists’, and to a lesser extent, to clients’ adult attachment styles. In conclusion, if the current findings can be replicated, it would suggest that therapist and client attachment characteristics may be critical in guiding the process of therapy, as well as attempting to uncover factors that may help us understand how clinical dyads develop a sense of “collaboration,” which appears to be a necessary component of successful therapy. If adult attachment orientations are indeed established as critical relational factors in counseling, it would have important implications for the training and development of future counseling psychologists. That is, it may become 116 necessary to educate counselor trainees about the importance of monitoring clients’, as well as their own, attachment dynamics. For example, as others have recently suggested, therapists may need to assess clients’ attachment styles and alter their interpersonal stance and interventions to enhance therapeutic working alliances (Dolan, Arnkoff, & Glass, 1993). Finally, continuing to apply attachment theory to the conceptualization of psychotherapy process and outcome appears to be warranted and may eventually help us to better understand how to best provide “secure bases” and “safe havens” to our therapy clients. As Bowlby stated, the initial goal of therapy “is to provide the patient with a secure base from which he can explore the various unhappy and painful aspects of this life, past and present, many of which he finds it difficult or perhaps impossible to think about and reconsider without a trusted companion to provide support, encouragement, sympathy, and, on occasion, guidance” (p. 138). 117 APPENDIX A CLIENT INSTRUMENTS Dear (Client) Participant: Thank you for your expressed interest in our study of the therapeutic process. This research is being conducted by Eric M. Sauer, a counseling psychology doctoral candidate. This project is under the direction of Frederick G. Lopez, a professor in the in the Department of Counseling, Educational Psychology and Special Education at Michigan State University. ' The purpose of this research is to learn more about characteristics that contribute to the development of therapeutic relationships. If you choose to participate, you will be asked to complete three packets containing a few short self-report questionnaires. Some questions will explore your attitudes and feelings about close relationships; others will ask you about your current therapy experiences. We expect that it should take you about 30 - 45 minutes to complete all of the questionnaires in the survey packets, and we do not anticipate that your participation will result in any physical or emotional risk to you. As a benefit for participation, you will receive a free video rental coupon for completing each of the three survey packets. Please answer all of the questions as honestly as possible. Please know that your responses to this survey will be kept completely anonymous and confidential. DO NOT put your names on any of the questionnaires. This way your name cannot be connected to any of your answers and your anonymity and confidentiality can be assured. Please note that neither counselors nor clients will have access to this information. Also, the agency will not see your responses. Your participation in this survey is strictly on a volunteer basis. You are free to withdraw your consent and to stop participation at any time. If you decide to participate, read the brief statement below and PRINT then SIGN your name and enter today's date on the appropriate lines. This form will be kept separate from you survey responses. If you have any questions, please feel free to contact Eric M. Sauer or Dr. Lopez at 355-8502. I agree to participate in the survey described above. I understand the nature of the project, the nature of my participation, that my participation is voluntary, and that I can terminate my participation at any time without penalty. Print your name Sign your name Today's date 118 CPDQ Thank you for participating in this study. The following questions ask you to supply background and current information. Please circle the appropriate number under each of the items listed below or enter the correct information in the appropriate blank spaces that are provided. 