MSU LIBRARIES .—.‘._- ’ NW 3 1293 10730 RETURNING MATERIALS: PIace in book drop to remove this checkout from your record. FINES wil] be charged if book is returned after the date stamped beIow. SE? 1 8 TI :3, ' MAGIC 2 AP§?0123ma r" E 06 WI \\\ I \ "N 0794 THERAPEUTIC ALLIANCE AND NEGATIVE INDICATORS IN OUTPATIENT PSYCHOTHERAPY BY Timothy Trell Eaton A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1987 ABSTRACT THERAPEUTIC ALLIANCE AND NEGATIVE INDICATORS IN OUTPATIENT PSYCHOTHERAPY BY Timothy Trell Eaton Therapeutic alliance has long been discussed in psychodynamic theory as a vital factor in producing a beneficial psychotherapeutic relationship, and has equally been recognized as an important variable by other therapeutic schools of thought. Given the paucity of research attempting to measure and analyze therapeutic alliance, this study addresses a variety of broad questions concerning the development of therapeutic alliance in outpatient psychotherapy; the relationship between therapeutic alliance and negative indicator variables; the relationship between therapeutic alliance and negative indicators with treatment outcome; and the impact of pretreatment patient symptomatology on therapeutic alliance and negative indicators. Forty cases from an outpatient community psychological clinic serve as the data base for all analyses. Cases are classified into short, medium, and long length treatment groups, and ratings are randomly obtained from the beginning, middle, and final phases of psychotherapy. The main findings are: 1) Strength of therapeutic alliance and negative indicators is established within the first three sessions of psychotherapy with little Timothy Trell Baton significant change over the remaining course of the therapeutic relationship. This result supports similar findings from previous research. 2) Negative indicators have a clear impact on therapeutic alliance, but their impact does not occur until later in the therapeutic relationship. There seems to be a beginning "grace“ period in which negative process variables do not impact on alliance formation as much as occurs later in therapy. Therapist positive contributions to the alliance also seem to be less affected by negative indicators. 3) A mild relationship between positive therapeutic alliance factors and treatment outcome is observed, with therapist positive contributions especially predicting the decline in patient symptoms of anxiety. 4) Patient contributions to therapeutic alliance are affected by pretreatment symptomatology. Patient positive contributions are lower given higher levels of pathology, while patient negative contributions increase. TO My Family ii ACKNOWLEDGMENTS My deepest appreciation is extended to all those who have made this work possible. My dissertation committee, Dr. Norman Abeles, Dr. Bertram Karon, Dr. Gershen Kaufman, and Dr. Ralph Levine, have continued to display concern for the improvement of the process of psychotherapy and interest in supporting research to meet their concern. I especially thank Dr. Abeles for his personal support and friendship throughout my graduate education. He has truly been a driving force behind my accomplishments. Special thanks are extended to my raters, Gary Gunther, Ann Isenberg, Michelle Klee, and David Rubin, for their precious time, effort, and friendship. We worked together in mutual support through all of our frustrations and “dissertation blues,” as well as through our pleasant surprises and happiness. This work could not have been done without them. Most importantly, I wish to thank my wife, Kay. She continued to offer her love and understanding when I gave her little reason to do so. As I have stated before, my personal accomplishments truly seem unimportant compared to the life we have together. iii TABLE OF CONTENTS Page LIST OF TABLESOOOOOOO00......OIOOOOOIOOOOOOOOOOOOOO. Vi. Chapter I. LITERATURE REVIEWOOOIOO0.00.000000000000000 1 Therapeutic Alliance: The Psychoanalytic Tradition............... 1 Therapeutic Alliance: Defining the conceptOOOOOOOOOOOOIOOOOOOIOOOOOOOO 8 Therapeutic Alliance and Psychotherapy Outcome: Introduction................. 13 Therapist and Patient Contributions to Successful Therapy.................. 14 Process Variables and the Importance of the Therapeutic Relationship: conCIuSionOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 2g The Contribution of Therapeutic Alliance to Successful Therapy......... 22 Therapeutic Alliance and Outcome: conCIUSionOOOIOOOOOOOOOOOOOOOOOOOOOOOO. 32 Therapeutic Alliance Formation and Pretreatment Variables................. 33 Negative Indicators in Psychotherapy..... 41 II. HYPOTHESESOOOOIO...OOOOOOOOOOOOOOOOOOOOOOOO 48 III. METHODOIOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO..0. 53 Therapy Cases............................ 53 Patients...I.O...OOOOOOOOOOOOOOOOOOOOOOO. 55 TherapiStSoo0.00.00.00.00...ooooooooooooo SS InstrumentSOOOOOOOOOOOOOOOOOOOIOOOOOO0.0. 56 Raters.OOOIOOOOOIOOOOOOOOOI0.0.00.0000... 60 Data samplinQOOOOOOOOOOOOOOOOOOOOOOOOO0.0 61 General Procedure and Description Of GroupSOOOOIOOO...OOOOOOOOIOOOOOOOOOI 62 IV. RESULTSOOOIOO0.00.00.00.0000000000000000... 67 Reliability of Measures.................. 67 iv Intercorrelations of TARS and VNIS SUbscales O O O O O O O O O O O O O O O O O O O O O O O O 0 Testing the Hypotheses. I O O I O O O O O O O O O O O O O O V. DISCUSSIONOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 1. Therapeutic Alliance Development Across Time.......................... 2. Level of Negative Indicators Across Time.......................... 3. The Relationship Between Negative Indicators and Therapeutic Alliance.. 4. Therapeutic Alliance and Psychotherapy Outcome................ 5. Negative Indicators and Psychotherapy Outcome................ 6. Therapeutic Alliance and Pretreatment Symptomatology.......... 7. Negative Indicators and Pretreatment Symptomatology.......... Limitations of the Study................. Implications for Research and Practice... APPENDICESCOOOOOOOO...0.0.0.0...OOOOOOOOOOOOOOOOOOO. REFERENCESOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOO0...... 71 72 112 113 117 119 122 126 128 130 131 135 140 185 Table 10. 11. 12. 13. 14. LIST OF TABLES Page Description of Groups...................... 140 Reliabilities of TARS and VNIS Subscales... 146 Intercorrelations of TARS Subscales........ 147 Intercorrelations of VNIS Subscales........ 148 TARS Phase x Length Analysis of varianceOOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOO 75 VNIS Phase x Length Analysis of varianceOOOOOOOO0.00000000000000000...... 77 Correlations Between TARS and VNIS SUbscaleSOOOOOOOOOOOOOOO0.0.00.00.00.000. 79 Correlations Between TARS and VNIS Subscales by Treatment Phase............. 81 Correlations Between TARS and Therapists' and Patients' Perspectives on Outcome.... 85 Correlations Between VNIS and Therapists' and Patients' Perspectives on Outcome.... 86 Correlations Between TARS and Absolute symptom ChangeOOOOOOOOOOOOOO0.0.0.0000... 88 Correlations Between VNIS and Absolute Symptom Change........................... 99 Correlations Between TARS and Percentage of Symptom Change from Pretreatment Level-8.00.00.00.000000000000000000.0...00 94 Correlations Between VNIS and Percentage of Symptom Change from Pretreatment Level-$000.00....OOOOOOOOOOOOOOOOOOOOOOOO. 96 vi is. 16. 17. 18. Significant Multiple Regression Equations for the Prediction of Posttreatment symptomat01°gy.ooooooooooooon...ooooooooo 99 Significant TARS and VNIS Subscales in Predicting Posttreatment Symptomatology.. 106 Correlations Between TARS and Pretreatment Symptomatology.............. 107 Correlations Between VNIS and Pretreatment Symptomatology.............. 109 vii Chapter I LITERATURE REVIEW Therapeutic Alliance: The Psychoanalytic Tradition The next question with which we are faced raises a matter of principle. It is this: When are we to begin making our communications to the patient? When is the moment for disclosing to him the hidden meaning of the ideas that occur to him, and for initiating him into the postulates and technical procedures of analysis? The answer to this can only be: Not until an effective transference has been established in the patient, a proper rapport with him. It remains the first aim of the treatment to attach him to it and to the person of the doctor. . .If one exhibits a serious interest in him, carefully clears away the resistances that crop up at the beginning and avoids making certain mistakes, he will of himself form such an attachment. . .It is certainly possible to forfeit this first success if from the start one takes any standpoint other than that of sympathetic understanding (Freud, 1913. PP. 139-140). Any historical review of the therapeutic alliance must begin with Freud's view of the alliance between patient and analyst that was part of his earliest thoughts on transference in psychoanalysis. Freud introduced the idea that a bond formed between the patient and analyst is critical to the success of psychoanalytic treatment (Freud, 1912). As the preceding quote emphasizes, he clearly understood the role of attention, warmth, and empathy in 1 2 establishing a therapeutic rapport necessary to especially work through early resistances in treatment. Freud's early concept of the transference included the ideas of warmth and empathy as necessary factors in establishing a “positive transference“ between patient and analyst, thus emphasizing the analytic situation as an important component of treatment. This positive rapport was largely a conscious interpersonal bond; however, the patient would also form an attachment to the analyst based on past interpersonal experiences, libidinal impulses, and frustrations of needs which would all be part of the patient's neurosis (Freud, 1912). The bond stemming from these factors would remain largely unconscious, and would become both 'positive' and "negative" components of the transference depending on the patient's neurosis. These ”erotic positive" and "negative“ transferences would become the focus of the interpretive process in analytic technique. With the structural development of the ego and superego in psychoanalytic theory, Richard Sterba (1934) became one of the first writers to recognize the importance of ego-identification in the analytic relationship and emphasized the role of the positive transference as central to allowing the ego-identifying process to occur. Freud joined Sterba in this position by stating: The analytic physician and the patient's weakened ego, basing themselves on the real external world, have to band themselves together into a party against the enemies, the instinctual demands of the id, and the conscientious demands of the superego. We 3 form a pact with each other. The sick ego promises us the most complete candour--promises, that is, to put at our disposal all the material which its self-perception yields it; we assure the patient of the strictest discretion and place at his service our experience in interpreting material that has been influenced by the unconscious. Our knowledge is to make up for his ignorance and to give his ego back its mastery over lost provinces of his mental life. This pact constitutes the analytic situation (1940, p. 173). Sterba's (1934) view was that the patient's ego—identification with the analyst would allow the analyst to have a significant outside influence on the ego, strengthening it to accept a more balanced, adult—like attitude. He stated: Thus we may describe the transference and the resistance which goes with it as the conflict laden final result of the struggle between two groups of forces, each of which aims at dominating the workings of the ego . . . In Opposition to this dual influence, the object of which is to inhibit the analysis, we have the corrective influence of the analyst, who in his turn, however, must address himself to the ego. He approaches it in its capacity of the organ of perception and of the testing by reality. By interpreting the transference-situation he endeavors to Oppose those elements in the ego which are focused on reality to those which have a cathexis of instinctual or defensive energy. What he thus accomplishes may be described as a dissociation within the ego . . . Hence, when we begin an analysis which can be carried to completion, the fate that inevitably awaits the ego is that of dissociation . . . The therapeutic dissociation of the ego is a necessity if the analyst is to have the chance of winning over part of it to his side, conquering it, strengthening it by means of identification with himself and opposing it in the transference to those parts which have a cathexis of instinctual and defensive energy (1934, pp. 119-120). 4 Sterba emphasized that ego-dissociation was promoted by the analyst's use of 'we' in referring to the analytic relationship, by the patient's wish for recovery, and by the positive transference in the relationship (Sterba, 1934; Sterba, 1940). Through the dissociative and interpretive processes set up in the analytic situation, an "assimilation" process begins in which the ego begins to move toward a harmony with reality (Sterba, 1934). Other analytic writers followed Freud and Sterba's work with varying emphases on the therapeutic relationship. Nunberg (1932) saw the patient as only being attached to the therapist through irrational, infantile wishes for love and attention. Strachey (1934) viewed the will to recovery and the prospect of health as being essential for the patient to be induced to engage in the difficult and demanding process of analysis. Fenichel (1941) recognized the patient's predispositions as emphasized by Nunberg, but also emphasized the patient's desire for recovery associated with his or her ”reasonable ego.” Anna Freud (1946) stated that rather than completely arising from patients' predispositions, positive transference was maintained by the flexible and gratifying activity of the analyst. Zetzel (1956) emphasized the therapeutic alliance as an essential component of any psychotherapeutic treatment. She saw it as a cooperative, realistic, stable aspect of the therapeutic relationship, requiring the therapist's participation and partnership with the mature functioning of 5 some part of the patient's ego to fully develop. Influenced by Sterba and Zetzel's work, Greenson (1964) introduced the term I'working alliance“ as a separate component of the therapeutic relationship in comparison to the more irrational, transferential components. He defined the working alliance as the rational rapport developed between patient and analyst, and saw its primary function as encouraging the patient to work in analysis. Much like the process discussed by Sterba, Greenson felt the working alliance developed through the identification of the patient's reasonable, analyzing ego with the analyst's analyzing ego. As Greenson understood, however, the working alliance would necessarily contain some irrational features depending on the patient's past experiences. It is the analyst's responsibility to maintain a warm, reasonable, and professional position toward the patient in order to promote the ego-identification process necessary for the working alliance to develop. Much of the work on the therapeutic relationship has emphasized the therapist's responsibility in promoting patient variables necessary for an alliance to occur. This position is again emphasized by Sandler, Dare, and Holder (1973) when they talk about the patient's capacity to deve10p a positive alliance given a relationship formed around basic trust and self-observation. Unfortunately, this position has almost by definition excluded schizophrenic and most characterological patients from the 6 possibility of analytic success since they, as classical analysts have long held, are unable to form a therapeutic alliance within the analytic situation. In Opposition to this classical idea, Stone (1961), Winnicott (1955), and Kohut (1971, 1977), among others, have argued for psychotherapeutic treatment of psychotic patients based on a more secure, need-gratifying relationship in which the therapist is responsible for providing “legitimate" gratification as a necessary means to develOp a positive transference. More recently, Karon and Vandenbos (1981) have emphasized that schizophrenic patients are unable to spontaneously develop a therapeutically usable transference based on a positive, trusting dependence on the therapist given their undependable past experiences, but the therapist is certainly able to help build a positive transference by being a strong, protective, and more gratifying figure. In fact, they suggest the first function of the transference relationship with schizophrenic patients is to provide sufficient protection and gratification, with the other two functions, the permission of insight and providing a model for identification, only possible following the first. As all of these writers recognize, it is the therapist's commitment to help the patient, their caring, concern, empathy, and security, that make up the basis for the “real“ therapeutic relationship (Lehrke, 1977). This real relationship becomes the rational rapport between patient and therapist necessary to support 7 ego—identification and the more transferential elements of the therapeutic situation. Based on research at the Menninger Clinic, Horwitz (1974) observed that the support and gratification recommended by Stone along with the object relations processes of introjection and projection seemed to be the most important in establishing a curative relationship. Horwitz saw the relationship built from these factors as one involving a transference (positive and negative) based upon the patient's perception of the therapist as genuinely helpful and caring (the real relationship). He called this relationship the "therapeutic alliance,“ and viewed it as the foundation from which change could occur. Even though the historical development of the concept of the therapeutic alliance has been dominated by the analytic school of thought, other therapeutic orientations have also stressed the importance of the therapeutic relationship to varying degrees. Clearly, Rogers (1957) emphasis on warmth, genuineness, and unconditional positive regard as the necessary and sufficient ingredients of successful psychotherapy places central importance for treatment success on a positive therapeutic relationship. The emphasis within this orientation is obviously on the real relationship as discussed previously, especially when comparing it to the work-oriented extreme found within the behavioral school of thought. Behaviorists have tended to place central emphasis on the technical aspects of their 8 work, but even here the therapeutic relationship has not gone unnoticed as an important factor in successful treatment (Wilson and Evans, 1976; Beck et a1., 1979). Comparing these two extremes offers a clear distinction between the relationship-oriented focus on the therapeutic situation and the work—oriented focus. As we have seen, almost all writers have attempted to define this distinction within their own models of the therapeutic relationship, and it becomes increasingly important when attempting a common conceptual definition of the therapeutic alliance. Therapeutic Alliance: Defining the Concept Generally speaking, the therapeutic relationship encompasses the activities of the psychotherapy process, the transference relationship, and the genuine elements of care, empathy, and understanding typical of any interpersonal helping relationship. These various components have traditionally been discussed separately when attempting to understand the process of psychotherapy, and each has been given its place of importance in leading to positive treatment outcomes. Dynamic psychotherapy has typically alligned the specific concept of the therapeutic alliance with the more rational, non-transferential, genuinely human elements of the therapeutic relationship. This concept is very similar to the relationship-oriented focus of client-centered therapy, and is clearly distinguished from the working 9 alliance and transference components of the therapy relationship emphasized by Greenson (1964, 1967). Dickes (1975) has suggested using the term "therapeutic allianceI to refer to the c00peration between therapist and patient in meeting therapeutic goals, and use the term "working alliance" to refer to the more rational, relationship-oriented elements of the therapeutic interaction. All of these distinctions serve to again separate the therapeutic relationship into its relationship-oriented elements and its work-oriented elements, with little consideration toward establishing the therapeutic alliance as a concept bridging all of the powerfully interactive elements of the therapy relationship. Horwitz's (1974) definition of what he called the "therapeutic alliance“ in analytic therapy noted both the work-oriented and relationship-oriented elements in his concept. He emphasized the mechanisms leading toward the internalization of the "good therapeutic relationship“ as the primary process of change in psychotherapy, and defined the growth of the internalization of the ”real relationship“ as the therapeutic alliance. As is typical when defining a therapeutically relevant concept, Horwitz's view of the therapeutic allianCe is very theory-specific. In contrast to previous views, Edward Bordin has attempted to remove the concept of the therapeutic alliance from the confines of theory-bound definitions in order to generalize it to all l0 psychotherapies. In his review of the psychotherapy research literature, Bordin (1974) prOposed that a common-denominator across all psychotherapies was what he termed the alliance between patient and therapist in undertaking a particular treatment approach. If an alliance could be formed, the specific interventions called for under the treatment approach in question would be effective. Bordin considered the therapeutic working alliance (as he called it) to be the key ingredient in the psychotherapy change process, and offered four propositions in conceptually differentiating various approaches to psychotherapy (Bordin, 1979): 1) All genres of psychotherapy have embedded working alliances and can be differentiated most meaningfully in terms of the kind of working alliance each requires. 2) The effectiveness of a therapy is a function in part, if not entirely, of the strength of the working alliance. 3) Different approaches to psychotherapy are marked by the difference in the demands they make on patient and therapist. 4) The strength of the working alliance is a function of the closeness of fit between the demands of the particular kind of working alliance and the personal characteristics of patient and therapist. The conceptual definition of the therapeutic alliance used in this study follows Bordin's attempt at integrating the relationship elements and working elements of psychotherapy in defining the positive working commitment between patient and therapist. Bordin (1979) emphasizes three features of the therapeutic relationship in defining what he terms the therapeutic working alliance: 11 1) Agreement on Goals: The central feature of establishing an alliance with an individual seeking help is the formation of a working agreement between the patient's help-seeking requests and the therapist's theory-specific treatment goals. By the nature of his or her training, the therapist offers a particular perspective from which to view the patient's requests that restricts the range of collaborative agreements possible between them. At the same time, the patient is in a position of accepting the therapist's perspective and establishing treatment goals within this perspective or, if not, finding the work of therapy potentially impossible. It is thus vitally important at the onset of therapy that both therapist and patient openly understand and agree to the nature and direction of the therapy process. Without such a mutual agreement, the patient will have little basis for an accurate perception of the therapy situation and little allegiance to the goals of treatment (Hartley, 1978). A strong therapeutic alliance thus depends on clear expectations and mutually established goals for therapeutic work. 2) Assignment of Tasks: As part of the collaboration between patient and therapist the methods of the therapy process need also to be explicitly defined. The patient needs to understand how therapy works and how the tasks of this process specifically apply to his or her problems. Bordin (1979, p. 254) strongly states that I'the 12 effectiveness of such tasks in furthering movement toward the goal will depend upon the vividness with which the therapist can link the assigned task to the patient's sense of his difficulties and his wish to change.“ If the patient does not understand how his problems will be helped by the methods employed by the therapist, the strength of the therapeutic alliance is obviously threatened. 3) Development of Bonds: Depending on the specified goals and tasks for the therapeutic relationship, a particular type of human bond must be established between patient and therapist for a strong alliance to be formed. As Bordin (1979) points out, if the working relationship is defined to last several years with several meetings a week in which inner experiences are explored and understood, a different type of bond is required in comparison to meeting for three months in which specific behavioral tasks are employed. The strength of the bond may not be different, but certainly the type of bond must be different. It is quite possible to deve10p a positive bond in a therapeutic relationship in which goals are only implicit and tasks are never explicitly specified. The client-centered focus on the real relationship is a good example of this type of therapeutic relationship, but even here a mutual collaboration and implicit agreement in the direction of the therapy process forms the initial basis for a human bond to develop. In summary, Bordin's model of the therapeutic alliance 13 can be viewed in Hartley's (1985, p. 534) terms as ”a goal-directed system, with the cognitive and interpersonal styles of the patient and therapist interacting with each other and with the chosen techniques of the therapy.“ This conceptual definition moves beyond the mere positive transference or real relationship elements of the therapeutic relationship to a view of the alliance between patient and therapist as a collaborative working partnership essential for the process of therapy to be effective, irregardless of the specific therapeutic orientation being used. Therapeutic Alliance and Psychotherapy Outcome: Introduction Establishing an effective therapeutic alliance clearly involves variables centered around both the individual therapist and patient. Both contribute to the ongoing process of therapy in ways that directly effect the development of a working relationship and the eventual success or lack of success of the therapeutic work. Many of the process variables from both therapist and patient perspectives have been identified and subject to a number of investigations. The following is a brief review of psychotherapy process variables consistent with building a therapeutic alliance as they pertain to therapy outcome. An important question to keep in mind when reviewing these variables is, ”What role do process variables specific to 14 the individual participants have in the overall success of a therapeutic relationship?" Therapist and Patient Contributions to Successful Therapy Many of the process variables thus far discovered and explored in psychotherapy research have been a direct result of the facilitative conditions espoused by Rogers (1957). Studies focusing on therapist empathy, unconditional positive regard, and self-congruence all come from the perspective of the therapist's interpersonal behavior having a direct impact on behavior change. A number of measures have been develOped in the attempt to define and understand these facilitative conditions, one of the most influential being the development of the rating scales by Truax and his colleagues (Truax and Carkhuff, 1967). Ratings for these scales are obtained through the use of nonparticipant observers rating recorded process segments of therapy, offering a research technique that has been adapted for use in a number of various process measures. A great deal of the research has been focused upon the relationship between therapist warmth, empathy, and therapeutic outcome. A study by Halkides (1958) showing this to be a positive relationship has seemingly stirred a wealth of studies. Truax and Mitchell (1971) discovered eleven studies on warmth and thirteen on accurate empathy in their review of the literature prior to 1970. Orlinsky and Howard's (1978) review added another twelve studies on 15 warmth (e.g., Truax, 1970a; Mullen and Abeles, 1971; Truax, Wittmer, and Wargo, 1971; Garfield and Bergin, 1971; Schauble and Pierce, 1974; Truax et a1., 1973) and twenty-two on empathy (e.g., Bergin and Jasper, 1969; Truax, 1970b; Mintz, Luborsky, and Auerbach, 1971; Kurtz and Grummon, 1972). As a whole, the studies cited here lead to the conclusion that empathy and warmth do not necessarily lead to a positive outcome, but they do significantly add to the combination of significant “ingredients" in therapy that as a whole lead to a positive outcome. In another study of therapists' interpersonal behavior, Crowder (1972) used Leary's (1957) model to describe the relationship between interpersonal process and outcome in successful and unsuccessful cases. Crowder discovered that therapists in both successful and unsuccessful cases were most frequently supportive-interpretive in their interpersonal behavior, but therapists in the successful cases were significantly more supportive-interpretive, less hostile-competitive, and less passive-resistant late in therapy; and more hostile-competitive and less passive-resistant early in therapy than the therapists in the unsuccessful cases. Agreeing with Orlinsky and Howard's (1978) assessment of these findings, active and positive participation by the therapist seems to be of positive therapeutic value. In support of this general conclusion, a number of studies have linked active and positive participation by the 16 therapist to successful outcome by studying the specific clinical techniques used by therapists. Ashby et a1. (1957) found leading or guiding behavior to be more associated with successful therapeutic outcome as compared with reflective behavior. Direct approval has been linked to successful outcome as Opposed to clarifying or interpretive statements (Sloane et a1., 1975). Likewise, confrontation has been associated with positive therapeutic outcome by establishing an active, engaging group process with chronic schizophrenics (Mainard, Burk, and Collins, 1965). In Opposition to some of these findings, Baker's (1960) study of leading and reflective techniques found no significant difference between the two, and Nagy (1973) found confrontation to be unrelated to therapeutic outcome with outpatient subjects. The amount and style of therapists' verbal activity have also been looked at as possible sources of positive therapeutic outcome. However, studies measuring the amount and rate of therapist speech have shown no relationship to therapeutic outcome (Barrington, 1961; Scher, 1975; Sloane et a1., 1975). On the other hand, research by Rice (1965) on therapists' voice quality has indicated that an expressive vocal style and use of fresh language are associated with greater positive outcome as Opposed to an artificial vocal style and use Of stereotypic language. Systematic case-study research by Strupp (1980a, 1980b, 1980c) suggested that although therapists' skills and 17 attitudes toward the patient do have some effect on the process of psychotherapy, the