‘ _.,.__, »...r-..-..'-_.-~_ _~..—.. — a.._._.... -.‘-_..w...*;;;,,', SUCCESS 0R FHLHFE TM 'FS'YCHo‘FHEHAFY‘ FFFE ‘ " EFFECTS OF COMPARABLE CLIENT FHFFFFFH AME) 5 V CHEN? 31%erch OTHER INTERACTION FF? 1* 5 V , UPON FHE PROCESS AND OUTCOME FF ; PSYLHOTHEFHFF " Thesis for Th5 W iof PF: 23 MICHIGAN ST ATE UNTVERSTTY THOMAS WAWE SPTERFJNG 1972 ' ...... HalHIIITIHIMHIIQIMQHZIUIlull/31W! l This is to certify that the thesis entitled SUCCESS OR FAILURE IN PSYCHOTHERAPY: THE EEFECTS OF COMPARABLE CLIENT-THERAPIST AND CLIENT-SIGNIFICANT OTHER INTERACTION PATTERNS UPON THE PROCESS AND OUTCOME OF PSYCHOTHERAPY presented by Thomas Wayne Spierling has been accepted towards fulfillment of the requirements for Ph.D. Counseling, Personnel Jegree in Services and Educational Psychology ABSTRACT SUCCESS OR FAILURE IN PSYCHOTHERAPY: THE EFFECTS OF COMPARABLE CLIENT-THERAPIST AND CLIENT-SIGNIFICANT OTHER INTERACTION PATTERNS UPON THE PROCESS AND OUTCOME OF PSYCHOTHERAPY By Thomas Wayne Spierling The purpose of this study was to evaluate whether the degree of comparability in client-therapist interaction patterns as related to client reports of previous interactions with other important persons provides a significant process dimension upon which to differentiate between successful and unsuccessful psychotherapy. Comparability was defined as the degree of difference between client reports of behavior used with and received from others and the actual behavior which clients used with and received from the therapist. Three fundamental questions were posed regarding client-therapist behaviors which were similar to client reports of interaction with others: (1) Would the comparability level between client reported interactions and actual client-therapist interactions for the total range of therapy discriminate between suc- cessful and unsuccessful psychotherapy cases? (2) Would the compara- bility level for the combined groups of clients and therapists vary over three stages of therapy? And, (3) Would successful and unsuc- cessful cases differ in the similarity of their interaction patterns across three stages of psychotherapy? Thomas Wayne Spierling To seek the answers to these questions twenty cases were selected from among thirty-six counseling and psychotherapy cases on file in the tape library at the Michigan State University Counseling Center. The thirty—six cases were selected on the basis of two criteria: (1) a minimum of nine sessions; and (2) the availability of pre- to post therapy MMPI profiles. These cases were divided into successful and unsuccessful groups on the basis of ratings on the MMPI profiles by three judges. Ten successful and ten unsuc- cessful cases (N = 20) were then randomly selected for study. The tapes were analysed in two ways. First, client reports of interactions with important others (others, parents, others plus parents) were analyzed from the early phase of therapy. The second fifteen-minute segment of each of two sessions from the early, middle and late stages of therapy were selected for analysis of client-therapist interaction patterns in the second scoring. The actual client-therapist interactions as well as client reports of interactions with others were rated by use of the Inter- personal Circumplex (Leary, 1957). The judges were two Ph.D. candi- dates. They were trained in the use of the Interpersonal Rating System and demonstrated the ability to use the system reliably. In order to test the three questions under investigation, comparisons of client-therapist and client-other behaviors were made in two different ways. The actual behaviors which the client exhibited with the therapist were contrasted with the client's reports of his Thomas Wayne Spierling behaviors with other important persons. The second contrast involved the similarity of the behaviors which therapists used with clients as compared to the reported reactions of others to clients. The test of the first question involved comparing total client and therapist behaviors in the successful group to client and therapist behaviors in the unsuccessful group. A univariate analysis of variance was used to test for differences in the degree of similar client- therapist interaction patterns vs. reported client-other interaction patterns over the entire range of therapy between outcome groups. The prediction that the behavior patterns of successful, as compared with unsuccessful cases, would be less similar to client reported inter- action with others, was tested and rejected. No significant differences were found between outcome groups in the degree of comparable interaction patterns used over the entire range of psychotherapy. Testing the second question involved combining both outcome groups in order to ascertain whether the degree of comparability between client and therapist behaviors and the reported interactions of clients with others fluctuated across the early, middle and late stage of therapy. Results of a two-way repeated measures analysis of variance which allowed analysis of the main effect for stages of therapy indi- cated that all clients, regardless of outcome did not evidence fluctua- tions in comparability level over the three stages of therapy. However, therapists in both outcome groups did evidence significant differences in comparability level across the early, middle and late stages of therapy. As therapy progressed from the early, through the middle, Thomas Wayne Spierling to the late stage, all therapists increased the frequency of their behaviors which were parallel to the reported behaviors of significant others with the client. The third question dealt with comparisons of client and thera- pist behaviors which paralleled client reported interaction patterns with others across three stages of therapy for each success group. A two-way repeated measures analysis of variance which allowed investi- gation of the interaction of stages with outcome was employed. The prediction that there would be no difference between the parallel behavior patterns of successful and unsuccessful client-therapist pairs at the early stage was accepted. Differences were predicted between successful and unsuccessful cases at the middle stage and it was predicted that successful, as com- pared with unsuccessful cases, would behave in ways which were less com- parable to client reported interactions with others at the late stage of therapy. These predictions were tested and failed to be accepted for client-therapist vs. client-parent and client—other plus parent comparisons. Significant differences were found between outcome groups when client behaviors were examined on the client-to-other (excluding parents) vs. client-to-therapist comparisons. Across the three stages of therapy successful clients behaved with the therapist in ways which were less similar to their reports of behavior with others than did unsuccessful clients. By the late stage of therapy successful clients, as predicted, behaved in ways which less frequently paralleled their reported behavior with others than their unsuccessful counterparts. Thomas Wayne Spierling Conclusions from the results of the research were that the level of comparability between the therapist's behavior toward clients and the reported reaction of others to clients, as operationalized in this study, cannot be regarded as a process variable which effectively differentiates between outcome groups. Likewise, the level of client comparability, as defined along two dimensions (client-to-parent; client-to-other plus parent vs. client-to-therapist) did not discrim- inate between successful and unsuccessful cases. It appears, however, that the degree to which the client's reaction to the therapist paral- lels his reported behavior with others (excluding parents) does provide a process variable which effectively discriminates between successful and unsuccessful psychotherapy cases. These results were discussed in terms of the fact that the population from which the sample was drawn consisted of college stu- dents whose primary concerns probably centered with mastering peer relationships. The possibility that differences in initial client reports of interaction with others may have accounted for the differences or lack of same found between the groups was also discussed. The need for further research encompassing a different method of selecting sessions for analysis was cited. Different selection procedures might allow investigation of whether client reports of behavior used with and received from others change as therapy progresses and, if so, whether changes in client reports illuminate differences in comparable behavior patterns between outcome groups. SUCCESS OR FAILURE IN PSYCHOTHERAPY: THE EFFECTS OF COMPARABLE CLIENT-THERAPIST AND CLIENT-SIGNIFICANT OTHER INTERACTION PATTERNS UPON THE PROCESS AND OUTCOME OF PSYCHOTHERAPY By Thomas Wayne Spierling A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services and Educational Psychology 1972 (.1 2.. FT (V , H. ,1 .‘f \l— I To Elaine and to Jeff ii who gave not only of their time and skill but of their comradery and friendship. To the staff of the Counseling Center who contributed to the establishment of the tape library, to Bill Mueller under whose leadership the library was developed and to the many therapists and clients who over the years have participated in the Center's research project. For their help in the conceptualization of this study as well as for their friendship and contribution to my personal and professional growth I wish to thank: Cecil Williams, supervisor and friend, for his participation in the conceptualization of this study and most importantly for his caring and his deep contribution to my personal and professional development. John Powell, supervisor and colleague, who was there during the struggling and sometimes faltering first steps of concep- tualization and from whom I learned about psychotherapy. My fellow interns--Jim Archer, Sam Dietzel, Tom Feister, Karen Rowe, Bob VanNoord, Kathy Scharf, Diane Borchelt who, each in their own way contributed so much and particularly, Ken Hall for the many hours of brainstorming together. Finally, I wish to thank a most “important other“, Elaine, my wife, for the many late hours and diligent typing and for her caring, support and enthusiasm. iv TABLE OF CONTENTS CHAPTER I. INTRODUCTION . . . . . . . . . . . . . Purpose. . . . . . . . . . . . . . Need. . . . . . . . . . . . . . . General Hypotheses . . . . . . . . . . Theory . . . . . . . . . . . . . . Freudian Theory--Transference and . Countertransference . . . . . . . . Interpersonal Implications of Freudian Theory . . . . . . . . Interpersonal Factors Influencing Personality Development. . . . . . . The Therapeutic Relationship. . . . . . Transference--the Client's Interpersonal EIICItationSo o o o o o o o o o Countertransference--Therapist Responsitiv' to Client Elicitations . . . . . . Therapist Responsitivity and Therapeutic Outcome . . . . . . . . . . . . Overview . . . . . . . . . . . . . ll. RELATED RESEARCH. . . . . . . . . . . . The Interpersonal Circumplex. . . . . . . Rating Parallel Modes of Behavior in Therapeutic Relationships . . . . . . . Parallel Client Behavior Patterns and _ Therapeutic Outcome . . . . . . . . . The Reciprocal Effects of Interpersonal Behavior . . . . . . . . . . . . . Reciprocality and Carson's Concept of Complementarity . . . . . . . . . . Interactional Complementarity During the Therapeutic Process and Psychotherapeutic Outcome A Synthesis. . . . . . . . . . . . . PAGE pug—1...; ll l3 l5 l5 I8 I9 20 21 22 23 CHAPTER Statement of Hypotheses . . . . . . . Comparability of Client-Therapist Interaction Patterns and Therapeutic Outcome. . . Comparability of Client-Therapist Interaction Patterns During Three Stages of Therapy Comparability of Client-Therapist Interaction Patterns During Three Stages of Therapy and Therapeutic Outcome. . . . . . Early Stage . . . . . . . . . Middle Stage. . . . . . . . . Late Stage . . . . . . III. METHOD . . . . . . . . . . . . . Source of Data . . . . . . . . . . Selection of Cases. . . . . . . . Therapeutic Outcome . . . . . . . . Reliability of MMPI Ratings. . . . . Selection of Sessions. . . . . . . . Client-Therapist Interaction. . . . . Client-Other InteraCtion o o o o o 0 Interaction Analysis System. . . . . . Rating Interpersonal Interaction . . . . Reliability Samples . . . . . . . . Experimental Design . . . . . . . . Preparation of Data . . . . . . . . Analysis of Data . . . . . . . . . IV. RESULTS . . . . . . . . . . . . . Operational Definition of Hypotheses, , , Client-Therapist Comparability over the Entire Range of Therapy and Therapeutic Outcome, , , Hypothesis I . . . . . . . . . Hypothesis 1a . . . . . . . . . Hypothesis 1b . . . . . . . . . Client-Therapist Comparability Level Across Stages of Therapy . . . . . . . . Hypothesis 2 . . . . . . Hypothesis 2a . . . . . . . . . Hypothesis 2b . . . . . . vi PAGE 25 25 26 CHAPTER PAGE Client-Therapist Comparability Level Across Early, Middle and Late Stages of Therapy and Therapeutic Outcome. . . . . . . . . . . . . . . . 55 Hypothesis 3 . . . . . . . . . . . . . 57 Hypothesis 3a . . . . . . . . . . . . . 59 Hypothesis 3b . . . . . . . . . . . . . 7] Hypothesis 3.1b . . . . . . . . . . . . 75 Hypothesis 3.2b . . . . . . . . . . . . 75 Hypothesis 3.3b . . . . . . . . . . . . 76 Status of Research Hypotheses . . . . . . . . . 78 V. DISCUSSION AND CONCLUSIONS . . . . . . . . . . . 80 Summary . . . . . . . . . . . . . . . . 80 Client Comparability . . . . . . . . . . . 82 Therapist Comparability . . . . . . . . . . 84 Discussion . . . . . . . . . . . . . . . 85 CIient-to-Other vs. Client-to-Therapist Comparability Level . . . . . . . . . . . 86 Client-to-Therapist Behaviors: An Exploratory Question. . . . . . . . . . . . . . . 87 Therapist-to-Client Behaviors: An Exploratory Question . . . . . . . . . . . . . . 89 Implications . . . . . . . . . . . . . . 91 Implications for Further Research . . . . . . . 94 BIBLIOGRAPHY o o o o o o o o o o o o o o o o o o 98 APPENDICES . . . . . . . . . . . . . . . . . . 104 Appendix A. Study Cases. . . . . . . . . . . . . . . . 105 B. MMPI Ratings . . . . . . . . . . . . . . . 107 C. Scoring Manual for the Interpersonal Behavior Rating System . . . . . . . . . . . . . . ' 109 D. Proportions of Behavior in the Octants of the Circumplex used by Both Outcome Groups in the Three Stages of Therapy Plotted Against the Proportions of Behavior which Clients Reported Using with Others and Receiving from Parents . . . . . . . . . . . . . . 127 vii 10. 11. LIST OF TABLES Sex of clients in the population and sample of two outcome groups . . . . . . . . . . . Client-Therapist characteristics and mean and range of sessions for two outcome groups (N = 20) . . Intra-judge reliability of MMPI ratings (N = 20) . . . Inter-judge reliability of MMPI ratings by three judges using Ebel's intraclass correlation formula (N = 20) . . . . . . . . . . . . . Percentage agreement scores and Dittman's R for client- therapist interactions based upon 40 15-minute tape reliability segments. . . . . . . . . . . . Percentage agreement scores and Dittman's R for client- parent and client-other significant person interaction based upon 20 50-minute tape reliability segments . . Cell mean Q_scores on the total level of comparability for successful and unsuccessful cases (N = 20) . . . Summary of the univariate ANOVAs on the total level of comparability for successful and unsuccessful cases . Cell mean D scores for the level of comparability between client-therapist and client-other interaction patterns Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy-- main effect of stage Hypothesis 2 [(C*O+P) - (C+T)] . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy-- main effect of stage Hypothesis 2 [(O+P+C) - (TFC)] . . . . . viii Page 35 36 37 38 45 46 53 54 57 58 59 12. 13. IA. 15. 16. 17. l8. Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns across three stages of therapy--main effect of stage Hypothesis 2a [(C+P) - (C+T)]. . . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns across three stages of therapy--main effect of stage Hypothesis 2a [(P+C) - (T+C)] . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2b [(C+0) - (C*T)] . . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2b [(0+C) - (T+C)] . . . . . . Cell mean and 2 scores for the level of comparability between client-therapist and client-other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy . . . . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome Hypothesis 3 [(CeO+P) - (CFT)] . . . . . . Repeated measures ANOVA table for the level of comparability between client-therapist and all client—significant other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy-~interaction effect of stage with outcome Hypothesis 3 [(O+PaC) - (T+C)] . . . . . . ix Page 61 62 64 65 67 68 69 I9. 20. 21. 22. Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome Hypothesis 3a [(C+P) - (C+T)] . Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome Hypothesis 3a [(P+C) - (T+C)] . . Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome Hypothesis 3b [(C90) - (CFT)] . . Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns across the early, middle and late stages of therapy-~interaction effect of stage with outcome Hypothesis 3b [(osc) - (Tec)] . . Page 70 71 72 73 LIST OF FIGURES Figure Page I. Pictorial representation of experimental deSign O O O O O O O O C O O O O O O O 47 2. Graph of therapist comparability level over three stages of therapy [(O+P+C) - (T+C)] . . . . 60 3. Graph of therapist comparability level over three stages of therapy [(P9C) - (T+C)] . . . . . 63 A. Graph of therapist comparability level over three stages of therapy [(04C) - (T4C)] . . . . . 55 5. Graph of client comparability level for two outcome groups over three stages of therapy for [(C+O) - (CeT)] interaction patterns . . . . . . . . . . . 74 xi CHAPTER I INTRODUCTION Purpose The main objective of this study was to contribute to the knowledge of whether differences in psychotherapeutic outcome are associated with factors of the therapist-client interaction patterns. One aspect of the client-therapist interaction was investigated in an attempt to determine whether differences could be found between successfully and unsuccessfully treated cases. Client-therapist inter- action patterns which parallel or are comparable to the client's past interaction patterns with significant others were investigated as those patterns occurred during the course of psychotherapy. The purpose of the study was to determine whether such comparable inter- action patterns varied with some predictability over the therapeutic process as well as to determine whether differences in the comparability phenomenon were related to outcome. Need Considerable effort has been expended by researchers in the area of counseling and psychotherapy in an effort to demonstrate the efficacy of psychotherapeutic intervention. Studies comparing treat- ment with no-treatment controls have abounded with varying and often discouraging results. As Keisler (1966) point out, what becomes evident from these studies is that clearly some clients working with some therapists and undergoing some treatments improve while others either show no improvement or deteriorate. Hence research aimed at demonstrating the efficacy of psychotherapy versus no therapy often ends up moot because the treatment of the groups remains undefined. More valuable than comparisons of therapy vs. no-therapy groups then are studies which focus upon considerations of why some thera- peutic experiences appear to have positive impact upon client growth while others do not. Such an approach has led some investigators to isolate certain elements of client or therapist dynamics in an effort to further illuminate differences between successful and unsuccessful treatments. However, this very isolation of client and therapist dynamics, while possibly sound in terms of controlling irrelevant research variables, runs head long into conflict with a significant body of clinical psychotherapeutic theory. According to this theore- tical position, the sources of the individual's maladjustment lie in earlier problematic encounters with family members and with other significant persons. Human neurosis is characterized, not solely as an intrapsychic phenomenon, but as a disturbance primarily fostered and maintained in interpersonal relationships (Horney, I939). If this theoretical position is sound, that is, if both the individual's adaptive and maladaptive behavior is learned from past interpersonal interactions and is maintained and enhanced in present interactions, then the curative power of psychotherapy most probably centers within the nature of the interpersonal interaction between client and therapist. The therapeutic relationship becomes viewed as the basic milieu in which and through which maladaptive interpersonal behavior may be changed. Following from these theoretical considerations, the focus of research regarding therapeutic effectiveness might well shift to inquiries about what happens in the interpersonal interaction which positively or negatively affects the outcome of the psychotherapeutic enterprise. Such questions are reflected in the recent thrust in psychotherapeutic research aimed at elucidating the complex process of psychotherapeutic interaction. The interaction between client and therapist becomes a primary variable of investigation. Empirical studies of the moment to moment behavior of both client and therapist occurring during the therapeutic process may well highlight relevant variables affecting therapeutic outcome. What are needed in psycho- therapeutic research thenarermn studies of therapist or client dynamics as isolated variables but further clarification of the interpersonal interaction of client and therapist during the thera- peutic process as that interaction affects the success or failure of psychotherapy. General Hypotheses 1. During psychotherapy, successful client-therapist pairs will engage in behavior toward each other which is less comparable to the client's previous interaction patterns with significant others than will unsuccessful pairs. 