1. Your sex: (circle one) (1) Male (2) Female 2. Your current age: 3. Your racial/ethnic background: (1) African-American (2) Asian-American (3) Caucasian/White (4) Hispanic/Latino(a) (5) Native-American (6) Multiracial 4. Your current marital status: (1) Single, never married (2) Married (3) Divorced, not remarried (4) Separated (5) Widowed 5. Highest degree obtained (1) High School Diploma (2) Associate's Degree (3) Bachelor's Degree (4) Master's Degree (5) Doctorate (6) Other (please describe): 6. Have you ever received counseling before this experience? (1) Yes (2) No 119 RQ Directions: Please read each of the descriptive paragraphs below and place a checkmark next to the ONE paragraph that best describes how you feel about close relationships. After this, using the scale provided, choose a number from 1 to 9 to rate how characteristic each paragraph is to your typical relationships. Scale: Not at all Slightly Moderately Very Extremely 1 ----- 2 ----- 3 ----- 4 ----- 5 ----- 6 ----- 7 ----- 8 ----- 9 Best Rating Description (1 to 9) (check one) 1. It is easy for me to become emotionally close to others. I am comfortable depending on others and having others depend on me. I don't worry about being alone or having others not accept me.-- > 1. 2. I am comfortable without close emotional relationships. It is very important to me to feel independent and self—sufficient, and I prefer not to depend on others or . have others depend on me. --------------- > 2. 3. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. ----------------------------- > 3. 4. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others. ------------------------------ > 4. 120 AAI Directions: Using the scale adjacent to each of the item below, indicate (by circling the appropriate number) to what extent the item describes how you have typically felt towards romantic partners in general. Strongly Strongly Disagree Agree 1. I find it relatively easy to get close to others ....... 1 2 3 4 5 6 7 2. I’m not very comfortable others depend on me ....... 1 2 3 4 5 6 7 3. I’m comfortable having others depend on me .............. 1 2 3 4 5 6 7 4. I rarely worry about being abandoned by others ....... 1 2 3 4 5 6 7 5. I don’t like people getting too close to me ........... 1 2 3 4 5 6 7 6. I’m somewhat uncomfortable being too close to others.1 2 3 4 5 6 7 7. I find it difficult to trust others completely ......... 1 2 3 4 5 6 7 8. I’m nervous whenever anyone gets too close to me ...... l 2 3 4 5 6 7 9. Others often want me to be more intimate than I feel comfortable being ......... 1 2 3 4 5 6 7 10. Others are often reluctant to get as close as I’d like..1 2 3 4 5 6 7 11. I often worry that my partner(s) don’t really love me ...... 1 2 3 4 5 6 7 12. I rarely worry about my partner(s) leaving me ..... 1 2 3 4 5 6 7 13. I often want to merge with others, and this desire sometimes scares them away .......... 1 2 3 4 5 6 7 121 SIS Directions: Please consider your experience in you most recently concluded counseling session when responding to the items below. Using the rating scale immediately below each item, circle the number whose corresponding descriptor best fits your experience. As a result of this session 1. I now have new insight about myself or have understood something new about me; I see a new connection or see why I did or felt something (Note: There must be a sense of “newness” as a result of something which happened during the session). Not at all Slightly Somewhat Pretty much Very much 1 2 3 4 5 2. I now have insight about another person or have understood something new about someone else or people in general. Not at all Slightly Somewhat Pretty much Very Much I 2 3 4 5 3. I have been able to get in touch which my feelings, thoughts, memories, or other experiences; I have become more aware of experiences which I have been avoiding; some feelings or experiences of mine which have been unclear have become clearer (Note: Refers to becoming clearer about what you are feeling rather than ypy you are feelings something). Not at all Slightly Somewhat Pretty much Very Much I 2 3 4 5 122 Working Alliance Inventory —- Client Form On the following pages are sentences that describe some of the different ways you might think or feel about your therapist. As you read the sentences, mentally insert the name of your therapist in place of each" " in the text. Below each statement is a seven point scale that looks like the one below: 1 2 3 4 5 6 7 never rarely occasionally sometimes often very often always If the statement describes the way you always feel (or think), circle the number 7; if it never applies to you, circle the number 1. Use the other numbers to describe variations between these extremes. Remember, there are no right or wrong answers. We realize that your thoughts or feelings about your therapist may undergo changes over a period of time, but we would like to know your views or feelings as of right now. Thanks again for your help. 1.I feel uncomfortable with . 