patient variables are really the key to therapeutic process associated with outcome. Many Of these client variables reflect the same investigative paths as therapist variables. For example, voice quality of clients who were successful in therapy has been studied showing that such clients use more Of an Open and expressive vocal style than less successful clients (Butler, Rice, and Wagstaff, 1962). Barrington's (1961) research suggests that clients who use more words with larger numbers Of syllables tend to have more positive outcome. Furthermore, Orlinsky and Howard (1978) report a number Of articles suggesting that successful clients not only have more to say in therapy, but also take the time to think about what they want to say (use Of silences). Crowder's (1972) study Of interpersonal behavior in psychotherapy reported earlier for therapist variables also found interesting client behaviors associated with more successful outcome. In the early phases of therapy, the more successful clients were more hostile-competitive, less passive-resistant, and more support-seeking. During the middle phases of therapy, the trends on the passive resistant and support-seeking scales continued. Studies focusing on client self-perceptions and self-experience have provided some interesting data beyond the “Objective” Observer perspective. Lorr and McNair (1964) found that clients who perceive themselves as acting 18 in a hostile-controlling manner had less successful outcomes than those not having such self-perceptions. In the same study, they also found that clients who perceive themselves as being actively involved in therapy are more successful. Supporting this finding, Gomes-Schwartz (1978a) also found that greater patient involvement most consistently predicted positive therapeutic outcome. Saltzman et a1. (1976) found that clients who felt a greater sense of responsibility for solving their problems and changing their behavior Offered higher self-rated outcomes. From a greater self-experiencing perspective, Cabral et al. (1975) found that clients in group therapy who perceive themselves as having intense emotional expressions were consistently more successful in their therapy. Client self-experiencing fits the theoretical model Of good therapy process espoused by Eugene Gendlin (1973), and stimulated by the client-centered approach develOped by Carl Rogers (1957). Gendlin dismisses the importance of any individual therapeutic perspective by concentrating on how therapy, no matter what the orientation, elicits an ”experiencing" response from the client. This response, which may best be described as an accurate and personal feeling about some event, situation, or thought expressed in cognitive and affective terms (Rice, 1974), is the key to successful therapeutic outcome in this model. Rice (1974) explains the “experiencing" response as a cognitive necessity to completely processing feelings that had been l9 denied or distorted (thus incompletely processed) when first encountered. Whatever description used for the "experiencing” response, its importance as an insight event in psychotherapy has generally been verified. Gendlin et a1. (1960) found that clients who move from talking about their feelings to experiencing them are more likely to improve within individual client-centered therapy, and no correlation was found between positive outcome and therapists' perceptions Of clients talking about therapy, the therapist, or the present. Kirtner and Cartwright (1958) differentiated successful and unsuccessful cases of client-centered therapy by identifying those clients who discussed their feelings in the first session as Opposed to those clients who spoke of their problems as being basically external to themselves. Truax and Wittmer (1971) Obtained similar results when they found clients' use of personal references to be correlated with positive outcome as Opposed to clients' use of nonpersonal references. These data are further verified by a study by Schauble and Pierce (1974) which identified an association between positive outcome and clients directly confronting their problems and feelings. A more recent study by Elliott et a1. (1982) using four different evaluative paradigms Of therapy process (process-outcome, sequential process, immediate process recall, and retrospective attribution) identified client experiencing and the therapist's direct reference to that 20 experiencing process as the primary helpful factors in a single case study. A number Of studies have used more of a quantitative approach in their study Of client experiencing based on the scales developed by Gendlin et a1. (1968). Studies by Tomlinson and Stoler (1967); Tomlinson and Hart (1962); van der Veen (1967); Kirtner et al. (1961); and Gendlin et a1. (1968) all relate successful outcome in psychotherapy to client personal referents, internal referencing, or what generally may be called the process of client experiencing. In contrast to these results, a more recent study by Eaton (1984) comparing the frequency Of client verbal responses most associated with client "experiencing“ in successful and unsuccessful insight-oriented therapy cases found no significant differences between groups. Process Variables and the Importance of the Therapeutic Relationship: Conclusion As is quickly emphasized by the preceding review, research on process variables specific to the therapist and patient are Often contradictory, inconsequential, and micrOSCOpic in focus. Many of these variables are important in their Obvious contribution to the therapeutic relationship, but often lose their relevance when artificially removed from the confines Of the interactive relationship. As Sullivan (1953) pointed out over fifty years ago, an understanding of psychotherapy will really 21 only be achieved at the interpersonal level of therapist and client interaction rather than focusing on their separate characteristics. This is certainly not to say that delineating effective ingredients Of psychotherapy process is not a very valuable avenue for research, but only to emphasize the interactive nature of these variables and their overall fit into the therapeutic relationship. Orlinsky and Howard (1978) stated that effective psychotherapy may best be distinguished by the collaborative bond built between client and therapist who both invest a great deal of effort and energy in making the relationship supportive and encouraging, but also challenging and stimulating. Saltzman et al.'s (1976) research on the formation of a therapeutic relationship in a university counseling center serves to highlight results typical for this area of study. They found that relationships characterized by therapist and patient endorsing mutual feelings of warmth, respect, caring, Openness, and understanding were associated with positive outcome, greater satisfaction, and fewer premature terminations. Within this relationship, patients who had a greater sense of responsibility for solving their problems and changing their behavior had higher self-rated outcomes. They also discovered that not only could a mutually caring and effective therapeutic relationship be formed, but disruptions and dissatisfactions in the relationship were noted as early as the third session in cases leading to poor 22 outcome or premature terminations. Measures of therapeutic alliance have attempted to capture the "collaborative bond“ between therapist and patient in order to quantitatively define differing levels of alliance formation and empirically determine its importance in psychotherapy. The approach places the interactive relationship, conceptually defined as the therapeutic alliance, in the role of primary importance within which all other factors Operate to produce a therapeutic outcome. The alliance is Obviously influenced by a host of factors, many of which were reviewed in the previous section, but a measure of the alliance itself becomes the key variable. The Contribution of Therapeutic Alliance to Successful Therapy Considering the recognized importance of the therapeutic relationship throughout the history of psychotherapy, it is somewhat surprising that the conceptual development of the therapeutic alliance was not really solidified until recent years. There have been a number of studies, of varying empirical validity, dealing with transference and therapeutic relationship issues, but studies specifically addressing therapeutic alliance as a measure of the therapeutic relationship have only been completed since Bordin's (1975) first statements of definition. Since this time, a number of alliance measures 23 have been developed by a variety of investigators, and a few studies are now available for review. An interesting note to the literature to date is that the different alliance measures have been develOped at different sites of investigation; consequently, the few studies available tend to be attempts at establishing the specific investigator's measure. The first outcome study in which results tended to lend support to the importance of therapeutic alliance as a specific concept was not actually designed to measure alliance (Horwitz, 1974). The longitudinal study on psychoanalysis at the Menninger Clinic begun in 1954 was designed to test the validity of uncovering and working through of intrapsychic conflicts as best predictors of treatment outcome. Data were collected from 42 patients assigned to analysis or supportive therapy (as defined in the study) before treatment, at termination, and at a two-year follow-up. Horwitz's (1974) interpretation of these data attributed the most significant change in supportive treatment and greater than expected change in insight-oriented treatment to what he termed the growth of the therapeutic alliance. Horwitz's view of alliance was very similar to Freud's mild positive transference, and found this "clear prevailing positive relationship“ to be quite evident in his successful treatment dyads. It is this relationship that he viewed as the foundation from which change could occur. 24 _Using one of the first systematic measures of therapeutic alliance, Luborsky (1976) distinguished two main types of what he termed the helping alliance in psychotherapy. Type 1 was defined as a therapeutic alliance based on the patient's experiencing the therapist as supportive and helpful within the therapeutic relationship. Type 2 was defined as a therapeutic alliance based on a sense of working together in a "joint struggle against what is impeding the patient." As Luborsky points out, the emphasis in a Type 2 alliance is on the shared responsibility for therapeutic work or the “collaborative bond“ built between therapist and patient. Luborsky viewed the Type 2 alliance as being stronger and more vital for long-term gain, with a transition from Type 1 to Type 2 often occuring around termination issues. In an attempt to emphasize the importance of therapeutic alliance to therapy outcome, Luborsky (1976) compared the ten most improved versus the ten least improved of the 73 cases in the Penn Psychotherapy Research Project (only cases of at least 25 sessions were chosen). Using a residual gain criterion in which the gain from pretreatment to posttreatment measures is scaled relative to patients starting at the same initial level, he decided on seven high-gain patients and eight low-gain patients for final comparison. Type 1 and Type 2 alliance levels were rated by two expert judges according to a counting signs method in which any examples of a Type 1 or Type 2 alliance in the 25 initial twenty minutes of selected therapy sessions were identified by the judges and totaled to determine alliance strength. Two sessions were selected for ratings from the early stages of therapy (sessions 3 and 5) and two from the later stages of therapy (at least 90% of therapy completed). Luborsky predicted that a helping alliance would be formed as early as the fifth session for high improvers but not for low improvers. He also predicted that at least some of the high improvers would develop Type 2 helping relationships by the later stages of therapy while the low improvers would not. As expected, Luborsky discovered that six of the seven high improvers and none of the eight low improvers develOped Type 1 helping relationships as early as the fifth session. In this sense, Luborsky speculated that the best predictor of later benefits may be signs of early benefits expressed in the early sessions. Also as expected, Luborsky found two high improvers who had some evidence of a Type 2 helping relationship by the end of therapy, while low improvers had no such evidence. The total number of high improvers developing Type 2 relationships was somewhat lower than anticipated, but represented an expected trend with such a low sample size. Following the counting signs approach to rating therapeutic alliance, LubOrsky developed rating scales for Type 1 and Type 2 alliances that allows raters a broader sc0pe from which to judge differences in qualitative strengths of alliance between therapy dyads. Each item in 26 the new scales is rated on a ten-point Likert-type range reflecting the degree to which the item is present. Evaluating the predictive validity of these new scales for therapy outcome, Morgan et a1. (1982) used the same cases as in Luborsky's previously reported study, but included the five cases originally excluded by Luborsky based on the patients' and therapists' overall ratings of change. As was found in the previous study, an analysis of variance showed a significant between group difference between Type 1 and Type 2 alliances for the ten most improved versus the ten least improved cases in the Penn Psychotherapy Project. The most improved cases had higher levels of Type 1 and Type 2 alliances than the least improved cases, and the most improved cases had a nonsignificant trend toward greater Type 2 ratings in later sessions. Interestingly, no significant differences in alliance strength were discovered between early versus late sessions in either group. The initial development and use of the Vanderbilt Therapeutic Alliance Scale was conducted by Hartley (1978) in her dissertation on therapeutic alliance and the success of brief psychotherapy. She collected alliance ratings from the 28 therapy dyads in the Vanderbilt Psychotherapy Project that included only trained therapists. The cases were divided into high outcome, low outcome, and premature termination groups based on three outcome perspectives (patient, therapist, and observer) and treatment length (therapy lasting less than five sessions was labelled as 27 premature terminations, all other dyads could not work longer than 25 sessions). The Therapeutic Alliance Scale developed in this study consisted of 44 six-point Likert—type items tapping six primary factors: positive climate, patient resistance, therapist intrusiveness, patient motivation, patient responsibility, and patient anxiety. Two advanced graduate students rated 15-minute segments selected from the beginning, middle, and end of sessions chosen from the quartile points (including first and last sessions) of the total number of sessions (all sessions were used from premature terminations). Hartley expected that therapeutic alliance would be greatest for the high outcome group, lowest for the premature termination group, and somewhere between these two for the low outcome group. Surprisingly, her results did not support her predictions. She discovered that all groups were initially equal in level of alliance, but alliance in the drOpout group actually increased by the end of therapy as opposed to decreasing for the other two groups. The high outcome and low outcome groups did not differ statistically, but a trend was observed in which the high outcome group seemed to peak at the first quartile point and then trailed off, while the low outcome group was a mirror image of this pattern. Looking closely at the factor ratings in this study, Hartley was able to make a few statements concerning these unexpected data. She observed that the therapists in the dropout group were less intrusive and more willing to allow 28 the patient independence than in the other groups, while the less successful patients tended to be more hostile, critical, and mistrustful of the therapist, and more anxious and awkward within therapy sessions than in the other groups. Research by Comes-Schwartz (1978a) leading to the development of the Vanderbilt Psychotherapy Process Scale (VPPS) identified a Patient Involvement index as the best measure of the patient's contribution to the therapeutic alliance derived from the VPPS. The Patient Involvement index is derived from scores on the patient participation and patient hostility subscales of the VPPS, and was found to be the most significant and consistent predictor of therapy outcome in Comes-Schwartz's (1978) original study. Even though this scale is not identified as a specific and overall measure of therapeutic alliance, it is identified as a sound measure of the patient's willingness to enter into a collaborative, trusting relationship with the therapist. In this respect, it may be viewed more specifically as a measure of the patient's contribution to therapeutic alliance. Following up on this initial research, O'Malley, Suh, and Strupp (1983) collected VPPS ratings from the first three sessions of 38 cases in the Vanderbilt Psychotherapy Project. They wanted to test the ability of the three broad therapy dimensions measured by the VPPS, Patient Involvement, Exploratory Processes, and Therapist-Offered 29 Relationship, in predicting overall therapy outcome from early session levels. Consistent with Comes-Schwartz's original finding, they discovered that Patient Involvement most consistently predicted overall outcome and change in target complaints assessed from therapist, patient, and independent clinician ratings. They further discovered that this predictive association increased across the first three sessions from almost no association in the first session to a consistent association in the third. Marziali, Marmar, and Krupnick (1981) attempted to consolidate the therapeutic alliance scales used by Luborsky and Hartley, as well as the Negative Indicators Scale developed by Comes-Schwartz (to be reviewed in a later section), into a new therapeutic alliance scale focusing specifically on the affective, attitudinal aspects of the therapeutic climate. Their scale consists of 42 six-point Likert-type items making up four subscales: therapist positive contribution, therapist negative contribution, patient positive contribution, and patient negative contribution. Therapist total contribution and patient total contribution scores can be obtained from these scales by collapsing across the appropriate subscales. In an initial study with their new scale, Marziali, Marmar, and Krupnick (1981) decided to test the ability of therapeutic alliance as measured by this scale in predicting therapeutic outcome. They selected the five patients with the highest amount of positive change and the five with the 30 lowest amount of positive change out of a sample of 25 cases treated in brief dynamic psychotherapy at the Center for the Study of Neuroses, Langley Porter Institute, University of California, San Francisco, for comparison. All patients were involved in lZ-session treatments, and all were classified as having neurotic-level reactions to traumatic life events. Two experienced raters were given 20-minute segments randomly selected from the 2nd, 5th, 8th, and 11th sessions of each case for alliance ratings. Comparing the two groups across the four subscales and the total contribution scores, they discovered that only the patient's contribution to therapeutic alliance had an association with final outcome. Patients rated as making a strong positive contribution to therapeutic alliance had more positive treatment outcomes than patients rated as making a negative contribution to therapeutic alliance. The therapists' contribution did not discriminate the two groups. This finding is in keeping with the previously reviewed studies by Gomes-Schwartz and O'Malley et al., but the small sample size limits the power of the findings. Following this initial work and addressing the sample size issue, Marziali (1984) used her new scale to rate 42 cases selected for brief psychotherapy (twenty sessions) at the Clarke Institute of Psychiatry, Toronto, Ontario, Canada. Rather than simply using independent clinician ratings in this study, she was especially interested in the comparison of ratings between the participant therapists, 31 the participant patients, and independent clinicians. She discovered that all three groups of raters agreed to a large extent in their ratings of therapeutic alliance, except for their ratings of therapist negative contributions. Marziali suggested this finding may represent a vulnerability of the therapist negative contribution subscale to subjective bias. She also discovered that both patient and therapist positive contribution subscales were significantly associated with treatment outcome across all three rating groups, and this association could be determined as early as the first or third sessions. The importance of the therapist's contribution is somewhat in contrast to previous research, but Marziali's sample in this study was larger and presumably more heterogeneous than past studies. Using a modified version of the Marziali, Marmar, and Krupnick scale, Horowitz et a1. (1984) studied the impact of therapeutic alliance on the outcome of brief dynamic psychotherapy with 52 bereaved patients (recent loss of a parent or husband). Their research was also conducted at the Center for the Study of Neuroses, and was only one part of a larger project involving numerous process and outcome variables. Looking at the direct correlations between the four alliance subscales and two measures of outcome, they discovered only a mildly significant relationship between the patient's negative contribution to alliance and symptom change. As predicted, the greater the patient's negative contribution the less the rate of decline in symptoms. 32 Since the relationship between therapeutic alliance and treatment outcome was not as great as expected in these analyses, Horowitz et a1. decided to use a hierarchical multiple regression analysis to explore the interaction between measured patient characteristics and therapeutic alliance ratings. They discovered a number of interesting results emphasizing the importance of the interaction between process variables: 1) exploring negative patient contributions with highly motivated patients fostered the alliance; 2) the expression of negative contributions in less well-motivated patients undermined the alliance; 3) the therapist's maintenance of positive contributions with poorly motivated patients fostered the alliance and promoted positive outcome; and 4) the failure to explore negative contributions with highly motivated patients who otherwise expressed consistently positive attitudes toward therapy undermined the alliance and fostered poor outcome. Therapeutic Alliance and Outcome: Conclusion Given the few studies that have attempted to measure therapeutic alliance directly, there are still many questions concerning the association between alliance and psychotherapy outcome. Generally, results have tended to support some association between the two, but low sample sizes and the paucity of research with any one measure limits the strength of this association. The question of variability within groups is not Often addressed, especially 33 looking at the necessity of a strong or weak alliance for a given outcome. An unpublished study by Morgan et a1. (1977) suggests that a poor alliance almost always predicts a poor outcome, but much variability exists with good outcomes. The differences between patient and therapist contributions to alliance, especially early in therapy, is intriguing, and suggests the possibility of at least some pretreatment patient variables having an influence on early alliance development. Considering the apparent importance of early alliance formation, these variables may have a strong impact on overall treatment outcome. Results of research focusing on alliance formation and pretreatment variable issues will be addressed in the next section. Therapeutic Alliance Formation and Pretreatment Variables As observed in a review of therapeutic alliance studies in the preceding section, ratings of therapeutic alliance seem to be well-established as early as the third to fifth session in brief dynamic therapy. The results of these studies are generally quite mixed, with a great deal of variance centering around the implications of early alliance formation on final treatment outcome. It seems to make more intuitive sense that some time would be needed for the patient and therapist to establish a strong working bond in which both individuals feel comfortable working together, but the trend of the few studies available does not seem to support this view. 34 Briefly re-stating the pertinent results reported in the preceding sections, Luborsky (1976) discovered that Type 1 helping relationships were established as early as the fifth session in successful cases of brief dynamic therapy, but not in unsuccessful cases. He further found evidence of Type 2 helping relationships in the final phases of therapy for successful cases, but not for unsuccessful cases. These results suggest some progression of alliance development may exist for some successful therapy cases, but is not necessary in predicting positive outcome. Using a modified version of Luborsky's original scale, Morgan et al. (1982) found that levels of Type 1 and Type 2 helping relationships remained significantly higher throughout therapy in successful cases as compared to unsuccessful cases, but there was no difference in alliance strength between early and final stages of therapy. The successful cases in this study established a strong alliance very quickly and maintained the same level of alliance throughout therapy. The unsuccessful cases never established a strong alliance early in therapy, and were never able to build alliance strength at any point in therapy. O'Malley, Suh, and Strupp (1983) discovered that a measure of the patient's contribution to therapeutic alliance predicted positive treatment outcome as early as the third session, but did not predict outcome in the first session. At least from the patient's perspective, some time was needed to build an alliance, but the amount of time needed was very short. 35 Unfortunately, measures were not collected for later sessions in this study. In contrast to the more positive findings that early alliance formation is associated with successful outcome, Hartley (1978) discovered that high and low outcome groups did not significantly differ in levels of alliance at any point in therapy, but the drOpout group actually increased alliance strength over the course of their few therapy sessions. Even though high and low outcome groups did not significantly differ, it was still observed that a stable alliance formation was achieved as early as the first quartile point in these therapy dyads, and the high outcome group showed a trend toward a peak of alliance strength at the first quartile point. Marziali (1984), grouping the ratings from 42 brief dynamic cases together, found lower positive alliance ratings in the early sessions as compared to ratings of the final session. These results still reflect the beginnings of alliance formation early in therapy, but strength of alliance seemed to grow more consistently in her cases than had been observed in other studies. A study by Horn-George and Anchor (1982) compared perceptions of the therapeutic relationship in long- versus short-term therapy in which alliance was not measured directly, but results are pertinent to this discussion. They collected data on therapist and patient perceptions of the therapeutic process from 30 dyads each designated as 36 either long or short in duration. A case was considered long-term if the participants had, or were expected by the therapist to have, twenty or more sessions together. A case was considered short-term if there had been fifteen sessions or fewer, and if the therapist initially expected it to be a brief therapy relationship. The results of data analyses suggested that therapists and patients in long-term therapy are more congruent in their feelings about therapy, each other, therapeutic goals, and direction of therapy than participants in short-term therapy. For many of the long—term cases, congruence Of perceptions existed from the beginning of the therapy relationship. By definition, the long-term therapy dyads in this study perceived themselves as having a stronger therapeutic alliance than the short-term dyads. These results seem to appear in contrast to Hartley's (1978) data, but it must be remembered that therapeutic alliance was not measured directly in this study, and the long- and short-term therapy distinctions were largely based on the therapist's initial impressions. The primary point of mentioning this study is to further highlight the potential impact of length of therapy on alliance formation. All previous studies on therapeutic alliance have used therapy cases in which length of therapy was predetermined to a great extent by the therapeutic modality. Since therapists and patients are well aware of time limitations from the outset in brief dynamic therapr it is conceivable 37 that early alliance formation may be the result of a direct and active push by both participants to form a working relationship as quickly as possible. What would be the effect on alliance formation if no or broader limitations were initially set? It must also be kept in mind that patients are screened for apprOpriate inclusion in brief dynamic therapy, with central criteria being the patient's capacity to actively work and establish a therapeutic relationship as quickly as possible without becoming too dependent (Malan, 1979; Davanloo, 1980; Strupp and Binder, 1984). What would be the effect on alliance formation if patients were not initially screened, but simply came to therapy with the variety of pretreatment variables typically seen in an outpatient clinic? Unpublished doctoral dissertations by Ryan (1973), Sarnat (1975), and Lehrke (1977) initially attempted to answer the latter question by studying the impact of patient pretherapy variables on early alliance development. Ryan was interested in demonstrating that pretherapy patient personality measures would predict quality of alliance and patient participation early in treatment. He discovered that ratings of patients' hOpefulness and capacity to form object relations were significantly correlated with his measures of therapeutic alliance and patient participation. Sarnat used Ryan's measures plus more client-centered measures with both psychoanalytic and client-centered cases to determine if different treatment orientations placed 38 varying demands on different patients. Among a variety of results, she tentatively supported Bordin's view that different treatment approaches make specific demands on patients, which must fit with the patients' capacities to work within these approaches for an effective alliance to be formed. Lehrke, using Ryan and Sarnat's cases, found both contrasting and supportive evidence of Bordin's views. She discovered that early alliance develOpment was not so much the result of the therapist's active efforts at enhancing the fit between the patient's pretherapy capacities/dispositions and treatment demands, but the indirect structurung of a safe and supportive climate for growth in the treatment situation. Furthermore, alliance formation was not enhanced by the patient's in-therapy communications and actions, but rather was based on the patient's pretherapy capacities toward relationship formation and ego functioning. These results strongly reflect the importance of pretreatment patient variables in alliance formation, and the therapist's capacity to provide an environment in which these variables can function. Studies looking directly at the effect of capacity or quality of pretherapy interpersonal relations on therapeutic alliance have been conducted. Moras and Strupp (1982) selected 33 cases of single, male college students from the Vanderbilt Psychotherapy Project in which pretherapy data on interpersonal relations were available. Independent clinicians had interviewed each patient and rated adequacy 39 of interpersonal adjustment based on their ability to form and sustain close or intimate relationships with family and friends, and prominence of conflicts or hostile attitudes toward family, friends, and peers. They discovered a significant correlation between these pretherapy interpersonal relation ratings and the patient's positive contribution to therapeutic alliance. Those patients who were rated as having positive interpersonal relationships upon entering therapy were more capable of forming collaborative bonds with their therapists within therapy. It should be noted that even though significant, this association did not account for a major portion of the variance in alliance formation. Marziali's (1984) previously reported study also measured patients' pretherapy self—reported social adjustment and its relation to therapeutic alliance, and found similar results to Moras and Strupp. Patients who reported a high level of social adjustment were more capable of forming a strong therapeutic alliance. The effect of patients' level of symptomatology on therapeutic alliance has likewise been addressed in a few studies. The preceding study by Moras and Strupp included a psychological health measure based on pretherapy ratings Of patients' severity of problems, intensity of subjective distress, and adequacy of functioning in social, work, and academic roles. A partial correlation between psychological health and their measure of patient's contribution to 40 therapeutic alliance (controlling for overlap with interpersonal relations) was insignificant. The partial correlation between interpersonal relations and alliance ratings remained significant. The preceding study by Marziali included a measure of severity of general symptomatology finding similar results. Alliance ratings in her study were not significantly influenced by patients' level of distress. The previously reported study by Morgan et a1. (1982) also included a correlation between alliance ratings and patients' initial level of pathology. They used the Health-Sickness Rating Scale to assess initial pathology and discovered that it did not significantly correlate with helping alliance ratings. The results of these three studies, even though quite limited in their measures of pathology, tentatively suggest that patients' pretherapy symptomatology does not significantly impact alliance formation. In summary, it appears that alliance formation may critically depend on the patient's capacity to form interpersonal relationships upon entering the treatment situation. If the patient has good interpersonal capacity and the therapist provides an environment that supports the growth of an interpersonal relationship, a strong therapeutic alliance can form very quickly with good prognosis for treatment success. As Marziali (1984, p. 422) states, "Conversely, patients with a history of highly conflicted relationships add stress to the alliance and 41 demand from therapists a capacity to absorb and manage responses which are ambivalent, confusing, and obstructionistic. Probably, it is the 'working through' of these treatment relationship conflicts that determines the outcome of psychotherapy for these more resistive patients.