2. During psychotherapy, the level of behavioral comparability between the client-therapist relationship and the client's past relationships with significant others will vary over time. 3. Successful therapy may be distinguished from unsuccessful therapy on the basis of differences in behavioral comparability during different stages of the therapeutic process. Theory Freudian Theory:éTransference and Countertransference Though centering his theoretical emphasis primarily upon an intradynamic, instinctual view of the nature of man, Freud recognized the relevance that past interpersonal encounters held for both the individual's future personality development as well as for his relationship with a therapist. The importance of historical inter- personal antecedents is most clearly evident in Freud's commentaries on the psychoanalytic process. Many of the difficulties confronted by the analyst when attempting to understand, interpret, and recon- struct the patient's ego emanated from the deleterious effects of the patient's transference reactions and, at times, from the analyst's own countertransference reactions. In general Freud defined transference as the patient's emo- tional reactions to the therapist derived from the patient's previous interpersonal experience, often with little reference to the thera- pist's personal reality. Precise definition regarding Freud's notion of transference is difficult, however, since he offered conflicting and sometimes contradictory views of the phenomenon. At varying times, Freud conceptualized transference as an indication of the patient's susceptibility to the therapist's suggestions, e.g. positive rapport; as an interpersonal example of the patient's repetition compulsion; and as a more general phenomenon prevalent in all inter- personal relationships. Transference was viewed as having both positive and negative effects upon the therapeutic endeavor. It was described both as a necessary neurosis or illness which, if handled properly, led to improved functioning as well as an example of unconscious resistance that interfered with patient memories and interrupted the interpretive, insightful work of analysis (ORR, 195h; Kepecs, I966; Crowder, I970). Freud was neither as prolific nor as inconsistent when dealing with the therapist's emotional reactions to the patient. His formu- lations regarding the phenomenon of therapist countertransference reflect what appears as a reluctant recognition that the patient's in-therapy behavior does have an impact upon the therapist. Essen- tially, countertransference is described as an instance of therapist transference of repressed infantile emotions onto the patient. This occurs primarily as a result of the “patient's influence” upon the therapist's unconscious and often narcissistic feelings (Freud, I959a). Since such therapist emotional reactivity to the patient is seen as harmful for the analytic process, the therapist is enjoined against acting upon these unconscious feelings. He is to conduct the work of analysis in a ”state of abstinence“ (Freud, 1959b). Complete and successful analysis for the therapist accompanied by continual self- analysis are offered as preventive measures designed to interfere with countertransference manifestations (Freud, I9S9a). Freud's major means of explaining the origins and intent of countertransference and transference behavior was in terms of libido theory, a theory fostering the view of both therapeuticzparticipants as rather isolated, self-contained, intradynamic entities. While Freud's use of libido theory seems to cloud the interdynamic elements of transference and countertransference, he did recognize the existence of these phenomena as interpersonal, interactional events. His formulations also underlined the significant role played by conflictual antecedent relationships in the development of an individual's future interpersonal behavior. As such, Freud's description of transference and countertransference behavior, his theoretical commentaries on their causality and handling, represents perhaps his major contribu- tion to psychotherapeutic theory. These notions have stimulated further theoretical conceptions regarding both the course and process of personality development as well as the process of therapy designed to correct maladaptive or truncated development. Interpersonal Implications of Freudian Theory. In his later writings, Freud (1948) implied that transference behavior is not isolated to the analytic relationship but is prevalent in all inter- personal relationships. The phenomenon is operative for all individuals whether adjusted or maladjusted. Transference, he states, ”is an unusual phenomenon of the human mind .,. . and in fact dominates the whole of each person's relations to his human environment . . . (Freud, 1948, p. 75).” Subsequent clinical observers expanded and elaborated upon this universal notion of transference and placed it in a central position from which whole theories of neurosis and personality develop- ment have generated. The interpersonal theorists, in particular, assumed a causal view of maladjustment as having at its core an interpersonal purpose stemming from the individual's previous inter- actional experience with significant people. Psychopathology is seen as conceived and perpetuated by and between persons and charac— terized by their interpersonal relationships (Malone, I970). Interpersonal Factors Influencing Personality Development Several theorists, notably Horney, Fromm and Erickson, have highlighted the interdynamic, social aspects of neurosis and person- ality development. The work of Harry Stack Sullivan, however, is perhaps most representative of interpersonalist theory. Sullivan (1953), defining human behavior in an interpersonal perspective, proposed that the basis for neurotic adjustment lies with the integra- tive patterns of behavior derived from the individual's previous interaction with important others during the development of his personality. As he devel0ps, the child strives for interpersonal integration and security. Thus he integrates his behavior in terms of a ”self-in-relation-to-A“ pattern, a ”self-in-relation-to-B” pattern according to”. . . the number of important people to whom he had to adjust in the course of his early development (Rioch, 1943, p. 1A9).” These interactional patterns, once specifically defined, become familiar to the individual and serve as models or “prototypes” for the individual's future interpersonal encounters (Sullivan, 1938). This tendency to reexperience other people using the original reference frame or integrative pattern is particularly prevalent if the nature of the child's earlier experience has been problematic or traumatic. In such instances spontaneity is stifled, further emotional development is truncated and the child's original integrative stance will likely persist (Rioch, p. 149). These persistent and often unrealistic integrative patterns form the basis for what Sullivan labels as ”parataxic distortions”--or transference reactions-~which characterize many interpersonal relationships. More recently Carson (1969) and earlier Leary (1957) have been particularly precise in describing interpersonal behaviors as purposeful attempts at security-maintenance. All interpersonal behaviors form elements of the individual's primary security-maintenance system. The intent of each individual engaging in an interpersonal interaction is to consolidate familiar interactional patterns in order to reduce interpersonal anxiety and achieve momentary security. The child then, during the course of his development, is forced to adjust or integrate his interpersonal behavior according to the interactional expectancies of the more powerful and important pe0ple in his environment. Through his interactions with these important others the child learns which interpersonal patterns are reinforced, decreasing his anxiety and safeguarding his security; which are punished, increasing his anxiety and lessening his security. He learns certain ways of behaving, certain modes of interpersonal action and ways of relating accompanied by certain expectancies regarding reactions from others upon which he bases, in part at least, all further interpersonal interaction. The Therapeutic Relationship Following from this interpersonally oriented concept of personality development are some significant propositions regarding the nature of the therapeutic process designed to correct neurotic interpersonal adjustment. The first is the prOposition that both client and therapist behavior have an effect upon each other. Wolstein (I959), Brody (I955), Macalpine (1950), and others indicate that the nature and quality of client transference behavior never occurs in a vacuum but is influenced by the nature of the therapist's personality and behavior within the therapeutic relationship. Neither does therapist behavior occur in isolation but it too takes place within a relationship and, as such, is influenced by client behavior. The therapeutic process thus becomes a special human relationship or special instance of interpersonal interaction designed to improve the interpersonal functioning of one of its members. Evolving from this interactional assumption regarding the therapeutic relationship are two propositions which focus upon the development and manipulation of transference and countertransference manifestations during the course of psychotherapy. Transference-~the Client's Interpersonal Elicitations. As the client enters the therapeutic relationship with some expectancies for help, he is likely to develop a relationship with the therapist which is significant to him. The client will also carry into the therapeutic relationship his stagnated interpersonal patterns and is apt to attri- bute emotions and motives to the therapist which are clouded by his 10 previous traumatic interpersonal experiences. Such perceptual distor- tions will probably lead to behavioral transference. As therapy progresses and as the client's anxiety increases he is likely to behave toward the therapist in terms of his past security-maintaining behaviors by attempting to elicit familiar interpersonal responses (Leary, I957; Rioch, I9H3; Kell and Mueller, 1966; Carson, 1969). This eliciting behavior represents a commonly observed phenomenon and pro- vides the therapist the opportunity within the therapeutic relationship to observe and experience the client's behavior as a living function of his maladjustment, an accurate reflection of his problem. . ... the tendency of the patient to reestablish the original reference frame is precisely because he is afraid to experience the other person in a direct and unreserved way. He has organized his whole system of getting along in the world, bad as that system might be, on the basis of the original distor- tion of his personality and its subsequent vicissitudes (Rioch, 1943, p. 152). Countertransference - Therapist Responsitivity to Client Elici- tations. In addition to reflecting the interpersonal variants of his maladjustment, the client's elicitations also affect the nature of the therapeutic relationship and may have a powerful impact upon the therapist's behavior. As the therapeutic relationship progresses and deepens, therapists often countertransfer emotions and behavior to their clients. Most theorists explain these countertransference reactions as emanating from the therapist's own conflictual experience (e.g. therapist transference) and/or as the result of the therapist's responsivity to the client's ”expertise” in eliciting reactions which are familiar because of their similarity to the behavior of significant others in the client's past conflicted relationships. 11 While accepting the proposition that the therapist's own personal problems may confound the therapeutic interaction, most interpersonal theorists view the latter eXplanation as probably the most frequent source of therapist countertransference reactions (Wolstein, I959; Heiman, 1950; Leary, 1957; Kell and Mueller, 1966; and Carson, 1969). Carson (1969) is perhaps most explicit in high- lighting the very real power resulting from the desperate quality of these client elicitations. This disordered person driven by powerful forces is likely to have acquired a very high degree of expertise in moving others into the position he needs them to be in, and he is often quite prepared, if necessary to go to very extreme lengths in the exercise of power to achieve his goals (p. 281). Clients then in order to reduce their anxiety and fortify their neuroses often succeed in eliciting familiar responses from the therapist. Therapist Responsitivity and Therapeutic Outcome The occurrence of client transference distortions and eliciting behavior also leads to the possibility of correcting the client's previous emotional learning through what may be experienced by the client as a unique relationship. An essential element of the therapeutic process then centers with the therapist's responses to the client's security-maintaining elicitations. Thus as therapy progresses, the therapist, armed with the knowledge and understanding of the client's past interactional patterns, is in a unique position to make use of his countertransference emotions and interfere with the client's neurosis-maintaining interpersonal behavior. 12 Many theorists suggestthattfluatherapist'snajor task_revolves around the development and manipulation of the client's transference elicitations. Alexander and French (1946), Weiss (19h6), and Wolstein (I959) offer the termination of transference elicitations and the learning of new interpersonal behavior as a major criterion for successful therapeutic outcome. As a consequence, Alexander (1965) instructs the therapist to assume a role opposite to that which the client attempts to evoke in order to interfere with the client's neurosis-maintaining patterns and provide a corrective emotional experience. Fenichel (I939) warns against therapist responsivity to client elicitations--“not joining in the game”--and Halpern (I965) identifies the essential ingredient in the therapeutic process as therapist avoidance of ”. . . becoming ensnared in the disturbance perpetuating maneuvers of his patient (p. 175).“ Carson (1969) shares similar assumptions regarding the deveIOpment and perpetuation of neurotic adjustment. He describes the behavioral necessities for the therapist in successful therapeutic relationships: . . . the therapist must be the one person in the client's life . . . who does not yield to the client's pressure to supply confirmatory information to the latter's crippled self. (The therapist must) . . . avoid the adopting an interpersonal position complementary to and confirmatory of the critical self-protective position to which the client will almost invar- iably attempt to move in the course of the therapeutic inter- action (p. 280). Thus, if the therapist is aware of both the client's trans- ference elicitations as well as his own countertransference responses he will probably resist entrapment or move toward resolution when old interaction patterns occur. In such therapeutic relationships 13 the client is likely to achieve new and more profitable self-enhancing, interpersonal behaviors. Successful therapeutic relationships, then, are characterized by client-therapist interaction patterns which are less frequently comparable to the client's earlier learned interactional; patterns. When the therapist consistently engages in countertransference behavior by responding to the client's interpersonal elicitations in ways comparable to the behavior of important persons in the client's past, the therapeutic process is likely to encounter difficulty. If the client is successful, for example,in maneuvering the therapist into providing responses similar to those sent to the client by his parents, the therapeutic relationship may become entrapped in a seem- ingly inevitable replay of the client's previous conflict-maintaining interactions. Such a therapeutic interaction, when continued, may be experienced as secure and comfortable by the client but unless the pattern is broken no new learning and only entrenchment is likely to result. Unsuccessful psychotherapeutic relationships, then, are likely to be characterized by client-therapist interaction patterns which are highly comparable to the client's past interactional experience with important others. Overview In Chapter II a review of the relevant research will be pre- sented. Research dealing with the empirical study of past and present interaction patterns, and the relationship of certain interaction 14 sequences to therapeutic outcome is included as well as a review of the use and appropriateness of the Interpersonal Circumplex (Leary, 1957) for research about human interaction. Also included in Chapter II is a statement of the study's research hypotheses accom- panied by their theoretical backdrop. The basic methodology is presented in Chapter III with a description of the population and sample, the reliability of the raters, the outcome criterion, presen- tation of the experimental design and method of analysis, and a detailed explanation of the behavior analysis system. In Chapter IV statistical analysis of the data is presented as well as the results for each research hypothesis. A discussion of the study results, summary and implications are contained in Chapter V. CHAPTER II RELATED RESEARCH Three bodies of empirical research converge upon the previously discussed theories regarding both the power and predictability of interpersonal elicitations as well as the similarity of therapeutic interactions to the client's past behavior with family members and other significant persons. The first centers in the usefulness of the Interpersonal Circumplex (Leary, 1957) as a method for studying interpersonal interactions. The second deals with investigations of the psychotherapeutic process directed at demonstrating the presence of parallel or comparable modes of behavior (past interactions with family members and significant others versus present therapy interactions) within different counseling or therapeutic relationships. Lastly, research is reported which focuses on the reciprocal effects of interpersonal behaviors in dyadic relationships. The Interpersonal Circumplex--A Method of Interaction Analysis in Psychotherapy Research The method of interaction analysis used in the present study involves the interpersonal diagnosis system of behavioral analysis developed by the Kaiser Research Foundation (Freedman, Leary, Ossorio and Coffey, 1951) and most clearly delineated by Leary (1957). Using this system, each communication unit (uninterrupted speech) of both client and therapist is scored and defined by one or more of 16 15 16 interpersonal reflexes arranged around a circumplex. Each reflex may be collapsed with three others and defined in terms of two major axes: a dominant-submissive and an affiliative-disaffiliative axis. The 16 reflexes are illustrated by the following verbs: boast, reject, punish, hate (disaffiliative-dominant); complain, distrust, condemn self, submit (disaffiliative-submissive); admire, trust, cooperate, love (affiliative-submissive); and support, give, teach, dominate (affiliative-dominant). The Interpersonal Circumplex has been successfully used by several researchers to investigate varying aspects of psychotherapeutic interaction. Mueller (I969a) used this method to map the psychothera- peutic process and to study transference and countertransference behavior (Mueller and Dilling, 1969; Mueller, 1969b). The interper- sonal behaviors of clients and therapists were rated on the circumplex by Crowder (1970) in an effort to study transference and identification hypotheses. Swenson (1967) and Deitzel (1971) used the circumplex to study the interpersonal stances of clients and therapists as their behavior affected therapeutic outcome. Of these studies, those authored by Mueller and Dilling, 1969; Mueller, I969b; Swenson, I967; Crowder, 1970; and Deitzel, 1971 are reported in greater detail below. This behavioral analysis system has also been used in other research settings having implications for psychotherapy. The inter- personal behavior of six “hyperaggressive” boys in a residential treat- ment program was studied by Rausch, Dittman and Taylor (1959). The behavior of each boy with adults and peers was observed and rated on 17 the circumplex during the early and later stages of treatment. By the later stages of treatment the interpersonal behaviors of the boys changed positively toward the expected direction. Using the same sample of hyperaggressive boys, Rausch, Farbman and Lewellyn (I960) employed the Interpersonal Circumplex to compare the interpersonal behavior of normal and hyperaggressive boys. Normal children were found to change their behavior according to variances in the stimulation of the social setting more frequently than disturbed boys. Disturbed boys behaved more like normals, particularly in their relationships with adults, as they reached the later stages of treatment. Heller, Myers and Kline (1963) used the circumplex to demon- strate the reciprocal impact of certain interpersonal stances. They trained four client-actors to assume the behavioral roles associated with the four major quadrants of the circumplex. The behavior of 3A interviewers with these four actors was observed and rated on the circumplex. These authors found that dominant actor behavior evoked dependent interviewer behavior; dependent behavior evoked dominance; aggressive behavior evoked aggression; and affiliative behavior evoked affiliation. In a study by MacKenzie (1968), the interaction differences between members of normal and clinic families were rated on the circumplex. The normal families were found to express more affilia- tive behavior than did clinic families. Clinic mothers were more dominant and more hostile than normal mothers and clinic sons more passive-aggressive than normal sons. In addition, clinic father-son 18 relationships did not evidence the extent of behavioral reciprocality as might be predicted from the Heller, Myers and Kline (1963) study. Rating Parallel Modes of Behavior in Therapeutic Relationships The therapeutic process has also been studied in an effort to provide evidence that client and therapist behavior often parallel the client's interactional experiences in previous important relationships as therapy progresses. The work of Mueller and Dilling (1969) and Mueller (1969b) provided a viable methodology and demonstrated that parallel interpersonal behavior patterns (transference and counter- transference)occurringduring the therapeutic process may be empirically_ studied. These investigators used the Interpersonal Circumplex to rate the client-therapist interaction as well as the client's in-therapy reports of his interactions with significant others. Mueller (I969b) defined transference as high similarity between client elicitations toward the therapist and client recalled client elicitations sent to significant others, particularly parents. Therapist countertransfer- ence behavior was defined as high similarity between therapist behavior toward the client and the behavior of significant others toward the client. He found that as therapy progressed clients sent behaviors to the therapist which were increasingly similar to the client's recalled past behavior sent to parents and significant others. Thera- pist's behaviors sent to the client in later interviews also became increasingly similar to client recalled behavior sent to the client by parents and other significant persons. 