1 2 3 4 5 6 7 2. and I agree about the things I will need to do in therapy to help improve my situation. 1 2 3 4 5 6 7 3.I am worried about the outcome of these sessions. 1 2 3 4 5 6 7 4.What I am doing in therapy gives me new ways of looking at my problem. 1 2 3 4 5 6 7 and I understand each other. 1 2 3 4 5 6 7 perceives accurately what my goals are.1 2 3 4 5 6 7 find what I am doing in therapy confusing. 1 2 3 4 5 6 \lmU‘ I l .I believe likes me. 1 2 3 4 5 6 7 .I wish and I could clarify the purpose of our sessions. 1 2 3 4 5 6 7 10.I disagree with about what I ought to get out of therapy. 1 2 3 4 5 6 7 11.I believe the time and I are spending together is not spent efficiently. 1 2 3 4 5 6 7 12. does not understand what I am trying to accomplish in therapy. 1 2 3 4 5 6 7 13.1 am clear on what my responsibilities are in therapy. 1 2 3 4 5 6 7 14.The goals of these sessions are important to me. 1 2 3 4 5 6 7 ——————— 15.I find what and I are doing in therapy are unrelated to my concerns. 1 2 3 4 5 6 7 \DCD 123 16.I feel that the things I do in therapy will help me to accomplish the changes that I want. 1 2 3 4 5 6 7 17.I believe is genuinely concerned for my welfare. 1 2 3 4 5 6 7 18.I am clear as to what wants me to do in these sessions. 1 2 3 4 5 6 7 19. and I respect each other. 1 2 3 4 5 6 7 20.1 feel that is not totally honest about his/her feelings toward me. 1 2 3 4 5 6 7 21.I am confident in 's ability to help me 1 2 3 4 5 6 7 22. and I are working towards mutually agreed upon goals. 1 2 3 4 5 6 7 23.1 feel that appreciates me. 1 2 3 4 5 6 7 24.We agree on what is important for me to work on. 1 2 3 4 5 6 7 25.As a result of these sessions I am clearer as to how I might be able to change. 1 2 3 4 5 6 7 26. and I trust each other. 1 2 3 4 5 6 7 27. and I have different ideas on what my problems are. 1 2 3 4 5 6 7 28.My relationship with is very important to me. 1 2 3 4 5 6 7 29.I have the feeling that if I say or do the wrong things, will stop working with me. 1 2 3 4 5 6 7 30. and I collaborate on setting goals for my therapy. 1 2 3 4 5 6 7 31.I am frustrated by the things I am doing in therapy. 1 2 3 4 5 6 7 32.We have established a good understanding of the kind of changes that would be good for me. 1 2 3 4 5 6 7 33.The things that is asking me to do don't make sense. 1 2 3 4 5 6 7 34.I don't know what to expect as the result of my therapy. 1 2 3 4 5 6 7 35.I believe the way we are working with my problem is correct. 1 2 3 4 5 6 7 36.I feel cares about me even when I do things that he/she does not approve of. 1 2 3 4 5 6 124 APPENDIX B THERAPIST INSTRUMENTS Dear (Therapist) Participant: Thank you for your expressed interest in our study of the therapeutic process. This research is being conducted by Eric M. Sauer, a counseling psychology doctoral candidate. This project is under the direction of Frederick G. Lopez, a professor in the in the Department of Counseling, Educational Psychology and Special Education at Michigan State University. The purpose of this research is to learn more about characteristics that contribute to the development of therapeutic relationships. If you choose to participate, you will be asked to complete three packets containing a few short self-report questionnaires. Some questions will explore your attitudes and feelings about close relationships; others will ask you about your current therapy experiences. We expect that it should take you about 30 - 45 minutes to complete all of the questionnaires in the survey packets, and we do not anticipate that your participation will result in any physical or emotional risk to you. As a benefit for participation, you will receive a free video rental coupon for completing each of the three survey packets. Please answer all of the questions as honestly as possible. This study is most interested in examining how a sample of therapists and clients respond to the survey questionnaires. Be assured that we will not be examining or evaluating your individual survey responses. Please know that your responses to this survey will be kept completely confidential. To ensure privacy, you will be given a research identification number so that your name will not appear on any of the survey instruments. Please not that neither instructors nor clients will have access to this information. Also, the agency will not see your responses. Your participation in this survey is strictly on a volunteer basis. You are free to withdraw your consent and to stop participation at any time. If you decide to participate, read the brief statement below and PRINT then SIGN your name and enter today's date on the appropriate lines. This form will be kept separate from you survey responses. If you have any questions, please feel free to contact Eric M. Sauer or Dr. Lopez at 355-8502. 