“ Even though the effects of pretreatment symptomatology on therapeutic alliance have not been adequately tested, initial indications suggest little overall symptom impact on alliance strength. Negative Indicators in Psychotherapy Broadly speaking, psychotherapy, particularly psychotherapy based on psychodynamic principles, appears to give rise to positive changes if in important respects the patient~therapist interaction becomes a collaborative endeavor, called variously therapeutic alliance, therapeutic working relationship, or working alliance. By the same token, deficiencies in the therapeutic alliance are increasingly being viewed as serious impediments to therapeutic progress. Unless corrected, such deficiencies may lead to an impasse, interminable therapy, intractable resistance, premature termination, power struggles between patient and therapist, anger or other countertransference reactions on the therapist's part, deterioration of the patient's condition or lack of significant therapeutic progress. For present purposes, any characteristics of the patient, the therapist, and their interaction leading to such develOpments are termed 'negative indicators.‘ (Strupp et a1., 1981, p. 1). By definition, negative indicators in psychotherapy are variables that serve a destructive role in the attempt at building a strong therapeutic alliance. Consequently, negative indicators are theoretically predictive of poor 42 therapy outcome. Bergin (1963, 1971) initially highlighted the potential of psychotherapy as not only a helpful process, but a destructive process as well. In reviewing the psychotherapy outcome literature, he noticed that treated groups of patients tended to show much greater variability in outcome than untreated controls. This observation led Bergin to conclude that psychotherapy is an effective process that has the potential for both positive and negative change. Even though Bergin's statements evoked criticism (see Franks and May, 1980), a number of studies have been reviewed suggesting consistent findings of negative effects in psychotherapy (Lambert, Bergin, and Collins, 1977). Many of these findings are methodologically suspect (Strupp, Hadley, and comes-Schwartz, 1977), but emphasize a growing awareness of the potential for adverse consequences as a result of the psychotherapeutic process. A survey of eminent theorists, clinicians, and researchers in the psychotherapy field (Hadley and Strupp, 1976) suggested a number Of broad areas that may include factors having a negative impact on psychotherapy. These areas are briefly identified as follows: 1) Patient Characteristics: As reviewed in the previous section, pretherapy patient variables may have an influence on alliance formation and the eventual outcome of psychotherapy. The personal capacities of the patient, especially the ability to form interpersonal relationships, can play a vital role in therapeutic process. The earlier 43 review of patient contributions to successful therapy also pointed out a number of factors that can have a negative impact. Lorr and McNair (1964) found that clients who perceive themselves as acting in a hostile-controlling manner had less successful outcomes than those not having such self-perceptions. Clients who cannot express or tolerate extreme affects in psychotherapy may be less likely to succeed (Cabral et a1., 1975; Huxster et a1., 1975). Likewise, patients who have difficulty moving from more of a cognitive realm in dealing with affect to more of a self-experiencing realm may find treatment success more limited (Gendlin et a1., 1960). Sandell (1981) also pointed out that patients with low levels of ego strength, poor reality testing, rigid defensive organizations, or poor motivation are more likely to have unsuccessful therapeutic experiences. 2) Therapist Personal Qualities: Freud (1910) pointed out the potential negative effects of countertransference reactions on the therapeutic relationship, especially when the therapist's needs and conflicts unconsciously become the primary focus of the treatment. On the other hand, a rigid interpretation of therapeutic neutrality in analytic therapy may lead a therapist to offer little emotional or interpersonal interaction to the patient. In this sense, either too much or too little personal investment by the therapist may have detrimental consequences to treatment outcome. A number Of therapist contributions to successful 44 therapy have already been reviewed. Some of the most important therapist qualities that have been relatively well-researched are Rogers' (1957) qualities of genuineness, empathy, and unconditional positive regard. The lack of any of these qualities as expressed in the treatment relationship may deter therapeutic progress, but the ultimate importance of such qualities is suspect. Suffice it to say that ignoring such therapist qualities would be ignoring variables with potentially significant impact, both positive and negative, on a patient's outcome. 3) Errors in Technique: Hadley and Strupp's (1976) survey pointed out two therapy errors having negative potential at least in dynamic therapy. One is a failure to accurately assess a patient's capacities for the type and level of work requested of them in therapy. This is very much in keeping with Bordin's view that a strong alliance will not be formed if the client cannot work within the prescribed therapy framework. A second error is the imprOper use of interpretations in therapy. Interpretations that are premature or beyond the patient's present level of tolerance can be quite destructive to therapeutic alliance, as can missed interpretations or interventions. Langs (1973) emphasized an imprOper focus of therapy as having a strong negative impact, leading to therapy that is either too threatening or too superficial. Also in keeping with Bordin's view mentioned above, Sandell (1981) Observed that a significant error in the Vanderbilt Psychotherapy Project 45 sample was the therapists' lack of flexibility in modifying their therapeutic approach to fit their patients' needs and abilities. 4) Patient-Therapist Interaction: Obviously, all of the preceding factors have an impact on the quality of the patient-therapist interaction, but more basic errors in interaction may also be made. Therapists who do not maintain sufficient professional distance (potentially overlapping with countertransference issues) can have a destructive impact on their patients, as can patients who become overly dependent on the therapeutic relationship. Likewise, interactions that are too emotionally distant can lead to ineffective therapy (Sandell, 1981). Using these categories, Comes—Schwartz (1978b) initially develOped a 25-item scale in an attempt to identify and measure negative indicators in psychotherapy leading to less successful change. She asked five experienced clinicians to each rate a positive and negative case from the Vanderbilt Psychotherapy Project using her new scale. She found good reliability among raters, and most importantly discovered that her scale was quite useful in differentiating these cases. Expanding on this work, Strupp et a1. (1981) studied a number of sessions in the Vanderbilt Project to include more clinically relevant items in order to make the scale more comprehensive and useful. Their current product is a 42-item version called the Vanderbilt Negative Indicators 46 Scale (VNIS), which includes five subscales: Patient Qualities, Therapist Personal Qualities, Errors in Technique, Patient-Therapist Interaction, and Global Factors. The items are grouped into categories within each subscale on a conceptual basis, and each item is explicitly defined in the rating manual. Two studies using the VNIS have been conducted as part of its develOpment and validation. Strupp et a1. (1980) had two experienced clinicians rate third sessions for ten cases in the Vanderbilt Project. These cases consisted of a high-change and a low-change client for each of five therapists. Interrater reliabilities for the subscales were at acceptable levels, except for the Errors in Technique subscale. In general, they discovered that VNIS total scores significantly differed between high-change and low-change cases, with low-change cases receiving higher ratings. The variability in subscale scores, however, led to a failure for these scores to discriminate between high and low outcome in these few cases. Improving upon this design, Sandell (1981) asked two experienced judges to rate the first three sessions of therapy for the 18 Vanderbilt Project patients who were seen by professional therapists. As in Strupp et al.'s study, interrater reliabilities were generally good, but Errors in Technique still did not meet generally acceptable levels of reliability. A composite measure of outcome was obtained for each case, and correlations were obtained with the VNIS 47 ratings. Significant negative correlations were discovered between all VNIS subscales and outcome by the third session of therapy, with the exception of Therapist Personal Qualities. Partial correlations Obtained between the same measures revealed that only the specific contribution of Errors in Technique was significant, but the sample size is too small for any conclusions to be drawn from this finding. In general, Sandell found that reliable ratings can be obtained with this scale, and measuring negative indicators in psychotherapy can potentially be useful in highlighting problems in the therapeutic relationship leading toward less successful outcomes. Chapter II HYPOTHESES The importance of the therapeutic alliance in effective psychotherapy has been recognized throughout the history of the field. Unfortunately, it has also been recognized as a concept that is easily discussed, but difficult to measure directly. Since Bordin's more integrated definition of therapeutic alliance moving the concept beyond a theory-specific domain, a number of researchers have develOped scales and made the attempt to study alliance more directly. As we have seen, however, the amount of research is still sparse, and the quality of the studies is highly variable. The research has been done exclusively with either time-limited dynamic therapy, or with a focus on the initial phases of alliance formation in longer therapy. The broad questions concerning therapeutic alliance development, its impact on treatment outcome, and variables affecting its formation still remain, especially with different patient populations and therapy modalities. Given the need to add more to our knowledge base about therapeutic alliance, the purposes of the present study are to: 1) Examine the strength and development of therapeutic 48 49 alliance over time, controlling for length of therapy by dividing the cases into low, medium, and high number of sessions with comparisons made between groups. 2) Examine the frequency and develOpment Of negative indicators over time using the same groups. 3) Investigate the association between strength of therapeutic alliance and frequency of negative indicators. 4) Explore the association between therapeutic alliance, negative indicators, and therapy outcome from the therapist's, patient's, and change in symptomatology perspectives. 5) Examine the association between therapeutic alliance, negative indicators, and pretreatment patient self-reported symptomatology. The review of the therapeutic alliance literature suggests several hypotheses to be tested in this study: 1) The cases used in this study are outpatient clients from the surrounding community seeking psychotherapy from a university-based psychological clinic with predominantly graduate student therapists. Generally, therapy in this setting is dynamic in orientation, but not designed from the outset to be time-limited. However, due to the nature of the situation (e.g., end of the academic year for the therapists, end of services for part of the summer, therapists leaving the clinic for other positions, etc.) the length of therapy may be restricted. Since a time-limited approach is not planned, it is hypothesized that therapeutic 50 alliance will develOp strength over time as the patient and therapist build their relationship. Therapy with a higher number of sessions is expected to have established a significantly stronger alliance by the end of therapy than therapy with the fewest sessions. Medium length therapy is expected to fall within the expected trend, but not significantly differ from longer therapy. As highlighted in previous research, alliance is not expected to differ significantly during the initial stage of therapy between groups. Since the lower length therapy, however, may include some premature terminations due to a lack of alliance formation, it is expected that problems initially establishing alliance may be more frequent in this group, suggesting lower, if not significant, alliance ratings. Within each group, a trend toward stronger therapeutic alliance by the end of therapy is expected, with a smaller trend observed for shorter therapy. 2) Rather than negative indicators being influenced by length of therapy, it is expected that frequency of negative indicators will influence therapy length by increasing premature terminations and dropouts. Since all of the cases in this study at least have the establishment of an initial therapeutic relationship (no case with fewer than ten sessions included), the difference between the shorter length group and the longer length group is not expected to be significant, but a trend should be observed. Within groups, negative indicators are not expected to change over 51 the course of therapy. 3) By definition, negative indicators are destructive to the establishment of a strong therapeutic alliance; consequently, it is expected that significant negative correlations between the VNIS subscales and the two positive alliance subscales and significant positive correlations with the two negative alliance subscales will be discovered in all groups. However, since the lower length group is not expected to have develOped as strong a therapeutic alliance due to situational variables more than poor therapy process, the correlations may be lower for this group. NO predictions are made concerning specific subscale correlations. 4) In keeping with the trend in the literature, therapeutic alliance (as measured by the therapist and patient positive subscales) is expected to be associated positively with all three outcome perspectives. Likewise, a negative association is expected between negative indicators and all three outcome perspectives, as well as between the negative therapeutic alliance subscales. It is also hypothesized that ratings of the patients' contribution to therapeutic alliance will be associated more highly with patients' subjective ratings of change, as will the therapists' contribution with therapists' subjective ratings of change. No predictions are made concerning specific negative indicator subscales or change in specific symptoms. 5) Previous studies have suggested that pretreatment 52 patient symptomatology does not affect alliance strength, but the quality of patients' pretherapy interpersonal relationships does have a significant impact on alliance formation. The studies measuring pretreatment symptomatology have only focused on broad assessments of general psychological health as opposed to specific symptom dimensions. When looking at more specific symptomatology, it is expected that those symptom dimensions most affecting the develOpment of interpersonal relations (in this study the dimensions of interpersonal sensitivity, hostility, and paranoid ideation) will be negatively associated with therapeutic alliance. Likewise, the same symptom dimensions are expected to lead to a higher frequency of negative indicators in the process of theraPY: consequently, a significant positive association is expected between these measures . Chapter III METHOD Therapy Cases This study used therapy cases collected over a four year period (September, 1978 through August, 1982) for research purposes at the Michigan State University Psychological Clinic, an outpatient clinic serving non-student members of the Michigan State University community and surroundings. The Clinic is a training and research agency of the Department of Psychology, and serves as a low cost clinic to adults, children, and families. The original data collection was based on the premise of obtaining relatively nonintrusive information on patients coming to the Clinic, and consists of pre- and post-therapy written measures as well as audiotapes from selected sessions. The purpose of the data collection was to provide a data source for research on the process and outcome of psychotherapy. Over the time period from which the cases used in this study were taken there were approximately 84 terminations of adult patients at the Clinic for which at least some research data were collected. Of the 84 terminations, 70 53 54 cases included both pre- and post-therapy measures pertinent to this study. Since one of the purposes of this study was to assess therapeutic alliance develOpment over time, one of the criteria for inclusion of a case was that it was at least ten sessions in length. Other criteria were that appropriate audiotapes be available for ratings and all pertinent pre- and post-therapy data were available and complete. Using these criteria, 40 cases were finally left for analysis in this study. A feature of this study was a comparison of alliance development across therapy of varying lengths. For these analyses, the 40 available cases were grouped into high, moderate, and low number of sessions as arbitrarily defined by the following criteria: a high session case included over 40 therapy sessions; a moderate session case included over twenty but fewer than 41 sessions; and a low session case included fewer than 21 therapy sessions. Obviously, what is considered longer therapy here would be considered time-limited in some therapeutic modalities, but the important comparison is between well-defined short-term dynamic therapy models and therapy length that extends beyond the common 25 session limit in such models. Using the preceding criteria for therapy length, 12 cases were included in the long length group, 15 cases were included in the medium length group, and 13 cases were included in the short length group. 55 Patients The forty patients whose therapy cases were used in this study agreed to allow the data to be used for research purposes. The patients were all community members who accepted the offer of therapy at the Clinic, and who completed all of the pre- and post—therapy forms correctly. Fourteen of the patients were male and twenty-six were female. Their average age at the beginning of therapy was 30.35 years (ranging from 20 to 57 years), and their average length of education was 14.68 years (ranging from 4 to 20 years). Looking at the groups separately, the average age of the long length group was 28.00 (20-41 years) and education was 15.42 (12-17 years); the average age of the middle length group was 29.87 (22-48 years) and education was 15.00 (4-20 years; and the average age of the short length group was 33.08 (24-57 years) and education was 13.62 (6-20 years). Therapists The therapists whose cases were used in this study all consented to take part in the Clinic's research. The group consisted of experienced clinical psychology graduate students with at least one year of psychotherapy experience and generally more than two years. Some therapists included in the study were second-year graduate students who were beginning their first practicum, but most of whom had some client contact before entering the program. Twenty-five of 56 the therapists were male and fifteen were female. Their average age, education, and experience level could not be calculated from the available data. The specific therapeutic orientation used in each case was not specified; however, the general orientation of the students in the Michigan State University Clinical Psychology program is insight-oriented and a previous assessment of therapist verbal response mode use on a sample of these cases found they most closely matched a broadly defined dynamic orientation (Eaton, 1984). The fact that therapist experience level is significantly lower than previous studies on therapeutic alliance is recognized as a potential confound, but may also prove to be of interest in and of itself. The impact of gender differences on therapeutic alliance is unknown, and is beyond the scope of this study. Instruments A. Measures of Therapeutic Alliance and Negative Indicators. 1. Therapeutic Alliance Rating Scale (Marziali et a1., 1981; Marziali, l984--Appendix B). The Therapeutic Alliance Rating Scale (TARS) is a 42-item instrument designed to measure both positive and negative contributions to alliance made by therapists and patients. Items have been drawn from work by Luborsky, Hartley and Strupp, and Gomes-Schwartz, with items pertaining to specific therapist actions and patient responses purposefully excluded leaving 57 a scale focused more specifically on the affective-attitudinal aspects of the therapeutic relationship, including Bordin's concept of the collaborative working bond as assessed from an attitudinal perspective. All items are rated on a six-point intensity of presence scale and are summed across the four subscales (therapist positive and negative contribution, and patient positive and negative contribution) to achieve subscale scores. Internal consistency and interrater reliability between independent raters have proved adequate in the initial work with this scale, and patient and therapist positive contribution subscales have especially proved valid in discriminating between outcome groups and predicting therapeutic success (see Marziali et a1., 1981 and Marziali, 1984). 2. Vanderbilt Negative Indicators Scale (VNIS, Strupp et a1., l981--Appendix C). The VNIS is a 42-item instrument designed to measure factors detrimental to the development of a strong therapeutic alliance and predictive of negative therapeutic outcomes. The items are grouped into five broad subscales: patient personal qualities and attitudes, therapist personal qualities and attitudes, errors in technique, patient-therapist interaction, and global session ratings. Items are further classified into conceptually distinct categories within each subscale. For the purposes of this study, a composite subscale identified as VNIS total was included in the analyses. All items are rated by 58 independent observers on a six-point intensity of presence scale and are summed across categories and subscales to achieve specific scores. Initial studies with the VNIS have established generally acceptable levels of interrater reliability, and the scale has generally been found useful in highlighting problems in the therapeutic relationship leading toward less successful outcomes (Strupp et a1., 1980; Sandell, 1981; Sachs, 1983). B. Measures of Outcome. 1. Post-therapy Client Questionnaire (Appendix D). A 56-item client form (Strupp, Lessler, and Fox, 1969, shortened version) was given to clients at the termination of therapy. This form tapped the clients' subjective beliefs about the effectiveness of their therapy. Four questions which appeared to be the best representatives of overall therapy outcome were selected for use in this study (questions 3, 4, 11, and 15). Level of success from the client's perspective was thus determined from the client's ratings on these four questions. For example, question 3 is, “How much have you benefited from your therapy?” The client can answer: 1) a great deal, 3) a fair amount, 5) to some extent, 7) very little, or 9) not at all. Ratings on all four questions were combined and averaged to get a total improvement rating. 2. Post-therapy Therapist Questionnaire (Appendix E). A 33-item therapist form includes ten questions from the therapist version of the HOpkins Symptom Checklist and 23 59 questions relating to the therapist's subjective beliefs about the effectiveness of therapy (Strupp, Lessler, and Fox, 1969, shortened version). Three questions which appeared to be the best representatives of overall therapy outcome were selected for use in this study (questions 22, 27, and 29). Level of success from the therapist's perspective was thus determined from the therapist's ratings on these three questions in the same manner as the client ratings. 3. Hopkins Symptom Checklist (SCL-90, Derogatis et a1., l976)--C1ient form (Appendix F). This measure consists of ninety statements of problems. The problems comprise and load on nine symptom dimensions and a global severity index. The symptom dimensions are somatization, obsessive-compulsiveness, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. Clients were instructed to check those statements that were current problems for them and to rate the degree of distress (0-4) associated with each problem before the beginning of therapy and at termination. Level of pathology for each symptom dimension was then determined by averaging the item ratings within each symptom dimension. For the purposes of this study, degree of change on each symptom dimension was assessed by subtracting clients' SCL-90 ratings at termination from their pre-therapy ratings. Even though this measure is still based on clients' self-report, it offers a different 60 outcome perspective from simply broadly asking a client whether therapy was helpful or not. Relatively recent research on the validity of the SCL-90 suggests that it is a valid measure of client distress from both a client and therapist perspective (Filak, 1982). C. Measure of Symptomatology. 