EM” ‘4'??? QEE‘EEEETEE? EEESE EESEENSE TU A1133? :EEEEEEE CLEEEE ESE. E‘ EbSESES EEE zEEE‘H’ EE3 3ES$ LEE 8%. EEETCEJ EEGE‘E . E _, THESES FOE EEE EEEEEE 0F EH. B. MTCHTTGAN STATE UNIVERSITY RGSAMOND MITCHELL -1 9 7 1 gut-.3”; This is to certify that the thesis entitled Relationship Between Therapist Response to Therapist-Relevant Client Expressions and Therapy Process and Client Outcome presented by Rosamond Mitchell has been accepted towards fulfillment of the requirements for Ph.D. degree in ng Major professor Date Januar 26, 1971 0-169 pOSiti 'J c Client pSYchn J L 1 pp» vces ABSTRACT RELATIONSHIP BETWEEN THERAPIST RESPONSE TO THERAPIST-RELEVANT CLIENT EXPRESSIONS AND THERAPY PROCESS AND CLIENT OUTCOME BY Rosamond Mitchell The present research was an investigation of the frequency and explicitness with which therapists interpret client statements in relation to themselves or the immediate therapeutic relationship and thereby focus on the client- therapist relationship. The Immediate Relationship Scale (IRS) was constructed to measure the explicitness with which therapists respond to client references to the thera- pist of varying degrees of overtness. In View of the concensus among therapists regarding the critical role in psychotherapy of the client's feelings about the therapist and the importance attached to the therapist's encouragement of client expression and explorations of such feelings, a positive relationship between therapists' IRS scores and client improvement was predicted. Two separate studies using tape recordings of actual psychotherapy sessions were conducted. The data for Study L a process study, was based on tape recordings of first Rosamond Mitchell therapy sessions of 56 different therapist—client dyads. Ratings on the IRS and the E, W, G, and DX scales by Carkhuff and his associates were obtained on five 3-minute segments excerpted from each of the 56 tapes. The results of Study I indicated that for the total sample of thera- pists, IRS scores were (1) positively related to the therapists' core condition scores and clients' DX scores, (2) ordered therapists according to Orientation from higher to lower as Relationship, Eclectic, Client-Centered, and Analytic, respectively, (3) differentiated Relationship therapists from Analytic therapists and tended to also differentiate Relationship therapists from Eclectic and Client-Centered therapists and (4) did not differentiate between inpatient and outpatient therapists or between high and low experience level therapists within any of the four orientations. The data in Study II, an outcome study, was based on tape recorded psychotherapy sessions of 40 outpatients seen by four psychiatric residents at Johns Hopkins Univer- sity. Ratings on the IRS and the E, W, and G scales by Truax were obtained on six 3-minute segments excerpted from the recordings of each client's sessions. Five measures of client outcome were used: global improvement ratings by clients and therapists, Discomfort, Target, and Social Ineffectiveness. The results indicated that although for the total sample of therapists IRS scores and core condition E1 __ 1.1-7? ' SCOI] for ti outccn indica Cf cli HOWE'JQ core ct IRS SC< therap; lesser I'l‘eaSLlre tiVely DESS W( outCOfie indica. Standi] Way in hi:Sel: Such r6 Client; Client Rosamond Mitchell scores were unrelated, IRS scores were positively related to the core condition scores for the high functioning therapists but negatively related for the low functioning therapists. The hypothesis predicting a positive relation- ship between client improvement and therapists' IRS scores for the entire sample was rejected for each of the five outcome measures; indeed, step-wise regression analyses indicated that higher IRS scores were actually predictive of client lack of improvement on one outcome measure. However, additional analyses, which took into account the core condition context of IRS scores, indicated that higher IRS scores in a context of relatively low levels of therapist—offered empathy and genuineness were related to lesser degrees of client improvement on three outcome measures. In contrast, IRS scores in a context of rela- tively high levels of therapist-offered empathy and genuine- ness were unrelated to client improvement on any of the outcome measures. These findings were interpreted as indicating that in the context of a therapist's deep under— standing and genuine responses to the client the particular 'way in which a therapist responds to client references to himself, e.g., whether he ignores or explicitly interprets such references, is relatively inconsequential to the client's improvement. But when a therapist interprets client statements in relation to himself and thus attempts to focus on the relationship in the context of the ther unde a thii ing 1, shoul of hi. Rosamond Mitchell therapist's failure to respond genuinely and to accurately understand the client's feelings, the client fails to improve or even deteriorates. Thus, the effectiveness of a therapist's IRS responses is dependent upon the accompany- ing level of his empathy and genuineness and, consequently, should be evaluated in conjunction with or in the context of his level of functioning on the core conditions. RELATIONSHIP BETWEEN THERAPIST RESPONSE TO THERAPIST-RELEVANT CLIENT EXPRESSIONS AND THERAPY PROCESS AND CLIENT OUTCOME r{-' \ Rosamond Mitchell A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1971 TABLE OF CONTENTS Page LIST OF TABLE S O I O O O O O O O O O O O O O O O O O 0 iv INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1 METHOD . . . . . . . . . . . . . . . . . . . . . . . . 