125 Therapist consent form, continued, pg. 2 I agree to participate in the survey described above. I understand the nature of the project, the nature of my participation, that my participation is voluntary, and that I can terminate my participation at any time without penalty. Print your name Sign your name Today's date 126 TPDQ Thank you for participating in this study. The following questions ask you to supply background and current information. Please circle the appropriate number under each of the items listed below or enter the correct information in the appropriate blank spaces that are provided. 1. Your sex: (circle one) (1) Male (2) Female 2. Your current age: 3. Your racial/ethnic background: (1) African-American (2) Asian-American (3) Caucasian/White (4) Hispanic/Latino(a) (5) Native-American (6) Multiracial 4. Your current marital status: (1) Single, never married (2) Married (3) Divorced, not remarried (4) Separated (5) Widowed 5. Highest degree obtained (1) Bachelor's Degree (2) Master's Degree (3) Doctorate 6. Are you currently enrolled in a graduate program? (1) Yes (2) no if no, skip to question 9) 7. What is the name of you graduate program? (1) MA—Counseling (2) MA-Educational Psychology (3) MSW—Social Work (4) PhD-Counseling Psychology (5) PhD-Clinical Psychology (6) Other (please describe): 8. Specify your year in graduate school ? 127 9. Numbers of years of counseling/therapy experience? (1) 0-1 years (2) 2—4 years (3) 5 or more years 10. What is your primary theoretical orientation? (1) Psychodynamic 2) Eclectic 3) Cognitive-Behavioral 4) Systems 5) Humanistic 6) Other (please describe): 11. Location of current training/employment (1) University counseling center (2) Community counseling agency (3) Other (please describe): 128 RQ Directions: Please read each of the descriptive paragraphs below and place a checkmark next to the ONE paragraph that best describes how you feel about close relationships. After this, using the scale provided, choose a number from 1 to 9 to rate how characteristic each paragraph is to your typical relationships. Scale: Not at all Slightly Moderately Very Extremely 1 ----- 2 ----- 3 ----- 4 ----- 5 ————— 6 ----- 7 ————— 8 ————— 9 Best Rating Description (1 to 9) (check one) 1. It is easy for me to become emotionally close to others. I am comfortable depending on others and having others depend on me. I don't worry about being alone or having others not accept me.-- > 1. 2. I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. --------------- > 2. 3. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them. ----------------------------- > 3. 4. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others. ------------------------------ > 4. 129 AAI Directions: Using the scale adjacent to each of the item below, indicate (by circling the appropriate number) to what extent the item describes how you have typically felt towards romanticgpartners in general. 10. 11. 12. 13. Strongly Digagree I find it relatively easy to get close to others ....... 1 2 I’m not very comfortable others depend on me ....... 1 2 I’m comfortable having others depend on me .............. 1 2 I rarely worry about being abandoned by others ....... 1 2 I don’t like people getting too close to me ........... 1 2 I’m somewhat uncomfortable being too close to others.1 2 I find it difficult to trust others completely ......... 1 2 I’m nervous whenever anyone gets too close to me ...... 1 2 Others often want me to be more intimate than I feel comfortable being ......... 1 2 Others are often reluctant to get as close as I’d like..l 2 I often worry that my partner(s) don’t really love me ...... 1 I rarely worry about my partner(s) leaving me ..... 1 2 I often want to merge with others, and this desire sometimes scares them away .......... 1 2 130 Strongly Agree 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 Working Alliance Inventory -— Counselor Form On the following pages are sentences that describe some of the different ways you might think or feel about your client. As you read the sentences, mentally insert the name of your client in place of each " " in the text. Below each statement is a seven point scale that looks like the one below: 1 2 3 4 5 6 7 never rarely occasionally sometimes often very often always If the statement describes the way you always feel (or think), circle the number 7; if it never applies to you, circle the number 1. Use the other numbers to describe variations between these extremes. Remember, there are no right or wrong answers. We realize that your thoughts or feelings about your client may undergo changes over a period of time, but we would like to know your views or feelings as of right now. Thanks again for your help. 1.I feel uncomfortable with . 