19 Mueller selected interviews in which to rate client-therapist interactions on the basis of high client anxiety level, identified by use of a semantic differential technique, as well as on the basis of a perceptual change occurring in the client's relationship to parents. Interviews rated were selected by these criteria on the assumption that transference and countertransference reactions would be more likely to occur when the client's anxiety was high and when he experienced perceptual changes regarding significant relationships. Consequently, Mueller's ratings did not sample the entire range or even similar intercase time sampling of the therapeutic process. Mueller did demonstrate that both the client's and therapist's behavior may be reliably rated from audio recordings of the therapeutic process and that transference and countertransference phenomena occurred. He did not, however, deal with any questions regarding the causality of these transference and countertransference phenomena or with the effects of these reactions and their relationship to outcome criteria. Parallel Client Behavior Patterns and Therapeutic Outcome Crowder (1970), employing comparable methodology, defined transference in the same way as Mueller and investigated the rela- tionship during certain stages of the therapeutic process of client transference behaviors and transference dissipation to outcome. Using the initial interview as a base-rate from which to measure increases in client parallel modes of behavior (transference) during the middle stages of therapy, Crowder failed to demonstrate empirically the occurrence of transference reactions. He did, however, uncover 20 some interesting and differing trends between successful and unsuc- cessful dyads. He found that, in the middle stage of the psychothera- peutic process, unsuccessful clients tended to evidence higher proportions of behavior which were similar to their reported behavior with both parents and significant others than did successful clients. In addition, by the later stage, Crowder's unsuccessful clients decreased the proportion of their behavior with the therapist which was similar to their behavior with non-parent, significant others. Successful clients during this later stage decreased more of their behavior which was similar to their reported behavior with parents. Crowder also studied the reciprocal nature of certain client and therapist reflexes rated on the circumplex as they related to therapeutic outcome. He did not, however, investigate the relation- ship of therapist parallel modes of behavior (countertransference) to therapeutic success or failure. The Reciprocal Effects of Interpersonal Behavior The third line of inquiry converging upon questions regarding the effect of parallel therapist-client interactions and its relation- ship to therapeutic outcome deals vfith the reciprocal effects which interpersonal behaviors have upon members of an interaction dyad. As early as 1928 Schilder suggested that an important psychological rule may govern all human relationships. He proposed that certain patient feelings will naturally elicit complementary feelings from the therapist. The work of Freedman, Leary, Ossorio and Coffey (1951), Leary (I957) and others has focused both theoretically and empirically upon the 21 notion that certain general classes of interpersonal behavior do elicit lawful and predictable responses from members of an interaction dyad. The latter work of Rausch, Dittman and Taylor (1959), Rausch, Farbman and Llewellyn (1960), Heller, Myers and Kline (1963) and MacKenzie (1968) has generally supported these propositions and demonstrated that the reciprocal effects of interpersonal behavior may be reliably observed, noted and classified on the Interpersonal Circumplex developed by Leary (1957). Most of this research supports the proposition that oppositional interpersonal behaviors are compatible or reinforcing on the dominant-dependent axis of the Circumplex whereas oppositional behaviors on the affiliative-disaffil- iative axis are incompatible and punishing. Reciprocality and Carson's Concept of Complementarity While failing to present any empirical evidence supporting his hypotheses, Carson (1969) presents a theoretical synthesis of the interpersonal concepts proposed by Sullivan (1953) and Leary (1957). Carson provides an excellent review of Sullivan's work and suggests what seems to be a general "rule'I regarding the origins of inter— personal maladjustment. Carson proposes that the individual's unique (learned) ways of behaving with significant others when transferred outside the immediate and original situation may be maladaptive causing increased stress to which the individual responds by restricted interpersonal elicitations and ”rule-breaking“ (Carson, p. 281). Noting Leary's contributions and development of the Interper- sonal Circumplex and borrowing some of Haley's (I963) concepts 22 regarding power strategies, Carson also suggests that there may be a central tendency for certain interpersonal stances to be reinforcing and to elicit certain predictable behavioral counter stances. He codified this central tendency under the concept of “complementarity”. Complementarity is defined in terms of the two major axes of the Interpersonal Circumplex and occurs in interpersonal interaction when behaviors are reciprocal on the dominance-submissive axis (dominance evoking dependence, and visa versa) and when they corres- pond on the disaffiliative-affiliative axis (affiliation evoking affiliation; disaffiliation evoking disaffiliation). Complementary interactions are rewarding and increase the individual's moment-to- moment security. Anticomplementary interactions are experienced as threatening and diminish security (Carson, I969, p. 14A). Interactional Complementaripijuring the Therapeutic Process and Psychotherapeutic Outcome Carson further suggests that the therapist, aware of the complementarity dimensions, is in a unique position to either reinforce maladaptive client behavior by offering complementary responses or to interfere with these patterns by responding in a non-complementary manner to client elicitations. Both Swenson (1967) and Dietzel (1971) have utilized the concept of behavioral complementarity as defined by the axes of the Circumplex to study the interpersonal stances of clients and therapists during the therapeutic process. Swenson (1967) proposed that successful therapeutic dyads would be characterized by high levels of interpersonal complementarity. Though Swenson finds support for this hypothesis, his methodology is questionable 23 (MMPI ratings taken prior to therapy were used to categorize client and therapist circumplex stances) and his outcome criterion (supervisor ratings) probably is invalid (Metzoff and Kornreith, I970). Dietzel (1971), following Carson's predictions, used an improved methodology and hypothesized that successful therapeutic dyads would be characterized by less complementarity than would unsuccessful dyads. Further, he proposed that the level of interactional complementarity would fluctuate during different stages of the therapeutic process and that successful and unsuccessful dyads would differ in comple- mentarity during different stages. Dietzel found support for the proposition that the complementarity level will fluctuate and that successful dyads will evidence less complementarity during the middle (working) stage of therapy. Though he indicates that successful dyads tended to interact at a lower complementarity level, he found no significant differences between unsuccessful and successful dyads on the complementarity dimension over the entire range of therapy. A Synthesip The proposition that past interpersonal interaction may be both anxiety producing and may modify later interpersonal behavior has, as Mueller indicates, ”. . . been repeatedly advanced and confirmed in clinical settings by practicing therapists of a variety of orientations” and is reflected in ”. . . most theories of person- ality development derived from clinical practice . . . (Mueller, l969b, p. 2).” In addition many theorists, as discussed previously, recognize the similarity between many of the client's in-therapy elicitations 24 to his past anxiety-reducing behaviors and caution against therapist responses which reward these repetitious behaviors. Perhaps, then, the general notion of behavioral complementarity as defined by Carson (1969) and studied by Dietzel (1971) may be refined to take into account the individual's unique past interpersonal experience. If the individual's chdice of interpersonal elicitations is the result of his past interaction with significant others, then a therapist response classified as complementary by Carson's system may, when compared with the client's past interaction patterns, be experienced by the client as only semi-complementary or non- complementary. MacKenzie's (I968) work on the interactional patterns of clinic and normal families suggests this may be accurate at least for clinic father-son relationships. The reward value (complementarity) of certain therapist responses to specific client elicitations, then, may be more precisely defined in terms of client expectancies derived from past experience than in terms of a more general notion regarding the reinforcement valence based cwnthe two axes of the Interpersonal Circumplex. Hence, the concept of ”comparability“--e.g. the comparability of client-therapist interaction patterns to past client-parent or client- significant other patterns--may provide a powerful relationship dimension upon which to base investigation of the process and outcome of therapy. The methodology developed by Mueller and Dilling (1968b), and used by Mueller (1969b) to study the process dimensions of trans- ference and countertransference and by Crowder (1970) to relate the transference dimension to therapeutic outcome would seem to provide 25 an effective tool with which to identify highly comparable client- therapist/client-parent-significant other interaction patterns. The same basic methodology was used in the present study. The intent was to assess whether behaviors sent back and forth between the client and therapist during the therapeutic process which are similar to the client's recalled past interactional experience with parents and significant others provide a significant process dimension which affects the eventual outcome of psychotherapy. Statement of Hypotheses Comparabiligy of Client-Therapist Interaction Patterns and Therepeutic Outcome It will be recalled that Mueller's (1969) research suggests that clients do behave with their therapists in ways which are similar to client recalled behaviors with significant others and that therapists may at times respond to client behaviors similarly to the way in which the client recalls his parents and other signif- icant persons responding. Though Mueller's study was not designed to relate interaction patterns to outcome criteria, Crowder (1970) did study the effect of client transference patterns on successful and unsuccessful dyads but failed to find significant differences between the outcome groups. Hence, the existence of client elicitations which parallel past elicitations may not in itself provide a process variable powerful enough to discriminate between outcome groups. Theoretically, therapist responses to these highly comparable client elicitations may prove deleterious or therapeutic. The 26 frequency of client-therapist interactions which prove highly compar- able to client recalled client-parent/significant other interactions over the entire range of therapy may discriminate on outcome criteria. A relationship characterized by high incidents of such interaction patterns may simply replay the client's earlier relationships with little new learning resulting. Such clients are likely to evidence little positive change or even negative change on outcome measures. Hypothesis 1: The level of comparability between client-therapist and all client-significant other interaction patterns will be lower in the successful, as opposed to the unsuccessful, therapy dyads. la: The level of comparability between client-therapist and client-parent interaction patterns will be lower in the successful, as opposed to the unsuccessful, therapy dyads. lb: The level of comparability between client-therapist and client-other significant person interaction patterns will be lower in the successful, as opposed to the unsuccessful, therapy dyads. Comperability of Client-Therapist Interaction Patterns During_Three Stages of Therapy Much of the psychotherapeutic literature regarding the ''trans- ference” phenomena suggests that comparability of client-therapist interaction patterns with the client's past interactional experiences may vary within certain phases of successful relationships (Alexander and French, 1946). Kell and Mueller (1966) indicate that the thera- pist's responses, by their similarity to reSponses of earlier signifi- cant persons, may encourage and induce client recollection of signifi- cant past interactions and may stimulate the reenactment of the generic conflict within the therapeutic relationship. They propose that such reenactment of the client's conflictual experience may in some 27 instances be a necessary pre-condition for conflict resolution (p. 138). Thus, ”Successful“ and “unsuccessful“ therapeutic relationships may both exhibit high and low levels of interactional comparability as therapy progresses. Hypothesis 2:. There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy. 2a: There will be differences in the level of comparability between client-therapist and client-parent interaction patterns across three stages of therapy. 2b: There will be differences in the level of comparability between client-therapist and client-other significant person interaction patterns across three stages of therapy. Comparability of Client-Therapist Interaction Patterns During Three Stages of Therepy and Therapeutic Outcome There seems to be substantial consensus for separating the process of successful therapy into three primary stages: 1) the early stage characterized by relationship-building behaviors; 2) the middle stage during which the client's transference increases and when the ”work” of therapy is done; and 3) the later stage characterized by integration, increased client adjustment and more reality-oriented relating (Alexander and French, 1946; Crowder, I970; Dietzel, 1971). In addition, the work done by Dietzel (I97I) suggests that ”successful“ therapy may be distinguished from ”unsuccessful'I therapy on the basis of differing interaction patterns during different stages of the process. It is expected then that the level of comparability will differ between successful and unsuccessful client-therapist pairs in accordance with the therapeutic task in the stage of therapy sampled. 28 Hypothesis 3: There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns for successful and unsuccess- ful client-therapist dyads across the early, middle and late stages of therapy. 3a: There will be differences in the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful client-therapist dyads across the early, middle and late stages of therapy. 3b: There will be differences in the level of comparability between client-therapist and client-other significant person interaction patterns for successful and unsuc- cessful client-therapist dyads across the early, middle and late stages of therapy. Early Stage It is expected that all therapists during the early stage of therapy will endeavor to establish a viable working relationship with their clients. Any sustained interference with the client's security- operations can be expected to increase his anxiety, causing early terminations (Carson, 1969). It is likely, then, that all therapists will operate so as to maintain client anxiety at moderate, relationship- maintaining levels. In addition, establishing and entering a new relationship, particularly one couched with change-inducing significance, will in itself be anxiety evoking for clients. Clients may, as a result, use at least a moderate level of their past anxiety-reducing eliciting behaviors with the therapist during the early stage. Hence, in early sessions, it is likely that both successful and unsuccessful relationships will be characterized by similar and moderate levels of interactional comparability. Hypothesis 3.1: There will be no difference, during the early stage of therapy, in the level of comparability between 29 client-therapist and all client-significant other interaction patterns for successful and unsuccessful therapy dyads. 3.1a: There will be no difference, during the earlx stage of therapy, in the level of comparability between client-therapist and client-parent interaction pat- terns for successful and unsuccessful therapy dyads. 3.1b: There will be no difference, during the earlx stage of therapy, in the level of comparabity between client-therapist and client-other significant person interaction patterns for successful and unsuccessful therapy dyads. Mipdle Stage It is during the middle or ”work” stage of therapy that client anxiety and, hence, the comparability of client elicitations in the present relationship with his elicitations in past relationships, can be expected to be at theirpeak. Since these elicitations may become more repetitious, more desperate, and more powerful, it is likely that therapist responses will also converge more frequently on responses which the client recalls receiving from significant others. Thus the client-therapist relationship can be expected to be most comparable to the client's previous interactions with significant others as he reenacts his generic conflict with the therapist. Kell and Mueller (1966) caution that it is not the counselor's entrapment in the client's conflicted experience but his continued entrapment and reinforcement of the conflict which leads to therapeutic failure. Since reenactment and continued reinforcement may be difficult to distinguish during this stage, both successful and unsuccessful relationships may be expected to evidence high and similar levels of interactional compar- ability lNIth the client's past significant relationships. 30 Yet, while Mueller's (I969b) results tend to support the assumption of equal comparability during this stage, other research evidence points to the possibility that differences may exist between successful and unsuccessful relationships. Though Crowder found no significant ”transference'l differences between his groups, he did report a tendency for unsuccessful clients to behave with their therapists in ways which paralleled their past behavior with parents and significant others more frequently than successful clients. In addition, Dietzel (1971) reports that successful relationships differed significantly from unsuccessful dyads during the middle stage of therapy. Dietzel hypothesized that successful therapists would attempt to interfere with the client's disturbance maintaining behaviors. Successful dyads were in fact observed to be interacting at lower levels of complementarity during this stage. Hence, it is likely that differences in the comparability of client-therapist/client- significant other interaction patterns may be found between outcome groups. Since both the theoretical backdrop and the research evidence are conflictual, the direction of the expected differencesissnot stated. Hypothesis 3.2: There will be differences, during the 919912 stage in therapy, in the level of comparability between client-therapist and all clientepignificant other interaction patterns for successful and unsuccessful therapy dyads. 3.2a: There will be differences, during the middle stage of therapy, in the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful therapy dyads. 3.2b: There will be differences, during the middle stage of therapy, in the level of comparability between client-therapist and client-significant other person interaction patterns for successful and unsuccessful therapy dyads. 31 Late Stage Assuming that the therapist has not reinforced the client's disturbance maintaining elicitations and that transference reactions have been resolved, the client can be expected to interact with the therapist as a real person during the later stages of therapy (Alexander and French, 19H6). In successful therapy, the client's need for his previously learned security enhancing elicitations has diminished. He has learned a broader variety of self-enhancing interpersonal behaviors. In such relationships the client-therapist interaction will reflect low levels of comparability with the client's previous interpersonal experiences. If, however, the therapeutic relationship during this later stage continues to reflect high or even moderate levels of comparability with the client's previous interactions, then the client's elicitations were probably reinforced by the therapist. In such cases, the client and therapist have simply re-established and replayed the client's previous interpersonal experiences from which new learning is unlikely. Thus, in the later stage of therapy, successful therapeutic relation— ships will evidence lower levels of interactional comparability than will unsuccessful dyads. Hypothesis 3.3: During the .late stage of therapy, the level of comparabiliEy—BEtween client-therapist and all client— significant other interaction patterns will be lower in the successful, as opposed to the unsuccessful, therapy dyads. 3.3a: During the late stage of therapy, the level of comparability between client-therapist and client- parent interaction patterns will be lower in the successful, as opposed to unsuccessful, therapy dyads. 