1. Hopkins Symptom Checklist (SCL-90, Derogatis et a1., l976)--Client form (Appendix F). Initial levels of symptomatology were Obtained from the SCL-90 as described above. Only pre-therapy measures were used in assessing the impact of specific symptom dimensions on therapeutic alliance and negative indicators. Raters Therapeutic alliance and negative indicator ratings were obtained from separate sets of judges. Each set of judges consisted of two advanced graduate students in the clinical psychology program at Michigan State University. There was one female and one male judge in each group, and each had a number of years of psychotherapy experience before entering the program and during their advanced training. The judges in both pairs were matched as closely as possible on standing in the program, years of experience, and therapeutic orientation; however, it was recognized that the judges rating the therapeutic alliance scales were more divergent in experience and orientation than the other pair. Training for all raters consisted of reading the 61 manuals for each rating system, rating practice segments of various therapy sessions used in previous research, and participating in group meetings in which the items for each rating system and practice transcripts were discussed to achieve consensus on definitions and rules for rating. The total length of the training period was 14 weeks in which 50 practice segments were rated and discussed in approximately 20 hours of research meetings. The training period ended when a criterion level of reliability of .70 (as assessed using averaged Pearson product-moment correlations) was reached on each of the separate TARS and VNIS subscales over at least the final half of the practice sample. The only subscale not to reach this criterion level during the training period was Errors in Technique from the VNIS which had already been identified in previous research as the least reliable of the VNIS subscales (Strupp et a1., 1980; Sandell, 1981). In addition to the training period described above, 16 hours of meetings were also conducted during the 20 weeks of actual research ratings. These meetings were held for each pair of judges to discuss any problems they were having and to help prevent rater drift during the actual rating process. Data Sampling Three fifteen minute segments were sampled from each of the 40 audiotaped cases in this study. In order to assess 62 therapeutic alliance develOpment over time, one segment was taken from each of the first, middle, and final phases of the total therapy process for each case as defined by the following criteria: first phase of therapy was any session falling within the first 20% of the total number of sessions; middle phase was any session falling within the middle 40-60% of the total number of sessions; and final phase was any session falling within the final 20% of the total number of sessions. An attempt was made to choose sessions with equal therapy length between first to middle phase and middle to final phase. Karl and Abeles (1969) showed that different segments of a therapy session can be characterized by markedly different content areas; consequently, segments were also randomly chosen from the first, middle, and final fifteen minutes of each session, with one of each segment chosen from each case. All segments were randomly mixed onto master tapes for rating. All ratings were obtained exclusively from the judges listening to the audiotaped segments. General Procedure and Description of Groups Original collection of pre- and post-therapy data was handled by the Clinic staff during the time that psychotherapeutic treatment was actually offered and performed. Clients requesting services at the Psychological Clinic were required to attend an intake interview session where they initially stated their concerns and Clinic 63 interns evaluated the appropriateness of the patient to be served by the Clinic. During this preliminary session, Clinic policies and procedures were also explained, and the patients were informed of the Psychotherapy Research Project. It was explained that participation in the research project was strictly voluntary, and required the completion of a few forms both before and after treatment, as well as consent to have a selected number of audiotapes stored in the data library. All data were kept strictly confidential, and a signed consent form was required before any data were collected. Patients volunteering for the study understood that they would be aiding researchers in understanding and improving the process and outcome of psychotherapeutic treatment. If, at this point, patients consented to participate in the project, they were given a pretreatment research packet consisting of the written consent form, HOpkins Symptom Checklist, and a demographic information sheet. They were asked to complete the material and return it to their first psychotherapy session where the therapist would collect it and give it to the research coordinator for the Clinic. During the course of psychotherapy, the research coordinator collected the first, third, and every fifth audiotaped session thereafter including the termination session. Since this also involved material offered by the therapist and the therapist would be required to fill out post-therapy questionnaires, voluntary written consent was 64 also obtained from the therapists in the project. If the therapist declined participation, the case was withdrawn from the research project. If there was a problem with any of the scheduled audiotaped sessions, the research coordinator asked for the next session to be included in the data library. If the termination of therapy was planned between therapist and patient, the research coordinator asked the therapist to give the patient a pre-stamped post-therapy research packet consisting of the HOpkins Symptom Checklist and a post-therapy client questionnaire. The instructions were to complete the forms and mail them back at the patient's convenience. If termination was unplanned or forms were simply forgotten, the post-therapy packet was mailed to the patient's residence with the same instructions. It was understood that all post-therapy data would remain confidential to all but the research coordinator and future researchers, including no access by the therapist involved. The latter criterion was applied to ensure honest self-reports and to keep the therapists blind from client reports. The therapists were also given a post-therapy research packet at termination consisting of a clinician's version of the Hopkins Symptom Checklist and a post—therapy therapist questionnaire. Stressing the importance of confidentiality, all patients were given code numbers that were used for all forms and audiotapes to ensure anonymity of data in future 65 use. Recognizing that voices from audiotapes may also be identified, a two year moratorium was placed on all data to help ensure that a researcher who may be a current acquaintance of a patient would have less availability to their data. The data collected through the above described process formed the data base from which all apprOpriate material was Obtained and analyzed in this study. Descriptions of data sampling, ratings, and measures used are found elsewhere in this chapter. In order to assess therapeutic alliance development and frequency of negative indicators over time, ratings from the first, middle, and final phases of therapy were combined and averaged across judges and across all cases within each treatment length group. A composite alliance and negative indicator score was thus obtained for each subscale, and appropiate statistical comparisons were made between and within groups for each subscale. Table 1 (see Appendix A) summarizes the data available for each group and for the combined group of all forty cases. Differences between therapeutic alliance and negative indicator subscales between groups will be discussed in the next chapter, although significant within group differences were not present. The only significant difference found between or within groups for SCL-90 symptom dimensions is a significantly higher level of interpersonal sensitivity in the long therapy length group as compared to 66 the short length group prior to beginning treatment. There were no significant differences between patient and therapist subjective ratings of outcome, although a trend toward higher outcome ratings as therapy length increases was noted. Patient and therapist gender data is included in Table l, with the long length group standing out as having a much larger distribution of male therapists. The impact of any gender differences between groups in this study, however, is unclear, and may be an area of focus in future research. Generally speaking, the minimal differences in raw data between groups in this study allow for comparison of groups as essentially equivalent. Chapter IV RESULTS Reliability of Measures For both the TARS and VNIS ratings, the judges rated three segments for each of the forty total cases bringing the total number of segments rated to 120. Interrater reliabilities were established for each of the measures based on the total sample and following the procedures recommended by Kiesler (1973) and used by Sandell (1981) and Hartley (1978). Interrater reliability measures the extent to which independent judges' ratings tend to covary, and all reliabilities in this study were assessed based on the Pearson product-moment correlation. The assessment of reliability for individual judges' ratings was the basic product-moment correlation established between each pair of judges across the 120 segments. The assessment of reliability for the average of the judges' ratings was obtained by applying the Spearman-Brown prOphecy formula to the individual Pearson product-moment correlations. The distinction between individual reliabilities and average reliabilities made here is especially noteworthy 67 68 considering the importance of each to this study. The individual reliability offers an estimate of the accurate consistency of either one of the judges' ratings if used alone. More importantly for this study, the average reliability offers an estimate of the accurate consistency of the two judges' ratings when used together. All statistics in this study were applied to the combined judges' ratings of the TARS and VNIS measures, thus the average reliabilities are the most pertinent of the two. The establishment of a level of acceptable reliability is a matter of some debate in the field of psychotherapy research, and is usually based on the precedent established within a specific area of study. For most recent psychotherapy research, level of acceptance has largely been based on the suggestions of Kraemer (1981) who emphasizes the necessity for realistic levels of reliability when dealing with clinical data. She argues that the 'moment-to-moment and day-to-day true lability of the subject's clinical condition,“ as well as the inherent subjectivity of measures will lower the achievable upper limit of reliabilities. In this regard, she suggests a reliability coefficient of .80 to be regarded as almost perfect, .60-.80 as satisfactory, .40-.60 as acceptable but possibly improvable, .20-.40 as demanding improvement, and below .20 as totally unacceptable and probably unimprovable. Based on this scale, a level of reliability of .60 in this study is considered clearly acceptable, while reliabilities 69 near this level may be considered cautiously acceptable. Table 2 (see Appendix A) lists the average and individual reliabilities for the TARS and VNIS subscales from the judges' ratings of the data in this study as outlined to this point. As can be seen, the average reliabilities generally fall well within the acceptable range used in this study. One subscale from the TARS, Therapist Positive Alliance, and two subscales from the VNIS, Errors in Technique and Global, fall below the .60 criterion, but are still within the cautiously acceptable range and will be included in all analyses with that understanding. Overall, the average reliability for the four TARS subscales of .65 and the six VNIS subscales of .70 is considered adequate. Comparing the reliability coefficients obtained in this study to previous research using the same measures suggests that higher reliabilities could possibly have been Obtained. Using the TARS, Marziali, Marmar, and Krupnick (1981) obtained a Finn's r composite reliability of .76 for their patient alliance subscales and .82 for their therapist alliance subscales. These compare to average reliabilities of .67 and .64 for patient and therapist subscales respectively obtained in the present study. Using the VNIS, Sandell (1981) Obtained an overall average reliability of .84 for her six subscales compared to a coefficient of .70 in the present study. It is difficult to assess what may account for these 70 differences, although a few possibilities exist. The period of time used for rater training in the present study was actually longer and involved more rated segments than either of the comparable studies, although real quality of training is difficult to assess and compare. Differences between the pairs of raters used in the studies may account for the greatest differences in reliability data. Comparable studies used highly trained (three of the four raters having their Ph.D.'s) and experienced raters chosen and paired for similarity of therapeutic orientation. Even though the raters in the present study were also paired as closely as possible for similarity of experience and orientation, clear differences in orientation existed between the raters in both pairs, and all four raters were less academically trained (none had their Ph.D.'s) and experienced than the raters in the comparable studies. While some may argue that there is an advantage in using raters who do not think alike, it is quite possible that this lack of similarity affects reliability. Given the basic differences in the judges between the studies, it may actually be considered rather remarkable that reliabilities were as high as they were in the present research. It must also be emphasized that even though the reliabilities are lower in this study, Marziali, Marmar, and KrlJpnick used 40 rated segments and Sandell used 54 as c<33fllpared to the 120 used in the present study. When comparing the power (the product of the sample size and 71 reliability coefficient) of the measured data used in the present study to the comparable studies, the present data have much higher power when used collectively. This is a very important point that further adds to the adequate reliability of the data in this study. Intercorrelations of TARS and VNIS Subscales Before reviewing the results of the predictions of this study, it is important to examine the relationships of the TARS and VNIS subscales to one another in these data. Previous research using the TARS (Marziali, Marmar, and Krupnick, 1981) had shown expected significant negative correlations between Therapist Positive and Negative subscales as well as Patient Positive and Negative subscales with no significant correlations between Therapist and Patient subscales. In contrast to these findings, subscales in the present study offer somewhat different interrelationships. Table 3 (see Appendix A) illustrates the intercorrelations of the TARS subscales in this study. As can be seen, all correlations are in the expected direction but represent far more interrelatedness than in previous research. Therapist Positive and Negative subscales do not have a significant negative correlation, and the Therapist and Patient subscales are significantly ifltercorrelated. The interrelationships of VNIS subscales have not been as closely examined in previous studies, but Table 4 (see 72 Appendix A) shows a similar pattern to that of the TARS subscales in this study. All subscales are highly intercorrelated in a positive direction representing strong interrelationships between scales. Obviously, the VNIS Total subscale should be significantly correlated with the other subscales since it is simply a composite scale. The internal consistencies of TARS and VNIS subscales have been established elsewhere (Hartley, 1978; Sandell, 1981), but even though the items in the subscales may form unified dimensions they may not necessarily represent strongly separate dimensions. In reality, it makes clinical sense that if a therapist is making a number of mistakes in therapy he or she is probably making them across a number of measurable categories, thus leading toward high intercorrelations between categories. It also makes sense that much of the behavior rated in psychotherapy is often difficult to assess into subscales of larger dimensions, with ratings of one scale influencing the ratings of another and overlap existing between how scales are perceived for rating certain types of clinical behavior. For these reasons, the TARS and VNIS subscale intercorrelations seen in Tables 3 and 4 suggest caution in interpreting these subscales as absolutely separate dimensions in this study. Effiiging the Hypotheses As a brief summary, the following hypotheses were Predicted in this study: 73 1) Positive therapeutic alliance was expected to be higher by termination for longer length therapy as compared to shorter length therapy with medium length therapy falling in the trend between these two groups. Within each group, a trend toward higher positive alliance and lower negative alliance was expected as therapy progressed. The groups were not expected to differ significantly on level of positive or negative alliance at the beginning of therapy, although a trend toward lower positive alliance and higher negative alliance ratings was expected for the shorter length group. 2) Negative indicators were not expected to be significantly different at any phase between groups, although a trend toward lower negative indicators as therapy length increased was expected. Within group differences were not expected across phase of therapy, although a trend toward lower negative indicators as therapy progressed was predicted. 3) A relationship between therapeutic alliance and negative indicators was predicted. It was expected that positive alliance subscales would be negatively correlated with negative indicator subscales, while negative alliance subscales would be positively correlated with negative indicator subscales. Significant relationships were expected across all three groups. 4) It was predicted that therapeutic alliance and negative therapy indicators would be related significantly 74 to the outcome of psychotherapy across all three outcome perspectives (patient, therapist, and change in patient symptomatology). Positive alliance subscales were expected to be positively related, while negative alliance subscales and negative indicators were expected to be negatively related. These relationships were predicted to exist across all three therapy length groups. 5) The pretreatment symptom dimensions most affecting the formation of interpersonal relationships (in this study the SCL-90 factors of interpersonal sensitivity, hostility, and paranoid ideation) were expected to be negatively related to positive alliance but positively related to negative alliance and negative indicators. Hypothesis 1: Two-way analyses of variance were performed in testing part of the first hypothesis with therapy length and treatment phase used as the two independent variables. As depicted in Table 5, there were no interaction effects within the four TARS subscales, and only one main effect for length in the Therapist Positive subscale. Considering that no main effects for phase and no interaction effects were discovered, the first part of hypothesis 1 is not supported. Positive alliance is not greater (i.e., statistically significant) at the end phase for longer length therapy as compared to shorter length therapy. As can be seen from Table l, the predicted trend toward higher positive alliance at the end of therapy as treatment 75 Table S TARS Phase x Length Analysis of Variance T Positive Alliance Phase Length P x L T Negative Alliance Phase Length P x L P Positive Alliance Phase Length P x L P Negative Alliance Phase Length P x L **p < .01 DF hNN #NN SS 42.25 143.53 46.08 130.20 95.55 1.03 48.33 20.76 14.01 75.43 33.42 72.86 MS 21.13 71.77 11.52 65.10 47.78 0.26 24.17 10.38 3.50 37.72 16.71 18.22 1.41 4.80 0.77 2.93 2.15 0.01 0.92 0.40 0.13 1.62 0.72 0.78 .25 .01** .55 .06 .12 .99 .40 .67 .97 .20 .49 .54 76 lengthens is observed for the Therapist Positive subscale, but not for the Patient Positive subscale. T-test comparisons of phases within groups also show no significant within-group differences, and the only consistent trend observed is that negative alliance ratings tend to increase as therapy progresses. It was also predicted that differences in positive alliance at the beginning of therapy would not be Observed. Student-Newman-Keuls comparisons of groups for the one main effect of Therapist Positive Alliance found the ratings for this subscale to be significantly higher (2, 3.83, p < .03) for the medium length group as compared to the shorter therapy length group in the beginning phase of therapy. There were no significant differences in comparisons with the longer treatment length group, although ratings for the short therapy length group remained the lowest. This finding cautiously suggests that therapy which may terminate sooner (without the intention of short-term therapy) may be marked by some problems in the therapist's contribution to positive alliance. Hypothesis 2: Two-way analyses of variance using therapy length and treatment phase as the independent variables were also used to test significant differences between negative indicator subscales. As with the therapeutic alliance analyses, Table 6 shows no main effects for treatment phase across the subscales and no interaction effects. These analyses support the predicted hypothesis VNIS Phase x Length Analysis of Variance 77 Table 6 or 55 MS F 2 Patient Qualities Phase 2 16.01 8.01 1.35 .27 Length 2 29.22 14.61 2.46 .09 P x L 4 23.57 5.89 0.99 .42 Therapistggualities Phase 2 4.29 2.14 0.30 .74 Length 2 41.99 21.00 2.92 .06 P x L 4 22.89 5.72 0.80 .53 Errors in Technique Phase 2 2.65 1.33 0.32 .73 Length 2 29.51 14.76 3.50 .03* P x L 4 19.95 4.99 1.18 .32 P-T Interaction Phase 2 2.22 1.11 0.97 .38 Length 2 2.87 1.44 1.25 .29 P x L 4 5.37 1.34 1.17 .33 Global Phase 2 5.20 2.60 0.69 .51 Length 2 8.32 4.16 1.11 .34 P x L 4 20.08 5.02 1.33 .27 Total Phase 2 102.93 51.47 0.79 .45 Length 2 298.55 149.27 2.30 .11 P x L 4 231.73 57.93 0.89 .47 *p < .05 78 that significant phase differences would not be evident in these data. As can be observed from Table 1, there are no consistent within-group trends toward lower negative indicators as therapy progresses. The only significant result in the analyses of variance is a main effect for therapy length in the Errors in Technique subscale. Student-Newman-Keuls comparisons between groups found the ratings for this subscale to be significantly higher (2, 5.68, p < .01) at the mid-phase of therapy for the shorter length group as compared to both the medium and long length groups. Considering that no other consistent trends are observed possibly suggesting higher rates of negative indicators in the shorter length group, and considering the cautious reliability of the Errors in Technique subscale, this finding is considered interesting but most probably spurious. In fact, the longer length group tends to have higher negative indicator ratings than the short length group, although these differences are insignificant. Hypothesis 3: As Table 7 indicates, negative indicator subscales are correlated with therapeutic alliance subscales in the expected direction. In looking at the overall results with all forty cases combined, it appears that hypothesis 3 is supported with a strong relationship observed between the establishment of a therapeutic alliance in psychotherapy and the level of negative process indicators existing in therapy. These correlations are 79 Table 7 Correlations Between TARS and VNIS Subscales Combined Group (N=40) Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance P Qualities -.21 .41** —.73*** .62*** T Qualities -.15 .73*** -.53*** .54*** Errors Tech. -.57*** .41** -.70*** .52*** P-T Inter. -.30* .74*** -.70*** .71*** Global -.31* .61*** -.62*** .51*** Total -.37** .69*** -.78*** .66*** Long Length Group P Qualities -.46 .55* -.82*** .73*** T Qualities -.28 .94*** -.61* .82*** Errors Tech. -.58* .86*** -.73** .81*** P-T Inter. -.61* .93*** -.88*** .94*** Global -.34 .81*** -.65** .78*** Total -.49* .94*** - 80*** .91*** Medium Length Group P Qualities -.17 .46* -.80*** .67** T Qualities -.40 .46* -.79*** .54* Errors Tech. -.26 .07 -.73*** .64** P-T Inter. -.25 .48* -.84*** .74*** Global -.37 .45* -.7l** .49* Total -.39 .46* -.88*** .64** Short Length Group P Qualities .06 .30 -.6l** .67** T Qualities -.04 .27 -.17 .04 Errors Tech. -.59* .35 -.55* .21 P-T Inter. -.36 .57* -.35 .25 Global -.30 .40 -.44 .11 Total -.25 .47* -.66** .48* *p < .05, **p < .01, ***p < .001 80 weaker for the Therapist Positive Alliance subscale, suggesting the possibility that the therapist's positive contribution to therapeutic alliance is less affected by negative process variables. Analyzing these results more closely, we observe that the relationships seen in the combined data are not consistent across therapy length groups. In fact, there seems to be a clear progression with the strength of the predicted relationships increasing as the overall length of therapy increases. The short length group shows some relationship between alliance and negative indicators, but the relationship is much weaker when compared to the medium length group. In turn, the relationship in the medium length group is somewhat weaker than in the long length group. Given these results, the obvious question that arises is whether therapeutic alliance is more affected by negative indicators as the course of psychotherapy progresses. Table 8 represents an a posteriori analysis of the data breaking the correlations down into each treatment phase within each group. As can be observed across all three groups, there is less of a relationship between TARS and VNIS subscales in the early phase of therapy, with only the therapist's negative contribution to alliance consistently increasing as negative indicators increase in the medium and long length groups. It is not Observed, however, that the strength of 81 Table 8 Correlations Between TARS and VNIS Subscales by Treatment Phase Long Length Group (N=12) Beginning Phase: Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance P Qualities -.31 .44 -.76** .03 T Qualities -.10 .78** -.22 .42 Errors Tech. -.28 .78** -.30 .48 P"T Inter. -027 075** ‘03“ 051 GIObal -014 071** -036 053 TOtal -026 087*** -046 .46 Middle Phase: P Qualities ‘023 0““ -032 .00 T Qualities -.54 .86*** -.65* .88*** Errors Tech. -.7l** .80** -.58* .76** P-T Inter. -07g** 072** -066* 086*** GlObal -034 061* -052 072** Total -.62* .82*** -.7l** .86*** Final Phase: P Qualities -.38 .59* — 82*** .82*** T Qualities -029 094*** -057* 056 Errors TECh. -044 074** -046 050 P-T Inter. -041 083*** -074** .73** GlObal -048 085*** .055* 05“ Total -.43 .88*** -.70** .70** Medium Length Group (N=15) Beginning Phase: P Qualities -038 052* -050 023 T Qualities _026 081*** -02“ .01 Errors TECH. -047 093 -009 .01 P-T Inter. -003 027 -035 054* GIObal -035 065** -027 .14 TOtal -042 064** -039 018* Middle Phase: P Qualities T Qualities Errors Tech. P-T Inter. Global Total Final Phase: P Qualities T Qualities Errors Tech. P-T Inter. Global Total 82 Table 8 (cont'd.). Therapist Therapist Patient Positive Negative Positive Alliance Alliance Alliance -.23 .64** -.47 -.12 .36 -.72** -.25 .25 -.80*** -.04 .45 -.57* -.38 .10 -.70** -.21 .42 -.76*** .03 -.03 -.27 -.18 .28 -.57* -.21 -.23 .04 -.40 .38 -.54* -.42 -.10 -.24 -.28 .02 -.37 Short Length Group (N=13) Beginning Phase: P Qualities T Qualities Errors Tech. P-T Inter. Global Total Middle Phase: P Qualities T Qualities Errors Tech. P-T Inter. Global Total Final Phase: P Qualities T Qualities Errors Tech. -0g5 .21 -053 -.06 -022 .23 -.40 -.54 -.26 -.28 -.37 -016 -062* .03 -041 -022 -016 -015 -068** 017 -028 -004 -063* 062* -053 064* -018 .40 .04 08g*** -039 093 -023 .72** -032 011 -072** 075** -006 -.14 -.49 Patient Negative Alliance .77*** .74** .72** .33*** .45 .82*** .22 .73** .02 .47 -.06 .33 .29 -054 -015 .04 .06 .64* .22 .01 .41 .10 .34 .45 .27 .04 83 Table 8 (cont'd.). Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance P-T Inter. -.55* .76** -.12 .33 Global -.34 .17 -.40 -.11 Total -.65* .40 -.66* .30 *p < .05, **p < .01, ***p < .001 84 the TARS and VNIS correlations continue to increase over the course of therapy. There is some hint of this progression in the short length group, but the medium and long length groups seem to achieve a peak in the middle phase of therapy, with an obvious decline in the final phase for the medium length group and little change from the middle phase to the final phase in the long length group. When looking at the average length of therapy at each phase across the treatment length groups, it appears that a stronger relationship between therapeutic alliance and negative indicators exists by about the 15-20 session point in therapy, with the strength of the relationship becoming more variable after this point. Hypothesis 4: The relationship between therapeutic alliance and negative indicators with the outcome of psychotherapy was assessed in a variety of ways. The therapists' and clients' subjective ratings of outcome were taken from their respective post-therapy questionnaires, and Pearson product-moment correlations were obtained between these perspectives on outcome and TARS and VNIS overall subscale ratings across groups. Tables 9 and 10 show the results of these analyses. No consistent relationships were observed between either outcome perspective and TARS and VNIS subscales when cases were combined, although two interesting results were noted within groups. The negative indicator subscale of Patient Qualities was significantly associated with both therapists' 85 Table 9 Correlations Between TARS and Therapists' Patients' Perspectives on Outcome Combined Group (N=40) Therapist Therapist Positive Negative Alliance Alliance Therapist Outcome .11 .18 Patient Outcome .22 -.04 Long Length Group (N=12) Therapist Outcome -.09 .60* Patient Outcome -.25 .22 Medium Length Group (N=15) Therapist Outcome .36 -.09 Patient Outcome .17 -.31 Short Length Group (N=13) Therapist Outcome -.01 -.14 Patient Outcome .38 -.36 *p < .05 Patient Positive Alliance -0“? .13 -048 .21 .44* .23 -0“? .02 Patient Negative Alliance .01 -016 .58* -.06 -024 -023 ‘023 -034 86 Table 10 Correlations Between VNIS and Therapists' and Patients' Perspectives on Outcome Combined Group (N=40) Patient Therapist Errors P-T Global Qual. Qual. Tech. Inter. Ther. out. -023 034* .01 .24 014 Pat. OUt. -026 013 -099 -001 _OG8 LongyLength Group (N=12) Ther. Out. .15 .73** .33 .55* .62* Pat. OUt. -008 021 037 015 002 Medium Length Group (N=15) There out. 017 -014 -0ge -017 -043* Pat. out. -034 -027 -014 ‘045* -029 Short Length Group (N=13) Ther. Out. -.56* .15 -.05 .20 .05 Pat. onto .05“. 009 -035 -061 -018 *p < .05, **p < .01 Total .10 -093 .55* .20 -024 -020 -024 -038 87 and clients' assessments of outcome in the short length group. It appears that as negative patient quality variables increase, therapists and clients observe therapy outcome to be worse in shorter length therapy. The other interesting result was the discovery that negative indicators and negative alliance subscales were significantly correlated in a positive direction with the therapists' perspective on outcome in the long length group. This suggests that as negative process variables increased and negative alliance increased the therapists in the longer length therapy viewed outcome as being better. This is a very odd finding that will be further discussed in the next chapter, but may suggest that therapists in part view patient improvement through the encounter and resolution of negative events. Another method of assessing psychotherapy outcome rather than simply asking therapists and clients their subjective Opinions on whether therapeutic goals were achieved was to Obtain pre- and post-therapy symptom checklists from the clients and look at symptom change ratings. Initial analyses using these data are shown in Tables 11 and 12. Pearson product-moment correlations were obtained between absolute symptom change (the actual amount a symptom changed in either direction) for each of the ten SCL-90 symptom dimensions and the TARS and VNIS subscales. As can be seen, there are few significant relationships between these variables and many of the significant 88 Table 11 Correlations Between TARS and Absolute Symptom Change Combined Group (N=40) Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance Somatization -.05 .09 .12 -.17 Obsessive-Comp. -.17 -.09 .30* -.38** Interpersonal Sen. -.03 -.13 .20 -.19 Depression -.24 .00 .15 -.34* Anxiety -.22 -.10 .02 -.20 Hostility -.06 .07 .06 .02 Phobic Anxiety -.20 -.08 .23 -.35** Paranoia .12 .10 .06 -.06 Psychoticism -.07 .11 .14 -.23 GPI -.16 -.03 .19 -.29* Long Length Group (N=12) Somatization -.06 .44 .01 .15 Obsessive-Comp. -.26 -.02 .33 -.28 Interpersonal Sen. -.05 .13 .10 -.02 Depression -.27 -.05 .ll -.16 Anxiety -.11 .01 —.04 .01 Hostility .08 .22 .17 -.00 Phobic Anxiety -.l4 -.14 .33 -.35 Paranoia .07 .30 -.04 .17 Psychoticism -.20 .24 .13 -.04 GPI -.15 .09 .17 -.12 Medium Length Group (N=15) Somatization .12 .21 .34 -.40 Obsessive-Comp. .10 .24 .41 -.52* Interpersonal Sen. .36 .12 .29 -.25 Depression .05 .12 .35 -.6l** Anxiety .12 .04 .27 -.45* Hostility .51* .ll .22 .00 Phobic Anxiety .05 .10 .38 -.S6* Paranoia .30 .ll .34 -.37 Psychoticism .24 .18 .19 -.42 GPI 023 015 041 -058** Short Length Group (N=13) Therapist Positive Alliance Somatization -.05 Obsessive-Comp. -.04 Interpersonal Sen. .00 Depression -.30 Anxiety -.35 Hostility -.38 Phobic Anxiety -.38 Paranoia .31 Psychoticism -.03 GPI -.14 *p < .05, **p < .01 Table 11 (cont'd.). 