1 2 3 4 5 6 7 2. and I agree about the steps to be taken to improve his7her situation. 1 2 3 4 5 6 7 3.1 have some concerns about the outcome of these sessions. 1 2 3 4 5 6 7 4.My client and I both feel confident about the usefulness of our current activity in therapy. 1 2 3 4 5 6 7 5.I feel I really understand . 1 2 3 4 5 6 7 6. and I have a common perception of her/his goals. 1 2 3 4 5 6 7 7. finds what we are doing in therapy confusing. 1 2 3 4 5 6 7 8.1 believe likes me. 1 2 3 4 5 6 7 9.I sense a need to clarify the purpose of our session(s) for . 1 2 3 4 5 6 7 10.I have some disagreements with about the goals of these sessions. 1 2 3 4 5 6 7 11.I believe the time and I are spending together is not spent efficiently. 1 2 3 4 5 6 7 12.I have doubts about what we are trying to accomplish in therapy. 1 2 3 4 5 6 7 13.I am clear and explicit about what '3 responsibilities are in therapy. 1 2 3 4 5 6 7 14.The current goals of these sessions are important for. 1 2 3 4 5 6 7 15.I find what and I are doing in therapy is unrelated to her/his current concerns. 1 2 3 4 5 6 7 16.I feel confident that the things we do in therapy will help to accomplish the changes that he/she desires. 1 2 3 4 5 6 7 17.1 am genuinely concerned for 's welfare. 1 2 3 4 5 6 7 131 18. 19. 20. 21. 22 23. 24 25 26. 27 28. 29. 30. 31. I am clear as to what I expect to do in these sessions. 1 2 3 4 5 6 7 and I respect each other. 1 2 3 4 5 6 7 I feel that I am not totally honest about my feelings toward . 1 2 3 4 5 6 7 I am confident in my ability to help. 1 2 3 4 5 6 7 .We are working towards mutually agreed upon goals. 1 2 3 4 5 6 7 I appreciate as a person. 1 2 3 4 5 6 7 .We agree on what is important for to work on. 1 2 3 4 5 6 7 .As a result of these sessions is clearer as to how she/he might be able to change. 1 2 3 4 5 6 7 and I have built a mutual trust. 1 2 3 4 5 6 7 .and I have different ideas on what his/her real problems are. 1 2 3 4 5 6 7 ’5 relationship with me is important to him/her. l 2 3 4 5 6 7 has some fears that if she/he says or does the wrong things, I will stop working with him/her. 1 2 3 4 5 6 7 and I have collaborated in setting goals for these sessions. 1 2 3 4 5 6 7 is frustrated by what I am asking her/him to do in therapy. 1 2 3 4 5 6 7 32.We have established a good understanding between us of the kind of changes that would be good for . 1 2 3 4 5 6 7 33.The things that we are doing in therapy don't make much sense to . 1 2 3 4 5 6 7 34. doesn't know what to expect as the result of 35. 36. therapy. 1 2 3 4 5 6 7 believes the way we are working with her/his problem is correct. 1 2 3 4 5 6 7 I respect even when he/she does things that I do not approve of. 1 2 3 4 5 6 7 132 REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testinggand interpreting interactions. Newbury Park: Sage Publications. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of Attachment: A Psychological study of the strange situation. Hillsdale, NJ: Erlbaum. Al-Darmaki, F., & Kivlighan, D. M. (1993). Congruence in client-counselor expectations for relationship and the working alliance. Journal of Counseling Psychology, 40, 379-384. Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-244. Bartholomew, K., & Shaver, P. R. (1998). Methods of attachment: Do they converge? In J. A Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships. NY: Guilford Press. Biringen, Z. (1994). Attachment theory and research: Application to clinical practice. American Journal of Orthopsychiatgy, 64, 404-420). Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research and Practice, 16, 252—260. Bowlby, J. (1969/1982). Attachment and loss. Vol I: Attachment. New York: Basic Books. Bowlby, J. (1977). The making and breaking of affectional bonds. British Journal of Psychiatry, 130, 201— 210. Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New York: Basic Books. Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46-76). NY: Guilford Press. Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58, 644-663. 