3.3b: 32 During the late stage of therapy, the level of comparability between client-therapist and client- other significant person interaction patterns will be lower in the successful, as opposed to the unsuccessful, therapy dyads. CHAPTER III METHOD Source of Data The psychotherapy cases for the present study were obtained from the research library at the Michigan State University Counseling Center. The research library contains test data and audio tape recordings from the counseling and psychotherapy cases of fifty-one clients. All clients were undergraduate self-referrals who sought help at the Center for personal and social problems and who agreed to participate in the Center's research project. Clients were assigned to therapists on the basis of matching client and therapist schedules. Therapists included senior staff members, interns and practicum students in counseling and clinical psychology. The senior staff therapists included 7 Ph.D. counseling and clinical psychologists with between 2 and 20 years of experience. The intern group included 3 second-year interns and 8 first-year interns. All had completed their practicum experience and averaged two years of supervised experience. The two therapists who were enrolled in an advanced practicum program at the Counseling Center had approximately one year of supervised experience. Descriptive data for_the cases used in this study are found in Appendix A. 33 34 Selection of Cases Two criteria were used to select cases from the tape library for this study. The first criterion was that the client must have continued in therapy for at least nine sessions. A minimum of nine sessions seemed necessary in order to sample and separate the three stages of therapy under investigation and to allow time for the process dimension to develop. The second criterion was that both pre- and post-therapy profiles be available for each case selected. These MMPI profiles were used to determine therapeutic outcome for each case and thus were necessary in order to test the study hypotheses. Therapeutic Outcome The outcome measure used in the present study was derived from clinical ratings of client change (i.e. degree of improvement or deterioration) evident from the pre-post psychotherapy MMPI profiles of clients. Rated profiles were available in the library for all clients who had taken both pre and post MMPI inventories. Available profiles had been rated by three judges who had graduate training and from 2 to 5 years experience with MMPI interpretation. The judges included two senior staff members at the Counseling Center and an advanced Ph.D. student in counseling psychology. The judges were given the following instructions for rating the profiles: Objective: To determine change in the MMPI as an indication of psychological change. 1. Compare pre-counseling and post-counseling profiled MMPI scores for each subject. 35 Consider the nine common scales (Hs + .SK, D, Hy, Pd + 4K, Pa, Pt + 1K, Ma + 2K, Sc + 1K). 2. Score the change as follows: 5 = satisfactory = partly satisfactory I. 3 = no change 2 = partly unsatisfactory I = unsatisfactory 3. In order to establish intra-rater reliability, please score each profile twice, one week apart. Each client, as a result of this scoring system, received six ratings--two ratings per judge. Appendix B contains the six individual ratings and average ratings for each case by three judges. The cases for this study were dichotomized into two groups (successful and unsuccessful) on the basis of the average of all ratings for each client. An average rating of 5. 3.00 represented the unsuccessful category. Clients whose average was > 3.00 were regarded as successful. TABLE 1.--Sex of clients in the population and sample of two outcome groups (N = 36). Number Sex of Client Group of Sample Population Cases M F M F Successful 20 1 9 5 15 Unsuccessful 16 A 6 6 IO 36 As reported in Table I, (N = 36) cases in the tape library met the criteria of 9 sessions and had MMPI profiles available. Twenty of these cases were judged as successful and sixteen as unsuccessful. Fifteen of the twenty successful clients were women, five were men; whereas ten of the sixteen unsuccessful clients were women and six were men. Stratified random sampling on the basis of outcome was used to select the cases for study. As a result of this procedure, the final sample (N = 20) consisted of 10 successful and 10 unsuccessful cases. Table 2 contains a summary of therapist and client character- istics as well as the mean number and range of sessions for each case. TABLE 2.--Client-therapist characteristics and mean and range of sessions for two outcome groups (N = 20). Group Number Client Therapist Mean Mean Range of of Sex Sex Experience Sessions Sessions Cases M F M F Level a Successful IO 1 9 6 4 2.HO 15.9 12-24 Unsuccessful IO 4 6 7 3 2.10 17.5 9-24 a. Experience Levels: 1 = senior staff; 2 = 2nd year intern; 3 = Ist year intern; 4 = practicum student Reliability of MMPI Judges Two reliability checks were made on the MMPI ratings: (I) an intra-judge reliability was obtained in order to detennine the agree- ment over time (one week apart) between the two ratings for a given 37 judge; and (2) an inter-judge reliability was obtained in order to determine the extent of agreement of the average of all ratings by the three judges for each client. The intra-judge reliability was tested by obtaining Pearson product-moment correlation coefficients between the first and second ratings (on the one to five scale) of each judge. The results of the intra-judge reliability data are listed in Table 3. TABLE 3.--Intra-judge reliability of MMPI ratings (N = 20). Judge Pearson Correlation Judge 1 .88 Judge 2 .82 Judge 3 .97 The inter-judge reliability was checked by use of the intra- class correlation formula developed by Ebel (1951). The inter-judge reliability data are listed in Table 4. The intraclass formula was used in order to check the reliability of the average of all ratings of all three judges for each client. This index was deemed appropriate because the categorization of cases into dichotomous groups was based upon the average of all ratings on each case. It is apparent from Tables 3 and 4 that both the inter-judge and intra-judge reliabilities are considerably greater than zero and that the judges gave consistent ratings on the measure used to assess client change from the beginning to the end of psychotherapy. 38 TABLE 4.--Inter-judge reliability of MMPI ratings by three judges using Ebel's intraclass correlation formula (N = 20). Reliability Reliability Source df SS MS Fs of of Average Ratings a Ratings b Clients . 19 161.09 8.48 .92 .97 Judges 2 6.95 3.48 Errorc 38 9.01 .24 Total 59 MS clients - MS error a. r = MS clients + (deudges) MS error MS clients - MS error MS clients c. The final ratings on which the decision to place a case in the successful or unsuccessful group was based upon averages of ratings from all judges. Therefore, the “between-judges“ variance was removed from the error term (Ebel, 1951). Selection of Sessions Client-Therapist Interaction Sessions were selected for analysis at three different points in the process in order to rate the client-therapist interaction patterns during the “early”, llmiddle“, and ”late” stages of therapy. The selection of sessions here was similar to the method used by Crowder (1970) and Dietzel (1971). Crowder selected the first three sessions, the pre-median, median, and post median sessions and the 39 last three sessions to represent the “early”, “middle”, and I'late” stages of psychotherapy. Dietzel (1971) selected two sessions--the first two; the pre-median and median; and the last two--as represen- tative of each of the three stages. Since some tapes selected in both studies proved inaudible, tapes from adjoining sessions or additional samples of the same session were rated. Such a procedure caused little difficulty with the Dietzel study since adjoining tapes were available and such substitution did not cause an overlapping of stages. Crowder, however, confronted difficulty when the over-all range of sessions was brief. His solution was to rate additional samples of the same session which caused some sessions to be more heavily represented for a stage than others. In this study the Dietzel method of selecting only two sessions as representative of each stage was used. It was felt that the loss of information encountered by rating two, as opposed to three, sessions per stage would be more than balanced by the gain in the session and stage representativeness. The first and second interviews, the pre- median and median interviews, and second last and last interviews were selected to represent the ”early“, ”middle” and “late” stages of therapy. Thus six sessions of each case, for a total of 120 sessions, were selected for analysis of client-therapist interaction patterns. Thirteen of the originally selected 120 tape recorded sessions were unratable because of poor sound reproduction and in one case a multiple (individual therapy using two therapists) ensued during the last two sessions. This necessitated selection of other sessions 40 for analysis. In these instances, adjacent sessions were substituted. In no case did this substitution result in loss of stage representation. For example, in one case sessions 17 and 18 were to be rated. Session 18 was, however, inaudible and session 16 was rated instead. Session 16 was still three sessions away from the boundary dividing the middle and late sessions and seven sessions from the median session. In both the studies discussed above as well as in earlier studies where the Interpersonal Circumplex was used, the usual proce- dure was to analyze a portion of a session and then to regard that portion as representative of the entire session. This procedure was used in the present study. A 15 minute segment of each selected session was rated. In order to avoid the normal greetings and leave- taking interactions and to maximize the probability of rating more significant interactions, the second 15 minutes of a typically fifty- minute session was rated. Client-Other Interaction It was also necessary to rate the tapes for client reports of interactions with parents and other significant persons. In two previous studies (Mueller, I969b; Crowder, 1970), the investi- gators pooled the ratings of client-other interactions from early and later stages in order to obtain an overall pattern of propor- tional client responses. It seemed likely, however, that client recollections of I1is interactions with others may change as therapy progresses and as his relationship with the therapist and with others outside of therapy change. A pooling of client-other 41 responses from early and later sessions might, then, obscure the client's original interpersonal dysfunction for which he seeks help. It also seemed likely that the two early stage 15 minute segments would contain relatively few client reports of his interaction with significant others. Therefore, the entire first two sessions of all selected cases were rated for client-significant other interaction as reported by the client. In addition, it was determined that a minimum of ten client reports each of behavior sent to and received from parents and other significant persons was necessary in order to provide an acceptable standard error for the prOportions in each octant of the circumplex. In those cases where this minimum number of reflexes was not reached within the first two sessions, the judges continued rating up to the pre-median interview until the minimum number of client reports in each of four categories was achieved. In one case the minimum was not achieved prior to the pre-median session. That case was discarded and another randomly selected from the same outcome category. As a result of this procedure, at least two entire interviews per case plus twenty additional sessions for a total of 60 sessions were selected from the early stage of the process for analysis of client-other interaction patterns. Of these sixty sessions, four tapes proved inaudible and adjacent sessions were substituted. No tapes, however, were rated for client-other behavior at the pre-median session or beyond. 42 Interaction Analypis System The method of tape analysis used in this study involved the interpersonal diagnosis system of behavioral analysis deveIOped by Freedman, Leary, Ossorio, and Coffey (I951), elaborated on by Leary (1957) and employed in several settings by Raush, g£_el. (1959), Raush, e§_§l. (1960), Mueller and Dilling (1968), Crowder (1970), and Dietzel (1971). Using this system, each communication unit (uninterrupted speech) of both client and therapist is scored and defined by one or more of 16 interpersonal reflexes arranged around a circumplex. Each reflex may be collapsed with three others and defined in terms of two major axes: a dominant-submissive and an affiliative-disaffiliative axis. The 16 reflexes are illustrated by the following verbs: boast, reject, punish, hate (disaffiliative-dominant); complain, distrust, condemn self, submit (disaffiliative-submissive); admire, trust, c00per- ate, love (affiliative-submissive); and support, give, teach, dominate (affiliative-dominant). A central aspect of this analysis system is that interpersonal behaviors are conceptualized as attempts on the part of each therapy participant to create an emotional state in the other which will evoke or elicit a predictable response. Raters, then, are to empathize with the person exhibiting the behavior from the position of the person to whom the behavior is directed (Freedman, e£_gl., 1951). The judges were instructed to rate each client and therapist response (uninter- rupted speech) first by locating it on the circumplex by quadrant 43 (e.g., affiliative-submissive, disaffiliative-dominant, etc.), octant, and then by specific reflex. When multiple reflexes occurred within the same response, the judges scored them sequentially. Ratingglntegpersonal Interaction In order to test the hypotheses under investigation two separate scorings were necessary. Client reports of interactions with parents and significant other persons were rated in the first scoring. For this scoring, the judges' task was to detennine: (1) whether the client statement was appropriate to be rated as client-other report; (2) the reflex sent; and (3) the target of the behavior. The potential targets were client, father, mother, brother, sister, male or female peer, male I or female authority figure and other. The client-therapist interaction was rated in the second scoring. Appendix G contains the scoring manual developed by Crowder (1970) and used by the judges in this study. Sixty complete sessions of client reports of interaction with others plus 120 tape segments of client-therapist interaction were randomly assigned to, and rated by, two judges. Both judges were advanced graduate students in counseling psychology with supervised psychotherapy experience and were presumed to be sensitive to the subtleties of human communication. The judges were extensively trained in the use of the interpersonal rating system. Training was done on non-study psychotherapy tapes and required approximately 45 hours. Two short training and review sessions were required following the completion of the client-other rating before rating of the client- therapist interaction could proceed. These review sessions were also conducted on non-study tapes. 44 Reliability Spmples Twenty of the 60 sessions (33%) rated for client reports of interaction with others and forty of the 120 tape segments (33%) rated for client-therapist interaction were selected to determine the reliability of the judges on the interpersonal scoring system. The tape segments were chosen so that the three stages of therapy under investigation (i.e. early, middle, late) would be approximately equally represented. Other than for this stipulation, both the selection of segments and the sequence of rating was random within the total sample. Independent ratings of both client reports of client-other behaviors and of client-therapist behaviors were made by the judges as they listened simultaneously to the psychotherapy tapes. For the client-other rating, each judge was randomly assigned to serve as criterion judge for one-half of the study sample in order to identify the specific client report to be rated. Within their random assign- ment, the judges alternated. Aside from selecting appropriate client reports, the only interaction permitted of the judges during rating was an occasional check of the specific ”response number” currently being rated. The inter-rater reliability was computed by use of Dittman's R (Dittman, 1958). The inter-rater reliability of the judges in scoring the client-therapist behaviors is reported in Table 5. The reliability of the judges in rating client reports of interaction with others is reported in Table 6. Of the (N = 3178) client-therapist 45 interaction ratings, the inter-rater reliability was +.75. 0f the (N - 721) ratings of client-other interaction, the judges achieved a reliability of +.84. These figures are somewhat higher than relia- bilities reported by Mueller (1969b) and Crowder (1970). Mueller reported reliabilities of +.64 on client-therapist reflexes and a reliability of +.73 for client-other ratings. Crowder reports reliabilities of +.62 and +.69 respectively. Thus the results in Tables 5 and 6 indicate a very acceptable inter-rater reliability on the interpersonal rating system. TABLE 5--Percentage agreement scores and Dittman's R for client-therapist- interactions based upon 40 IS-minute tape reliability segments. Agreement Unity of % of Cumulative Dittman's Dittman's Discrepancy a Agreement Agreement % 5b RC 0 - o 2206 .694 .694 0 ii = +.76 I - D 207 .065 .759 207 2 - D 214 .067 .826 428 3 - D 154 .048 .874 462 4 - D 197 .062 .936 788 5 - D 37 .012 .948 185 6 - D 115 .036 .984 690 7 - D 37 .012 .995 259 8 - D 11 .004 1.000 88 Total 3178 Sum (a) 3107 a. 0 - D = perfect interjudge agreement, 8 - D = bipolarity of interjudge agreement. b. 5 = number of categories between the ratings ofnthe judges. 6 .2 /n c. For a 16 variable circumplex, Dittman's -_ "' R’ 1. 46 TABLE 6.--Percentage agreement scores and Dittman's R for client-parent and client-other significant person interaction based on 20 50-minute tape reliability segments. Agreement Unity of % of Cumulative Dittman's Dittman's Discrepance a Agreement Agreement % 6b 'fic 0 - D 533 .739 .739 0 I - 0 79 .109 .848 79 E = +.84 2 - D 43 .060 .908 86 3 - D 24 .033 .941 72 4 - o 16 .022 .963 54 5 - D 8 .011 .974 40 6 - D 14 .019 .993 84 7 - D 2 .003 .996 14 8 - 0 2 .003 .999 16 Total 721 Swn ( 6) 455 a. O - D = perfect interjudge agreement, 8 - D = bipolarity b. C. of interjudge agreement. 6 = number of categories between the ratings of the judges. n .2 9/n 1_ l-I For a 16 variable circumplex, Dittman's R'= l. 47 Experimental Design A two-group comparative design with repeated measures over time (stages) was used in this study. Figure 1 provides a pictorial representation. O H H O H N Hoe-cocoa. N 0...... O [.4 10 Figure 1.--Pictorial representation of experimental design. :0 II Random assignment Client ('3 —I II 01 = Successful Therapeutic Outcome 02 = Unsuccessful Therapeutic Outcome $1'= Early Stage of Psychotherapy $2 = Middle Stage of Psychotherapy S3 = Late Stage of Psychotherapy T = Total behavior patterns calculated over the entire range of therapy sampled. 48 Preparation of Data for Analysis The degree of similarity or ”level of comparability” of the client-therapist and client-other behavior patterns was measured by the Q_statistic developed by Cronbach and Gleser (1953). Since the frequency of rated behaviors in several of the 16 circumplex cate- gories was sometimes sparse, the circumplex was collapsed into octants (BC--boast, regject; DE--punish, hate; FG--complain, distrust; HI-- condemn-self, submit; JK--admire, trust; LM--cooperate, love; NO-- support, give; PA--teach, dominate). The proportion of reflexes which each client reported using in response to parents, other significant persons, and all others in each of the circumplex octants was obtained. Similar proportions were obtained for the behaviors which each client reported receiving from others [parent + client (P+C); other significant person + client (0+C); all others + client (O+P+C)]. These proportions were then deviated, scale by scale, from the pr0portions of actual behavior which each client used with and received from the therapist (client + therapist; therapist + client) over the entire range of therapy and for each of three stages. Each scale by scale deviation for the eight behavioral categories was squared and summed across all scales and the square root of the summed squared differences was derived. A 2 score was obtained for the six comparisons of client-other and client-therapist behavior (C+T-C+P; C+T-C+0; 0+T-C+P+0; T+C-P+C; T+C-O+C; T+C-P+0+C) for each client-therapist pair across the three stages and for the entire range of therapy sampled. 