89 Therapist Negative Alliance -.43 -.32 -.39 .17 -.19 -.18 -.21 -004 -.02 -.14 Patient Positive Alliance .14 .37 .31 .19 -.07 -.16 .08 -.01 .14 .19 Patient Negative Alliance -.20 -.23 -.16 .11 .11 .21 -.11 .23 -.06 .05 Correlations Between VNIS and 90 Table 12 Absolute Symptom Change Combined Group (N=40) Pat. Qual. Somatization -.13 Obsessive-Comp. —.24 Inter. Sen. -.24 Depression -.09 Anxiety —.08 Hostility .02 Phobic Anxiety -.31* Paranoia .14 Psychoticism .02 GPI -.13 Ther. Qual. .08 -.18 -.18 -.17 -.24 -.05 -.23 .03 .03 -.17 Long Lengph Group (N=12) Somatization .07 Obsessive-Comp. -.21 Inter. Sen. -.04 Depression -.04 Anxiety .23 Hostility .00 Phobic Anxiety -.20 Paranoia .29 Psychoticism .09 GPI -.00 .33 -.15 .06 -.22 -.21 .00 -.30 .22 .10 -.09 Medium Length Group (N=15) Somatization —.23 Obsessive-Comp. —.37 Inter. Sen. -.l7 Depression -.16 Anxiety -.25 Hostility .04 Phobic Anxiety -.30 Paranoia -.10 Psychoticism -.06 GPI -.23 -029 -.26 -.39 -.17 -.36 -023 -026 -.13 .11 -.27 Errors Tech. -.09 -.16 -.21 -.08 -.07 .02 -.22 -.11 .00 -.15 .50* .16 .32 .23 .25 -.02 -.05 .31 .37 .24 -.60** _.66** -.64** -.56** “048* -011 -OS3* -038 -035 -.63** P-T Inter. .06 .28* .26* .22 .22 .06 .26 -006 .03 .23 .30 .06 .04 .03 .05 .05 .16 .20 .17 .03 Global -.07 -.21 -.25 -.17 -030* -018 .030* .02 -0gl -023 .16 -.35 -.07 -.25 -.27 -.16 -.46 .36 .02 -.20 -.15 -.08 -.13 -.14 -.33 -.38 -.20 -.13 .10 -.21 Total -092 -.20 -.24 -012 -.17 -.02 -.28* .04 .05 -.17 .34 -.12 .09 -.08 -.00 .00 .25 .33 .20 .01 -030 ‘03“ -022 -034 -017 -021 .00 -.31 91 Table 12 (cont'd.). Short Length Group (N=13) Pat. Ther. Errors P-T Global Total Qual. Qual. Tech. Inter. Somatization -.30 -.06 -.43 -.31 -.38 -.41 Obsessive-Comp. -.24 .05 -.27 -.23 -.23 -.26 Inter. Sen. -.35 -.17 -.41 -.38 -.59* -.51* Depression -.l9 -.13 -.07 -.14 -.30 -.22 Anxiety -.27 -.04 -.19 -.11 -.44 -.31 Hostility. -.11 -.03 -.14 -.04 -.30 -.16 Phobic Anxiety -.44 -.25 -.19 -.22 -.36 -.45 Paranoia .14 -.14 -.54* -.29 -.43 -.24 Psychoticism .06 -.02 -.18 -.20 -.16 -.08 GPI -.24 -.09 -.38 -.24 -.42 -.37 *p < .05, **p < .01 92 correlations scattered throughout the data may be spurious given the number of correlations involved in the analyses. Two patterns of significant results did emerge, however, both involving the medium length group and both representing unexpected findings. In the TARS analyses, the Patient Negative Alliance subscale was significantly related to a number of symptom dimensions in the direction suggesting that as the patients' negative contributions to therapeutic alliance increased absolute symptom change increased. The strength of this relationship in the medium length group also carried over to the combined data analyses. In the VNIS analyses, the Errors in Technique subscale was significantly related to a number of symptom dimensions in the medium length group also in the direction suggesting that as the frequency and intensity of technique errors increased absolute symptom change increased. These clearly represent contrasting findings than those predicted by hypothesis 4, but are also weak given their specificity to a single group and single subscales. One of the problems with using absolute symptom change for the outcome analyses in this study is that it does not account for the level of symptomatology at the beginning of therapy. Cases in which symptoms are rated higher at the beginning of therapy will be weighted more heavily since there is a greater range for symptom ratings to change. With few cases in each group, correlational analyses using absolute change can be biased or simply uninformative. One 93 way of accounting for this problem is to transform the data into percent change data in which absolute change ratings are represented as a percentage of the beginning symptom ratings. Analyses involving these data are represented in Tables 13 and 14. The previous unexpected findings in the medium length group were not upheld for the VNIS subscale Errors in Technique, but the Patient Negative Alliance subscale from the TARS still suggests the same relationship, albeit somewhat weaker, to symptomatology as in the previous analyses. The only truly consistent finding in these analyses was the relationship between the symptom dimension of somatization and both TARS and VNIS subscales within the long length group. These correlations were in the expected direction, suggesting that as positive alliance increases and negative alliance decreases, somatization improves (declines). Likewise, as negative process indicators decline somatization improves. Once again, these findings were only within the long treatment length group, and no other expected relationships between therapeutic alliance, negative indicators and symptom change were evident using percent change data. Considering the fact that a number of predictor variables of symptom change were involved in the preceding analyses, including the possibility of pretreatment symptom levels and therapy length, all interacting in unknown ways, stepwise multiple regression analysis was also chosen as a 94 Table 13 Correlations Between TARS and Percentage of Symptom Change from Pretreatment Levels Combined Group (N=40) Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance Somatization .23 .05 .28* .02 Obsessive-Comp. .23 .14 -.12 .32* Interpersonal Sen. .04 .14 -.04 .19 Depression .07 .ll -.16 .30 Anxiety .23 .04 .18 -.01 Hostility -.ll .04 -.18 .08 Phobic Anxiety .23 .16 .09 .18 Paranoia -.17 -.06 .06 -.07 Psychoticism .08 .03 .02 .19 GPI .20 .ll -.01 .19 Long Length Group (N=12) Somatization .10 -.59* .49* -.57* Obsessive—Comp. .19 -.06 -.22 .16 Interpersonal Sen. .06 .06 .16 -.07 Depression -.10 .24 -.14 .20 Anxiety .25 -.36 .53* -.48 Hostility -.25 .06 -.28 .15 Phobic Anxiety .15 .27 -.24 .27 Paranoia -.21 -.33 .08 .24 Psychoticism .22 -.34 .46 -.41 GPI .01 -.05 .02 -.02 Medium Length Group (N=15) Somatization .17 -.26 -.15 .36 Obsessive-Comp. -.00 -.18 -.18 .43 Interpersonal Sen. —.22 -.09 -.18 .37 Depression -.27 .07 -.30 .53* Anxiety .04 -.20 .07 .16 Hostility -.28 .04 -.24 .17 Phobic Anxiety .24 -.24 .13 .30 Paranoia -.46* .01 -.33 .30 Psychoticism -.50* .19 -.25 .44* GPI -031 002 -032 054* 95 Table 13 (cont'd.). Short Length Group (N=13) Therapist Therapist Patient Patient Positive Negative Positive Negative Alliance Alliance Alliance Alliance Somatization .21 .36 .37 .05 Obsessive-Comp. .24 .36 -.21 .31 Interpersonal Sen. -.01 .27 -.02 .00 Depression .25 -.18 -.l3 -.l3 Anxiety .16 .26 .07 -.06 Hostility .29 .25 .08 -.05 Phobic Anxiety .16 .22 .25 .04 Paranoia -.17 .03 .32 -.44 Psychoticism .52* -.17 .16 -.05 GPI .31 .06 .12 -.19 *p < .05 96 Table 14 Correlations Between VNIS and Percentage of Symptom Change from Pretreatment Levels Combined Group (N=40) Pat. Ther. Errors P-T Global Total Qual. Qual. Tech. Inter. somatization -067 '003 -008 063 -0g6 ‘0“7 Obsessive-Comp. .16 .05 -.05 .13 -.03 .05 Inter. Sen. .15 .14 .15 .17 .13 .16 Depression .12 .14 .07 .16 .09 .12 Anxiety .01 .04 -.03 .04 .13 .02 Hostility .ll .18 .20 .17 .27* .21 Phobic Anxiety .17 .18 .08 .19 .14 .14 Paranoia -.08 -.01 .12 .09 .01 .00 Psychoticism -.12 .06 -.04 .02 .02 -.03 GPI .06 .14 .09 .18 .19 .13 Long Length Group (N=12) Somatization -.60* -.48 -.62* .55* .56* -.64** Obsessive-Comp. .26 -.00 -.23 .00 .23 .04 Inter. Sen. -.22 .24 -.28 .03 .25 .01 Depression .10 .27 -.14 .20 .19 .15 Anxiety -.73** -.13 -.51* .47 .21 -.44 Hostility .21 .22 .28 .21 .35 .28 Phobic Anxiety .39 .35 .25 .24 .50* .40 Paranoia -.14 -.29 -.23 .17 .34 -.28 Psychoticism -.65** -.13 -.51* .45 .18 -.41 GPI -.10 .15 -.23 .04 .22 .01 Medium Length Group (N=lS) Somatization .11 .06 .51* .16 .06 .16 Obsessive-Comp. .06 -.02 .37 .09 .30 -.02 Inter. Sen. -.08 .12 .40 .17 .09 .07 Depression -.02 .10 .34 .18 .03 .09 Anxiety -.02 .02 .18 .12 .09 .02 Hostility .05 .40 .14 .19 .36 .24 Phobic Anxiety -.16 -.30 .14 .03 .33 -.23 Paranoia .06 .12 .32 .20 .02 .13 Psychoticism -.04 .02 .23 .13 .00 .04 GPI -.03 .18 .42 .27 .15 .18 97 Table 14 (cont'd.). Short Length Group (N=13) Pat. Ther. Errors P-T Global Total Qual. Qual. Tech. Inter. Somatization .07 .17 .08 .27 .13 .15 Obsessive-Comp. .23 -.08 -.16 .14 .05 .08 Inter. Sen. .32 .03 .27 .21 .39 .35 Depression .24 -.16 -.03 -.09 .18 .ll Anxiety .28 .02 .20 .16 .47* .32 Hostility .15 .30 .31 .28 .37 .34 Phobic Anxiety .18 .25 .13 .30 .21 .27 Paranoia -.ll .30 .39 .33 .50* .27 Psychoticism -.ll .21 -.23 .14 .06 -.04 GPI .15 .02 .17 .10 .37 .22 *p < .05, **p < .01 98 way to represent the possible impact of therapeutic alliance and negative indicators on psychotherapy outcome. Through the use of semi-partial correlations, stepwise multiple regression allows one to statistically control for the influence of preceding predictor variables and demonstrate the amount of variance uniquely accounted for by the 'kth' predictor at the 'kth' step of the analysis (Schaffer and Abeles, 1982; Nie et a1., 1975). The first step of the regression equation uses the predictor variable with the largest correlation to the criterion variable. The predictor with the second largest correlation is entered onto the second step, once the variance accounted for by the first predictor has been removed from the second predictor but not from the criterion variable. This process continues for as many steps as defined by the analysis. The analysis used in this study included all of the subscales from the TARS and VNIS, therapy length, and pretreatment symptom ratings as predictor variables for posttreatment symptom ratings. Table 15 represents the stepwise multiple regression analysis for each of the ten SCL-90 symptom dimensions for which posttreatment ratings were obtained in this study. Included in the table is the stepwise list of predictors included in the longest significant regression equation for each symptom dimension. For each predictor, R2 change represents the amount of variance of the criterion uniquely accounted for by that particular step, and whether this represents a significant 99 Table 15 Significant Multiple Regression Equations for the Prediction of Posttreatment Symptomatology Somatization Step Variable R2 ch. Sim. r II. T Neg. Alliance .04 -.24 III. VNIS Total .03 -.01 IV. P-T Interaction .04 -.16 Sum a: .24 Adj R .15 F 2.76* Obsessive-Compulsive Step Variable R2 ch. Sim. r I. Global .04 .21 II. P Neg. Alliance .06 -.10 III. Pre-Depression .03 .14 IV. Pre-Somatization .l0* -.14 V. T Pos. Alliance .04 -.15 VI. P-T Interaction .04 -.00 VII. T Qualities .04 .21 VIII. P Qualities 003 -032 Sum R: .38 Adj R .22 F 2.34* Interpersonal Sensitivity Step Variable R2 ch. Sim. I. Pre-Inter. Sen. .26*** .51 II. Pre-Anxiety .10* -.02 III. Pre-Paranoia .06 .51 IV. Pre-Obsessive-Comp. .05 .15 V0 PIG-GPI .02 034 VI. Pre-Somatization .03 .15 VII. P Neg. Alliance .01 .26 VIII. Pre-Depression .01 .28 IX. T Neg. Alliance .02 .02 X. Pre-Phobic Anxiety .03 .04 XI. Pre-Psychoticism .01 .35 XII. Errors in Technique .01 .07 XIII. P Pos. Alliance .01 -.32 XIV. Pre-Hostility .01 .36 Step XV. XVI. XVII. XVIII. Depression Step I. II. III. IV. V. VI. VII. VIII. IX. x. XI. XII. Anxiety Step II. III. IV. V. VI. VII. VIII. IX. X. XI. 100 Table 15 (cont'd.) Variable T Pos. Alliance VNIS Total Length Global Sum R; Adj R F Variable Pre—Hostility Pre-Phobic Anxiety Pre-Depression Pre-Obsessive-Comp. T Pos. Alliance P Qualities Pre-Anxiety P-T Interaction T Neg. Alliance Global VNIS Total Pre-Paranoia Sum R; Adj R F Variable Pre-Hostility T Pos. Alliance Length Pre-Anxiety Global T Neg. Alliance Pre-Phobic Anxiety VNIS Total Errors in Technique P-T Interaction Pre-Depression NN Sum R Adj R F 2 R ch. .01 .01 .00 10.9 .65 .38 2.21* R ch. .19** .96 .05 .06 .03 .01 .01 .02 .01 .01 .02 It; .49 .27 2.19* R ch. .14* .12* .04 .04 .03 .05 .01 Cal .02 .01 It; .47 .26 2.25* Sim. .04 .11 .07 .01 Sim. .43 -.16 .26 -.08 -.06 -.02 .02 -.04 -.01 .00 .00 .11 Sim. .37 -.31 -.24 .26 .08 .05 .15 .07 .11 .04 .11 r. r I 101 Table 15 (cont'd.). Hostility Step Variable R2 ch. Sim. r I. Pre-Anxiety .12* -.34 II. Pre-Hostility .18** .31 III. Pre-Inter. Sen. .02 .04 IV. Pre-Obsessive-Comp. .06 -.31 V. T Qualities .02 -.00 VI. P-T Interaction .02 -.06 VII. Pre-Phobic Anxiety .01 -.31 VIII. P Neg. Alliance .01 -.07 IX. T Neg. Alliance .01 -.03 X. Pre-GPI .01 -.17 XI. Global :01 -.01 Sum R: .47 Adj R .27 F 2.30* Phobic Anxiety Step Variable R2 ch. Sim. r I. Pre-Phobic Anxiety .37*** .61 II. Pre-Obsessive-Comp. .09* .09 III. T Pos. Alliance .07* -.32 IV. Errors in Technique .04 .11 V. Pre-Inter. Sen. .01 .07 VI. Pre-Depression .01 .06 VII. Length .01 —.15 VIII. P-T Interaction .01 -.03 IX. P Pos. Alliance .03 -.21 X. P Qualities .01 .09 XI. Pre-Psychoticism .01 .14 XII. Pre-Paranoia .01 .09 XIII. Pre-Hostility .01 .09 XIV. Pre-Anxiety .01 .32 XV. Pre-GPI .02 .20 XVI. P Neg. Alliance .01 .12 XVII. T Qualities .00 -.12 XVIII. VNIS Total .00 .00 XIX. Global .01 .03 XX. T Neg. Alliance LEE -.06 Sum R; .71 Adj R .40 F 2.32* 102 Table 15 (cont'd.). Paranoia Step Variable R2 ch. Sim. r I. Pre-Inter. Sen. .29*** .54 II. P Qualities .l4** .54 III. Pre-Hostility .14** .42 IV. T Pos. Alliance .04 .15 V. P Pos. Alliance .04 -.53 VI. Errors in Technique .03 .15 VII. Pre-Phobic Anxiety .01 .27 VIII. Pre-Anxiety .01 .26 IX. Global .01 .28 X. P-T Interaction .00 .36 XI. P Neg. Alliance .01 .40 XII. Pre-Obsessive-Comp. .00 .35 XIII. Pre-GPI .00 .48 XIV. Pre-Depression .00 .36 XV. Pre-Psychoticism .01 .42 XVI. Pre-Somatization .01 .28 XVII. Pre-Paranoia .00 .51 XVIII. T Neg. Alliance :2! .24 Sum R3 .71 Adj R .47 F 2.90** Psychoticism Step Variable R2 ch. Sim. r I. P Pos. Alliance .28*** -.53 II. Pre-Hostility .09* .33 III. Errors in Technique .05 .18 IV. P Neg. Alliance .04 .23 V. Pre-Depression .03 .31 VI. Pre-Obsessive-Comp. ,05 .15 VII. Pre-Inter. Sen. .05 .43 VIII. VNIS Total .04 .33 IX. Pre-Psychoticism .02 .37 X. Pre-Paranoia .03 .31 XI. Global .01 .28 XII. T Qualities .01 .17 XIII. Pre-Phobic Anxiety .01 .07 XIV. Pre-Anxiety .02 .15 XV. Pre-Somatization .01 .15 XVI. T Pos. Alliance .00 -.07 XVII. Length .00 .12 XVIII. P Qualities .00 .41 XIX. P-T Interaction .00 .26 XX. T Neg. Alliance L22 .13 103 Table 15 (cont'd.). Sum R2 2 .73 Adj R .46 F 2.63* Global Pathology Index Step Variable R2 ch. Sim. r I. Pre-Hostility .23** .48 II. P Qualities .03 .15 III. Pre-Phobic Anxiety .04 -.02 IV. Pre-Inter. Sen. .03 .30 V. T Pos. Alliance .03 -.13 VI. P-T Interaction .03 .03 VII. Pre-Paranoia .02 .23 VIII. Pre-Anxiety .02 .10 IX. Pre-Depression .01 .23 X. Pre-Obsessive-Comp. .02 .04 XI. P Pos. Alliance pygl -.20 Sum R3 .46 Adj R .25 F 2.16* *p < .05, **p < .01, ***p < .001 104 amount of additional variance. The simple r for each predictor provides information about the direction of the relationship to the criterion variable. Each analysis also includes the total amount of variance accounted for by the 2 2 given predictor variables (sum R change), the adjusted R which corrects for the inflation of R2 in the sample as compared to the overall population, and the overall F-test for the given regression analysis. As can be seen, the majority of significant predictor variables for posttreatment symptom ratings were pretreatment symptom ratings. These were not always the pretreatment ratings for the same symptom criterion, but it is generally clear from these analyses that the higher the pretreatment symptomatology the higher the posttreatment symptomatology. Interestingly, therapeutic alliance predictors were significant for three symptom dimensions as Opposed to only one negative indicator predictor. Therapist Positive Alliance significantly predicted lower posttreatment symptomatology for the symptom dimensions of anxiety and phobic anxiety. Patient Positive Alliance significantly predicted lower posttreatment symptomatology for the symptom dimension of psychoticism. From the negative indicator subscales, an increase in Patient Qualities ratings (more negative patient quality process variables) significantly predicted an increase in posttreatment symptomatology for the symptom dimension of paranoia. Interestingly, the pretreatment symptom dimension 105 of hostility was the only significant predictor of overall symptomatology (posttreatment ratings of GPI) at the end of therapy. Table 16 summarizes the results of these analyses for the TARS and VNIS subscales, including the number of times each subscale appears as part of a significant multiple regression equation and the number of times each subscale appears as a significant predictor variable within the equations. Hypothesis 5: The relationship between therapeutic alliance, negative indicators, and pretreatment symptomatology was tested using Pearson product-moment correlations between each of the TARS and VNIS subscales and pretreatment symptom ratings from the SCL-90. Tables 17 and 18 represent these analyses. As is seen in the combined group analyses, the majority of symptom dimensions show an expected relationship to the patient subscales from the TARS, but no relationship to the therapist subscales. These analyses suggest that the higher the patient's symptomatology upon entering therapy the lower the patient's positive contribution and the higher the patient's negative contribution to therapeutic alliance. The significant results include the predicted relationships between the symptom dimensions of interpersonal sensitivity and paranoia with TARS patient subscales. In contrast to the extensive relationship between many symptom dimensions and patient subscales from the TARS, only the symptom dimension of obsessive-compulsiveness shows a 106 Table 16 Significant TARS and VNIS Subscales in Predicting Posttreatment Symptomatology Therapeutic Alliance Rating Scale Sig. Equation Sig. Predictor T Positive Alliance 8 2 P Postive Alliance 5 l T Negative Alliance 8 0 P Negative Alliance 6 0 Vanderbilt Negative Indicators Scale Sig. Equation Sig. Predictor Patient Qualities 6 1 Therapist Qualities 4 0 Errors in Technique 5 0 P-T Interaction 9 0 Global 8 0 Total 6 0 107 Table 17 Correlations Between TARS and Pretreatment Symptomatology Combined Group (N=40) Therapist Therapist Negative Alliance Positive Alliance Somatization .03 Obsessive-Comp. .09 Interpersonal Sen. .05 Depression .23 Anxiety .05 Hostility .10 Phobic Anxiety -.04 Paranoia -.0l Psychoticism .03 GPI .10 Long Length Group (N=12) Somatization .16 Obsessive-Comp. .20 Interpersonal Sen. -.02 Depression .42 Anxiety .13 Hostility .19 Phobic Anxiety .10 Paranoia -.09 Psychoticism .08 GPI .21 -.24 .14 .13 -.01 .13 -.11 .03 .07 -.04 .04 .26 -096 -.l8 -.08 -.35 -.12 -.04 -.09 -.11 Medium Length Group (N=15) Somatization -.40 Obsessive-Comp. -.39 Interpersonal Sen. -.35 Depression -.44* Anxiety -.41 Hostility -.55* Phobic Anxiety -.09 Paranoia -.45* Psychoticism -.36 GPI -.50* -.39 -.16 -.11 -.05 .07 .07 .09 -022 -012 Patient Positive Alliance -.l3 ... 35** -.4a** .22 -.14 -.07 -.31* _.39** -.38** -033* -033 Patient Negative Alliance .04 .38** .35** .31* .22 -.08 .36** .32* .33* .32* -.16 .45 .29 .11 .05 .22 .30 .28 .21 .03 .36 .32 .41 .31 -.03 .42 -.44* -044* .37 108 Table 17 (cont'd.). Short Length Group (N=13) Therapist Therapist Patient Positive Negative Positive Alliance Alliance Alliance Somatization .04 .20 -.36 Obsessive-Comp. .08 .10 -.33 Interpersonal Sen. .06 .32 -.49* Depression .22 -.14 -.34 Anxiety .15 .41 .02 Hostility .36 -.25 .13 Phobic Anxiety -.11 .64** -.34 13315311013 -00“ -013 -033 Psychoticism -.00 .03 -.38 GPI .11 .16 -.38 *p < .05, **p < .01 Patient Negative Alliance .24 .27 .38 .27 .15 -.27 .43 .07 .25 .28 Correlations Between VNIS and Pretreatment Symptomatology Combined Group (N=40) Pat. Qual. Somatization .12 Obsessive-Comp. .26* Inter. Sen. .34* Depression .09 Anxiety .10 Hostility -.07 Phobic Anxiety .31* Paranoia .23 Psychoticism .18 GPI .21 Long Length Group (N=12) Somatization -.06 Obsessive-Comp. .25 Inter. Sen. .44 Depression -.00 Anxiety -.12 Hostility -.13 Phobic Anxiety .08 Paranoia .24 Psychoticism .21 GPI .11 109 Table 18 Medium Length Group (N=15) Somatization -.04 Obsessive-Comp. .28 Inter. Sen. .15 Depression .22 Anxiety .28 Hostility .04 Phobic Anxiety .21 Paranoia .29 Psychoticism .17 GPI .21 Errors P-T Ther. Qual. Tech. -.11 .ll .33* .23 .16 .23 .18 .ll .23 .14 005 -0g3 .10 .25 all 019 .06 .09 .18 .18 -.34 -.38 .41 .08 -017 OG3 -.03 -.ll .15 -015 -.l9 -.14 .02 -.12 -.05 .03 -094 -017 .02 -.11 .19 .38 .36 {.64** .36 .53* .49* .66** .60** .52* .30 .07 .29 .52* .28 .42 .21 .53* .43* .60** Inter . .04 .33* .32* .21 .25 .01 .18 .28* .15 .24 .33 .25 .16 .10 .04 .26 .a2 .16 .06 .00 .14 .42 .38 .40 .55* .22 .39 .40 .26 .42 .09 .37** .22 .19 .35** .19 .25 .17 .14 .29* .06 .57* .02 .05 .28 .07 .19 .04 .07 .17 Global Total .01 033* .29* .14 .22 .02 .22 .20 .12 .23 .30 .34 .04 .07 .02 .21 .01 .03 .01 .00 .16 .38 .41 .48* .56* .30 .31 .41 .31 .46* 110 Table 18 (cont'd.). Pat. Qual. Ther. Qual. Short Length Group (N=13) Somatization Obsessive-Comp. Inter. Sen. Depression Anxiety Hostility Phobic Anxiety Paranoia Psychoticism GPI .45 .28 .39 .23 .20 .12 .62** .23 .11 .34 *p < .05, **p < .01 -065 -.03 .23 .22 -008 .30 .22 .07 -.11 .08 Errors Tech. .54* .21 .40 .14 .14 .13 .41 .46 .21 .38 P-T Inter. .15 .14 .37 .18 .08 .03 .36 .14 .06 .19 .42 .26 .44 .25 .26 .12 .46 .28 .14 .38 Global Total .49* .28 .51* .30 .19 .07 .63** .36 .14 .42 111 clear relationship to VNIS subscales in the combined group analysis, while the symptom dimension of interpersonal sensitivity shows a relationship to those VNIS subscales that most include patient variables. These results suggest that more negative process indicators will generally be present in psychotherapy in which the patients' Obsessive-compulsive symptoms are higher, and more negative patient variables (including patient-therapist interaction variables) will be present when interpersonal sensitivity symptoms are higher. In both the TARS and VNIS analyses, it is interesting to note that the most significant relationships to pretreatment symptomatology occurred in the medium length group. This was especially true for the VNIS subscales where the symptom dimensions of depression and anxiety also consistently predicted higher negative indicators, and the subscale Errors in Technique was consistently elevated with greater pretreatment symptomatology. In the TARS analysis, positive alliance subscales were somewhat more affected by pretreatment symptoms than negative alliance subscales. Stronger relationships between subscales and symptom dimensions in a given treatment length group were not expected, although may in large part be due to small N's when breaking the analyses down to the group level. Chapter V DISCUSSION The present study was designed to address a number of the important issues of therapeutic alliance and negative indicators in outpatient psychotherapy. What little research that has been specifically focused on these variables has tended to Offer mixed results or failed to address certain questions that could be asked of the data available for this study. Seven specific questions were addressed in the present study: 1) Does strength of therapeutic alliance change over time and differ in psychotherapy of varying lengths?; 2) Do the levels of negative indicators change over time and differ in psychotherapy of varying lengths?; 3) DO negative indicators actually have an impact on therapeutic alliance?; 4) Does therapeutic alliance have an impact on the outcome Of psychotherapy?; 5) Do negative indicators have an impact on the outcome of psychotherapy?; 6) Does pretreatment symptomatology impact on the strength of therapeutic alliance?; and 7) Does pretreatment symptomatology impact on the level of negative indicators in psychotherapy?. 112 113 1. Therapeutic Alliance Development Across Time The development of therapeutic alliance in psychotherapy has generally been observed and compared in differing outcome groups, but has never been compared across cases of equal outcome for therapy of varying lengths. Luborsky (1976) found that strength of alliance as measured by his counting signs method indeed differed between high and low outcome groups, but significant differences in frequency of Type 1 or Type 2 alliance was not observed between the beginning and end phases of therapy. A trend was observed for the high outcome group in which alliance strength nonsignificantly increased from beginning to end, but it was essentially observed that alliance strength remained consistent throughout therapy. Using the same measures, Morgan et al. (1982) discovered the same within group consistency in high and low outcome groups. Hartley (1978), on the other hand, found no differences between groups for alliance strength at the beginning of therapy, but discovered that alliance strength actually increased over time in her dropout group as compared to a decline over time in her high and low outcome groups. The present study focused attention on therapeutic alliance development in treatment of varying lengths, and found results consistent with Luborsky and Morgan et al.'s research with outcome groups. Strength of therapeutic alliance did not significantly differ across phases of therapy within different treatment length groups. The level 114 of therapeutic alliance established within the first three sessions remained statistically consistent throughout the course of psychotherapy. This finding emphasizes what appears to be a growing consistency in the literature in that therapeutic alliance seems to be established quickly in psychotherapy and is maintained throughout the course of therapy around this initial level. This appears to be true not only for therapist and patient positive contributions to alliance, but also for their negative contributions. Intuitively, one would expect that it would take time for positive alliance to be firmly strengthened and established between two people meeting each other for the first time. In the same sense, one would expect that negative alliance would decline as the relationship grew stronger and matured. The results of this study suggest that psychotherapy may be a unique type of human relationship in which such intuitive predictions are not well-founded. The importance of establishing therapeutic alliance early in psychotherapy is further exemplified by the only significant between group difference found in the results. The strength of the therapists' positive contributions to alliance was higher in the beginning phase of therapy for the medium length group as compared to the short length group. Therapist Positive Alliance in the long length group fell between these two. Even though between group differences did not significantly exist for the later phases 115 of therapy, the strength of Therapist Positive Alliance remained consistently lower in the short length group, but in the high length group the trend was toward an increase to levels equal to the medium length group. This very cautiously suggests that therapy which may terminate within the first twenty sessions (without being planned as short-term therapY) may be marked by some problems in the therapist's contribution to positive alliance early in therapy. This finding not only suggests the importance of early alliance formation, but emphasizes the therapist's positive contributions as being of potentially critical significance to the develOpment of a working, collaborative therapeutic relationship. This finding must be cautiously interpreted at this time due to the lower reliability of the Therapist Positive Alliance subscale, but the trends in the data lend to the interpretation. If therapeutic alliance, both positive and negative, is established early in therapy, what are the variables contributing to this quickly develOping relationship? Even though this study did not focus on specific process variables other than negative indicators, the process areas measured by the TARS Offer some suggestions to address this question. Therapists' positive contributions revolved around areas of conveyance of hope, commitment, trust, warmth, support, liking, confidence, and active involvement in a shared working relationship. Therapists' negative contributions included such variables as criticalness, 116 judgmental, dependency, lack of empathy, lack of understanding, disappointment, and ignoring the patient's immediate concerns. Looking at patient variables, positive areas of contribution included confidence, trust, importance of therapy, introspection, exploration, and active involvement in a shared collaborative relationship. Negative variables revolved around dependency, anger, self-criticalness, mistrust, suspiciousness, resistance, demandingness, and denial. As discussed in Chapter 1, these variables are frequently mentioned process variables related to successful or unsuccessful psychotherapy. The present study does not address how these variables may be established in psychotherapy, but it does support the growing conclusion that they are established very early in the therapeutic process and maintained throughout the course of the relationship. The results discussed here begin to support Luborsky's (1976) comment that what happens early in therapy may be the best predictor of what happens late in therapy, including the prediction of outcome. This concept is not exactly original. Freud (1912) had emphasized the importance of early alliance formation to serve as the foundation for the difficult work in the middle phases of therapy, and Langs (1973, 1974) suggested that after alliance is initially established it fades into the background only to be made explicit again if there is some disruption. Empirically, it does not appear that therapeutic alliance really fades into 117 the background of the therapeutic relationship, but the present study certainly supports the importance of early alliance formation. 2. Level of Negative Indicators Across Time There have been no previous studies looking specifically at the level of negative indicators at different points in psychotherapy, or comparing these ratings with different treatment length groups. Intuitively, it was expected that the types of negative process variables measured by the VNIS (Appendix C) would be evident throughout the course of psychotherapy. It was not expected that the negative personal qualities brought to the therapeutic relationship by either the therapist or patient would show much change over the time periods in this study, nor would the errors made by the therapist change within any given course of psychotherapy (obviously, it would be hoped that changes would occur given more experience, but again not generally within the time periods given in the present study). It was expected that a trend toward lower negative indicators as therapy length progressed would be observed, but any differences would be insignificant within these treatment lengths. As with the development of positive and negative therapeutic alliance, the results of this study also showed that frequency of negative indicators are established within the first three sessions of psychotherapy, and, as 118 predicted, are generally maintained at these levels across the course of psychotherapy. No consistent trends were observed in the data. Once again, the therapeutic relationship seems to be established very quickly with little overall change over time even in the types and frequency of negative variables that are initially observed. In this sense, if the therapist fails to focus the session or fails to address the resistance early in therapy, he or she will probably continue making these errors throughout the course of the therapeutic relationship. If the patient displays initial passivity or evasiveness in the therapeutic interaction, these qualities are likely to continue. This does not mean that change in the patient's problems is not occurring over the course of psychotherapy, only that the same level of “therapeutic stumbling blocks” are being faced throughout the course of the relationship. The one between group difference that was discovered in the data could initially be seen as hinting at the possibility of higher negative indicators in shorter length therapy (significantly greater errors in technique at the mid-phase of therapy for the short length group), but closer inspection shows no supporting trends in the data. If anything, the long length group tends to have higher negative indicator ratings than the short length group. In addition, the lower reliability of the Errors in Technique subscale further suggests the spuriousness of this finding. 