46 Study I . . . . . . . . . . . . . . . . . . . . . . 46 Study II . . . . . . . . . . . . . . . . . . . . . . 50 RESULTS . . . . . . . . . . . . . . . . . . . . . . . 59 Study I . . . . . . . . . . . . . . . . . . . . . . 59 Study II . . . . . . . . . . . . . . . . . . . . . . 93 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 139 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 158 REFERENCES . . . . . . . . . . . . . . . . . . . . . . 163 APPENDICES A. Immediate Relationship Scale . . . . . . . . . . 174 B. Empathic Understanding in Interpersonal Processes Scale by Bernard G. Berenson, Robert R. Carkhuff, and J. Alfred Southworth . 180 C. Respect or Positive Regard in Interpersonal Processes Scale by Robert R. Carkhuff, J. Alfred Southworth, and Bernard G. Berenson . . . . . . . . . . . . . . . . . . . 183 D. Facilitative Genuineness in Interpersonal Processes Scale by Robert R. Carkhuff . . . . 186 E. Self-Exploration in Interpersonal Processes Scale by Robert R. Carkhuff . . . . . . . . . 189 ii A Tentative Scale for the Measurement of Accurate Empathy by Charles B. Truax . . A Tentative Scale for the Measurement of Nonpossessive Warmth by Charles B. Truax A Tentative Scale for the Measurement of Therapist Genuineness or Self-Congruence by Charles B. Truax . . . . . . . . . . A Tentative Scale for the Measurement of Depth of Self-Exploration by Charles B. Truax O O O O O O O D O O O O I O O O O Discomfort Scale . . . . . . . . . . . . . Target Symptom Scale . . . . . . . . . . . Patient Global Improvement Scale Completed by Patient 0 C O O O O O O O O O O O O 0 Patient Global Improvement Scale Completed by Therapist . . . . . . . . . . . . . . Social Ineffectiveness Scale . . . . . . . Table 36. Y's, S.D.'s, Ranges of IRS Scores and Percent of IRS Ratings at each Stage for High and Low IRS Groups and IRS Quartile Groups in Studies I and II . . . . . . . iii Page 192 197 200 202 206 211 213 214 215 218 LI ST OF TABLES Table Page 1. Therapist and Client Responses Measured by IRS O O O O O O O O O O O O O O O O I O O 20 2. Summary of Therapists' Characteristics . . . . 48 3. r's among W, G, and Core Condition Scores . . 60 4. X's and S.D.'s of E, W, G, and Core Condition Scores for High and Low Core Therapists . . . . . . . . . . . . . . 62 5. 7'5 and S.D.'s of IRS and Core Scores for High and Low Core and Core Quartile Therapists . O O C C I O O I C I C O O O O O 64 6. Percent of IRS Scores at Each IRS Stage for Core Quartile Therapists . . . . . . . . 66 7. Percent of IRS Ratings at Each IRS Stage for High and Low Core and Core Quartile Therapists I I O O O O O O O O O O O O O O O 71 8. Y's and S.D.'s of IRS Scores for In— and Out- Patient and IRS Quartile Therapists . . 75 9. Percent of IRS Ratings at Each Stage for In- and Out- Patient and High and Low IRS Therapists O I I O O O O O O O O O O O O 77 10. Percent of Total IRS Ratings at Each Stage for Inpatient and Outpatient Therapists . . 78 ll. Y's and S.D.'s of IRS and DX Scores for High and Low DX and DX Quartile Groups . . . 80 12. Percent of IRS Ratings at Each Stage for High and Low DX and DX Quartile Groups . . . 81 13. R Years of Experience and Number of Inpatients and Outpatients for Therapists Classified According to Orientation and Experience Level . . . . . . . . . . . . . . . . . . . 84 iv Tablz H L11 l6 17. 18. 19. 20. 21. 22. 23, 24. 25. 26, Table 14. 15. 16. l7. l8. 19. 20. 21. 22. 23. 24. 25. 26. Page 7's and S.D.'s of Both Transformed IRS Scores and IRS Scores Prior to Transformation for Therapists Classified According to Orientation and Experience Level . . . . . . 85 Summary of Orientation x Experience Level AOV O O O O O O O O O O I O I O O O O 87 Percent of IRS Ratings at Each Stage for Experience Level and Orientation Groups . . 87 Significant Differences in Proportion of IRS Ratings at IRS Stages for Orientation and Experience Level Groups . . . . . . . . 89 Percent of IRS Ratings at Each Stage for Experience Levels Within Orientation Groups 0 O O I O O O O O O O O O O O O O I O 91 Y's, S.D.'s and r's for 12 Predictor Variables (N = 40) . . . . . . . . . . . . . 95 X's and S.D.'s of E, w, G, ZEWG, EEG and IRS Scores for Each Therapist and for Two High and Two Low Core Therapists . . . . 97 r's Among IRS, E, W, G, ZEWG, and ZEG Scores for High and Low Core Therapists . . 99 Percent of IRS Scores at Each IRS Stage for the High and Low Core Therapists . . . . 