49 _. _ 2 0- WP... 9..» _ 2 _ where (PC4T PC+P) _ 2 BC(PC+T PC+P) + DE . . . AP for: (C+T)-(C+P) (C+T) - (C+0 (C+T) - (C+P+O for Early, Middle and Late Stages (T+C) - (P->C (T+C) - (0+C (T+C) - (P+O+C (C+TT) - (c_,P) (C+TT) - (C+O) (C+TT) - (C+O+P for Total - Entire Range of Therapy (T+CT) ' (9+C Sampled (T+CT) - (0+0 (T+CT) - (0+P+C These p scores served as measures of comparability between two profiles. The lpgeg the 2 score, the highep is the level of compar- ability between two profiles and the greater is the similarity of the two behavior patterns compared. For example, hypotheses 1, 1a and 1b deal with the differences in comparability level between the two outcome groups over the entire range of therapy sampled. Hypotheses I, Ia and 1b predict that the successful outcome group will interact at lower levels of comparability (higher 2 scores) than will the unsuc- cessful group. The test for this hypothesis involved two comparisons: (I) [(C+O) - (C+T)]; and (2) [(O+C) - (T+C)]; for three different types of interactions (C/P; C/O; C/O + P). The greater the 50 comparability level between the client-other and client-therapist interaction patterns, the smaller were the mean 2 scores for each group. Thus for hypothesis 1, 1a and 1b to be accepted, the mean 52 scores for the successful group would have to be significantly higher than the mean 2 scores for the unsuccessful group. The data for the remaining study hypotheses may be interpreted in the same way. Analysis of Data A repeated measures analysis of variance using early, middle, and late stages of therapy as repeated measures was employed. This analysis allowed for consideration of both outcome groups together in a test for variations in behavior patterns across stages and also allowed for a test for differences in behavior patterns across stages by outcome, or stage with outcome interaction. Tukey post hoc compar- isons were used to test the differences between stages and Scheffe comparisons were used to test the differences for outcome groups between stages when significant interaction was detected. In addition, univariate ANOVA's were performed to test for differences in behavior patterns between outcome groups for the total range of therapy. One- way ANOVA's were necessary because the data for total behavior was calculated separately. A basic statistics computer program was utilized and a test for skewness showed no great nor consistent variations from zero. The data were, therefore, assumed to represent a symetrical and approxi- mately normal distribution. The repeated measures ANOVA design also assumes that the repeated measures, the stages in this case, have 51 approximately equal pair wise correlations. This assumption was not met. However, use of the Geisser and Greenhouse Conservative F test (1958) permitted analysis of the data. CHAPTER IV RESULTS Operational Definition of Hypotheses While in therapy clients not only interact with the therapist but also may recall interactions which have taken place with others. They often report the way in which they have behaved with others as well as the way in which others have interacted with them. The similarity or comparability between these two types of reported interactions and the actual behavior which the client used with and received from the therapist comprised the focus of this study. The "level of comparability" between these four sets of behaviors was studied at three different points in time as well as over the total range of therapy. The study hypotheses were operationalized in the following way: Each of the hypotheses involved comparing two sets of reactions for the two outcome groups: (1) comparison of client reported reactions £g_others with his actual reactions §p_the therapist; and (2) comparison of client reports of reactions received frpm_others, with the actual reactions the client received frpm_the therapist. The hypotheses were constructed so as to test separately the level of comparability between the client's reported behavior with others, parents, and all others (parents plus other significant persons) and the client's behavior with the therapist as well as the level of 52 53 comparability between the client's reports of the reactions of others, parents, and all others to him and the therapist's reactions to the client. Thus, each hypothesis was tested by comparing two sets of interactions. Client Therapist Comparability Over the Entire "Ragga of Therapy and Therapeutic Outcome Hypotheses I, la and lb deal with the main effect for outcome for total client-therapist behavior patterns summed over the entire range of therapy. Since the total Q_scores for each group were calculated separately, one-way ANOVAs were used. In this presenta- tion of results a summary of two abbreviated ANOVA tables for each of the three hypotheses will be presented followed by an evaluation of the applicable hypotheses. TABLE 7.--Cell mean Q_scores on the total level of comparability for successful and unsuccessful cases (N = 20). Total Mean Q_Scores Variable O1 02 (Successful) (Unsuccessful) H 1 [(C+0+P) - (0+TTII .358 .340 [(O+P+C) - (T+CT)] .515 .508 H 'a IIC+PI - (mp1 .468 .451 [(P+C) - (T+CT)] .445 .540 ”lb [(c»0) - IC+TTII .340 .294 [(O+C) - (T+CT)] .532 .540 Hypothesis 1 H 54 01: The level of comparability between client-therapist and all clientfisignificant other interaction patterns for the successfhl group will equal (or be higher than) the comparability level for the unsuccessful group. H “01 0' [(C+o+P) - (C+TT)] [(0+P->c) - (1+ch1 5- D02 1 D02 Al: The level of comparability between client-therapist and all clientL-significant other interaction patterns will be lower for the successful, as opposed to the unsuccessful, clientétherapist dyads. HAl H [(C+0+P) - (C+TT)] A' [(0+P+C) - (14.11 > D02 > ”02 TABLE 8.--Summary of the univariate ANOVAs on the total level of comparability for successful and unsuccessful cases. Mean Square Mean Square Variable Between Error F(df:l,18) H1 a [(C+O+P) - (C+TT)] .001538 .007147 .21519 [(O+P+C) - (1+cT)] .000240 .017471 .01373a ”la a [(C+P) - (0+1T)] .001443 .010595 .13618' [(P+C) - (1+cT)] .044566 .015139 2.75138a Hlb a [(C+O) - (C+TT)] .010347 .007324 1.41275 [(O+C) - (T+CT)] .008724 .027668 .32253a a Not significant 55 Results.--It can be seen in Table 8 that the F ratios of (.21519) and (.01373) were not significant on the main effect of outcome for [(C+0+P) - (C+TT)] and [(O+P+C) - (T+CT)] comparisons; therefore, the null hypothesis was not rejected. This hypothesis stated that the client-therapist behavior patterns in successful, as compared with unsuccessful cases, would be as similar or more similar to the reported interaction patterns of clients with pgpe£§_ and parents. Hypothesis la ”Ola: The level of comparability between client-therapist and client-parent interaction patterns for the successfuT’group will equal (or be higher than) the comparability level for the unsuccessful group. H01a [(C+P) - (C+T )] - 5' <'5 T ' Ol - 02 H 0'9 [(p+c) - (T+CT)] : 0'0, 5502 ”Ala: The level of comparability between client-therapist and client-parent interaction patterns will be lower for the succesSful, as opposed to the unsuccessful, client-therapist dyads. .— H A” [(0—49) - Imp] : b’ > 1302 01 H Ala , - - [(P+C) - (T+CT)] . Do] > 002 Results.--The F ratios of (.13618) and (2.76138) contained in Table 8 were not significant on the main effect of outcome for [(C+P) - (C+TT)] and [(P+C) - (T+CT)] comparisons. Therefore, the null hypotheses that the degree of similarity for successful cases 56 between client-therapist interaction patterns and reported client- parent behavior patterns would be equal to (or higher) than for unsuccessful cases, was not rejected. Hypothesis lb HOlb: The level of comparability between client-therapist and client—other significant person interaction patterns for the successtl group will equal (or be higher than) the comparability level for the unsuccessful group. H 0'” [(c->0) - (C+TT)] : 5m _<_ H 01b U 5- D02 0+0) - (1+cT)] : 501 HAlb: The level of comparability between client-therapist and client-other signifiCant person interaction patterns will be lower in the successful, as opposed to the unsuccessful. client-therapist dyads. HA1 b [(C+0) - (cap) : b‘ > no, 01 H A”) [(04) - (1+cT)] : 501 > “0'02 Results.--The F ratios of (1.41275) and (.32253) as listed in Table 8 were not significant on the main effect of outcome for [(C+0) - (C+TT)] and [(O+C) - (T+CT)] comparisons; therefore, the null hypothesis was not rejected. This hypothesis stated that the degree of similarity between the client and therapist behaviors and the reported behaviors of others (excluding parents) for the successful group would be equal to (or higher than) the degree of similarity for the unsuccessful group. 57 Client-Therapist Comparability Level Across *Three Stages of Therapy, Following are tables of the Cell Means and ANOVA results for the two-way repeated measures ANOVA. This information will be used in evaluating hypotheses 2, 2a, and 2b. TABLE 9.--Cell mean Q_scores for the level of comparability between client-therapist and client-other interaction patterns across three stages of therapy. Stages (Repeated Measures) Variable S1 (Early) 52 (Middle) 53 (Late) H2 [(C+0+P) - (C+T)] .366 .369 .377 [(0+P+C) - (T+C)] .588 .530 .471 H28 [(C+P) - (C+T)] .468 .474 .457 [(P+C) - (T+C)] .565 .528 .450 sz [(C+0) - (C+T)] .345 .336 .341 [(O+C) - (T+C)] .633 .587 .527 Hypothesis 2 H02: There will be no difference in the level of comparability between client-therapist and all client- significant other interaction patterns across three stages offtherapy. H 02 [(C+O+P) — (can : 551' l D II C3 52 S3 H02 -— -— -— [(0+P+C) - (T+C)] : as] - $2 — $3 I U I U 58 HA2: There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of' therapy. HA2 [(C+0+P) - (C+T)] : H02 is false HA2 [(O+P+C) - (T+C)] : H02 is false TABLE lO.--Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2 [(C+O+P) - (C+T)]. Source SS df MS F (df) Total .519 59 .008798 Outcome .004 1 .004192 Clients within outcome .327 18 .018140 b Stages .001 2 .000659 .13424a (1,18) Stages by outcome .010 2 .005174 Stages by clients within outcome .177 36 .004909 aNot significant. bRepeated measures ANOVA assumes that the measures (in this case Stages) have like pair wise correlations between and among them- selves. There was, however, no basis for making this assumption. Therefore, the Geisser and Greenhouse(l958) Conservative F test was used which allowed violation of this assumption. With this method the computation procedures for F are identical but reduced degrees of freedom are used for determining the critical value (a : F%T‘ r = no. of repeated measures; df = e(df1), e(df2). The liberal degrees of freedom would have been 2 and 36. CSix repeated measures ANOVAs were used to test for differ- ences in comparability level for the six comparisons: C+O+P; O+P+C; C+P; P+C; C+O; 0+C. Therefore, the mean squares listed in Table 9 through 15, which list the F ratios for the main effect of stage, are identical in a pair-wise fashion for the same comparison as the mean squares listed in Tables 17 through 22. Tables 17 through 22 list the F ratios for the interaction effect of stage by outcome. For example, the mean squares listed in Table 10 for the [(C+0+P) - (C+T)] comparison are identical to the mean squares listed in Table I7. 59 TABLE ll.--Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2 [(0+P+C) - (T+C)]. Source SS df MS F (df) Total 1.308 59 .022178 Outcome .000 1 .000409 Clients within outcome .893 18 .049600 Stages .137 2 .068490 9.07873b (1.18) Stages by outcome .007 2 .003359 Stages by clients within outcome .272 36 .007544 p < .10 on both Liberal and Conservative Tests Results.--It can be seen in Table 10 that the F ratio of (.13424) on the main effect of stages for the [(C+0+P) - (C+T)] comparison was not significant. Therefore, the null hypothesis which stated that there would be no fluctuation in the degree of compara- bility between client behaviors with the therapist and reported behaviors of clients with others and parents across three stages of therapy was not rejected. However, the F ratio of (9.07873) listed in Table 11 on the main effect of stages for [(O+P+C) - (T+C)] behavior was significant. Therefore, the null hypothesis was rejected in favor of hypothesis A2. This hypothesis stated that there would be fluctuations in the degree of comparability between therapist-to-client behaviors and other plus 6O parent-to-client behaviors across the three stages of therapy for the combined outcome groups. Comparability Level Means Low .600 T -.588 -.541 '500 " -.471 High .400 - Early Middle Late Figure 2.--Graph of therapist comparability level over three stages of therapy [(O+P+C) - (T+C)]. A Tukey post hoc analysis on the main effect of Stages for [(O+P+C) - (T+C)], using the conservative degrees of freedom, indicated the following: 1. Row mean for 52 (Middle Stage) was greater than the row mean for S3 (Late Stage) with p .10. Row mean for S2 (Middle Stage) was pr_greater than the row mean for S1 (Early Stage) with p .10. Row mean for S3 (Late Stage) was greater than the row mean for S1 (Early Stage) with p .10. Hypothesis 2a H 02a: There will be no difference in the level of comparability between client-therapist and client- parent interaction patterns across three stages of therapy; H02 H02 HA 61 [(c») - (cm) : D's [(P+C) - (T+C)] : fig 5' D 52 - $2 - $3 53 2a: There will be differences in the level of comparability between client-therapist and client- parent interaction patterns across three stages of therapy. HA2a [(C+P) - (C+T)] : H02a is false HA2a [(P+C) - (TeC)] : H02a is false TABLE 12.--Repeated measures ANOVA table for the level of compara- bility between client-therapist and client-parent interaction patterns across three stages of therapy--main effect of sta Hypothesis 2a [(C+P) - (C+T) e Source SS df MS F (df) Total .753 59 .012771 Outcome .005 I .005334 Clients within outcome .521 18 .028930 Stage .004 2 .001882 .30369a (1.18) Stages by outcome .001 2 .000259 Stages by clients within outcome .223 36 .006197 aNot significant 62 TABLE l3.--Repeated measures ANOVA table for the level of compara- bility between client-therapist and client-parent interaction patterns across three stages of therapy-~main effect for stage Hypothesis 2a [(P+C) - (T+C)] Source SS df MS F (df) Total 1.357 59 .023005 Outcome .126 1 .126491 Clients within outcome .800 18 .044468 Stage .136 2 .068136 8.45674b (1.18) Stage by outcome .004 2 .002027 Stage by clients within outcome .290 36 .008057 b p < .10 on both Liberal and Conservative Tests Results.--In Table 12 the F ratio of (.30369) for the [(C+P) - (C+T)] comparison on the main effect of stages is listed. This value was not significant. Thus, the null hypothesis which stated that there would be no fluctuation in the degree of similarity between client-to-therapist behaviors and reported client-to-parent behaviors across three stages of therapy was not rejected. However, the F ratio of (8.45674), listed in Table 13, on the main effect of Stages for the [(P+C) - (T+C)] comparison was sig- nificant. Therefore, the null hypothesis was rejected in favor of hypothesis A28. This hypothesis stated that there would be fluctuations in the degree of comparability between therapist-to- client behaviors and reported parent-to-client behaviors across the three stages of therapy for both outcome groups. Comparability Level Low .600 F 63 Means 565 .528 .500 r .450 High .400 . A. Early Middle Late Figure 3.--Graph of therapist comparability level over three stages of therapy [(P+C) - (T+C)]. A Tukey post hoc analysis on the main effect of stages for [(P+C) - (T+C)], using the conservative degrees of freedom, yielded the following: 1. Hypothesis 2b Row mean for 52 (Middle Stage) was greater than the row mean 53 (Late Stage) with p .10. Row mean for S2 (Middle Stage) was pp§_greater than the row mean for 51 (Early Stage) with p .10. Row mean for S3 (Late Stage) was greater than the row mean for S1 (Early Stage) with p .10. H 02b: There will be no difference in the level of compara- bility between client-therapist and client-significant- other person interaction patterns across Three stages of therapy. H 02" moo) - (C+T)] : '51 — 5 = '5 U I H 02b[(0>C)-(T+C)l : ‘5 -‘ -‘ I U 64 A2b: There will be differences in the level of comparability between client-therapist and client-significant other person interaction patterns across three stages of therapy. HA2b [(C+0) - (C+T)] : H02b is false HAzb [(0+C) - (T+C)] : H02b is false TABLE l4.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2b [(C+O) - (C+T)]. Source SS df MS F (df) Total .531 59 .008997 Outcome .028 1 .028032 Clients within outcome .336 18 .018673 Stage .001 2 .000379 .10190a (1,18) Stage by outcome .032 2 .016012 Stage by clients within outcome .134 36 .003719 aNot significant Results.--Table 14 contains the F ratio of (.10190) for the main effect of stages on the [(CFO) - (CFT)] comparison. This value was not significant. The null hypothesis which stated that there would be no fluctuation in the degree of comparability across the three stages of therapy between client-to-therapist behavior and client-to-other (excluding parents) behavior was, therefore, not rejected. 65 TABLE 15.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns across three stages of therapy--main effect of stage Hypothesis 2b [(O+C) - (T+C)]. Source SS df MS F (df) Total 1.805 59 .030594 Outcome .020 2 .019729 Clients within outcome 1.418 18 .078458 Stage .113 2 .056502 8.01446b (1,18) Stage by outcome .006 2 .003135 Stage by clients within outcome .254 36 .007050 p < .10 on both Liberal and Conservative Tests However, the F ratio of (8.01446) listed in Table 15 on the main effect of stages for the [(O+C) (T+C)] comparison was signifi- cant. Therefore, the null hypothesis was rejected in favor of hypothesis 2b which stated that fluctuations would occur in the degree of comparability between therapist-to-client behavior and other-to-client (excluding parents) behaviors across the three stages of therapy for the combined outcome groups. 66 Means Low .700 . >~. :3 'F .633 E '3; o 600 AL 2 6 87 (U _l D. h. E .527 “’ High .500 Early Middle Late Figure 4.--Graph of therapist comparability level over three stages of therapy [(O+C) - (T+C)]. A Tukey post hoc analysis on the main effect of Stages for [(O+C) - (T+CIL using the conservative degrees of freedom, yielded the following: 1. Row mean for 32 (Middle Stage) was gg§_greater than the row mean for S3 (Late Stage) with p. 10. 2. Row mean for S2 (Middle Stage) was pg§_greater than the row mean for S1 (Early Stage) with p .10. 3. Row mean for 53 (Late Stage) was greater than the row mean for S1 (Early Stage) with p .10. Client-Therapist Comparability Level Across the Early, Middle and LateTStages of Thengpy anHTTherapeutic Outcome Following are tables of the Cell Means and ANOVA results for the two-way repeated measures ANOVA. This information will be used in evaluating hypotheses 3, 38 and 3b as well as in investigating the interaction effects of 3.1b, 3.2b, and 3.3b. 67 TABLE 16.--Cell mean D scores for the level of comparability between client- -therapist and client-other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy. 0 0 Variable 1 2 S1 S2 S3 S1 52 S3 H3 [(G+O+P) - (0+T)] .363 .371 .404 .370 .366 .351 [(O+P+C) - (TFC)] .603 .530 .474 .572 .551 .468 H3a [(CeP) - (CFT)] .478 .479 .499 .457 .468 .474 [(P4C) - (TFC)] .529 .472 .405 .601 .584 .496 H3b [(C40) - (CaT)] .343 .350 .394 .347 .322 .288 [(04C) - (TaC)] .655 .595 .542 .601 .580 .511 Key:' 01 = Successful 02 = Unsuccessful S1 = Early Stage S2 = Middle Stage S3 = Late Stage Hypothesis 3 H03: There will be no difference in the level of comparability between client- -therapist and all client- significant other interaction patterns for the successtl and unsuccessful client- -therapist dyads across the early, middle and late stages of therapy. H 03 [(C+0+P) - (C+T)] : There will be no interaction of S x O H 03 [(O+P+C) - (T+C)] : There will be no interaction of S x O 68 A3: There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns for the succeséfUT'and' unsuccessful client-therapist dyads across the early, middle and late stages of therapy. H A3 [(C+O+P) - (C+T)] : There will be interaction of S x O H A3 [(O+P+C) - (T+C)] : There will be interaction of S x 0 TABLE l7.--Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome. Hypothesis 3 [(C+O+P) - (04T)] Source SS df MS F (df) Total .519 59 .008798 Outcome .004 1 .004192 Clients within outcome .327 18 .018140 Stage .001 2 .000659 Stage by outcome .010 2 .005174 V 1.05398a (1,18) Stage by clients within outcome .177 36 .004909 aNot significant Results.--It can be seen in Tables 17 and 18 that the F ratios of (1.05398) and(.44525) for the interaction of stage with outcome on the [(C+O+P) - (C+T)] and [(O+P+C) - (T+C)] comparisons were not significant. Therefore, the null hypothesis which stated 69 TABLE 18.--Repeated measures ANOVA table for the level of comparability between client-therapist and all client-significant other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome. Hypothesis 3 [(O+P+C) - (T+C)] Source SS df MS F (df) Total 1.308 59 .022178 Outcome .000 1 .000409 Clients within outcome .893 18 .049600 Stage .137 2 .068490 Stage by outcome .007 2 .003359 .44525a (1,18) Stage by clients within outcome .272 36 .007544 aNot significant that there would be no difference between outcome groups across the early, middle and late stages of therapy in the degree of comparability between client-therpist and client-othergplus_parent behavior patterns was not rejected. Hypothesis 3a HO3a: There will be no difference in the level of comparability between client-therapist and client- parent interaction patterns for the successful and unsuccessful client—therapist dyads across the early, middle and late stages of therapy. H 03a [(CSP) - (C+T)] : There will be no interaction of S x O H 03a [(p+c) - (T+C)] : There will be no interaction of S x’O 7O HABa: There will be differences in the level of comparability between client-therapist and clienteparent interaction patterns for the successful and unsuccessful client- therapist dyads across the early, middle and late stages of therapy. H A39 [(csp) - (C+T)] : There will be interaction of S x O H A33 [(P+C) - (T+C)] : There will be interaction of S x 0 TABLE l9.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy--interaction effect of stage with outcome. Hypothesis 38 [(C+P) - (C+T)] Source SS df MS F (df) Total .753 59 .012771 Outcome .005 1 .005334 Clients within outcome .521 18 .028930 Stage .004 2 .001882 Stage by outcome .001 2 .000259 .04179a (1,18) Stage by clients within outcome .223 36 .006197 aNot significant Results.--In Tables 19 and 20 the F ratios of (.04179) and (.25158) are listed for the interaction effect of stages with outcome for the [(C+P) - (C+T)] and [(P+C) - (T+C)] comparisons. These values were not significant. Thus, the null hypothesis which stated that 71 TABLE 20.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-parent interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy-iinteraction effect of stage with outcome. Hypothesis 3a [(P+C) - (T+C)] Source SS df MS F (df) Total 1.357 59 .023005 Outcome .126 1 .126491 Clients within outcome .800 18 .044468 Stage .136 2 .068136 Stage by outcome .004 2 .002027 .25158a (1,18) Stage by clients within outcome .290 36 .008057 aNot significant there would be no differences across the early, middle and late stages of therapy between outcome groups in the degree of similarity of client-therapist as compared With client-parent behavior patterns was not rejected. Hypothesis 3b H03b‘ There will be no difference in the level of comparability between client-therpist and client-other significant person interaction patterns for the successfhl and unsuccessful client-therapist dyads across the early, middle and late stages of therapy. H 03b [(c+o) - (C+T)] : There will be no interaction of S x 0 H 039 [(O+C) - (T+C)] : There will be no interaction of S x O 72 HA3b: There will be differences in the level of comparability between client-therapist and client-other si nificant person interaction patterns for the successful ana unsuccessful client-therapist dyads across the early, middle and late stages of therapy. H A3b [(cso) - (C+T)] : There will be interaction of S x O H A3b [(osc) - (T+C)] : There will be interaction of S x 0 TABLE 21.