119 3. The Relationship Between Negative Indicators and Therapeutic Alliance In Strupp et al.'s (1981) introduction to the Vanderbilt Negative Indicators Scale, they emphasize that negative indicators are variables that serve a destructive role in the attempt to build a strong therapeutic alliance. In this sense, they serve to create problems in the therapeutic relationship leading toward a working environment that is not conducive to positive outcome. Given this overall definition of negative indicators, a natural interest in this research was to see whether negative indicators actually do impact on therapeutic alliance as obviously expected. Looking at the correlations for all forty cases combined, it is readily observed that a strong relationship exists between therapeutic alliance and negative indicators as rated by these measures. Positive alliance seems to be adversely affected by negative indicators, and negative alliance seems to be promoted by negative indicators. The latter result is certainly not surprising given the large overlap between negative process variables being measured by each of these scales. A somewhat surprising result was that therapist positive contributions to alliance seem to be less vulnerable to negative indicators than patient positive contributions. This may in part be due to therapist expectations of problems in therapy and the therapist's ability to continue working toward a positive relationship. 120 In short, the therapist has more experience working in therapeutic relationships and has trained skills in dealing with potential problems. The therapist's negative contributions may remain the same throughout therapy and be intertwined with negative indicators, but he or she may still be able to work toward a collaborative, positive relationship with the patient. Predictably, errors in technique then become the negative indicators that have the most impact on therapist positive alliance. Interestingly, the relationship between alliance and negative process variables did not remain consistent across therapy length groups. Looking at the correlations with data combined across phases, there initially appeared to be an obvious increase in the relationship as therapy progresses. In an attempt to further decipher this possible relationship, it was found that the strength of the relationship between therapeutic alliance and negative indicators does not seem to increase over the entire course of therapy, but only through around 15-20 sessions before it becomes more variable. This finding may in fact be related to the nature of the data. The points in which data were collected may be adding to the variability of the results and to their over-interpretation. The fact that a decline in the strength of the relationship at the end of therapy in the medium length group is Observed in these data may simply be related to the small sample size of the group. However, 121 even if a more conservative approach is taken, the data are still suggestive of far less of a relationship between therapeutic alliance and negative indicators at the beginning of therapy than at later phases. Why should the relationship between these variables be any different early in therapy than later, especially after establishing the fact that they are equally present at all phases of therapy? It is possible that both the therapist and patient bring enough initial motivation to the psychotherapeutic relationship that they are both very forgiving of any negative variables that may arise. They may initially overlook or attempt to work within the initial problems of establishing a relationship, but as the negative process variables continue they become more impactful on the relationship and serve to diminish positive alliance and promote negative alliance. In this sense, the alliance is very tolerant of negative indicators early, but is far less tolerant at later points in therapy. It is not clear that alliance continues to be less tolerant throughout the remaining course of therapy, but may instead ebb and flow in relation to negative events. This interpretation of the data is reasonably consistent with a transference model of psychotherapy. Freud (1912) stated that the alliance is established in the beginning of therapy as a foundation for later work. This positive bond is largely a conscious working relationship in which both therapist and patient are attempting to form a 122 real relationship with one another. However, the bond is also based on past interpersonal experiences, impulses, and needs that are part of the individual's reasons for seeking psychotherapy. This transference relationship ebbs and flows around positive and negative components as therapy progresses, and the relationship will naturally be more affected by negative events that may become part of the interpretive process in analytic technique. 4. Therapeutic Alliance and Psychotherapy Outcome The research on the relationship between therapeutic alliance and psychotherapy outcome has by far been given the most attention, but as with the entire literature on therapeutic alliance there are still few studies using any single measure. Generally, previous research has supported a relatively consistent relationship between strength of therapeutic alliance and treatment outcome, usually using a research design directly comparing alliance ratings in separate outcome groups (see Chapter 1). The present study approached the outcome issue somewhat differently by using different treatment perspectives and using both direct correlational and multiple regression analyses. For each treatment length group, TARS subscales were initially correlated with therapists' and patients' subjective ratings of outcome. These analyses yielded no consistent relationships, and the few significant results that were observed were probably related to chance. 123 Using symptom change as an outcome measure, Pearson product-moment correlations were obtained between TARS subscales and absolute change ratings. These results also Offered no consistent relationship with outcome, and the unusual findings between Patient Negative Alliance and symptom change in the medium length group is considered spurious. These analyses, by using absolute symptom change, also ignored the potential impact of the initial level of symptomatology on symptom change ratings. Transforming the data to percent change data to account for initial levels of symptomatology still did not yield a consistent finding that therapeutic alliance is related to outcome in this study. A relationship was observed between the specific symptom dimension of somatization and TARS subscales in the long length group, but this certainly does not offer solid support for therapeutic alliance having impact on treatment outcome. Finally, considering the number of interacting variables in this study all having potential impact on therapy outcome, stepwise multiple regression was chosen to more precisely clarify the impact of therapeutic alliance on symptom change as compared to all other measured variables (including initial levels of symptomatology). Interestingly, only the positive alliance subscales accounted for a significant amount of the variance, and then for only three of the ten symptom dimensions. Therapist Positive Alliance predicted lower posttreatment 124 symptomatology for the anxiety and phobic anxiety dimensions, and Patient Positive Alliance predicted lower symptomatology for the psychoticism dimension. Rather than attempting to represent a relationship with actual symptom change as in the previous analyses, the multiple regression analysis identified variables that predicted posttreatment symptomatology that statistically controlled for the effects of pretreatment symptomatology. Even though therapeutic alliance subscales only significantly predicted lower symptomatology for three of ten symptom dimensions, this is still considered mildly supportive of a relationship between alliance and outcome. Specifically, these results mildly support previous research by Marziali (1984) in which she found that both therapist and patient positive contributions to alliance significantly predicted positive outcome. It is especially interesting that Therapist Positive Alliance was a predictor for both anxiety dimensions, suggesting a specific and possibly powerful relationship between positive therapist variables with the reduction of anxiety symptoms. It is equally interesting that positive patient alliance predicted lower psychotic symptoms, suggesting the importance of the patient's capacity to contribute to a positive working relationship in reducing such severe pathology along with the therapist's contributions (Karon and Vandenbos, 1981). This study chose multiple approaches in attempting to understand and further confirm the relationship between 125 therapeutic alliance and treatment outcome for a variety of reasons. The data did not readily lend themselves to treatment group comparisons, and the outcome measures available for the data were not considered powerful tools. It was hoped that a direct correlational approach would be sensitive to the subtle changes in outcome observed in the data, although this did not prove to be the case. It was also of interest to compare the different ways of using the outcome data, especially in comparison with the multiple regression approach. Horowitz et al. (1984) had shown that the direct correlation between alliance ratings and outcome ratings (including an SCL-90 change score) proved insignificant, probably due to the mediating impact of the other variables in their study. When using a multiple regression approach to understand the interactive nature of their variables, they discovered significant results (see Chapter 1). The present study supports this approach to data analysis when using a number of process variables in predicting outcome. A general correlational approach was not sensitive enough to identify statistically significant relationships, especially when trying to represent symptom change, and, in retrospect, a partial correlational approach with the same data would not have yielded any better results. A final interesting note from these anlyses is worth mentioning here. The only significant predictor of lower 126 overall symptomatology (GPI) found in the multiple regression analysis was pretreatment hostility. The higher the level of patient hostility at the beginning of therapy, the less likely overall symptoms would diminish. 5. Negative Indicators and Psychotherapy Outcome In the two studies using the VNIS to assess the relationship between negative process indicators and outcome, Sandell (1981), as expected, had discovered significant negative correlations between these variables and outcome. In a previous study, Strupp et a1. (1980) had found that the subscales from the VNIS were not useful in differentiating outcome groups, but their study involved so few cases they still held expectations that future results would establish this relationship. The present study approached the analysis of this issue in the same manner as that used for therapeutic alliance. It was expected that negative process indicators would indeed have a significant impact on therapeutic outcome, and a variety of measures and correlational approaches were used to assess this relationship. Just as with the therapeutic alliance data, the direct correlations provided little useful evidence of a consistent relationship with outcome across all treatment length groups, and only one consistent relationship with the symptom dimension of somatization in the long length group. Using the multiple regression analysis as described 127 previously, only one significant negative indicator predictor was discovered out of the ten symptom dimensions. Patient Qualities significantly predicted higher symptomatology at post-therapy for the symptom dimension of paranoia. This finding is both interesting and intuitive given the content of the Patient Qualities subscale, but hardly displays even a mildly consistent relationship between negative indicators and outcome in these data. Given the fact that negative alliance subscales were also not predictive of outcome, it appears that at least in this set of therapy cases negative process variables seemed to have much less impact than positive process variables on therapy outcome. As previously quoted, Marziali (1984, p. 422) states, '...patients with a history of highly conflicted relationships add stress to the alliance and demand from therapists a capacity to absorb and manage responses which are ambivalent, confusing, and obstructionistic. Probably, it is the 'working through' of these treatment relationship conflicts that determines the outcome of psychotherapy for these more resistive patients." Given the fact that we have already discovered that therapist positive variables were less impacted by negative variables in these data, it is possible that the therapists in this sample were skilled at working with negative process variables in psychotherapy to either reduce their impact, or turn them to positive gain. In essence, this would simply be a mark of good 128 psychotheraPY, and may suggest that negative variables may not necessarily cause psychotherapy to be unsuccessful. 6. Therapeutic Alliance and Pretreatment Symptomatology Previous research has shown a significant relationship between a patient's pretherapy capacity for relationship formation and development of therapeutic alliance (Ryan, 1973; Lehrke, 1977). Moras and Strupp (1982) and Marziali (1984) discovered a positive relationship between patient pretherapy social adjustment and patient positive contribution to alliance, but also found that their measures of initial patient pathology seemed to have no impact on alliance formation. The latter finding was supported by Morgan et a1. (1982), establishing what seemed to be a consistent finding that the patient's level of symptomatology at the beginning of therapy does not affect the formation of a strong therapeutic alliance. In contrast to these findings, the present study showed a consistent relationship between pretreatment symptomatology and alliance formation when the cases were combined across groups. Specifically, all SCL-90 symptom dimensions except somatization, anxiety, and hostility were significantly related in the expected direction to the patients' positive and negative contributions to alliance, but not to the therapists' contributions. In other words, the higher the patient's symptomatology upon entering therapy the lower the patient's positive alliance and the 129 higher the patient's negative alliance. The results for the overall analyses were primarily based on the stronger relationship between these variables in the medium length group. In this group, both therapist and patient positive alliance showed consistent relationships to pretreatment symptoms. The overall results may reflect a spurious finding based on a very select group of cases. However, observing the trends in all groups suggests that a more likely explanation is that pretreatment symptoms do have at least a mild impact on patient alliance formation, but the significance is generally lost when looking at small numbers of cases. The stronger findings in the medium length group reflect a somewhat stronger relationship in this group of cases, with the relationship to therapist positive alliance more likely an atypical result. Why should these data reflect a relationship in contrast to previous research findings? For one, the previous studies only measured general levels of pathology rather than specific symptom dimensions. It is likely the measures used in the present study were more sensitive to establishing subtle relationships. For another, the previous research focused on more homogenous groups involved in short—term therapy. The data used in the present study reflect a more general clinic population that has not generally been screened for specific factors related to the patients' capacities to form relationships. In this sense, 130 it is likely that patients in the previous research were either less pathological, or were screened to exclude patients with specific types of pathology that may have impacted on therapeutic alliance. 7. Negative Indicators and Pretreatment Symptomatology The impact of pretreatment symptomatology on level of negative process indicators had not previously been studied, but was expected to show a consistent relationship. In this sense, a higher amount of pathology at the beginning of therapy was expected to create more negative events in the psychotherapeutic relationship. In contrast to the therapeutic alliance results, little relationship between pretreatment symptoms and negative indicators was observed. In the overall analyses, only the symptom dimension of obsessive-compulsiveness shows a clear and consistent relationship with VNIS subscales. The higher the level of Obsessive-compulsiveness upon entering therapy, the greater the level of negative indicators. In addition, the interpersonal sensitivity dimension seemed to impact on negative patient events (Patient Qualities and Patient-Therapist Interaction) with no relationship to primarily therapist indicators. Once again, the medium length group displays stronger and different findings than in the overall analyses. As discussed for the therapeutic alliance results, this most likely reflects a specific difference in these cases rather 131 than a more generalized finding. The most general conclusion from the overall analyses is that pretreatment symptomatology tends to have little impact on negative indicators. Limitations of the Study As with any research, the issue of generalization of results is always in question. The data used in the present study were collected in an outpatient clinic with a relatively heterogeneous patient population and psychotherapists at various levels of training and experience. These elements of the research design were considered improvements over previous research that focused on very specific patient groups and specific time-limited psychotherapy. Given the fact that the data were drawn from a population more closely resembling a typical private or public clinic, it is expected that these results are more generalizeable than previous findings. However, the fact that graduate student therapists-in-training were used potentially places more serious limits on the more broad based implications of the results. In reality, the results are probably more conservative estimates of what would have been found with more experienced therapists; although, it should be kept in mind that most community clinics employ therapists with no more training than those used in the present study. This remains, however, a limit in generalization at the present time. 132 A potentially more serious problem with these data is the level of reliability established for the subscales. This was discussed in depth at the beginning of Chapter IV, and it was argued that the reliabilities were all at least cautiously acceptable. This argument remains, but the stress is still placed on the ”cautiously." It would naturally be hoped that in future research with these measures stronger reliabilities could be Obtained. It seemed that the more data we rated, the lower the reliabilities fell despite doing everything known possible to prevent rater drift. The general conclusion may be that the process variables being measured in this study are subjectively difficult to rate, and more experienced, closely matched raters may be necessary to Obtain truly solid reliabilities. As also pointed out in the beginning of Chapter IV, the TARS and VNIS subscales seem to display considerable overlap in the concepts they are measuring. The symptom dimensions from the SCL-90 also display the same overlap (Derogatis, 1977). From this aspect, it may be difficult to make specific comments about the impact or relationships of any specific subscale, and discussion of specific relationships was limited. Given the fact that process and symptom variables in psychotherapy are naturally going to show much interaction, it may not be possible to measure these variables more specifically. Simply combining them using a factor analytic method would lose the potentially 133 interesting data they still may provide even with their interaction. As in Horowitz et al.'s (1984) research, the answer to understanding the interaction between process variables probably lies with the use of more soPhisticated statistical analyses. The fact that therapeutic alliance and negative indicators were observed to remain statistically consistent throughout the course of psychotherapy may have more to do with the number of data points used than with any true consistency. Hartley (1978) had found fluctuations in strength of therapeutic alliance by using five data points as Opposed to three in this study, and, as stated, the results here may simply reflect too broad a sampling approach. Given the fact that the consistency in this study was found across all three treatment length groups in which smaller N's could be more affected by single case fluctuations and in which different average session points were being rated, it is still considered unlikely that any fluctuations would have been discovered using a broader sampling design. As previously stated in this chapter, the outcome measures were considered somewhat weak. It is always difficult to assess psychotherapeutic outcome from pure self-report symptom change measures, and the few subjective outcome questions asked of the therapists and patients were considered highly variable and not very powerful. In addition, the patient population sampled had a relatively 134 homogeneous treatment outcome, with few clearly unsuccessful cases but many clearly successful cases. In this sense, the psychotherapy represented in this study was basically good treatment. This was fortunate for the patients, but without more variability in the outcomes it was difficult to establish a solid relationship between the variables measured and treatment success. It was hoped that the correlational approach to data analysis would pick up subtle relationships, but it was discovered that the relationships were apparently not strong enough for this to be the case. As discussed, the multiple regression approach proved to be the most useful analysis in accurately understanding the nature of the relationships that did exist. Finally, the present study was confronted with an all too common psychotherapy research problem--small N's. The forty total cases represented a very solid sample, especially as compared to previous research, but when separated into treatment length groups the N's dropped to questionable sizes. As it turned out, treatment length played a relatively insignificant role in the overall conclusions, but it was very helpful in understanding the results leading to those conclusions. In this sense, the small N's for the separate groups were not considered a serious problem in this study, but it would always be nice to have more data. 135 Implications for Research and Practice The following is a broad summary of the general conclusions Offered by this study as they relate to the practice of outpatient psychotherapy: 1) Therapeutic alliance, both positive and negative, appears to be established within the first three sessions of psychotherapy with little significant change over the remaining course of the therapeutic relationship. This result reflects the importance of being aware of alliance issues early in the treatment process, especially as they form the foundation for later therapeutic work. It also suggests that whatever variables are entering into the development of the therapeutic alliance they are being brought to the relationship by both patient and therapist very early, with little change over time. These variables may reflect therapist and patient characteristics that are important for a working 'match' between the two. If this match does not exist early, these results suggest that it may be difficult for a strong alliance to develOp. This issue, however, requires more focused research. 2) The level of negative indicators observed early in psychotherapy also seem to remain consistent throughout the course of that therapy. The therapists in this sample were even being closely supervised, suggesting that the characteristics of the therapists and patients leading toward negative process events are difficult to change over the course of a single therapeutic relationship. This 136 result suggests more focused supervision of training therapists may be especially necessary at the very beginning of the therapeutic process. It was also discovered, however, that even though negative indicators are established early, their impact on therapeutic alliance does not occur until later in the relationship. There seems to be a beginning "grace” period in which negative process variables do not impact on alliance formation as much as occurs later in therapy. However, the relationship also appears to be variable later in therapy, suggesting that there may be certain times or certain negative events that have more impact on the therapeutic alliance. Future research focused on these process events may help to further define this result and aid therapists in knowing especially when to work with these events. 3) Therapist positive contributions to the alliance seem to be less affected by negative variables than other alliance factors. This may suggest that the therapist has, if you will, a special power in the relationship in which positive elements remain strong and persistent. This finding may help to explain why negative indicators seem to have relatively little impact on therapy outcome, and also emphasizes the responsibility of the therapist to maintain his or her positive role in working with negative events or negative transference issues. 4) Following with this idea, the only mild relationship between these variables and treatment outcome came from the 137 positive alliance factors, especially therapist positive contributions. This result further focuses responsibility on the therapist in maintaining and using the positive process elements in therapy (especially with anxious patients). It also suggests that at least in contrast to these elements, negative process variables from either therapist or patient have little impact in predicting outcome. In this sense, even if a therapist is making mistakes or dealing with a difficult patient, it does not necessarily mean that therapy will fail. Of course, the limitations on the strength of these results have already been discussed, but the finding remains interesting for further exploration. 5) Finally, both patient positive and negative contributions to alliance were mildly affected by the level of symptomatology at the outset of therapy. Symptoms did not generally impact on the level of negative indicators found in therapy, and they had no significant impact on therapist contributions to alliance. These results highlight the fact that a patient with more severe pathology may have a more difficult time contributing to the formation of a positive working relationship with the therapist. The therapist must be aware of this problem, and have the skills and resources necessary to do the primary work in building the relationship. The results of this study are broadly based and represent a further exploration of the general questions 138 from the therapeutic alliance and negative indicator literature. Given the paucity of the research field and the data available for study, it was felt these questions needed further testing. Indeed, the results presented here are interesting. They both support and contradict past research, as well as provide data on questions never having been asked before. Agreeing with a more recent review by Frieswyk et a1. (1986), the next step in therapeutic alliance research is to focus on more specific process variables that have definite impact on alliance formation, and to delineate more clearly how therapeutic alliance interacts with the multitude of psychotherapy process variables to impact on treatment outcome. This research will require fresh and innovative approaches to research design and data analysis, including the use of single-case designs (see Elliott et a1., 1982) to more clearly focus on specific process events that are critical to the therapeutic interaction. As Frieswyk et al. also pointed out, research designs can also be extended to specific patient pOpulations (as in their research on borderline patients) to understand the nature, development, and impact of the therapeutic relationship with specific individuals seeking psychotherapeutic help. Frieswyk et a1. argue the necessity of separating the patient and therapist contributions to alliance as was done in the present study, but they also argue for a distinct separation of experience variables versus collaboration 139 variables. Conceptually, this would be a nice distinction to make in attempting to understand therapeutic alliance more clearly. Realistically, it will obviously be very difficult to accomplish, and may reflect a tendency in the field to separate process variables so far that the vital significance of their interaction is lost. Future research will attempt to address this issue. The impact of negative indicators on psychotherapy was less than expected in this study, but this does not deny their importance to the therapeutic relationship. It has long been understood that psychotherapy is a powerful process that can have both positive and negative outcome. It is likely that negative outcome does not simply result from negative process indicators, and in fact it may be the working through of negative events in psychotherapy that leads to more successful outcome. The present study suggests that a future research direction will be to understand when and how negative indicators impact on the therapeutic relationship, and the process that occurs for the relationship to continue to grow and mature despite, or due to, these events. APPENDICES APPENDIX A PRELIMINARY TABLES Description of 140 Table 1 Groups Combined Group (N=40) Therapeutic Alliance Rating Scale Beg. Mid. End Tot. M (SD) M (SD) M (SD) M (SD) TPos. 26.41 (4.37) 27.86 (3.44) 27.23 (4.06) 27.14 (2.73) Alliance TNeg. 11.26 (3.89) 12.46 (4.13) 13.81 (5.79) 12.51 (3.17) Alliance PPos. 29.33 (4.93) 30.01 (4.35) 30.79 (4.62) 29.87 (3.42) Alliance PNeg. 13.15 (3.48) 14.74 (5.22) 14.91 (5.45) 14.18 (3.20) Alliance Vanderbilt Negative Indicators Scale Pat. 1.84 (2.35) 2.48 (2.16) 2.70 (2.84) 2.37 (1.90) Qual. Ther. 1.39 (2.72) 1.64 (2.60) 1.85 (2.83) 1.63 (2.14) Qual. Errors 2.09 (2.17) 1.83 (1.99) 2.15 (2.14) 2.07 (1.42) Tech. P-T 0.36 (0.82) 0.63 (1.26) 0.68 (1.01) 0.57 (0.68) Inter. Global 2.41 (1.73) 2.75 (2.23) 2.93 (1.87) 2.74 (1.33) Total 8.05 (7.44) 9.33 (8.06) 10.31 (8.82) 9.19 (6.11) SCL-90 Pre. Post. Chg. % Chg. M (SD) M (SD) M (SD) M Somatization 0.72 (0.59) 0.40 (0.36) -0.32 (0.59) 0.44 Inter. Sen. 1.75 (0.80) 1.17 (0.54) -0.58 (0.70) 0.33 Depression 2.08 (0.85) 1.21 (0.75) -0.87 (0.97) 0.42 Anxiety 1.60 (0.87) 0.94 (0.51) -0.67 (0.89) 0.42 141 Table 1 (cont'd.). SCL-90 Pre. Post. Chg. % Chg. M (SD) M (SD) M (SD) M Hostility 1.12 (0.85) 0.79 (0.72) -0.33 (0.93) 0.29 Phobic Anx. 0.68 (0.78) 0.36 (0.50) -0.32 (0.62) 0.47 Psychoticism 0.96 (0.71) 0.47 (0.35) -0.49 (0.66) 0.51 GPI 1037 (“061) 9082 (“036) ‘ness (0064) “04“ 3.90 (0.77) Patient Outcome Therapist Outcome 6.22 (1.59) Gender Pt. F = 26 Th. F = 15 M = 14 M = 25 Average Sessions Beg. = 2.37, Mid. = 15.60, End = 28.73 Long Length Group (N=12) Therapeutic Alliance Rating Scale Beg. Mid. End Tot. M (SD) M (SD) M (SD) M (SD) TPos. 26.33 (4.17) 28.67 (4.01) 28.21 (3.34) 27.68 (3.07) Alliance TNeg. 12.25 (4.57) 13.67 (3.47) 15.00 (7.24) 13.63 (4.00) Alliance PPos. 29.08 (6.26) 30.25 (4.40) 30.42 (6.29) 29.62 (4.52) Alliance PNeg. 13.17 (2.25) 14.58 (3.57) 17.29 (8.37) 14.74 (3.22) Alliance Vanderbilt Negative Indicators Scale Pat. 2.29 (2.68) 2.25 (1.56) 4.08 (4.16) 2.86 (2.35) Qual. Ther. 3.00 (4.45) 1.75 (2.93) 2.83 (4.60) 2.53 (3.43) Qual. Average Sessions Beg. = 2.50, Mid. = Table 1 (cont'd.). 