133 Dolan, R. T. (1992). Attachment styles and the quality of the therapeutic alliance in psychotherapy (Doctoral dissertation, The Catholic University of America, 1992). Dissertation Abstracts International, 53, 9217094. Dolan, R. T., Arnkoff, D. B., & Glass, C. G. (1993). Client attachment style and the psychotherapist's interpersonal stance. Psychotherapy, 66, 408—412. Dozier, M., Cue, K., & Barnett, L. (1994). Clinicians as Caregivers: The role of attachment organization in treatment. Journal of Consulting and Clinical Psychology, 62, 793-800. Dozier, M., & Tyrell, C. (1998). The role of attachment in therapeutic relationships. In J. A Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp.221- 248). NY: Guilford Press. 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. Dykeman, C., & LaFleur, N. K. (1996). Counselors’ adjective correlates of working alliance. Psychological Reports, 78, 563-570. Elliot, R., & Wexler, M. M. (1994). Measuring the impact of session in process—experimental therapy of depression: The session impacts scale. Journal of Counseling Psychology, 11, 166-174. Feeney, J. A., & Noller, P. (1990). Attachment style as a predictor of adult romantic relationships. Journal of Personality and Social Psychology, 66, 281-291. Feeney, J. A., & Noller, P, & Callan, V. J. (1994). Attachment style, communication and satisfaction in the early years of marriage. Advances in Personal Relationships, 6, 269-308. Fuendeling, J. M. (1998). Affect regulation as a stylistic process within adult attachment. Journal of Social and Personal Relationships, 16, 291-322. Gelso, C. J., & Carter, J. A. (1985). The relationship in counseling and psychotherapy: Components, consequences, and theoretical antecedents. The Counseling Psychologist, 16, 155-243. 134 Gelso, C. J., & Carter, J. A. (1994). Components of the psychotherapy relationship: Their interaction and unfolding during treatment. Journal of Counseling Psychology, 41, 296- 306. _— Hatcher, R. L., & Barends, A. W. (1996). Patients' View of the alliance in psychotherapy: Exploratory factor analysis of three alliance measures. Journal of Consulting and Clinical Psychologyy 61, 1326-1336. Hazan, C. H., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 66, 511-524. Heppner, P. P., Kivlighan, D. M., & Wampold, B. E. (1992). Research design in counseling psychology. Pacific Grove, CA: Brooks/Cole. Hill, C. E., Diemer, R. A., & Heaton, K. J. (1997). Dream interpretation sessions: Who volunteers, who benefits, and what volunteer clients View as most and least helpful. Journal of Counseling Psychology, 16, 53-62. Holmbeck, G. H. (1997). Toward terminological, conceptual, and statistical clarity in the study of mediators and moderators: Examples from the child-clinical and pediatric psychology literature. Journal of Consulting and Clinical Psychology, 66, 599-610. Horowitz, L. M., Rosenberg, S. E., & Bartholomew, K. (1993). Interpersonal problems, attachment styles, and outcomes in brief dynamic psychotherapy. Journal of Consulting and Clinical Psychology, 61, 549-560. Horvath, A. O. (1994). Research on the alliance. In A. Horvath & L. Greenberg (Eds.). The Working Alliance: Theory, Research and Practice. New York: Wiley and Sons. Horvath, A. O., & Greenberg, L. (1989). Development and validation of the Working Alliance Inventory. Journal of Counseling Psychology, 66, 223—233. Horvath, A. O. & Greenberg, L. (1994). Introduction. In A. Horvath & L. Greenberg (Eds.). The Working Alliance: Theory, Research and Practice. New York: Wiley and Sons. Horvath, A. O., & Marx, R. W. (1990). The development and decay of the working alliance during time-limited counselling. Canadian Journal of Counselling, 61, 240-259. Horvath, A. O., & Luborsky, L. (1993). The role of the therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology, 61, 561-573. 135 Horvath, A. O., & Symonds, B. D. (1991). Relation between the working alliance and outcome in psychotherapy: A meta-analysis. Journal of CounselinggPsychology, 38, 139— 149. — Kirkpatrick, L. A., & Hazan, C. (1994). Attachment styles and close relationships: A four-year prospective study. Personal Relationships, 1, 123-142. Kivlighan, D. M. (1990). Relation between counselors’ use of intentions and clients’ perceptions of working alliance. Journal of Counseling Psychology, 61, 27-32. Kivlighan, D. M., & Schmitz, P. J. (1992). Counselor technical activity in cases with improving working alliances and continuing-poor working alliances. Journal of Counseling Psychology, 66, 32-38. Kivlighan, D. M., & Shaughnessy, P. (1995). Analysis of the development of the working alliance using hierarchical linear modeling. Journal of Counseling Psychology, 66, 338-349. Kobak, R. R., & Hazan, C. (1991). Attachment in Marriage: Effects of security and accuracy of working models. Journal of Personality and Social Psychology, 6, 861-869. Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulations, and representations of self and others. Child Development, 66, 135-146. Kokotovic, A. M., & Tracey, T. J. (1990). Working alliance in the early phase of counseling. Journal of Counseling Psychology, 1, 16-21. Lopez, F. G. (1995). Contemporary attachment theory: An introduction with implications for counseling psychology. The Counseling Psychologist, 66, 395-415. Lopez, F. G., Gover, M. R., Leskela, J., Sauer, E. M., Shirmer, L., & Wyssmann, J. (1997). Attachment styles, guilt, shame, and relationship problem-solving orientations. Personal Relationships, 66, 187—199. Mallinckrodt, B. (1991). Clients' representations of childhood emotional bonds with parents, social support, and formation of the working alliance. Journal of Counseling Psychology, 66, 401-409. 136 Mallinckrodt, B. (1993). Session impact, working alliance, and treatment outcomes in brief counseling. Journal of Counseling Psychology,66, 25-32. Mallinckrodt, B. (1999). Attachment, social competencies, and interpersonal process in psychotherapy. Unpublished manuscript. Mallinckrodt, B., Coble, H. L., & Gantt, D. L. (1995). Working alliance, attachment memories, and social competencies of women in brief therapy. Journal of Counseling Psychology, 66, 79-84. Mallinckrodt, B., & Nelson, M. L. (1991). Counselor training level and the formation of the psychotherapeutic alliance. Journal of Counseling Psychology, 66, 133-138. Mann, J. (1973). Time limited psychotherapy. Cambridge, MA: Harvard University Press Mikulincer, M., & Florian, V. (1995). Appraisal of coping with a real-life stressful situation: the contribution of attachment styles. Personality and Social Psychology Bulletin, 61, 406-414. Mikulincer, M., & Florian, V. (1998). The relationship between adult attachment styles and emotional and cognitive reactions to stressful events. In J. A Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 143-165). NY: Guilford Press. Mikulincer, M., & Nachshon, O. (1991). Attachment styles and patterns of self-disclosure. Journal of Personality and Social Psychology, 61, 321-331. Mikulincer, M., & Orbach, I. (1995). Attachment styles and repressive defensiveness: The accessibility and architecture of affective memories. Journal of Personality and Social Psychology, 66, 917-925. Paterson, R., & Moran, G. (1988). Attachment theory, personality development, and psychotherapy. Clinical Psychology Review, 6, 611-636. Pistole, M. C. (1989). Attachment: Implications for counselors. Journal of Counseling and Development, 66, 190— 193. Pistole, M. C. (1993). Attachment Relationships: Self- Disclosure and Trust. Journal of Mental Health Counseling, 15, 94-106. 137 Pistole, M. C. (1995). College students' ended love relationships: Attachment style and emotion. Journal of College Student Development, 66, 53-60. Rutter, M. (1995). Clinical implications of attachment concepts: Retrospect and prospect. Journal of Child Psyghology and Psychiatry and Allied Disc1plines, 66, 549- 571. Satterfield, W. A., & Lyddon, W. J. (1995). Client attachment and perceptions of the working alliance with counselor trainees. Journal of Counseling Psychology, 66, 187-189. Sexton, T. L., & Whiston, S. C. (1994). The status of the counseling relationship: An empirical review, theoretical implications, and research directions. 166 Counseling Psychologist, 66, 6-78. Scarfe, E., & Bartholomew, K. (1995). Accommodation and attachment representations in young couples. Journal of Social and Personal Relationships, 6, 389-410. Simpson, J. A. (1990). Influence of attachment styles on romantic relationships. Journal of Personality and Social Psychology, 66, 971-980. Simpson, J. A., Rholes, W. R., & Nelligan, J.S. (1992). Support seeking and support giving within couples in an anxiety provoking situation: The role of attachment style. Journal of Personality and Social Psychology, 66, 434-446. Stiles, W. B., Reynolds, S., Hardy, G. E., Rees, A., Barkham, M., & Shapiro, D. A. (1994). Evaluation and description of psychotherapy sessions by clients using the session evaluation questionnaire and the session impacts scale. Journal of Counseling Psychology, 61, 175-185. Tichenor, V. & Hill, C. E. (1989). A comparison of six measures of the working alliance. Psychotherapy, 66, 195-199. Tyron, G. S., & Kane, A. S. (1993). Relationship of working alliance to mutual and unilateral termination. Journal of Counseling Psychology, 66, 33-36. 138