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns for successful and unsuccessful cases across the early, middle and late stages of therapy-~interaction effect of sta e with outcome. Hypothesis 3b [(C+0) - (C+T)1 Source SS df MS F (df) Total .531 59 .008997 Outcome .028 1 .028032 Clients within outcome .336 18 .018673 Stage .001 2 .000379 Stage by outcome .032 2 .015012 4.30545a (1,18) Stage by clients within outcome .134 36 .003719 ap <.lO on both Liberal and Conservative tests. Results.--Table 21 contains the F ratio of (4.30545) for the comparison of [(C+O) - (C+T)] on the interaction effect of stage with outcome. This value was significant at p < .10. Therefore, the null hypothesis was rejected in favor of hypothesis A3b which stated that 73 TABLE 22.--Repeated measures ANOVA table for the level of comparability between client-therapist and client-other interaction patterns for successful and unsuccessful cases across three stages of therap -- interaction effect of stage with outcome. Hypothesis 3b [(O+C) - (T+C)] Source 55 df MS F (df) Total 1.805 59 .030594 Outcome .020 2 .019729 Clients within outcome 1.412 18 .078458 Stage .113 2 .056502 Stage by outcome .006 2 .003135 .44468a (1,18) Stage by clients within outcome .254 36 .007050 aNot significant successful clients would differ from unsuccessful clients in the degree of comparability between their behavior with the therapist and their reported behavior with others (excluding parents) across the early, middle and late stages of therapy. In Table 22 the F ratio of (.44468) on the interaction effect of stage with outcome for the [(O+C) - (T+C)] comparisons is listed. This value was not significant. Thus, the null hypothesis which stated that there would be no differences across the early, middle and late stages of between outcome groups in the degree of similarity of therapist-to-client as compared with otheretoeclient (excluding parents) behavior patterns was not rejected. 74 Mean E Low .400 1’ d) _l 41- .5’ F- .300 ‘I’ .E «‘5 5- q- E- 8 High .200 Early Middle Late --1v———- Successful -——r— Unsuccessful Figure 5.--Graph of client comparability level for two outcome groups over three stages of therapy for [(C+0) — (C+T)] interaction patterns. A scheffé post hoc analysis for complex comparisons was used to investigate differences of both groups between each stage. The Scheffe was conducted using the conservative degrees of freedom and yielded the following results: 1. The difference between outcome groups for [(C+O) - (C+T)] interaction patterns at 53 (Late Stage) was greater than the difference between outcome groups at 52 (Middle Stage) with l, 18 degrees of freedom. 2. The difference between outcome groups for [(C+O) - (C+T)] interaction patterns at S3 (Late Stage) was greater than the difference between outcome groups as S1 (Early Stage) with l, 18 degrees of freedom. 3. The difference between outcome groups for [(C+O) - (C+T)] interaction patterns at 52 (Middle Stage) was Hg§_greater than the difference between outcome groups at S1 (Early Stage) with 1, 18 degrees of freedom. 75 There was no interaction between Stages and Outcome for either of the behavior patterns compared in hypotheses 3 and 3a. Therefore, hypotheses 3.1, 3.1a; 3.2, 3.2a; and 3.3, 3.3a which dealt with pre- dicted differences in comparability level over three stages of therapy could not be tested. However, hypothesis 3b dealing with [(C+O) - (C+T)] behavior patterns was supported and the results of the Scheffe post hoc indicated that differences existed between the Late and Early and the Late and Middle stages of therapy. The Scheffe technique allowed testing the interaction for differences in comparability level between stages. However, it was also of interest to test for differences in comparability between outcome groups at each of the three stages of therapy. Therefore, the Tukey post hoc technique was used. This technique allowed testing of hypotheses 3.1b, 3.2b and 3.3b on the [(C+O) - (C+T)] behavior patterns only. The results of this analysis follow. Hypothesis 3.1b H03 1b: There will be no difference, during the earl ‘ stage of therapy. in the level of compara i ity between client-therapist and client-other si nificant person interaction patterns for the success 01, as opposed to the unsuccessful, client-therapist dyads. H03.1b — [(C+O) - (C+T)] : 501 = 0 S1 Hypothesis 3.2b ”03.2b‘ 43.2b‘ Hypothesis 3.3b “03.3b‘ “A3.3b‘ 76 There will be no differences, during the middle stage of therapy, in the level of comparaBiIity " between client-therapist and client—other significant person interaction patterns for the successful, as opposed to the unsuccessful, client-therapist dyads. C+O) - (C+T) : D0 = 002 H 03.2b [( 1 5 s2 2 There will be differences, during the middle stage of therapy, in the level of comparability between client-therapist and client-other sjgnificantgperson interaction patterns in the successful, as opposed to the unsuccessful, client-therapist dyads. H A3.Zb [ (0+0) - (C+T)] : 501 ,1 602 S S 1 1 During the late stage of therapy, the level of comparability between client-therapist and client- other significant person interaction patterns for the successf T group will be equal to (or higher than) the level of comparability for the unsuccessful group. H . 03.3b , — — [Io->0) - (C+T)] . ”01 5002 S S 3 3 During the late stage of therapy. the level of comparability between client-therapist and client- other significant person interaction patterns will be waer f0r the successful, as opposed to the unsuccessful, client-therapist dyads. H A33” [(90) - (can : 501’ 002 S S 3 3 77 Results.--A Tukey post hoc analysis of the differences in comparability level across the early, middle and late stages of therapy for the client-to-other (excluding parents) vs. the client-to- therapist comparison was conducted using the conservative degrees of freedom. This test yielded the following results: 1. The mean Q_score for the unsuccessful group at S1 (Early Stage) was pp§_greater than the mean Q_score for the successful group at S1 (Early Stage) with p = .10. Therefore, the hypothesis which stated that there would be no differences at the early stage of therapy between the outcome groups in the degree of similarity of client-to-therapist and reported client-to-other (excluding parents) behavior patterns was accepted. 2. The mean Q_score for the successful group at S2 (Middle Stage) was pr_greater than the mean 9 score for the unsuccessful group at S2 (Middle Stage) with p = .10. Therefore, the null hypothesis which stated that there would be no differences between outcome groups at the middle stage of therapy in the degree of similarity of client- to-therapist as compared with client-to-other (excluding parents) behavior patterns was not rejected. 3. The mean H_score for the successful group at S3 (Late Stage) was greater than the mean 9 score for the unsuccessful group at S3 (Late Stage) with p = .10. Therefore, the null hypothesis was rejected in favor of hypothesis A3.3b which stated that successful clients at the late stage of therapy would behave with the therapist in ways which were less similar to their reported behavior with others (excluding parents) than would unsuccessful clients. Al‘ A2‘ 43‘ Ala‘ Alb‘ A2a: Azb‘ A3a: 78 Status of Research Hypothesis The level of comparability between client-therapist and 311_ client-significant other interaction patterns will be lower for the successful, as opposed to the unsuccessful client- therapist dyads. Not Confirmed The level of comparability between client-therapist and client- parent interaction patterns will be lower for the successful as opposed to the unsuccessful client-therapist dyads. Not Confirmed The level of comparability between client-therapist and client- significant other person interaction patterns will be lower for the successful, as opposed to the unsuccessful client-therapist dyads. Not Confirmed There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns across three stages of therapy. Not Confirmed for [(C+O+P) - (C+T)] interaction patterns. Chhfirmed for [(O+P+C) - (T+C)] interaction patterns. There will be differences in the level of comparability between client-therapist and clienteparent interaction patterns across three stages of therapy. Not Confirmed for [(C+P) - (C+T)] interaction patterns. Confirmed for [(P+C) - (T+C)] interaction patterns. There will be differences in the level of comparability between client-therapist and client-significant other person inter- action patterns across three sta es of therapy. Not Confirmed for [(C+O) - ?C+T)] interaction patterns. Confirmed for [(O+C) - (T+C)] interaction patterns. _There will be differences in the level of comparability between client-therapist and all client-significant other interaction patterns for the successful andihnsuccessful’client-therapist dyads across three stages of therapy. Not Confirmed There will be differences in the level of comparability between client-therapist and client-parent interaction patterns for the successful and unsuccessful client-therapist dyads across three stages of therapy. Not Confirmed 79 HA3b: There will be differences in the level of comparability between client-therapist and client-significant otherpperson inter- action patterns for the successful and unsuccesstT'client- therapist dyads across three stages of therapy. Confirmed for [(C+O) - (C+T)] interaction patterns. Not Confirmed for [(O+C) - (T+C)] interaction patterns. HA3 1 Hypotheses 3.1 and 3.1a predicted no ° difference in comparability level HA3 la between outcome groups at the early ‘ stage of therapy. These hypotheses were HA3 1b not directly tested. However, due to the ' [(O+C) - (T+C)] absence of interaction effect they were supported by the data. Confirmed. HA3.2 HA3.2a HA3 2b Since there were no applicable interaction ' [(O+C) - (T+C)] effects covering these hypotheses, they could not be tested. They were, therefore, H Not Confirmed. A3.3 HA3.3a HA3.3b [(0:0) - (T+C)] HA3 lb: There will be no difference, during the early stage of therapy, ’ in the level of comparability between client-therapist and client-other significant person interaction patterns for the successful, as opposed to the unsuccessful client-therapist dyads. Confirmed for [(C+0) - (C+T)] interaction patterns. HA3 2b: There will be differences, during the middle stage of therapy, ' in the level of comparability between client-therapist and client-other significant person interaction patterns for the successfhl, as opposed to the unsuccessful, client-therapist dyads. Not Confirmed for [(C+O) - (C+T)] interaction patterns. HA3 3b: The level of comparability, during the late stage of therapy, between client-therapist and client-significant other person interaction patterns will be lbwer for the successful, as opposed to the unsuccessful, client-therapist dyads. Confirmed for [(C+O) - (C+T)] interaction patterns. CHAPTER V DISCUSSION AND CONCLUSIONS Summary This study was an attempt to evaluate whether the degree of comparability in client-therapist interaction patterns as related to client reports of previous interactions with other important persons provides a significant process dimension upon which to differentiate between successful and unsuccessful psychotherapy. Comparability was defined as the degree of difference between client reports of behavior used with and received from others and the actual behavior which clients used with and received from the therapist. Three fundamental questions were posed regarding client-therapist behaviors which were similar to client reports of interaction with others: (I) Would the comparability level between client reported. interactions for the total range of therapy and actual client-therapist interactions discriminate between outcome groups? (2) Would the comparability level for the combined groups of clients and therapists vary over three stages of therapy? And, (3) Would successful and unsuccessful cases differ in the similarity of their interaction patterns across three stages of psychotherapy? To seek the answers to these questions, twenty cases were selected from among thirty-six of the cases on file in the tape 80 81 library at the Michigan State University Counseling Center. (The tape library contains test data and audio tapes from the counseling and psychotherapy cases of fifty-one clients.) The thirty-six cases were selected on the basis of two criteria: (1) a minimum of nine sessions; and (2) the availability of pre-post MMPI profiles. These cases were divided into successful and unsuccessful groups on the basis of ratings on the MMPI profiles by three judges. Ten successful and ten unsuccess- ful cases (N = 20) were then randomly selected for study. The cases were analyzed in two ways. In the first scoring, client reports of interaction with others (others, parents, others plus parents) were analyzed from the first two sessions of each case. Because the number of client reports was insufficient to establish a minimal standard error, the third session and additional sessions up to the pre-median session were rated for several cases. For the other scoring, the second fifteen- minute segment of each of two sessions from the early, middle and late stages of therapy were selected for analysis of client-therapist interaction patterns. The actual client-therapist interactions as well as client reports of interaction with others were rated by use of the Interper- sonal Circumplex (Leary, I957). The judges were two Ph.D. graduate students. They were trained in the use of the Interpersonal Rating System and demonstrated the ability to use the rating system reliably. In order to test the three basic research hypotheses, compari- sons of client-therapist and client-other behaviors (other and parent) were made in two different ways. The actual behaviors which the client ex- hibited with the therapist were contrasted with the client's reports of 82 behaviors with the client by other important persons. The second con- trast was of the similarity of the behaviors which therapists used with clients as compared to the reported reactions of others to clients. Client Comparabilipy The first question dealt with the comparison of outcome groups on the total comparability level of client-to-other vs. client-to- therapist behaviors. A univariate analysis of variance was used to test the differences in similar interaction patterns of successful and unsuccessful cases. The prediction that over the entire range of therapy, the behaviors which successful, as compared with unsuccessful clients, used with the therapist would be less similar to the client's reports of behavior with others, was tested and rejected. Inspection of the total mean Q scores for both outcome groups indicated that there was a tendency for successful clients to respond to the therapist at lower levels of comparability than unsuccessful clients. But the dif- ferences were too small to reach statistical significance. The level of comparability between client responses to thera- pists and the reported responses of clients to others across the early, middle and late stages of therapy was the subject of the second question. A two-way repeated measures analysis of variance was employed which allowed analysis of the main effect for stages of therapy. It was predicted that all clients, regardless of outcome, would evidence fluctuations over the three stages of therapy in the similarity of their reactions to the therapist as compared with their reactions to others. This prediction was not supported by statistical analysis of the data. 83 The third question was about client responses to the therapist which paralled client reported responses to others across three stages of therapy for each success group. A two-way repeated measures analysis of variance was also used. This test allowed investigation of the interaction of stages with outcome. The prediction that there would be no differences between the parallel behavior of successful and unsuccessful clients at the early stage was accepted. Differences were predicted between successful and unsuccessful clients at the middle stage of therapy. In addition, it was predicted that successful clients, compared with unsuccessful clients, would in- teract with the therapist in ways that were less like their reported behavior with others at the late stage. These predictions were tested and failed to be accepted for the client-to-parent vs. client-to-therapist and the client-to-other plus parent vs. client-to-therapist comparisons. Significant differences were found, however, when both outcome groups were examined on client-to-other (excluding parents) vs. client-to-therapist comparisons. Across the three stages of therapy, successful clients behaved with the therapist in ways which were less similar to the reports of their behavior with others than did unsuccessful clients. Though not significant, differences were found between the outcome groups at the middle stage, with the suc- cessful clients less comparable than the unsuccessful. By the late stage of therapy, it was predicted that successful clients would demonstrate new interpersonal learning by behaving in ways which paralleled less frequently their reported behavior with others 84 than their unsuccessful counterparts. This prediction was accepted for comparisons of client-to-other (excluding parents) vs. client-to- therapist behavior patterns. Therapist Comparability The similarity of the behaviors which therapists used with clients with the behaviors which clients reported others using with them was an additional consideration of the first question. A univariate analysis of variance was used to test the differences in comparability between outcome groups. It was predicted that over the total range of therapy successful therapists would behave less similarly to the reported behaviors of others toward the client than would their unsuccessful colleagues. This prediction was not accepted. Inspection of the total mean 2 scores for each group indicated that for two of the therapist-to-client contrasts, other-to-client and other plus parent-to-client, successful therapists evidenced fewer parallel interaction patterns than unsuccessful. For the therapist- to-client vs. parent-to-client contrast, successful therapists behaved more like the reported behavior of others more frequently than unsuccessful therapists. These tendencies were not, however, statistically significant. In order to test the second question for therapist-to-client interaction patterns, the data were analyzed by repeated measures two-way ANOVAs which tested the main effect for stages of therapy. The degree of similarity between the behavior of therapists in both 85 outcome groups with the clients' reports of others' behavior was tested across the early, middle and late stages of therapy. The prediction that therapist behavior which paralleled the reported behavior of others would fluctuate across the three stages was accepted. As therapy progressed from the early, through the middle, to the late stage, all therapists increased the frequency of their behaviors which were parallel to the reported behaviors of significant others with the client. The third question was tested for therapist-to-client behaviors by means of repeated measures two-way ANOVA's. The contrast of interest here was the interaction of stages of therapy with therapeutic outcome. No differences were predicted between successful and unsuc- cessful cases in the amount of therapist behavior which was similar to the reported behavior of others at the early stage of therapy. Since there were no significant interaction effects, this prediction was accepted. The predictions that there would be differences in parallel therapist behavior patterns between outcome groups at the middle stage, and that successful therapists would evidence fewer responses which were similar to the responses of others than unsuccessful therapists at the late stage, were not accepted. Discussion As defined and operationalized in this study, differences in the level of comparability between the therapist's behavior toward clients and the reported reaction of others to the client cannot be regarded as a process variable which effectively differentiates 86 between successful and unsuccessful cases in psychotherapy. Likewise, the level of client comparability, as defined along two dimensions (client-to-parent; client-to-other plus parent/vs. client-to-therapist), did not effectively discriminate between outcome groups. Significant differences were found for neither therapist comparability nor for client comparability on the two variables cited when the outcome groups were compared across the entire range of therapy and across each of three stages. However, the degree to which the client's reaction to the therapist became more like his reported behavior with others (excluding parents) appears to provide a process variable which effectively dif- ferentiates between successful and unsuccessful psychotherapy cases. Significant differences were found when the outcome groups were compared across the early, middle and late stages of therapy on this dimension. Successful clients were observed to change their behavior with the therapist as therapy progressed and behaved with the therapist in ways which were significantly different from their reported reactions to others (excluding parents) by the late stage of therapy. CIient-to-Other vs. Client-to-Therapist Comparability Level The finding that successful clients behaved in ways which were less comparable to their previous reports of behavior with others is consistent with both Freudian and interpersonal theories of transfer- ence.. According to these theories of psychotherapy, by the end of therapy, transference reactions should be resolved. Successful clients, then, will decrease the usage of old behavior patterns and reSpond to 87 the therapist in a more reality-oriented fashion. This is exactly the behavior change which occurred for the successful clients in this study. Those clients behaving less comparably on this dimension were successful. Those increasing their comparability were unsuc- cessful. If, as Freud (1959c) suggests, the handling of transference is the therapist's most difficult problem, then it would make sense that the degree of client transference may be inversely