142 2.29 (3.03) 0.96 (1.80) 2.50 (2.23) 12.67(14.23) Beg. Mid. M. (SD) M (SD) M Errors 2.67 (3.06) 1.29 (1.79) Tech. Inter. GLobal 3.46 (1.95) 2.92 (2.20) Total 12.21 (10.79) 8.63 (7.07) SCL-90 Pre. Post. M (SD) M (SD) Somatization 0.73 (0.58) 0.40 (0.30) ObSESS‘Comp. 2069 (“074)a 009“ (0027) Inter. Sen. 2.20 (0.62) 1.31 (0.54) Depression 2.23 (0.61) 1.17 (0.68) Anxiety 1.74 (0.72) 0.79 (0.33) Hostility 1.26 (0.92) 0.94 (0.86) Phobic Anx. 0.64 (0.72) 0.26 (0.38) Paranoia 1.43 (0.90) 1.03 (0.66) Psychoticism 1.18 (0.85) 0.53 (0.35) GPI 1.18 (0.59) 0.88 (0.32) Patient Outcome = 4.27 (0.56) Therapist Outcome - 6.89 (1.20) Gender Pt. F a 9 Th. F = 2 M = 3 M = 10 23.60, End (SD) Chg. % Chg. M. (SD) M -go33 (0059) 0045 -1.19 (0.75) 0.57 -0.90 (0.42) 0.41 -1005 (“068) “047 ‘go95 (0075) “055 -6040 (“077) 9028 -9064 (“069) 0054 -0074 (0058) 9047 End = 43.80 M Tot. (SD) 2.09 (1.74) 0.75 (0.95) 3.08 (1.83) 11.18 (8.98) Table 1 (cont'd.). 143 Medium Lengph Group (N215) Therapeutic Alliance Rating Scale TPOS. Beg. M Alliance TNeg. Alliance PPOS. Alliance PNeg. Alliance (SD) 28.43 (2.63) 11.50 (3.87) 30.33 (4.13) 13.53 (4.67) M Mid. (SD) 28.33 (3.16) 12.43 (4.04) 30.97 (3.97) 14.73 (5.89) M End (SD) 27.43 (3.95) 13.83 30.93 14.23 Vanderbilt Negative Indicators Scale Pat. Qual. Ther. Qual. Errors Tech. P-T Inter. Global Total SCL-90 Somatization Obsess-Comp. Inter. 1.30 (1.47) 0.97 1.50 0.27 1.63 5.63 Sen. Depression Anxiety Hostility Phobic Anx. (1.16) (1.39) (0.46) (1.29) (4.15) 2.03 (1.79) 1.40 0.90 2.47 7.87 Pre. M 0.77 1.66 1.65 2.32 1.64 1.14 0.68 (SD) (0.62) (0.82) (0.77) (0.95) (1.04) (1.00) (0.81) (2.08) (1.12) (1.70) (2.06) (7.46) 1.83 1.43 Post. M 0.41 0.81 1.14 1.30 “.91 0.57 0.37 (SD) (0.42) (0.41) (0.81) (0.58) (0.54) (6.01) (3.26) (2.76) (1.62) (1.66) (1.84) (1.65) (5.55) M Tot. (SD) 28.05 (1.34) 12.59 30.75 14.17 1.82 1.28 0.59 2.37 7.26 Chg. M -6036 ‘0.87 -0052 -l.02 '0.73 '0.58 '0.31 (SD) (0.52) (0.89) (0.63) (1.18) (0.98) (0.78) (0.63) (2.96) (2.59) (3.59) (0.97) (1.30) (1.11) (0.63) (1.14) (4.28) % Chg. M 0.47 0.52 0.32 0.44 0.45 0.51 0.46 144 Table 1 (cont'd.). SCL-90 Pre. Post. Chg. 8 Chg. M. (SD) M, (SD) M (SD) M Paranoia 1.20 (0.85) 0.78 (0.50) -0.42 (0.80) 0.35 Psychoticism 0.91 (0.56) 0.45 (0.39) -0.46 (0.72) 0.51 GPI 1039 (0063) “078 (0039) -0061 (“064) 0044 Patient Outcome = 3.83 (0.66) Therapist Outcome = 6.16 (1.08) 10 Th. F 5 M Gender Pt. F M 7 8 Average Sessions Beg. = 2.70, Mid. = 14.90, End = 26.40 Short Length Group (N=13) Therapeutic Alliance Rating Scale Beg. Mid. End Tot. M (SD) M (SD) M (SD) M (SD) TPos. 24.15 (5.24) 26.58 (3.05) 26.08 (4.77) 25.59 (3.11) Alliance TNeg. 10.08 (3.15) 11.39 (4.77) 12.69 (3.97) 11.38 (2.26) Alliance PPos. 28.39 (4.57) 28.69 (4.72) 30.96 (4.53) 29.08 (3.13) Alliance PNeg. 12.69 (2.96) 14.89 (6.03) 13.50 (3.89) 13.69 (2.87) Alliance Vanderbilt Negative Indicators Scale Pat. 2.04 (2.87) 3.19 (2.90) 2.42 (2.09) 2.55 (2.22) Qual. Ther. 0.39 (0.58) 1.81 (2.98) 1.42 (1.44) 1.20 (1.01) Qual. Errors 2.23 (1.92) 3.19 (2.34) 2.46 (1.49) 2.64 (1.28) Tech. Beg. M (SD) Inter 0 Global 2.35 (1.57) Patient Outcome Therapist Outcome Gender Pt. F a 7 M 6 Average Sessions Beg. = 1.90, Mid. = ap < .05 by the two-tailed t-statistic Table 1 (cont'd.). M 0.54 (1.13) 2.92 (2.57) 11.65 (9.56) 145 Mid. Total 7.00 (5.21) SCL-90 Pre. M (SD) Somatization 0.64 (0.60) Obsess-Comp. 1.45 (0.76)a Inter. Sen. 1.43 (0.83) Depression 1.66 (0.82) Anxiety 1.44 (0.82) Hostility 0.95 (0.62) Phobic Anx. 0.71 (0.86) Paranoia 0.83 (0.81) Psychoticism 0.80 (0.72) GPI 1.15 (0.58) (SD) M End (SD) 0.54 (0.59) 3.27 (1.82) 10.15 (4.66) Post. M “.39 “.98 1.08 1.14 1.10 0.91 0.45 0.66 0.43 0.80 3.64 (0.96) 5.67 (2.18) Th. F 8 6 M = 7 (SD) (0.37) (0.70) (0.60) (0.73) (0.54) (0.75) (0.74) (0.51) (0.31) (0.38) 8.30, M TOt . (SD) 0.36 (0.36) 2.84 (0.92) 9.59 (4.17) Chg. M '0.25 '0.47 -0.35 '0.52 '0.34 -0.04 -0027 '0.18 '0.37 -003“ End (SD) (0.70) (1.01) (0.89) (0.88) (0.85) (0.60) (0.74) (0.73) (0.58) (0.67) = 16.00 % Chg. M “.39 0.32 0.24 0.31 0.24 0.04 0.38 0.22 0.46 0.26 146 Table 2 Reliabilities of TARS and VNIS Subscales Therapeutic Alliance Rating Scale AVG. REL. T Positive Alliance .54 T Negative Alliance .73 P Positive Alliance .61 P Negative Alliance .73 Vanderbilt Negative Indicators Scale AVG. REL. Patient Qualities .72 Therapist Qualities .84 Errors in Technique .56 Pat.-Ther. Interaction .71 Global .53 Total .81 IND. REL. (.37) (.58) (.44) (.57) IND. REL. (.56) (.73) (.39) (.55) (.36) (.68) 147 Table 3 Intercorrelations of TARS Subscales T Pos. T Neg. P P08. P Neg. Alliance Alliance Alliance Alliance T Pos. Alliance 1.00 T Neg. Alliance -.22 1.00 P Pos. Alliance .44** -.51*** 1.00 P Neg. Alliance -.15 .65*** —.76*** 1.00 **p < .01, ***p < .001 148 Table 4 Intercorrelations of VNIS Subscales Pat. Ther. Errors P-T Global Tot. Qual. Qual. Tech. Inter. Pat. Qual. 1.00 Ther. Qual. .32* 1.00 Errors Tech. .56*** .59*** 1.00 P-T Inter. 053*** 082*** 067*** 1.00 Global .51*** .80*** .68*** .73*** 1.00 Total .71*** .85*** .83*** .85*** .89*** 1.00 *p < .05, ***p < .001 APPENDIX B THERAPEUTIC ALLIANCE RATING SCALE APPENDIX B THERAPEUTIC ALLIANCE RATING SCALE (TARS) Instructions: Listen to complete audio segment of session. Then read each question carefully, and circle the number which best represents your estimate of evidence for your rating. not a Therapist Items: at all lot 1. To what extent did the therapist 0 1 2 3 4 5 convey hope that the patient can make progress in therapy? 2. To what extent did the therapist 0 1 2 3 4 5 criticize the patient? 3. To what extent did the therapist 0 1 2 3 4 5 refer to information or experiences that he and the patient had shared in previous sessions? 4. To what extent did the therapist 0 1 2 3 4 5 pursue his own ideas in this session, ignoring what the patient wanted to talk about? 5. To what extent did the therapist 0 1 2 3 4 5 convey commitment to helping the patient get better? 6. To what extent did the therapist's 0 1 2 3 4 5 comments help the patient feel good about himself? 7. To what extent did the therapist 0 1 2 3 4 5 encourage the patient to be curious about his problems and the role that he could play in overcoming them? 149 10. 11. 12. 13. 14. 15. 16. 17. 18. 150 To what extent did the therapist convey that it was very important that the patient depend on him for finding solutions to his problems? To what extent did the therapist have difficulty in empathizing with the patient's problems and needs? To what extent did the therapist help the patient understand how therapy works and what he might hope to gain from it? To what extent did the therapist support the patient's ability to reflect on his thoughts and feelings and come up with solu- tions to his problems? To what extent was the therapist wanting the patient to comfirm his expertise? To what extent did the therapist convey a sense that they were working together in a joint struggle to find solutions to the patient's problems? To what extent did the therapist convey disappointment or annoyance, because he felt the patient was not making sufficient progress? To what extent was there a discre- pancy between what the therapist said to the patient and what he seemed to be thinking and/or feeling? To what extent did the therapist convey that he liked the patient during this particular hour? To what extent did the therapist communicate with the patient in such a way that would ensure that the patient understood him? To what extent did the therapist make unnecessary comments? 19. 20. 21. 151 To what extent was the therapist judgmental, i.e.: communicating that his way of viewing things was better than the patient's? To what extent did the therapist convey confidence in being able to help the patient with his problems? To what extent did the therapist convey disappointment and im— patience with the fact that the patient continues to do things which get him into trouble, or make him unhappy? 152 Patient Items: l. 10. 11. To what extent did the patient let the therapist know that therapy was helping him? To what extent did the patient convey confidence in his own ability to solve his problems? To what extent was the patient angry with the therapist, letting him know that he wasn't giving him what he wanted? To what extent did the patient convey that he was working to- gether with the therapist in a joint effort to solve his problems? To what extent did the patient hOpe that the therapist would have all the answers to his problems? To what extent did the patient seem to like the therapist? To what extent did the patient talk freely about himself? To what extent did the patient convey the impression that it is important for him to come on a regular basis to his therapy sessions? To what extent was the patient self-critical, feeling hopeless about his ability to get over his problems? To what extent did the patient express uncertainty about continuing in therapy? To what extent did the patient demonstrate a willingness to explore his own contributions to his problems? 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 153 To what extent did the patient avoid discussing concerns and worries because he seemed to feel the therapist wouldn't understand? To what extent did the patient argue with the therapist in an effort to bring him around to his way of thinking? To what extent did the patient's way of thinking about his problems fit with the therapist's? To what extent did the patient resist efforts to understand the connections between his feelings and his behavior? To what extent did the patient say and do things in the session that were similar to what the therapist said and did? To what extent does the patient feel hOpeful about the future because he views his experience in therapy as positive? To what extent did the patient not want to examine his behavior and how it gets him into the same trouble over and over again? To what extent was the patient mistrustful and suspicious of the therapist? To what extent did the patient examine his thoughts, feelings, and behavior and thus learn more about himself? To what extent did the patient hope that the therapist could make him feel better without his having to do much about solving his problems? APPENDIX C VANDERBILT NEGATIVE INDICATORS SCALE Copyright (c) Hans H. Strupp: Ph.D., et a1., 1981 A. LACK Lack APPENDIX C VANDERBILT NEGATIVE INDICATORS SCALE (VNIS) RATING MANUAL PATIENT PERSONAL QUALITIES AND ATTITUDES OF MOTIVATION of MOTIVATION for psychotherapy which may include: Lack of distress This patient does not express the concern, discomfort, or pain about his situation that can reasonably be expected from a person in psycho- therapy. This item should be rated on the basis of the patient's perception and presentation of his problems, rather than on whether you as a clinical judge feel that he needs help. Failure to take responsibility for problems This patient seems unwilling to consider the role he plays in his problems. ApprOpriate evidence for this item includes: a. Patient externalizes his problems by blaming others or environmental events. b. Patient believes that he can do little to influence or control his situation. Passivity in the therapeutic interaction The patient relies on the therapist to elicit information from him and to maintain the flow of the session. Although the patient seems to have a c00perative attitude and responds to direct questions from the therapist, the patient takes little or no initiative in introducing or elaborating upon topics. 154 155 (This item is to be distinguished from item 1.12, "Negative Attitudes.” If there is an ”oppositional," negativistic, or passive- aggressive quality to the patient's lack of initiative in introducing material, item 1.12 should be rated.) 1.4 Ambivalence about therapy The patient is conflicted about being in therapy. Some ambivalence regarding one's therapy is to be expected. Therefore, when rating this item rely on explicit evidence that the patient has reservations regarding his continued participation in therapy. Examples: 1. Patient directly questions the utility of therapy (evidence of this type merits a high rating.) 2. Patient discusses factors that make it difficult for him to participate in therapy such as other time commitments or the social stigma attached to psychotherapy. 3. Patient misses a session with no or inadquate justification. 1.5 Lack of introspection or curiosity about oneself The patient seems either unable or unwilling to expand his self-awareness. Rather than examining his behavior, thoughts, or feelings, the patient consistantly tends to discuss his experiences on a superficial level, gloss over incidents and/or minimize the significance of his behavior. This item is meant to assess a generalized tendency to avoid self-examination rather than "normal resistance.“ Consequently, this item should not be used to rate signs that the patient only avoids self-examination in circumscribed areas or with respect to specific issues. B. INSUFFICIENT SELF-DISCLOSURE 1.6 Evasiveness The patient consistently avoids exploring 156 certain feelings, experiences or issues. The patient tends to deflect his responses to neutral areas or answers in vague terms when addressing questions posed by the therapist. Two primary considerations determine whether or not this item should be rated: (1) It requires evidence that the patient's reluctance to explore feelings, etc., appears after the therapist has made probing interventions: (2) It applies when the patient's reluctance is ”localized“ or limited to certain areas. Apparent dissimulation The content of the patient's communications is contradictory or incongruous. Inconsistancies in the material presented by the patient suggest falsification, distortion, or intentional omission of information. VERBAL COMMUNICATION PROBLEMS 1.8 Problems with verbal self-expression The patient has notable difficulty articulating his feelings, thoughts, concerns, or experiences. The patient's descriptions of internal and external events are unusually unclear or ambiguous from the listener's perspective. PROBLEMS WITH AFFECT 1.9 Problems with affect This item is designed to tap various problems related to experiencing or expressing feelings and emotions. Relevant types of problems in this area include: a. Patient is unable to identify his feelings, put them into words, or acknowledge their presence or importance. b. Patient's manifest affect is incongruent with the content of the material that he presents. This incongruence may reflect over-control (e.g., obsessional tendencies) or undercontrol (e.g., histrionic tendencies) of affect. c. Patient displays a notable flatness of affect; that is, the patient's attitude has a remote, 157 withdrawn or bland quality. d. Patient displays a restricted range of affect. A restricted range of affect may be indicated if the patient's voice quality and feeling state do not seem to vary. E. SELF-DEROGATION 1.10 Self-rejection tendencies The patient expresses shame, self-hatred, or a sense of failure in excess of what the situation would suggest. This item is meant to capture a masochistic or hopeless quality (in contrast to a more constructive, help-seeking attitude) in the patient's negative statements about himself. Persons seeking psychotherapy often express dissatisfaction with various aspects of their behavior and/or have problems with self-esteem. This item requires the rater to distinguish between expected difficulties of this type and self-rejection that is associated with either a masochistic or hopeless attitude. F. UNREALISTIC EXPECTATIONS OR GROSS MISCONCEPTIONS OF PSYCHOTHERAPY 1.11 Inapprgpriate expectations The patient expresses unrealistic expectations about the potential outcome of psychotherapy or has misconceptions about the therapeutic process. Signs of inappropriate expectations include: a. Patient makes comments suggesting that he expects therapy to alleviate his distress with little or no effort on his part. This general attitude may be manifested by signs that the patient views the therapist as somone who will dispense a “cure." b. Patient seems to view the therapist as all- knowing. Consequently the patient believes that his distress will be relieved if he merely adopts the therapist's ideas. c. Patient expects to experience dramatic and enduring change within a short time (i.e., after a few therapy sessions). G. 158 NEGATIVE ATTITUDES 1.12 Negative attitudes toward the therapist or therapy The patient manifests attitudes toward the therapist or therapy that undermine therapeutic collaboration. The following considerations apply when rating this item: (1) It should not be rated when the patient's negative attitudes are reasonable reactions to the therapist's behavior. In such cases, select an appropriate item from one of the therapist-oriented sections of the manual. (2) Negative attitudes toward the therapist or therapy may be expected to appear at some point in a psychotherapeutic relationship and can be therapeutically useful if the patient is able to observe and examine them. Hence, this item should be rated if the patient manifests negative attitudes and seems either unable or unwilling to examine them. This item is clearly appropriate when (a) the patient manifests negative attitudes in the rater's judgment, (b) the therapist makes some attempt to explore these attitudes, and (c) the patient is unreceptive to this. Alternatively, when the patient manifests negative attitudes and neither the patient nor the therapist explicitly comment on this, rate this and item 3.5. Examples of negative attitudes include: a. Mistrustfulness: The patient appears to be suspicious of the therapist's motives or integrity. b. Negativism: The patient continually doubts or challenges the therapist's motives or integrity. c. Anger: The patient directs anger, resentment, or frustration specifically toward the therapist or therapy. DEFICIENT EGO RESOURCES The patient has weaknesses in facets of personality functioning that are commonly assumed to be highly relevant to a person's ability to benefit from brief, dynamic psychotherapy. The specific ego functions to be noted are: 1.13 Deficient sense of self and/or reality 1.14 1.16 159 The patient makes statements indicating that he experiences himself or things outside himself as unreal, inconsistant, or lacking cohesion. Disordered thought, language, and communication The patient manifests signs of what has traditionally been referred to by the global concept of “formal thought disorder." Appropriate evidence for rating this item includes: a. 'Derailment,‘ e.g., ideas expressed in sequence are unrelated or only obliquely related. b. Tangentiality, e.g., the patient's replies to questions are so irrelevant that they seem bizarre to the listener. c. Loss of goal, e.g., the patient's train of thought wanders to the extent that he never returns to the original subject. (DSM-III may be referred to for more comp- prehensive definitions of thought, language, and communication disorders.) Deficient impulse control The patient tends to engage in bahaviors of an aggressive, sexual, or self-destructive nature (e.g., drug or alcohol abuse) that maintain or create his problems. Lack of emotional ties to others, schizoid tendencies The patient expresses a sense of isolation, inner emptiness, a lack of empathy, and/or disdain for most other peOple. Lack of emotional ties may also be inferred if the patient's descriptions of his relationships suggest distance and the absence of mutual caring. 22 not rate this item if the patient des- cribes a sense of iEElation and/or poor or few relationships, but is distressed about this and seems interested in improving his interpersonal functioning. Deficiencies in autonomous functioning The patient makes statements suggesting that he is consistently unable to meet the responsi- 160 bilities asociated with his typical social roles (e.g., spouse, employee, student, or parent). THERAPIST PERSONAL QUALITIES AND ATTITUDES A. DEFICIENCIES IN THERAPEUTIC COMMITMENT 2.1 2.2 Lack of warmth The therapist's responses to the material pre- sented by the patient suggest insufficient concern and caring. Although objectivity, neutrality and some degree of personal distance are appropriate, the therapist seems cold and impervious to the patient's suffering. Exploitative tendencies The therapist makes interventions that seem to primarily serve his own psychological needs rather than therapeutic aims. Signs of exploitativeness include: a. Therapist's interest in the patient's inter- personal relationships or sexual behavior has a voyeuristic quality. b. Therapist makes self-disclosures of an overly intimate or otherwise inappropriate nature. c. Therapist's interventions seem to primarily serve a personal desire to be liked. From an observer's perspective, the therapist may appear to be overly solicitous or may seem to be trying to be a ”nice guy.” d. Therapist's interventions seem to primarily serve a personal need to be admired or noticed. The therapist's demeanor may have a theatrical quality or he may make self- aggrandizing remarks. B. REJECTING ATTITUDES: INSUFFICIENT REGARD FOR THE PATIENT'S SELF-ESTEEM 2.3 Lack of respect for the patient The therapist behaves in ways that suggest he views the patient as inferior to himself. Signs of insufficient respect for the patient include: 161 a. Therapist's interventions or offhand remarks have a condescending quality. b. Therapist makes comments that could readily be construed by the patient as signs that the therapist views him as abberrant or abnormal. c. Therapist makes comments or behaves in ways that would be considered rude in most social situations. Example: Therapist abruptly terminates a session by interrupting the patient. Critical tendencies The therapist's comments convey disapproval of the patient's actions, intentions, feelings, or ideas. The therapist's observatins regarding the patient's behavior or attitudes have a fault- finding, blaming quality. CONTROLLING TENDENCIES: INSUFFICIENT REGARD FOR PATIENT'S AUTONOMY 2.5 2.6 Moralistic tendencies The therapist seems highly committed to particular values and/or a particular life-style. Commitment to this personal philOSOphy reduces the therapist's ability to allow the patient to explore his own values and attitudes. Evidence for this item includes: a. Therapist's interventions have a moralizing or lecturing quality. b. Therapist offers unsolicited advice to the patient. Tendency to dominate the patient The therapist seems to assume an overbearing stance vis-a-vis the patient. Evidence for this item includes: a. Therapist presses the patient to accept his point of view or observations. Example: Therapist interrupts the patient specifically to reiterate an observation or interpretation that 162 the patient had previously rejected during the session. b. Therapist fails to allow patient to participate in decisions regarding the treatment. Example: Therapist tells patient that he had decided to increase the number of sessions per week, rather than presenting more intensive therapy as a recommendation that requires the patient's consideration. DEFICIENT ENACTMENT OF THERAPIST'S ROLE 2.7 2.9 Lacks confidence The therapist is notably tense, indecisive or hesitant Defensive behavior The therapist attempts to justify some aspect of his behavior Insufficient understanding The therapist does not seem to comprehend the significance that the patient's feelings, thoughts, or experiences have to the patient. The therapist seems insufficiently attuned to the perspective from which the patient views his problems. Rate this item on the basis of the therapist's actual statements rather than on 'omissions.‘ Evidence for insufficient understanding includes: a. Therapist does not grasp the elements of the patient's communications that seen most meaningful or crucial to the patient: The therapist seems to “miss the point.” b. Therapist's assessments of the patient's affective states seem inaccurate. 163 ERRORS IN TECHNIQUE A. FAILURE TO MAKE INTERVENTIONS 3.1 Failure to structure or focus the session The session seems aimless or lacks coherence. The therapist fails to make interventions that would help to organize the content and/or process of the therapy session. Evidence for this item includes: a. Therapist fails to identify focal therapeutic issues in the material presented by the patient. b. Therapist fails to integrate the material presented by the patient. The therapist does not identify themes or patterns in the patient's communications, reported behaviors, or manners of interacting with the therapist. c. Therapist lets the patient flounder, ramble, and/or repeatedly pursue tangents. Failure to address maladaptive behaviors or distorted apperceptions The therapist fails to call attention to the patient's maladaptive behavior or seriously distorted apperceptions. Evidence for this item includes. a. Therapist fails to comment on aspects of the patient's bahavior and/or attitudes that are likely to maintain his problems or create add- itional difficulties for him. Example: Patient says that he intends to quit his job without notice and the therapist does not attempt to explore with the patient either the likely consequences of this action or the dynamic processes that prompt the patient's intended action. b. Therapist fails to address probable distortions in the patient's apperceptions of self or others. Example: Patient says that whenever he walks by a group of strangers, he knows that they ridicule him after passing. 164 The therapist fails to make an intervention that would encourage the patient to examine his evidence for this assertion. Insufficient examination of potentially harm- ful behaviors or attitudes The therapist does not direct the patient's attention to behaviors and attitudes that could either induce severe psychological distress in others or result in physical harm to the patient or someone else. Failure to address signs of resistance The therapist fails to draw attention to patient behaviors that are commonly interpreted as signs of avoidance of the therapeutic task. Example: Patient repeatedly moves to tOpics apprOpriate for social conversation (e.g., the weather, politics) and the therapist fails to comment on this tendency. Failure to examine the patient-therapist interaction Therapist's interventions reflect inadequate attention to problematic aspects of the patient- therapist interaction. Evidence for this item include: a. Therapist fails to address obvious manifes- tations of the patient's attitude towards him (e.g., neither the patient's idealization of the therapist nor the patient's negative attitudes are noted). b. Therapist becomes embroiled in the patient's maladaptive patterns of relating to others. He seems unable to maintain sufficient distance from the relationship to respond in an objective, neutral manner. Example: The therapist responds to patient's passivity or resistance (frequent silences, terse remarks) by in- creasing his verbal activity (e.g., by asking more questions). 165 B. INAPPROPRIATE OR INADEQUATE INTERVENTIONS 3.6 Superficial interventions The therapist's interventions do not seem to enhance the patient's self-awareness, promote a new level of understanding, or provide the patient with a new perspective on a problem. Signs of superficial interventions include: a. Therapist's interventions consist mainly of restatements of the patient's responses (i.e., he seems to be merely 'parroting' the patient's statements). b. The therapist's interventions seem more like social responses than observations likely to enhance the patient's self-understanding. 3.7 Poorly-timed interpretations The therapist offers interpretations or obser- vations that are ”premature“ or otherwise poorly- timed. That is, although the therapist's interpretations seem valid, they are unlikely to be received by the patient due to his current affective state and/or view of his situation. C. POTENTIALLY HARMFUL INTERVENTIONS 3.8 Destructive interventions Therapist's persistent probes or direct confron- tations seem to excessively raise the patient's anxiety or heighten defensiveness. 3.9 Inappropriate use of silence The therapist relies on the use of silence to the extent that it seems to unduly raise the patient's anxiety or impair his ability to resume communication. The tension produced by these protracted silences may serve to disrupt therapeutic rapport. D. INFLEXIBLE USE OF THERAPEUTIC TECHNIQUES 3.10 Inflexible use of therapeutic techniques The therapist's commitment to a particular conceptual framework or therapeutic procedure prevents him from responding sensitively, A. POOR 4.1 166 flexibly, or in a manner that is meaningful to the patient. a. Therapist's interest in pursuing a particular dynamic formulation of the patient's problems leads him to be insufficiently attentive to alternative conceptualizations of focal issues. b. Therapist fails to adapt his standard techniques to the special requirements of certain situations or the unique character- istics of the patient. Examples: 1. Therapist persistently makes trans- ference interpretations without giving sufficient attention to pressing reality issues. 2. Therapist's interventions are too abstract or his language is too technical to be understood by the patient. PATIENT-THERAPIST INTERACTION THERAPEUTIC RELATIONSHIP Problems in the therapeutic relationship There seems to be disturbances in the affective climate of the patient-therapist relationship. Relevant evidence for problems in the relation- ship includes: a. Patient and therapist seem to be engaged in a power struggle to gain control or supremacy in the session. b. Patient and therapist appear to be relating in a detached or perfunctory manner; a sense of rapport and engagement is missing. c. The quality of the interaction appears to exceed the limits of intimacy considered appropriate for a therapeutic relationship. This problem may be manifested by the emotional intensity of the encounter, the content of the discussion, or by exchanges which seem to foster excessive mutual dependence. 