related to success in psychotherapy (Crowder, 1972).. While this finding is consistent with the theoretical expec- tations, the fact that comparable client-to-parent interaction patterns did not discriminate between the outcome groups is unsettling. Both groups were rather similar in client-to-parent comparability for each stage of therapy. However, the population from which the study sample was drawn may make a difference. This sample consisted of college undergraduates. The primary daily interpersonal interaction of these subjects was most probably with peers and not parents. Since according to Erickson (1965), a college student's personal social growth revolves around mastering his needs for intimacy and identity, it is likely that these subjects had similar social concerns with peer relationships. Thus, decreased client-to-other comparability as therapy progressed would be a logical place to look for changes in the therapeutic relationship which paralleled successful outcome. Client-to-Tpeggpist Behaviors: An E§ploratory_Question Since changes in the client's behavior toward the therapist in terms of the response patterns which paralleled his behavior with 88 others differentiated between successful and unsuccessful cases, it seemed appropriate to investigate which behaviors were changed or remained constant for both groups. In order to accomplish this, the proportions of client behaviors at each stage of therapy for each group were plotted on a graph against the proportion of behaviors which clients reported using with others during the early phase of treatment (See Appendix D, graphs D.1, D.2). An examination of graphs 0.1 and D.2 reveals that successful clients increased in the proportion of self-stimulating and competitive (i.e. boasting, intel- lectualized and accusing, argumentative) behaviors used from early to late stages. Unsuccessful clients decreased their use of these behaviors. Both groups decreased the proportions of informing-dominant (i.e. teaching, informing and dominating, directing) behaviors from early to middle stages, then increased from middle to late stages. Successful clients, however, used higher proportions of these behaviors at the late stage of therapy. In addition, successful clients report the use of more passive-resistant (self-condemning and submitting) behavior in their interactions with others than do unsuccessful clients. Yet successful clients exhibited less passive-resistant behavior with their therapists than unsuccessful clients. This factor plus the increased use of dominant and competitive behaviors by successful clients might well account for the observed changes in comparability. These exploratory findings are consistent with those of Crowder (1972) who found that successful clients were more hostile and competitive than unsuccessful clients during the early stage 89 and that unsuccessful clients were more passive-resistent in the middle stage than were successful clients. As Crowder (1972) suggests, it may be that hostile-competitive clients--those who express their anger clearly--are easier to treat than are passive-resistant clients. This conclusion would appear to be supported by the proportions of be- havior used by both outcome groups in this study. Therapist-to-Client Behaviors: An Exploratopy Question The finding that both groups of therapists reacted to clients with increasing comparability (or increased their countertransference reactions) as therapy progressed was at first puzzling. Most inter- personal and classical theories of psychotherapy and those upon which the hypotheses for this study were based suggest that therapist countertransference reactions are related inversely to therapeutic success. That is, therapist behavior which is comparable to the client's reports of the reactions of parents and others toward him are seen as detrimental since they contribute to the re-enactment of the original conflictual relationships. Therapist countertransference, particularly at the later stages, would likely be detrimental since it would be expected that in successful cases a more reality-oriented, less parallel relationship would have evolved. Yet for the cases studied here, successful therapists, like their unsuccessful colleagues, countertransfered as they progressed. 0n the parent-to-client dimension successful therapists countertransferred more than unsuccessful therapists, though this difference was not statistically significant. 90 In order to investigate this finding further, the proportions of therapist behaviors at the early, middle and late stages were derived for each outcome group. These were plotted against the pr0portions of behavior which clients in each outcome group reported receiving from parents during the early phase of treatment. It was thought that differences in client reports of parental behavior for each group may have accounted for these theoretically puzzling results. That is, if successful clients reported receiving more friendly, nurturent, helpful behavior from their parents than did unsuccessful clients and if successful therapists evidenced large proportions of these same behaviors at the late stage of therapy (e.g. the expected therapeutic role), then successful therapists would evidence greater countertransference behavior on the original comparisons used. Successful clients did indeed report receiving more nurturent, helpful behavior from parents than did unsuccessful clients. These proportions are plotted in graphs 0.3 and 0.4 contained in Appendix D. A check of the other behavior categories revealed that there was relatively little difference in the two groups in terms of client reports of the reaction of others (excluding parents) toward the client. In addition, therapists in both groups used similar proportions of behavior when reacting to the client. The difference in parental countertransference, then, might well be accounted for by the differences in the initial client reports given during the early phase of treatment. 91 Implications None of the major study hypotheses dealing with therapeutic outcome were supported on the therapist comparability variable. Only one of the two outcome hypotheses dealing with client compara- bility received support. These two facts suggest either that therapist comparability and two indices of client comparability are not variables which differentiate between successful or unsuc- cessful psychotherapy, or that limitations exist in the study methodology. The selection of sessions for analysis at three different intervals was based upon the assumption that therapeutically significant interactions would be obtained at what were assumed to be, from a time sampling technique, the three major stages of therapy. A more precise selection of sessions based upon the procedure which Mueller (1969) suggests may have produced more therapeutically significant sessions for analysis. Mueller used two criteria to select sessions: (1) high client anxiety; and (2) changes occurring in the client's perception of his parents. Mueller, by use of these criteria, may have been more successful in sampling client-therapist interactions containing frequent ”critical incidents” which could affect client growth. Thus, the possibility exists that the theory upon which the cases for this study were selected limited inter- pretation. That is, instead of client growth taking place within the gradual progression of the transference relationship, significant client change may result from crucial therapist and/or client behaviors at specific critical points in the process. 92 Both the client and therapist comparability indices were based upon client reports of interactions with others during the early phase of treatment. Since these reports came solely from the client, there was obviously no guarantee that they were valid. Given the nature of both the therapeutic endeavor and the sample subjects, the validity of these client reports was difficult to assess. Researchers concerned with this question might have greater success if the subjects were children and could be observed inter- acting with parents and others in more controlled situations. In addition, the practice of using client reports from only early sessions as the basis upon which to assess client and therapist comparability levels may have introduced further invalidity to the data. Clients reports of interactions with others may change as they become freer to experience increased negative recall. Both groups of clients reported their parents' behavior as fairly laudatory. If these reports changed as therapy progressed, that is, if client reports became ”more valid”, then use of the initial client reports may have obscured differences in therapist and client comparability at the middle and late stages of therapy. A further limitation may result from the procedure used to rate client-therapist interactions. Rating was done only on the second fifteen minutes of a typically fifty-minute session. Though raters were instructed to listen to several minutes of interaction prior to the second 15 minutes, this procedure may have introduced some invalidity. For example, seemingly calm and attentive listening 93 on the part of the therapist to client reports of past adventures may have initially been rated as "L'I or ”cooperative.” Yet, examin- ation of the preceding interaction may have revealed that the therapist had attempted on several occasions to interrupt the client from his reverie. Subsequent therapist responses then would likely be rated as deferring or submissive. These passive-resistant therapist behaviors might as a consequence, go unreported. The ratings of client-therapist behavior used in this study were taken from audio-tapes of the therapeutic interaction. While it was demonstrated that audio-tapes could be rated reliably on the Interpersonal Rating System, audio recordings do not allow for assess- ment of non-verbal behavior. Ratings of non-verbal behavior, while probably more complex, might add a significant qualifying dimension to the audio analysis of interpersonal interactions. A sixth and major limitation may lie with the method used to rate the interpersonal behaviors of the therapy participants. The Interpersonal Rating System may not be sensitive to the interpersonal dimensions which differentiate successful from unsuccessful psycho- therapy cases. Successful and unsuccessful therapeutic relationships may vary more in the quality of behaviors used than in the frequency of behaviors used. This consideration may be particularly signifi- cant for assessment of therapist behaviors. As may be observed from the graphs in Appendix D, there appears to be only minor variances in therapist behavior between successful and unsuccessful cases. Both groups of therapists behaved with moderate proportions of 94 ”cooperative” behavior, relatively higher proportions of ”teaching“ behavior, high proportions of “nurturent” behavior, etc. Yet ten cases were successful and ten unsuccessful. Assuming that therapist behavior does have a negative or positive impact upon clients, it would appear that the interpersonal rating system was not sensitive to the qualitative differences in the nurturent, teaching and cooperative behaviors used by therapists in both outcome groups. Implications for Further Research I. The selection of sessions for analysis in this study may not have resulted in the assessment of sessions containing highly significant therapeutic interactions. One possible improvement in the methodology would be for future: researchers to select sessions upon the basis of (1) high client anxiety level; (2) perceptual changes occurring in the client's perceptions of parents, as Mueller suggests; Egg (3) high therapist anxiety; as well as (4) changes in the therapist's perception of the client's perception of parents. The physiological methodology suggested by Archer, g£_al. (1972), may provide a valuable aid. In addition, if significant changes occur for clients at a crucial point in time, these critical points might be found and rated by consecutive scoring of each interview. For instance, use of the Client Growth Scales (Kagan, et al., 1967) might be helpful in identifying crucial interviews where clients both commit themselves to change and begin to differentiate various human stimuli. 5. 95 If the selection procedure suggested in recommendation number one were used, this procedure would probably result in the selection of whole sessions for analysis. The necessity for considering 15-minute segments as representative of the whole session as well as the possible invalidity in ratings introduced by rating only part of an extended interaction would thus disappear. Because the analysis of client and therapist comparability was derived from client reports from the early phase of treatment, some invalidity may have been introduced. A study comparing client and therapist behaviors with the client's reports of his interactions with others at the 2292 phase of treatment may produce more significant results. An investigation of whether client reports of interactions with others do, in fact, change as therapy progresses would produce a major addition to the theory and research on psychotherapy. Rating client-therapist behaviors from video recordings, as opposed to audio tapes, would introduce increased validity to the assessment of interpersonal behaviors on the Interpersonal Rating System. Though such rating would of necessity be complex, a method of rating could conceivably be designed which would take into account discrepancies between client and therapist verbal and non-verbal behavior. It is probable that the Interpersonal Rating System, while sensitive to quantitative differences in behavior, is not sensitive to differences in the quality of behavior. A further 96 study might be designed which takes into account the quantitative differences, by use of the Interpersonal Rating System, and adds ratings on qualitative differences. These qualitative differences for the therapist might, for example, beassessed by the ”Empathetic Understanding Scale“ developed by Carkhuff. The quality of client interpersonal response might be assessed on the ”Owning of Feelings in Interpersonal Processes” scale developed by Schauble and Pierce. Client comparability level for client-to-other (excluding parents) comparisons appears to be a variable which effectively discriminates between outcome groups. It would be an interesting and valuable addition to investigate whether client behavior with others outside of the therapeutic relationship also changes. That is, if changes in the behavior of successful clients with others parallel changes in client-to-other comparability, this variable would indeed provide an effective discriminator between successful and unsuccessful therapy cases. The procedure suggested by Cabush (1971) and by Archer (1971) of investigating peer reports of client behavior might provide a viable methodology for such an undertaking. The cases in this study were defined as successful or unsuccessful on the basis of their average rating by three judges on a five point change scale based upon pre-post MMPI profiles. Those cases at or below the mathematical mid—point (3.00) were assigned to the unsuccessful group. Those cases whose average ratings were above this midpoint were assigned to the successful group. Because 97 several cases clustered around the mid-point of the five point scale, it is possible that the procedure used to assign cases into outcome groups served to mask differences between successful and unsuccessful cases. A method of selecting only those cases from the upper and lower quartiles of the distribution would distinguish more sharply between outcome groups. Hence, any differences which may exist in the level of comparability between successful and unsuccessful cases might become evident. BIBLIOGRAPHY 98 BIBLIOGRAPHY Alexander, F. ”Psychoanalytic contributions to short term psychotherapy,“ in Wolberg, ed., Short-term Psychotherapy. New York: Grune and Stratton, 1965. Alexander, F., and French, T. Psyphoanalytic Therapy. New York: Ronald Press, 1946. Archer, J.. Jr. “Undergraduate as Paraprofessional Leaders of Interpersonal Communication Skills Training Groups Using an Integrated IPR Videotape Feedback/Affect Simulation Training Model.” Unpublished Ph.D. dissertation, East Lansing: Michigan State University, 1971. Archer, J.; Feister, T.; Kagan, N.; Rake, L.; Spierling, T.; and VanNoord, R. I'A New Methodology for Education, Treatment and Research in Human Interaction.‘' Journal of Counseling Psychology, 19: (I972). Brody, Morris W. ”Transference and Countertransference in Psychotherapy.” Psychoanalypjc Review, 42: 88-94 (1955). Cabush, D. W. ”Outcome Effects of Training College Student Clients in Facili- tative Self-Responding.“ Unpublished Ph.D. dissertation, East Lansing: Michigan State University, 1971. Carkhuff. R. R., and Berenson, B. G. Beyond Counseling and Psychotherapy. New York: Holt, Rinehart and Winston, 1967. Carson, R. C. Interaction Concepts of Personalipy. Chicago: Aldine Publishing Company, 1969. Cronbach, L. J., and Gleser, G. ”Assessing similarity between Profiles.‘I ngchological Bulletin, 50: 456-473 (1953). Crowder, J. E. ”Transference, Transference Dissipation and Identification in Successful vs. Unsuccessful Psychotheraphy.” Unpublished Ph.D. dissertation, East Lansing: Michigan State University,197o, 99 100 Crowder, J. E. ”Relationship Between Therapist and Client Interpersonal Behaviors and Psychotherapy Outcome.” Journal of Counselipg Psychology, 19: 68-75 (1972). Dietzel, C. S. “Client-Therapist Complementarity and Therapeutic Outcome.” Unpublished Ph.D. dissertation, East Lansing: Michigan State University, 1971. Dittman, A. T. ”Problems of Reliability in Observing and Coding Social Inter- actions.” Journal of Counseling Psychology, 22: 430 (1958). Ebel, R. L. “Estimation of the Reliability of Ratings.” Ppychometrika, 16: 407-424 (1951); also in, W. A. Mehrens, and R. L. Ebel, eds., Principles of Educational and Psychological Measurement. Chicago: Rand McNally and Company, 1967. Erickson, E. H. The Challepge of Youth. New York: Anchor. Doubleday, 1965. Fenichel, 0. Problems of Psychoanalytic Techniqp_. Albany, N. Y.: Psycho- analytic Quarterly, Inc., 1941, pp. 71-75. Freedman, M. B.; Leary, T. F.; Ossorio, A. G.; and Coffey, H. S. ”The Interpersonal Dimension of Personality.” Journal of Personality, 20: 143-161 (1951). Freud, S. An Autobiographical Study. London: Hogarth Press, 1948, pp. 75-77. Freud, S. ”The Future Prospects of Psychoanalytic Therapy.“ Collected Papers. Vol. 2. New York: Basic Books, 1959 (b). Freud, 5. ”Observations on Transference Love.“ Collected ngers. Vol. 2. New York: Basic Books, 1959 (b). Freud, S. ”Fragnent on an Analysis of a Case of Hysteria.” Collected Papers, Vol. 3. New York: Basic Books, 1959 (c). Geissner, S. and Greenhouse, S. W. ”An Extension of Box's Results on the Use of the F Distribution in Multivariate Analysis.“ Annals of Mathematical Statistics, 29: 885-981 (1958). 101 Haley, J. Strategies of Psychotherapy. New York: Grune and Stratton, I96—. Halpern, H. M. ”An Essential Ingredient in Successful Psychotherapy.“ Psyghothergpy, 2: 177-180 (1955). Harter, L. H., Order Statistics and Their Uses in Testing and Estimation. Vol. 1. Washington, D. C., U. S. Gov't Printing Office, Aerospace Research Laboratory, Office of Aerospace Research USAF, 1969. Hays, W. L. Statistics for Psychologists. New York: Holt, Rinehart and Winston, 1963. Heiman, P. ”On Counter-Transference.“ International Journal of Psycho- é“ analysis, 31: 81-84 (1950). Heller, K.; Myers, R. A.; and Kline, L. V. “Interviewer Behavior as a Function of Standardized Client Role.” Journal of Consulting Psychology, 27: 117-122 (1963). Horney, K. ' New Ways in Psychoanalysis. New York: W. W. Norton, 1939. Kagan, N., and Krathwohl, D. R. Studies in Human Interaction Final Report Grant No OE 7 3 o o -2- OHO-270. Educational Publication Services,,1967. Kell, B. L., and Mueller, W. J. impact and Change: A Study of Counseling Relationships. New York: Appleton-Century-Crafts, 1966. Kepecs, J. G. IITheories of Transference Neurosis.” PSXEhoana'Yt'°.QEarter'Y’ (T. 35: 497-521 (1966). Kiesler, D. J. ”Some Myths of Psychotherapy Research and the Search for a Paradim.“ Psychological Bulletin, 63: 110-136 (1966). Leary, T. Integpersonal Diggnosis of Personalipy. New York: Ronald Press, 1957. MacAlpine, I. “The Development of Transference.“ P§ychoanlytic Quarterly. 19: 501-519 (1950)- 102 MacKenzie, M. H. ”The Interpersonal Behavior of Normal and Clinic Family Members.” Unpublished Ph.D. dissertation, East Lansing: Michigan State University, 1968. Malone, T. P. ”Encountering and Groups,” in A. Burton, ed., Encounter. San Francisco: Jossey-Bass Inc., (1970). Meltzoff, J. and Kornreich, M. Research in Psychotherapy. New York: Atherton Press, 1970. Mueller, W. J. “Mapping the Psychotherapeutic Process through Analysis of Interpersonal Behavior.” Symposium presented at the meeting of the American Psychological Association, Washington, D. C., September, 1969 (a). Mueller, W. J. _ “Patterns of Behavior and Their Reciprocal Impact in the Family and in Psychotherapy.“ Journal of Counseling Psychology, 16: (2. Pt. 2) (1969) (b). Mueller, W. J. and Dilling, C. A. ”Studying Interpersonal Themes in Psychotherapy Research.” Journal of Counseling Psycholggy, 16: 50-58 (1969). Mullahay, P. Psychoanalypis and Interpersonal Psychology. New York: Science House, 19700 Orr, D. ”Transference and Countertransference: A Historical Survey.“ Journal of the American Psychoanalytic Association, 2: 621-670 T1954). Raush, H. L.; Dittman, A. T.; and Taylor, T. J. “The Interpersonal Behavior of Children in Residential Treatment.” Journal of Abnormal and Social Psychology, 58: 9-26 (1959). Reich, Annie ”Further Remarks on Countertransference.” The International Journal of Psychoanalysis, XLI: pts. IV-V, 299-395 (1960). Rioch, J. M. ”The Transference Phenomenon in Psychoanalytic Therapy.“ Psychiatry, 6: 147-156 (1943)- Schilder, P. Introduction to a Psychoanalytic Psychiatry. New York: Nervous and Mental Disease Publishing Company, 1928. 103 Sullivan, H. S. “Psychiatry: Introduction to the Study of Interpersonal Rela- tions.” Psychiatpy, 2: 121-135 (1938). Sullivan, H. S. The Inteppersonal Theory of Psychiatry, in Helen S. Perry and Mary L. Garvel, eds., New York: W. W. Norton and Company, 1953. Sullivan, H. S. The Psychiatric Interview. New York: W. W. Norton and Company, 1954. Swensen, C. H. I'Psychotherapy as a Special Case of Dyadic Interaction: Some Suggestions for Theory and Research.“ Psychotherapy: Theory, Research and Practice, 4: 7-13 (1967). ‘— Weiss, E. “Manipulation of the Transference Relationship,” in F. Alexander and T. French, eds., P§ychoanajytic Therapy. New York: Ronald Press, 41-54 (1946). Wolstein, B. Countertransference. New York: Grune and Stratton, I959. Wolstein, B. Transference: Its Structure and Function in Psychoanalytic Therapy. New York: Grune and Stratton, 1954. APPENDICES 104 APPENDIX A STUDY CASES 105 APPENDIX A STUDY CASES Client Therapist Client Number Therapist MMPI Code No. , Sex Sex of Experience Outcome Sessions Levela Rating 011 M F 18 I1 5 016 M F 22 I2 U 017 F F 16 11 s 024 F M 12 S S 026 F F 18 I2 5 031 M M 19 11 u 040 F F 17 I1 u 042 M F 18 S U 043 M M 16 S U 046 M M 9 P U 047 M F 12 P S 801 M F 24 S S 812 M F 12 S S 817 F F 20 S U 818 M F 16 11 S 823 M F 21 12 U 830 F F 9 I1 U 831 M M 24 S U 845 F F 15 I1 S 849 M F 16 I S a. II = first-year intern; 12 = second-year intern; S = senior staff; P = practicum student b. S = successful; U = unsuccessful 106 APPENDIX B MMPI RATINGS 107 APPENDIX B MMPI RATINGS Client Judge 1 Judge 2 Judge 3 Average Code No. Ratings Ratings Ratings Ratings of Judges lst 2nd lst 2nd Ist 2nd 011 5 5 5 5 5 5 5.00 016 4 3 2 2 2 2 2.50 017 4 4 3 4 2 2 3.17 024 5 5 5 5 5 5 5.00 026 5 5 5 5 5 5 5.00 031 4 4 l 2 3 3 2.83 040 3 2 3 3 3 3 2.83 042 3 3 4 I 2 2 2.50 043 2 3 4 2 3 3 2.83 046 5 4 2 2 2 2 2.83 047 5 5 5 5 5 5 5.00 801 5 5 2 2 4 4 3.67 812 4 5 3 3 3 3 3.50 817 l 2 3 3 2 2 2.17 818 5 5 5 5 5 5 5.00 823 2 2 I l 2 2 1.67 830 3 3 3 2 4 3 3.00 831 2 I 1 l 2 3 1.67 845 5 5 5 5 5 5 5.00 849 5 4 5 4 4 4 4.23 Satisfactory; 4 = Partly Satisfactory; 3 = No Change; NU‘I II Partly Unsatisfactory; 1 = Unsatisfactory. 108 APPENDIX C SCORING MANUAL FOR THE INTERPERSONAL BEHAVIOR RATING SYSTEM 109 APPENDIX C INTRODUCTION This scoring manual was deveIOped by James E. Crowder (1970). It was used in training and relating the psychotherapy sessions sampled in the present study. The scoring procedure contained herein was followed with but one exception. This exception was that the raters were instructed to rate only client recollections of actual interactions with others and to refrain from rating fantasized and projected material. 110 111 Scoring Manual for the Interpersonal Behavior Rating System1 General Considerations The interpersonal circumplex consists of 16 reflexes (cate- gories) of interpersonal behavior, into which all interpersonal behaviors may be rated. It is divided into quadrants by orthogonal axes. The vertical axis covers the dimension of dominance-submission, while the horizontal axis represents the affiliative-disaffiliative (or love-hate) dimension. In rating behaviors into categories, the behaviors are first judged in terms of the axes, and thus the behaviors are placed into quadrants of the circumplex. Then, a behavior is judged into a specific category within the quadrant by matching it with the descrip- tive terms of those categories. .Statements sometimes include behaviors of more than one category, in which case multiple scorings should be used. Problems arise because (1) the categories are not mutually exclusive, (2) the meaning of behaviors are determined party by the context in which they occur, (3) affect and content (i.e., words) are sometimes incongruent, and (4) raters may use different levels of interpretation. These problems are demonstrated below by the use of a few examples. Consider the client statement: ”I like you.” If this state- ment were genuine, it would be rated ”M”. If it were said I. Freedman, M. B., Leary, T. F., Ossorio, A. G., and Coffey, H. S. The interpersonal dimensions of personality. Journal of Personality, I951, 20, I43-I62. 112 sarcastically, it would be rated I'D”. If it came after an interpre- tation which the client did not want to deal with, it would be rated "F". For another example, consider the following client statement: ”You look tired today.“ If this statement connoted genuine sympathy, it wOuld be rated ”N”. If it came out of the client's guilt for seeking help from the therapist, it is possible to argue that it should be rated ”H”, but this rating would require deeper interpretation than the sympathetic “N”. The client statement, “I don't trust you,“ implies distrust ”G” and rejection “C”. It is necessary to choose one or the other in this rating system. In rating the client and therapist behaviors, the following priorities are listed so that the above problems will be minimized: (l) Context takes precedence over affect, (2) affect takes precedence over content, and (3) interpretation does not go beyond the immediate context. Three types of reported client-to-other behavior is scored. These are (I) client's reports of actual interaction with others, (2) client's fantasized interaction with others, (includes wishes, desires, should-haves, and fears), and (3) client's feelings about others as reflected in his statements about them. The following examples illustrate these categories: (1) C: ”My parents told me that I shouldn't get serious about any girls while I'm here. I told them to stay out of my affairs.“ (2) (3) 113 C: ”I wish I had some close friends.” C: ”I'm afraid that people will reject me.“ C: ”I should have told her off.“ C: ”I distrust my parents.II C: llThey are selfish people,II Below, examples of behavior for each category are listed, and, where deemed helpful, explanatory statements are included. It is (impossible to provide examples for some of the meanings of some reflexes, because the meanings are sometimes very dependent on the tone of voice, e.g., sarcastic behavior (reflex llD”). Exemples of Behavior for each Category Reflex “B” (Boastipg, Self-Stimulating, Narcissistic, Intellectualizing Behavior) ‘IHergpist and client ”B”. I. Therapist or client is boastful. Examples: C: ”I made the highest score on the final examination.” C: ”Looks like I really helped you.” Wandering, free-associating, conversation in which the speaker provides his own stimulation. This cate- gory usually applies more to the client statements in which a ”list” of activities since the previous session is covered without emotion, and without a previous therapist eliciting question. This is generally along, rambling statement, which may have been started by a therapist question, but which continued with the client providing his own stimulation. In this case, the 114 client's statement would be rated in two parts, the answer to the therapist's question would be rated an ”L“, and the rest of the client's statement a “B”. 3. Therapist or client intellectualizes. Therapist example: C: ”I feel really affectionate toward you.” T: ”That's because you once had that feeling toward your father.” Client example: T: ”What is it that's troubling you?” C: ”I haven't worked out my Oedipus complex.“ Client-to-other ”B”. 1. Client reports boasting to others. C: ”I told him how wonderful I am.” 2. Client reports having been narcissistic with others. C: ”I took advantage of her.” Reflex “C” (Rejecting, Withholding, Cpmpeting, Accusing) Therapist and Client I'C”. 1. Client or therapist rejects previous statement (regardless of whether previous statement was true). Examples: C: ”No, that isn't right. What bothers me is that no one seems to really care for me.“ Inthis example, the ”No, that isn't right” would be rated ”C”. The second part would be rated “P“ if no strong emotions were attached to it. Of course, if the client expressed 2. 3. 115 feelings of hurt or sadness, the second part may be rated “K”. A “no” statement following a therapist question with no point of view attached (i.e., where therapist does not make a positive statement that is subsequently rejected) should be rated ”L“ instead of “C”. Client and therapist are arguing, competing, usually with an undercurrent of hostility. Examples: T: “You can find people like that in New York.” C: ”I've looked and there are no peOple like that here.'I T: I'You haven't looked in the right places. You've met only a few people here.” C: I'I know I can't find people like that here. I need to go somewhere else.” The first therapist statement in this interchange may not be rated a “C”, depending on the previous client statement that elicited it. For instance, if the previous client statement had been ”I need to find some people that I could trust,“ the first therapist statement above might be rated ”P”. Client or therapist refused a previous suggestion, direc- tive, etc. T: “I will not see you twice a week.‘l C: ”No matter what you say, I won't stay here.” 116 Client-to-other HES. 1. Client reports rejection of others. C: “I don't like him.“ 2. Client reports competing with others. C: ”I tried to beat him at his own game.” Reflex ”DU_(Sarcastic, Threatening, Punishing Behavior) Therapist and Client ”D”. T: ”If you don't get out of that relationship, I'll stop seeing you.” C: ”PeOple are going to keep bugging me until I kill myself.” Client-to-other ”D“. C: ”I told him that if he continued to harass me that I wouldn't see him anymore.“ ‘Eeflex ”E“ (Hate, AttaoE,Disaffiliate). Therapist and Client ”E”. T: “Get out of my office.ll C: ”Go to hell.‘' T: ”Your're an idiot.” Client-to-other HES. C: ”She's nothing but a whore.“ C: “I broke up with him.“ C: ”I hate my mother.” Reflex ”F” (gomplain, Rebel, Nag, Sulk, Passively Resist) Therapist and Client ”F”. 1. Client passively resists therapist's interpretation put 117 in the form of statement or question. Examples: a.T: b.T: c.T: ”Sounds like you get anxious around competent females.ll ”I don't know.” “Is it that your boyfriend reminds you of your father in some ways?” ”I don't know. (pause) One thing that really disturbs me is that I can't concentrate when I study.” “Do I hear some resentment in there?” I'I don't know. (pause) You may be right. Yeah, I wasn't aware of it but I really do resent him for that.” Note: In example a, the client's l'I don't know” is rated “F”, because it indicates passive resis- tance to the therapist's statement. In these cases, the client is demonstrating an unwil- lingness to even consider the validity of the statement, but at the same time is not flatly rejecting it either. In example b, the ''I don't know“ is followed by the change of subject. In this case, it is rather obvious that the change of subject is a defensive maneuver, seemingly unrelated to the therapist's question. The “I don't know“ should be scored ”F“, and the change of subject should be scored “A”. In 2. 118 example c, the “I don't know” was intended to indicate thoughtfulness, an attempt to deal with the therapist's question, which is validated by the rest of the client's statement. In this example, the ”I don't know“ is not scored, but the remainder of the statement should be enclosed in parentheses and scored ”L”. Sometimes the therapist or client angrily withdraws (sulks), with some such comment as ”I don't know”. These should be scored as ”F”. Client-to-other “F” C: C: ”I resented his saying that, but I didn't say anything.” ”When Dad yelled at me, I went to my room and didn't come out for hours.” Reflex “G” (Distrust, Suspect, Be Skeptical) Therapist and Client ”G”. I. Therapist or client expresses skepticism at the previous statement of the other party. Examples: ”What?” ”What do you mean?” ”Maybe.” The first two examples would be scored ”G” when the previous statement and its meaning was perfectly clear. The “maybe” expresses incomplete acceptance, or, better, neither rejection nor acceptance, but does express skepticism. 119 2. Therapist or client is suspicious of feelings, motives, etc. expressed by the other party. Examples: C: ”I don't think you really like me.“ T: ”Are you sure you're dealing with the thing that's really bugging you?” Note: If the statement is an unconditional rejection or accusation (e.g., “You don't like mel”), it should be rated ”C”, not ”G”. Client-to-other ”G” C: ”I didn't believe her.” C: ”Sometimes, it seems like no one can be trusted.“ Reflex ”H“ (Condemn SelfHHWithdraH) Therapist and client ”H”. C: “I feel worthless.” T: ”You wouldn't feel that way if I were a good therapist.’I Client-to-other ”H”. C: ”I guess I should have confronted him, but I didn't know what to say, so I left.“ Reflex ”I” (Submit, Defer; Obey) Therapist and Client ”I” 1. Client or therapist submits more to avoid confrontation than to accept a statement because of its validity. This sometimes occurs after an argument, or to end an argument. 2. Client expresses extreme helplessness, inability to c0pe, without underlying belief that change is possible, that therapist will help. 120 3. ”I guess so,” and ”yeah” responses, which are total re- sponses, when the therapist is actually trying to elicit elaboration on something, or after therapist has made a statement about something. Client-to-other ”I”. C: ”I didn't want to go to college, but Mom insisted.ll C: ”They take advantage of me.” Reflex “J” (Ask Opiniopi Praise, Admire) Therapist and Client ”J”. C: “What should I do?” C: “You're the best therapist in the Counseling Center.” Client-to-other ”J”. C: “I asked her what she would do if she were me.” C: ”They're all so great--intelligent and sensitive.” Reflex ”K” (Ask for HelpLHDepend,HTrust) Therapist and Client “K”. C: ”This problem arose which I hope you will help me with . . .“ Client-to-other ”K”. C: ”I trust her.“ C: III depend on them.“ C: ”I asked him to help me repair the car.” Reflex ”L” (Cooperate, Confide,_§ollaborate, Agree). Therapjst and Client ”L”. 1. Client cooperates with therapist, works on problems, answers questions, elaborates on reflective or inter- pretive statements. Examples: 121 T: ”How old is your sister?” C: I'She's 18.” T: “It sounds like you have difficulty in accepting positive feelings.“ C: ”Yeah, I think you're right. The other day my roommate said she liked me, and . . .” Note: a. Sometimes its difficult to discriminate between elaboration and self-stimulating conversation. In general, self-stimulating conversation is much longer, and less affect-laden. Also, the focus of self-stimulating conversation shifts frequently. b. When the client's agreement comes after an argu- ment, is less sincere, and without elaboration to support it, ”I“ instead of “L” should be scored. 2. Client's ”Yeah” statements which merely lubricate comments coming from the therapist. Examples:' T: ”You remember last week when we were talking about sex,“ C: “Yeah.” T: ”You got very angry with me” C: "Yeah'I T: ”Well, I was wondering why that made you mad.“ Client-to-other ”L”. C: ”I went over and started a conversation with her.“ 122 C: ”We told each other our problems.“ Reflex ”M” (Affiliate, Identify With, Love) Therepist and Client “M”. T: “I really like you.‘I C: ”I feel close to you today.‘I Client-to-other ”M". C: “I dated him for two years.” C: ''I care a lot about my Dad.“ C: ”We seem to have the same feelings about everything.” Reflex “N” (Supporty_Sympathize, Reflect Feelingpy Reassure1 Generalize Conscious Feelings, Approve, Nurture, Therapeutic Probe) Therapist and Client ”N”. C: ”I'm sure you're intelligent, and capable of making it here.“ (Support, reassure) T: ”Sounds like you're very lonely, and feeling incapable of establishing any real friendships.‘| (Reflect feelings) T: ”You said that your father really preferred your brother?” (Therapeutic probe) C: ”Looks like you're very tired today.” (Sympathize) ('1 ”Well, I think you're doing a very good job.” (Support) Note: a. The above therapist statements are rated ”M” only if he is responding to data and feelings in the previous client statements. For instance, if the third therapist statement above had come after a client had said ”I had final exams yesterday,“ the therapist statement would be rated ”A” (Directive). 123 As a rule of thumb, reflecting feelings, therapeu- tic probes, generalize feelings, when rated “N” must come after a client statement which contained that data that is reflected, generalized, etc. Of course, support and reassurance, to be rated, does not suffer this limitation. The client statement above is rated ”N” if it seems genuinely sympathetic; the fact that it may be prompted by guilt over receiving help is irrelevant to the rating system. b. Reassurance occasionally turns into an argumenta- tive, competitive exchange, in which the first therapist statement should be rated ”N”, but the following ones should be rated ”C”: Example: T: ”I know you can handle it.” (Supportive) C: I'I know I can't!” (Angry) T: ”No, you don't £925 to, but I know you can!” Client-to-other ”N”. C: ”I told her that everything would turn out alright.“ T: “I can underStand.her feelings about that.” Reflex ”0“ (Give Help, Interppet Beyond Conscious Feelings) Therapist and Client “0”. T: “If you feel up tight next week, we could meet twice.“ T: “Your relationship with your boyfriend appears to be similar to the one you had with your father.“ Client-to-other ”O”. C: ”Mom had her hands full, so I helped her with the dishes.” C: 124 “I wish I could help him feel better about himself.” Reflex ”P” (Advise, Teachi Give Opinion, Inform) Therapist and Client ”P”. l. Therapist or client gives opinion, acts as authority on the state of things in the world. Examples: T: ”The way I see myself as being helpful to you is in trying to understand you, and in the process, helping you to understand yourself.“ ”To get some information about your interests, you should take the Strong.” ”You may have that feeling, but not be aware of it. It may be unconscious.II ”Innnlexperience, I've found that people in this society are like that.” “To make money farming, you have to do most of the work yourself. If you hire people to work for you, your expenses will be greater than your income.” Note: a. “P” is often scored after "C” in the same statement (example: ”No, I don't really feel that way. The way I feel is . . .”). Of course, if rejection is not followed by expla- nation, “P” would not be scored. If the whole statement is a rejection of the previously stated point of view, with an argument as to why the speaker's point of view is correct, or just an assertion that he is right, the whole 125 thing should be scored I'C”. ”C“. . .“A” or “C” . . .”8” might also be scored (i.e., rejection might be followed by a change of subject or self-stimulating conversation). b. Sometimes, statements of the way things are in the world is made to reassure, and should there- fore be scored “N” instead of ”P”. Example: C: “I really feel like I'm coming apart!” T: I'When people begin to change, they often feel like they're disintegrating. That seems to be what's happening to you.“ Client-to-other ”P“. C: “I taught him how to water ski.’I C: ”When he asked for my advice, I told him what I would do.'I Reflex I'A" (Dominate, Direcpl_Command, Diagnostic Probe, Independent Behavior) Therapist and Client “A”. 1. Therapist or client changes subject, begins new topic. Note: Occasionally, a change of subject should not be rated ”A”. Example: C: “Yes, I do have finals next week. (pause) I hate you.” In this example, strong emotion is expressed in the change of subject. In this case, the rating would be “L”. . .“E”. 2. Therapist asks questions of an information-gathering kind. 126 Example: T: ”How old are you?” 3. Therapist or client is dominating, bossy. Example: T: ”Do your studying between three and six o'clock.” (When no advice was asked for.) Client-to-other ”A”. C: I'I said, 'Judy, quit school and go to work.'“ C: I'l decided to leave my parents, because I felt like it was time for me to stop depending on them so much.” APPENDIX D PROPORTIONS OF BEHAVIOR IN THE OCTANTS OF THE CIRCUMPLEX USED BY BOTH OUTCOME GROUPS IN THE THREE STAGES OF THERAPY PLOTTED AGAINST THE PROPORTIONS 0F BEHAVIOR WHICH CLIENTS REPORTED USING WITH OTHERS AND RECEIVING FROM PARENTS 127 128 58 I .. (0+0) as 54 1 (C+T) Early—— —A———- 50 «F (C->T) Middle—-—U——--— _ (C+T) Late 8 I, 46 . 42 2 F 38 T 34 «i 4 3O .. 26 4F 22 ‘F 18 .. 14 j: 10 ‘ 08 <1- 06 u 04 ‘- 02 OO . . . - . . , BC DE FG HI JK LM N0 PA Graph D.l.--Proportion of behaviors in each octant of the circumplex which successful clients reported using with others (ex- cluding parents) and which successfhl clients used with the therapist in the early, middle and late stages of therapy [(C+O) - (C+T)]. .M. Key: BC = Self-stimulating-Competitive JK = Admire-Depend DE = Punish-Hate LM = Cooperate-Love FG = Complain-Distruct N0 = Support-Help MI = Withdraw-Submit PA = Teach-Dominate 129 I 54 -~ 50 L (C+O) o . -. (C+T) Early ———A———-— 7 46 .- (c-r) Middle—-—G—-——— 4" ' (C+T) Late IE 42 -- 1‘ 38 A ./ ~11— , \ 34 "' I o \ 30 -— \ 26 : \l .. \\ 22 "F I \ .. " 0‘ 1 18 ~~ ' \ .. I 10 a \ // ~1- //'." 08 i \ // HF 06 ,_ y / . -. I \ // a / 04 a» . ' / 7- .1). / ./ 02 or \\ I -I- “\lzb 00 t 4H 5 1 F 9L 3 4— BC DE FG HI JK LM NO PA Graph D.2.--Proportion of behaviors in each octant of the circumplex which unsuccessful clients reported using with others (ex- cluding parents) and which unsuccessful clients used with the therapist in the early, middle and late stages of therapy [(C+O) - (C+T)]. E,M,L Key: BC = Self-stimulating-Competitive JK = Admire-Depend DE = Punish-Hate LM = Cooperate-Love FG = Complain-Distrust N0 = Support-Help HI = Withdraw-Submit PA = Teach-Dominate 58 56 54 50 46 42 38 34 30 26 22 18 14 10 08 04 02 00 130 :: a... p A i : (T+C) Early — — -A——— l .. (1+0) Middle—--—CI—-— / \ .. (1+0) Late 8 I \ «- I :: /. ‘\ I I " \\ ‘_ I \\ \ F P/‘ _ 1 I: : p I j: «I I. P P\P I I . \p I j; 0\ ’ A ‘* ., ' A / .. . \ 4‘1, A.“ .5; y W: % .3 , BC DE FG HI JK LM NO PA Graph D.3.--Proportion of behaviors in each octant of the circumplex which successful clients reported receiving from parents and which successful therapists used with clients in the early, middle and late stages of therapy [P+C) - T;C)]. E, ,L Key: BC = Self-stimulating-Competitive JK = Admire-Depend DE = Punish-Hate LM = Cooperate-Love FG = Complain-Distrust N0 = Support-Help HI = Withdraw-Submit PA = Teach-Dominate l3l 58 ;_ “ (P+C) p 56 v (T+C) Early-—-—-—¢3_.____ 54 1: (T+C) Middle—~—c1_-_ (PC) Late 8 50 46 I: 42 :: 38 t w- 34 1- 3o 4- 26 -. 22 4. l8 4. 14 3- 10 ~- 08 .. 06 3. 04 :- 02 I 00 Graph D.4.--Proportion of behavior in each octant of the circumplex which unsuccessful clients reported receiving from parents and which unsuccessful therapists used with clients in the early, middle and late stages of therapy [(P+C) - (T+C)]. Key: BC = Self-stimulating-Competitive JK = Admire-Depend E’M’L DE = PunishtHate LM = Cooperate-Love FG = Complain-Distrust N0 = Support-Help . HI = Withdraw-Submit PA = Teach-Dominate "ll?lllllll7llllllllllls