167 B. POOR COLLABORATION 4.2 Inadequate therapeutic collaboration The character of the patient-therapist interaction suggests that there is inadequate agreement or clear disagreement regarding identification of a therapeutic focus, the goals of the treatment or the therapeutic procedures employed to achieve these goals. In contrast to the preceeding item (4.1), the emphasis here is on the task-related dimensions of the alliance. Relevant evidence for this type of problem includes: a. Patient and therapist seem to be pursuing incompatible strategies for ameliorating the patient's distress. b. Patient and therapist hold differing ex- pectations regarding the goals of treatment. Example: Patient suggests that his primary goal is to resolve a circumscribed problem (his poor study habits), whereas the therapist insists that the focal issue is the patient's maladaptive patterns of relating to others. GLOBAL SESSION RATINGS A. POOR MATCH BETWEEN PATIENT CHARACTERISTICS AND THE THERAPEUTIC APPROACH 5.1 Poor match between patient and therapeutic approach Ratings on this item should reflect your impression as to whether the patient posesses the qualities and capacities commonly cited as necessary for success in brief, dynamic therapy. That is, to what extent do you see the negative aspects of this session as a function of a poor match between the characteristics of the individual patient and the therapeutic approach offered by the therapist? B. DULL INTERACTION 5.2 168 Dull interaction When rating this item, focus on your personal reaction (as an external observer) to the content of the session and the nature of the exchange. A high rating is merited if the interaction strikes you as plodding, humdrum, laborious or in general fails to stimulate your interest. C. NEGATIVE IMPACT OF THE THERAPY SESSION 5.3 Destructiveness of the therapy session Ratings on this item reflect the extent to which you consider this therapy session to have had a detrimental effect on the patient's psychological well—being. In making this judgment consider the potentially damaging impact of single interventions within the context of the total therapy session. Ineffectiveness When rating this ductivity of the facilitating the ratings indicate of the therapy session item, evaluate the overall pro- therapy session in terms of goals of treatment. Higher your clinical judgment that the session was unlikely to further therapeutic aims. APPENDIX D POST-THERAPY CLIENT QUESTIONNAIRE APPENDIX D POST-THERAPY CLIENT QUESTIONNAIRE For each item choose the answer which you feel best describes your therapy experience. 1. How much in need of further therapy do you feel now? No need at all Slight need Could use more Considerable need Very great need 2. What led to termination of your therapy? My decision My therapist's decision Mutual agreement External factors *3. How much have you benefitted from your therapy? A great deal A fair amount To some extent Very little Not at all *4. Everything considered, how satisified are you with the results of your psychotherapy experience? Extremely dissatisified Moderately dissatisified Fairly dissatisified Fairly satisified Moderately satisified Highly satisified Extremely satisified *Questions used in this study. 169 170 What impression did you have of your therapist's level of experience? Extremely inexperienced Rather inexperienced Somewhat inexperienced Fairly experienced Highly experienced Exceptionally experienced How well did you feel you were getting along before therapy? Very well Fairly well Neither well nor poorly Fairly poorly Very poorly Extremely poorly How long before entering therapy did you feel in need of professional help? Less than 1 year 1 - 2 years 3 - 4 years 5 - 10 years 11 - 15 years 16 - 20 years How severly disturbed did you consider yourself at the beginning of your therapy? Extremely disturbed Very much disturbed Moderately disturbed Somewhat disturbed Very slightly disturbed How much anxiety did you feel at the time you started therapy? A tremendous amount A great deal A fair amount Very little None at all 10. *11. 12. 13. 171 How great was the internal “pressure" to do something ' about these problems when you entered psychotherapy? Extremely great Very great Fairly great Relatively small Very small Extremely small How much do you feel you have changed as a result of psychotherapy? A great deal A fair amount Somewhat Very little Not at all How much of this change do you feel has been apparent to others? (a) People closest to you (husband, wife, etc.) A great deal A fair amount Somewhat Very little Not at all (b) Close friends A great deal A fair amount Somewhat Very little Not at all (c) Co-workers, acquaintances, etc. A great deal A fair amount Somewhat Very little Not at all On the whole, how well do you feel you are getting along now? Extremely well Very well Fairly well Neither well nor poorly Extremely poorly 14. *15. 16. 17. 172 How adequately do you feel you are dealing with any present problems? Very adequately Fairly adequately Neither adequately nor inadequately Somewhat inadequately Very inadequately To what extent have your complaints or symptoms that brought you to therapy changed as a result of treatment? Completely disappeared Very greatly improved Considerably improved Somewhat improved Not at all improved Got worse How soon after entering therapy did you feel any marked change? weeks of therapy (approximate) How strongly would you recommend psychotherapy to a close friend with emotional problems? Would strongly recommend it Would mildly recommend it Would recommend it, but with some reservations Would not recommend it Would advise against it Please indicate to what extent each of the following statements describes your therapy experience. Disregard that at one point or another in therapy you may have felt differently. 18. 19. 20. - Strongly agree - Mildly agree Undecided - Mildly disagree - Strongly disagree \DxlUIwH I The following questions were rated on the above scale. My therapy was an intensely emotional experience. My therapy was often a rather painful experience. I remember very little about the details of my 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 173 therapeutic work. My therapist almost never used technical terms. On the whole I experienced very little feeling in the course of therapy. There were times when I experienced intense anger toward my therapist. I feel the therapist was rather active most of the time. I am convinced that the therapist respected me as a person. I feel the therapist was genuinely interested in helping me. I often felt I was “just another patient.“ The therapist was always keenly attentive to what I had to say. The therapist often used very abstract language. He very rarely engaged in small talk. The therapist tended to be rather stiff and formal. The therapist's manner was quite natural and unstudied. I feel that he often didn't understand my feelings. I feel he was extermely passive. His general attitude was rather cold and distant. I often had the feeling that he talked too much. I was never sure whether the therapist thought I was a worthwile person. I had a feeling of absolute trust in the therapist's integrity as a person. I felt there usually was a good deal of warmth in the way he talked to me. The tone of his statements tended to be rather cold. The tone of his statements tended to be rather neutral. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 174 I was never given any instructions or advice on how to conduct my life. The therapist often talked about psychoanalytic theory in my sessions. A major emphasis in treatment was upon my attitudes and feelings about the therapist. A major emphasis in treatment was upon my relationship with peOple in my current life. A major emphasis in treatment was upon childhood experiences. A major emphasis in treatment was upon gestures, silences, shifts in my tone of voice and bodily movements. I was almost never given any reassurances by the therapist. My therapist showed very little interest in my dreams and fantasies. I usually felt I was fully accepted by the therapist. I never had the slightest doubt about the therapist's interest in helping me. I was often uncertain about the therapist's real feelings toward me. The therapist's manner of speaking seemed rather formal. I feel the emotional experience of therapy was much more important in producing change than intellectual understanding of my problems. My therapist stressed intellectual understanding as much as emotional experiencing. APPENDIX E POST-THERAPY THERAPIST QUESTIONNAIRE APPENDIX POST-THERAPY THERAPIST QUESTIONNAIRE HOPKINS PSYCHIATRIC RATINGS Categories: - None - Slight - Mild Moderate - Marked - Severe - Extreme O‘U’Ithl-‘Q I Somatization Obsessive-Compulsive Interpersonal Sensitivity Depression Anxiety Hostility Phobic Anxiety Paranoid Ideation Psychoticism Global Pathology Index 175 E H P‘ re H h‘ WU‘UIU‘U'IU‘UIU'IW U'l mmmmmmmmmm 176 Please rate each of the following items, comparing the client with other clients whom you see in psychotherapy using the following scale: 1 - Very little 3 - Some 5 - Moderate 7 - Fairly great 9 - Very great 11. Defensiveness 1 3 5 7 9 12. Anxiety 1 3 5 7 9 13. Ego strength 1 3 5 7 9 14. Degree of disturbance 1 3 5 7 9 15. Capacity of insight l 3 5 7 9 16. Overall adjustment 1 3 5 7 9 17. Personal like for patient 1 3 5 7 9 18. Motivation for therapy 1 3 5 7 9 19. Improvement expected (Prognosis) 1 3 5 7 9 20. Degree to which counter- 1 3 5 7 9 transference was a problem in therapy. 21. Degree to which you usually 1 3 5 7 9 enjoy working with this kind of patient in psychotherapy. *22. Degree of symptomatic improvement 1 3 5 7 9 23. Degree of change in basic 1 3 5 7 9 personality structure. 24. Degree to which you felt warmly l 3 5 7 9 toward the patient. 25. How much of an “emotional 1 3 5 7 9 investment" did you have in this patient. 26. Degree to which you think 1 3 5 7 9 the patient felt warmly toward you. 177 *27. Overall success of therapy 1 3 5 7 9 28. How would you characterize 1 3 5 7 9 your working relationship with this patient? 1 - Extremely poor 3 - Fairly poor 5 - Neither good nor poor 7 - Fairly good 9 - Extremely good *29. How satisified do you think 1 3 5 7 9 the patient was with the results of his therapy? 1 - Extremely dissatisified 3 - Fairly dissatisified 5 - Neither satisified nor dissatisified 7 - Fairly satisified 9 ~ Extremely satisified 30. How would you characterize 1 3 5 7 9 the form of psychotherapy you conducted with this patient? 1 3 5 7 9 Largely supportive Intensive analytical 31. Do you recall any strikingly pleasant experiences that you had during the therapy sessions with this patient? If yes, please mark the number that best indicated the degree of pleasantness. Otherwise mark '0' for No. l 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Mildly pleasant Extremely pleasant 178 32. Do you recall any strikingly unpleasant experiences you had with this patient? If yes, please mark the number that best indicates the de- gree of unpleasantness. Otherwise mark '0' for no. 1 2 3 4 5 6 7 8 9 l 2 3 4 5 6 7 8 9 Mildly pleasant Extremely pleasant 33. Overall, how would you characterize your exper- ience with this patient? 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Unpleasant Pleasant *Questions used in this study. APPENDIX F SYMPTOM DISTRESS CHECKLIST CLIENT FORM (SCL-90) Copyright (c) Leonard R. Derogatis, Ph.D. and Nick Mellisaratos, 1976 PLEASE NOTE: Copyrighted materials in this document have not been filmed at the request of the author. They are available for consultation, however, in the author's university library. These consist of pages: P. 179-184 University. MICl'OfIImS lntemational 300 N. ZEEB RD. ANN ARBOR, Ml 48106 (3.13) 7614700 APPENDIX F SYMPTOM DISTRESS CHECKLIST CLIENT FORM (SCL-90) INSTRUCTIONS: Below is a list of problems and complaints that pe0p1e sometimes have. Please read each one carefully. After you have done so please circle one of the numbers to the right that best describes how much that problem has bothered or distressed you during the past couple weeks including today. Circle only one number for each problem and do not skip any items. Please read the example before beginning. CATEGORIES: 0 - Not at all 1 - A little bit 2 - Moderately 3 - Quite a bit 4 - Extremely EXAMPLE: How much were you bothered by 1. Backaches. By circling #1, this person answered that he/she was a little bit bothered by backaches. l. Headaches 0 l 2 3 4 2. Nervousness or shakiness inside 0 1 2 3 4 3. Unwanted thoughts, words, or 0 1 2 3 4 ideas that won't leave your mind. 4. Faintness or dizziness 0 l 2 3 4 5. Loss of sexual interest or 0 1 2 3 4 pleasure 6. Feeling critical of others 0 l 2 3 4 7. The idea that someone else 0 1 2 3 4 can control your thoughts 179 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 180 Feeling others are to blame for most of your troubles Trouble remembering things Worried about sloppiness or carelessness Feeling easily annoyed or irritated Pains in heart or chest Feeling afraid in open spaces or on the streets Feeling low in energy or slowed down Thoughts of ending your life Hearing voices that other peOple do not hear Trembling Feeling that most peOple cannot be trusted Poor appetite Crying easily Feeling shy or uneasy with the opposite sex Feeling of being trapped or caught Suddenly scared for no reason Temper outbursts that you could not control Feeling afraid to go out of your house alone Blaming yourself for things Pains in lower back Feeling blocked in getting things done 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 181 .Feeling lonely Feeling blue Worrying too much Feeling no interest in things Feeling fearful Your feelings being easily hurt Other people being aware of your private thoughts Feeling others do not under- stand you or are unsympathetic Feeling that people are unfriendly or dislike you Having to do things very slowly to ensure correctness Heart pounding or racing Nausea or upset stomach Feeling inferior to others Soreness of your muscles Feeling that you are watched Trouble falling asleep Having to check and double— check what to do Difficulty making decisions Feeling afraid to travel on buses, subways, or trains Trouble getting your breath Hot or cold spells Having to avoid certain things, places, or activities because they frighten you Your mind going blank MN N 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 182 Numbness or tingling in parts of your body A lump in your throat Feeling hopeless about the future Trouble concentrating Feeling weak in parts of your body Feeling tense or keyed up Heavy feelings in your arms or legs Thoughts of death or dying Overeating Feeling uneasy when peOple are watching or talking about you Having thoughts that are not your own Having urges to hurt, injure, or harm someone Awakening in the early morning Having ideas or beliefs that others do not share Sleep that is restless or disturbed Having urges to break or smash things Having ideas or beliefs that others do not share Feeling very self-conscious with others Feeling uneasy in crowds such as shopping or at a movie Feeling everything is an effort 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 183 Spells of terror or panic Feeling uncomfortable about eating or drinking in public Getting into frequent arguments Feeling nervous when you are left alone Others not giving you prOper credit for your achievements Feeling lonely even when you are with peOple Feeling so restless you couldn't sit still Feeling of worthlessness Feeling that familiar things are strange or unreal Shouting or throwing things Feeling afraid you will faint in public Feeling that peOple will take advantage of you if you let them Having thoughts about sex that bother you a lot The idea that you should be punished for your sins Feeling pushed to get things done The idea that something serious is wrong with with your body Never feeling close to another person Feelings of guilt 184 90. The idea that something is wrong with your mind L I ST OF REFERENCES REFERENCES Ashby, J. D., Fort, D. H., Guerney, B. B., and Guerney, L. F. (1957). Effects on clients of a reflective and a leading type of therapy. Psychological Monographs, 71:24. Baker, E. (1960). The differential effects of two psychotherapeutic approaches on client perception. Journal of Counseling Psychology, 1, 46-50. Barrington, B. (1961). Prediction from counselor behavior of client perception and of case outcome. Journal of Counseling Psychology, 8, 37-42. Beck, A. T., Rush, A. J., Shaw, B. F., and Emery, G. (1979). ngnitive therapy of depression. New York: Guilford. Bergin, A. E. (1963). The effects of psychotherapy: Negative results revisited. Journal of Counseling Psychology, 10, 244-250. Bergin, A. E. (1966). Some implications of psychotherapy research for therapeutic practice. Journal of Abnormal Psychology, 11, 235-246. Bergin, A. E., and Jasper, L. G. (1969). Correlates of empathy in psychotherapy: A replication. Journal of Abnormal Psychology, 14, 477-481. Bordin, E. S. (1974). Research strategies in psychotherapy. New York: John Wiley. Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research and Practice, 16, 252-260. Butler, J. M., Rice, L. N., and Wagstaff, A. K. (1962). On the naturalistic definition of variables: An analogue of clinical analysis. In H. H. Strupp and L. L. Luborsky (Eds.), Research in psychotherapy, Vol. 2. Washington, D. C.: American Psychological Association. 185 186 Cabral, R. J., Best, J., and Paton, A. (1975). Patients' 'and observers' assessments of process and outcome in group therapy: A follow-up study. The American Journal of Psychiatry, 132, 1052-1054. Crowder, J. E. (1972). Relationship between therapist and client interpersonal behaviors and psychotherapy outcome. Journal of Counseling Psychology, 12, 68-75. Davanloo, H. (1980). Short-term dynamic psychotherapy. New York: Jason Aronson. Derogatis, L. R. (1977). The SCL-90R manual 1: Scoring, administration and procedures for the SCL-90R. Baltimore: Johns Hopkins University School of Medicine, Clinical Psychometrics Unit. Derogatis, L. R., Abeloff, M., and McBeth, C. (1976). Cancer patients and their physicians in the perception of psychological symptoms. Psychometrics, ll, Dickes, R. (1975). Technical considerations of the therapeutic and working alliances. International Journal of Psychoanalytic Psychotherapy, 4, 1-24. Eaton, T. T. (1984). An analysis of therapist and client verbal responses in successful and unsuccessful insight-oriented psychotherapy. Unpublished thesis, Department of Psychology, Michigan State University. Elliott, R., Cline, J., and Shulman, R. (1982). Effective processes in psychotherapy: A single case study using four evaluative paradigms. Unpublished manuscript, Department of Psychology. The University of Toledo, Toledo, Ohio. Fenichel, O. (1941). Problems of psychoanalytic technique. Albany: Psychoanalytic Quarterly. Filak, J. (1982). Congruence of perception on client symptoms and therapy outcome. Unpublished thesiE, Department of Psychology. Michigan State University. Franks, L. M., and May, D. T. (1980). Negative effects revisited: A rejoinder. Professional Psychology, 2' 191-105. Freud, A. (1946). The psychoanalytical treatment of children. London: Imago. Freud, S. (1910). The future prospects of psychoanalytic therapy. Standard Edition, 11. London: Hogarth Press, 1953. 187 Freud, S. (1912). Dynamics of transference. Standard Edition, 43. London: Hogarth Press, 1953. Freud, S. (1913). On beginning the treatment. Standard Edition, 44. London: Hogarth Press, 1953. Frieswyk, S. H., Allen, J. G., Colson, D. B., Coyne, L., Gabbard, G. 0., Horwitz, L., and Newsom, G. (1986). Therapeutic alliance: Its place as a process and outcome variable in dynamic psychotherapy research. Journal of Consulting and Clinical Psychology, 44, 32-38. Garfield, S. L., and Bergin, A. E. (1971). Therapeutic conditions and outcome. Journal of Abnormal Psychology, 11, 108-114. Gendlin, E. T. (1973). Experiential psychotherapy. In R. Corsini (Ed.), Current psychotherapies. Ithaca, Ill.: Peacock. Gendlin, E. T., Beebe, J., Cassens, J., Klein, M. H., and Oberlander, M. (1968). Focusing ability in psychotherapy, personality and creativity. In J. M. Schlien, H. F. Hunt, J. D. Matarazzo, and C. Savage (Eds.), Research in psychotherapY. Vol. 3. Washington, D. C.: American Psychological Association. Gendlin, E. T., Jenney, R., and Schlien, J. (1960). Counselor ratings of process and outcome in client- centered therapy. Journal of Clinical Psychology, 44, 210-213. Gomes-Schwartz, B. (1978a). Effective ingredients in psychotherapy: Prediction of outcome from process variables. Journal of Consulting and Clinical Psychology, 44, 1023-1035. Comes-Schwartz, B. (1978b). Development of a scale to predict negative change inppsychotherapy. Paper presented at the annual meeting of the Society for Psychotherapy Research, Toronto. Greenson, R. R. (1964). The working alliance and the transference neurosis. Psychoanalyticiguarterly, 25. 155-181. Greenson, R. R. (1967). The technique and practice of psychoanalysis, Vol. 1. New York: International Universities Press. Hadley, S. W., and Strupp, H. H. (1976). Contemporary views on negative effects: An integrated account. 188 Archives of General Psychiatry, 44, 1291-1302. Halkides, G. (1958). An investigation of therapeutic success as a function ofifour therapist vafiables. Unpublished doctoral dissertation, University of Chicago. Hartley, D. E. (1978). Therapeutic alliance and the success of brief individual psychotherapy. Unpublished doctoral dissertation, Vanderbilt University. Hartley, D. E. (1985). Research on the therapeutic alliance in psychotherapy. In R. E. Hales and A. J. Frances (Eds.), Annual review, Vol. 4. Washington, D. C.: American Psychiatric Press. Horn-George, J. B., and Anchor, K. N. (1982). Perceptions of the psychotherapy relationship in long- versus short-term therapy. Professional Psychology, 44, 483-491. Horowitz, M. J., Marmar, C., Weiss, D. B., DeWitt, K. N., Rosenbaum, R. (1984). Brief psychotherapy of bereavement reactions. Archives of General Psychiatry, 44, 438-448. Horwitz, L. (1974). Clinical prediction in psychotherapy. New York: Jason Aronson. Huxster, H., Lower, R., and Escoll, P. (1975). Some pitfalls in the assessment of analyzability in a psychoanalytic clinic. Journal of the American Psychoanalytic Association, 44, 90-106. Karl, N. J., and Abeles, N. (1969). Psychotherapy process as a function of the time segment sampled. Journal of Consulting and Clinical Psychology, 44, 207-212. Karon, B. P., and Vandenbos, G. R. (1981). Psychotherapy of schi20phrenia: The treatment of choice. New York: Jason Aronson. Kiesler, D. J. (1973). The process of psychotherapy. Chicago: Aldine. Kirtner, W. L., and Cartwright, D. S. (1958). Success and failure in client-centered therapy as a function of initial in-therapy behavior. Journal of Consulting Psychology, 44, 329-333. Kirtner, W. L., Cartwright, D. S., Robertson, R. J., and Fiske, D. W. (1961). Length of therapy in relation to outcome and change in personal integration. 189 Journal of Consulting Psychology, 44, 84-88. Kohut, H. (1971). The analysis of the self. New York: International Universities Press. Kohut, H. (1977). The restoration of the self. New York: International Univeréities Press. Kraemer, H. C. (1981). Coping strategies in psychiatric clinical research. Journal of Consulting and Clinical Kurtz, R. R., and Grummon, D. L. (1972). Different approaches to the measurement of therapist empathy and their relationship to therapy outcomes. Journal of Consulting and Clinical Psychology, 44, 106-115. Lambert, M. J., Bergin, A. E., and Collins, J. L. (1977). Therapist-induced deterioration in psychotherapy. In A. S. Gurman and A. M. Razin (Eds.), Effective psychotherapy: A handbook of research. New York: Pergamon Press. Langs, R. (1973). The technique of psychoanalytic psychotherapy, Vol. 1. New York: Jason Aronson. Langs, R. (1974). The technique of psychoanalytic psychotherapy. Vol. 2. New York: Jason Aronson. Leary, T. (1957). Interpersonal diagnosis of personality. New York: Ronald Press. Lehrke, S. A. (1977). Working alliance development ear4y in psychotherapy. Unpublished doctoral dissertation, University of Florida. Lorr, M., and McNair, D. M. (1964). The interview relationship in therapy. Journal of Nervous and Mental Disease, 139, 328-331. Luborsky, L. L. (1976). Helping alliances in psychotherapy. In J. L. Claghorn (Ed.), Successful Psychotherapy. New York: Brunner/Mazel. Mainard, W. A., Burk, H. W., and Collins, L. G. (1965). Confrontation versus diversion in group therapy with chronic schizophrenics as measured by a 'positive incident' criterion. Journal of Clinical Psychology, 44, 222-225. Malan, D. H. (1979). Individual psychotherapy and the science of psychodynamics. London: Butterworth. Marziali, E. (1984). Three viewpoints on the therapeutic 190 alliance. The Journal of Nervous and Mental Disease, 172, 417-423. Marziali, E., Marmar, C., and Krupnick, J. (1981). Therapeutic alliance scales: Development and relationship to psychotherapy outcome. American Journal of Psyehiatry, 138, 361-364. Mintz, J., Luborsky, L., and Auerbach, A. (1971). Dimensions of psychotherapy: A factor-analytic study of ratings of psychotherapy sessions. Journal of Consulting and Clinical Psychology, 44, 106-120. Moras, R., and StruPp. H. H. (1982). Pretherapy interpersonal relations, patients' alliance, and outcome in brief therapy. Archives of General Psychiatgy, 44, 405-409. Morgan, R., Luborsky, L., Crits-Christoph, P., Curtis, H., and Solomon, J. (1982). Predicting the outcomes of psychotherapy by the Penn Helping Alliance rating method. Archives of General Psychiatry, 44, 397-402. Morgan, R., Luborsky, L., Curtis, H., and Solomon, J. (1977). Therapeutic alliance, therapist behaviors, insight, and resistance: All predictors of outcome in psychotherapy? Unpublished manuscript, University of Pennsylvania. Mullen, J., and Abeles, N. (1971). Relationship of liking, empathy, and therapist's experience to outcome of therapy. Journal of Counseling Psychology, 44, 39-430 Nagy, T. F. (1973). Therapist level of functioning and change in clients' quantifiable anxiety level and verbal behavior. 24§sert§tion Abstracts International, 44, 8783-8798. Nunberg, W. (1932). Princip4es of psychoanalysis. New York: International Universities Press, 1955. O'Malley, S. S., Suh, C. S., and Strupp. H. H. (1983). The Vanderbilt Psychotherapy Process Scale: A report on the scale development and a process-outcome study. Journal of Consultingand Clinical Psychology, 44, 581-586. Orlinsky, D. E., and Howard, K. I. (1978). The relation of process to outcome in psychotherapy. In S. L. Garfield and A. E. Bergin (Eds.), Handbook of psychotherapy and behavior change: An empirical analysis (2nd. Ed.). New York: Wiley. 191 Rice, L. N. (1965). Therapist's style of participation and case outcome. Journal of Consulting Psychology, 44, 155-160. Rice, L. N. (1974). The evocative function of the therapist. In D. A. Wexler and L. N. Rice (Eds.), Innovations in client-centered therapy. New York: Wiley. Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting Psychology, 44, 95-103. Ryan, E. R. (1973). The capacity of the patient to enter an elementary therapeutic relationship in the initial psychotherapy interview. Unpublished doctoral dissertation, University of Michigan. Sachs, J. S. (1983). Negative factors in brief psychotherapy: An empirical assessment. gournal of Consulting and Clinical Psychology, 44, 557-564. Saltzman, C., Luetgert, M. J., Roth, C. H., Creaser, J., and Howard, L. (1976). Formation of a therapeutic relationship: Experiences during the initial phases of psychotherapy as predictors of treatment duration and outcome. Journal of Consulting and Clinical Psychology, 44, 546-555. Sandell, J. A. (1981). An empirical study of negative factors in brief psychotherapy. Unpublished doctoral dissertation, Vanderbilt University. Sandler, J., Dare, C., and Holder, A. (1973). The patient and the analyst: The basis of the psychoanalytic process. New York: International Universities Press. Sarnat, J. E. (1975). A comparison of psychoanalytic and client-centered measures of initial in-therapy patiént participation. Unpublished doctoral dissertation, University of Michigan. Schauble, P. G., and Pierce, R. M. (1974). Client in-therapy behavior: A therapist guide to progress. Psychotherapy: Theory, Research and Practice, 44, 229-234. Scher, M. (1975). Verbal activity, sex, counselor experience, and success in counseling. Journal of Counseling Psychology, 44, 97-101. Sloane, R. B., Staples, F. R., Cristol, A. H., Yorkston, N. J., and Whipple, K. (1975). Psychotherapy versus 192 behavior therapy. Cambridge: Harvard University ‘Press. Sterba, R. F. (1934). The fate of the ego in analytic therapy. International Journal of Psycho-Analysis, 44, 117-126. Sterba, R. F. (1940). The dynamics of the dissolution of the transference resistance. Psychoanalytic Stone, L. (1961). The psychoanalytic situation. New York: International Universities Press. Strachey, J. (1934). The nature of the therapeutic action of psycho-analysis. International Journal of Psycho-Analysis, 44, 127-159. Strupp, H. H. (1980a). Success and failure in time-limited psychotherapy: A systematic comparison of two cases: Comparison 1. Archives of General Psychiatry, 41, 595-603. Strupp: H. H. (1980b). Success and failure in time-limited psychotherapy: A systematic comparison of two cases: Comparison 2. Archives of General Psychiatry, 41, 708-716. StrupP. H. H. (1980c). Success and failure in time-limited psychotherapy: Further evidence: Comparison 4. Archives of General Psychiatry, 41, 947-954. StruPp, H. H., and Binder, J. L. (1984). Psychotherapy in a new key: A guide to time:limited dynamic psychotherapy. New York: Basic Books. Strupp, H. H., Hadley, S. W., and Gomes-Schwartz, B. (1977). Psychotherapy for better or worse: An analysis of thepproblem of negative effects. New York: Jason Aronson. StruPp, H. H., Keithly, L., Moras, R., Samples, 8., Sandell, J., and Waterhouse, G. (1980). Toward the measurement of negative effects in psychotherapy. Paper presented at the annual meeting of the Society of Psychotherapy Research, Pacific Grove, California. Strupp. H. H., Lessler, R., and Fox, K. (1969). Patients view their psychotherapy. Baltimore: The Johns Hopkins University Press. Strupp, H. H., Moras, K., Sandell, J., Waterhouse, G., O'Malley, S., Keithly, L., and Gomes-Schwartz, B. 193 (1981). Vanderbilt Negative Indicators Scale: An instrument for the identification of deterrents to progress in time-limited dynamic psychotherapy. Unpublished manuscript, Vanderbilt University. Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: Norton. Tomlinson, T. M., and Hart, J. T., Jr. (1962). A \ validation study of the process scale. Journal of Consulting Psychology, 44, 74-78. Tomlinson, T. M., and Stoler, N. (1967). The relationship between affective evaluation and ratings of therapy process and outcome with schizophrenics. Psychotherapy: Theory, Research and Practice, 4, 14-18. Truax, C. B. (1970a). Therapist's evaluative statements and patient outcome in psychotherapy. Journal of Clinical Psycholpgy, 44, 536-538. Truax, C. B. (1970b). Length of therapist response, accurate empathy and patient improvement. Journal of Clinical Psychology, 44, 539-541. Truax, C. B., Altman, H., Wright, L., and Mitchell, K. M. (1973). Effects of therapeutic conditions in child therapy. Journal of Community Psychology, 4, 313-318. Truax, C. B., and Carkhuff, R. R. (1967). Toward effective counseling andppsychotherapy. Chicago: Aldine. Truax, C. B., and Mitchell, K. M. (1971). Research on certain therapist interpersonal skills in relation to process and outcome. In A. E. Bergin and S. L. Garfield (Eds.), Handbook of psychotherapy and behavior change: An empirical analysis. New York: Wiley. Truax, C. B., and Wittmer, J. (1971). Patient non-personal reference during psychotherapy and therapeutic outcome. Journal of Clinical Psychology, AZ! 300-302. Truax, C. B., Wittmer, J., and Wargo, D. G. (1971). Effects of the therapeutic conditions of accurate empathy, nonpossessive warmth, and genuineness on hOSpitalized mental patients during group therapy. Journal of Clinical Psychology, 41, 137-142. Van der Veen, F. (1967). Basic elements in the process 194 of psychotherapy. Journal of ConsultingyPsychology, 44, 295-301. Wilson, G. T., and Evans, 1. M. (1976). Adult behavior therapy and the therapist-client relationship. In C. M. Franks and G. T. Wilson (Eds.), Annual review of behavior therapy: Theory and practice, Vol. 2. New York: Brunner/Mazel. Winnicott, D. W. (1955). Metapsychological and clinical aspects of regression within the psycho-analytic set-up. International Journal of Psycho-Analysis, fig. 16-26. Zetzel, E. R. (1956). Current concepts of transference. International Journal of Psycho-Analysis, 41, 369-376.