100 Percent of IRS Ratings at Each IRS Stage for the High and Low Core Therapists . . . . 101 Summary of 2 x 2 AOV on IRS Ratings of Middle and Last Segments Selected from First Therapy Session for High and Low Core Therapists . . . . . . . . . . . . . . 105 X's and S.D.'s of Available IRS Scores on Segments Selected from First and Tenth Sessions for High and Low Core Therapists . . . . . . . . . . . . . . . . . 106 Summary of 2 x 2 x 2 AOV on Available IRS Ratings on Segments Selected from the Middle and Last Thirds of the First and Tenth Therapy Sessions for High and Low Core Therapists . . . . . . . . . . . . 107 Tabl fx) R) \O 30 31. 32. 33. 34. 35. 35. Table 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. Summary of AOV on IRS Scores Obtained on First, Tenth, and Fifth from Final Sessions for Therapists of 13 Clients for Whom Data was Available . . . . . . Y's, S.D.'s, and r's for Five Outcome Measures for 34 Cases . . . . . . . . . r's of 12 Predictor Variables with Five Outcome Measures (N = 40) . . . . . . . Summary of Significant Predictions of Each Outcome Measure Obtained from Step—Wise Regression Analyses . . . . . . . . . . X's and S.D.'s of Predictor Variables and Outcome Measures for the High and Low EEG Groups . . . . . . . . . . . . . . . Pearson r's Among Predictor Variables and Outcome Measures for the High and Low EEG Groups 0 I O O O O O O I O I O O O 0 Zero Order r's Between Each Predictor Variable and Bach Outcome Measure Within the High and Low EEG Groups . . . . . . Summary of Significant Predictions of Each Outcome Measure Obtained from Step-Wise Regression Analyses of the High and Low ZEG Groups . . . . . . . . . . . . . . . X's and S.D.'s of IRS, EEG, and Outcome Scores for 34 Cases Classified According to High, Moderate, or Low Levels of IRS and EEG Scores . . . . . . . . . . . . . Y's, S.D.'s, Ranges of IRS Scores and Percent of IRS Ratings at Each Stage for High and Low IRS Groups and IRS Quartile Groups in Studies I and II . . . . . . . vi Page 108 110 114 116 121 122 125 127 131 218 INTRODUCTION There is a concensus among therapists of all theo- retical orientations regarding the central importance of the client-therapist relationship for the successful outcome of therapy: this concensus consists of the prOposition that a "good client-therapist relationship" is essential to client improvement (Hobbs, 1962; Schofield, 1967; Patterson, 1967; Shoben, 1949). Therapists are not in agreement, however, regarding precisely what is meant by a good client-therapist relationship, and there have been diverse formulations of the nature of such a relationship in the past (Freud, 1959; Fromm-Reichman, 1950; Sullivan, 1954; Rogers, 1957; Rosen, 1953; Whitaker and Malone, 1953). More recently, however, several factor analytic studies have provided a compre- hensive yet parsimonious basis for specifying what therapists differing with respect to orientation and training actually denote by a good client-therapist rela— tionship. At the same time, these studies have empirically demonstrated the existence of a concensus among therapists regarding the central role accorded to the therapeutic relationship (McNair and Lorr, 1964; Sundland and Barker, 1962; Wallach and Struppr 1964). 2 In each of these studies factor analytic procedures were applied to therapist ratings of the extent of their agreement with a number of statements representing basic views and attitudes about therapeutic processes and tech- niques typically associated with different orientations, principally Freudian, Sullivanian, Client-Centered, and Experiential orientations. In each study only a few dimen— sions were sufficient to describe these therapist attitudes, and although some slight differences among the studies were apparent, the factors which emerged were markedly congruent. The following three factors were found in each study: the extent to which the therapist becomes personally involved in the treatment and the interpersonal relationship with the client, the extent to which the therapist uses classical psychoanalytic techniques, and the extent to which the therapist assumes active control of the treatment process. In addition, Sundland and Barker (1962) found a general factor, Analytic-Experiential, which included these three more specific factors and which they considered to be the single most important dimension upon which to describe therapists. Integrating the results of all three studies indicates that this or a quite similar dimension can be used to describe and differentiate therapists on the basis of their beliefs and attitudes about basic psychotherapy processes and techniques. The Analytic pole of this dimension describes therapists who stress the conceptuali- zation and planning of therapy, unconscious processes, 3 insight into childhood experiences, interpretation and analysis of dreams, resistance and transference, and the restriction of spontaneity and personal involvement of the therapist. In contrast, the Experiential pole describes therapists who de-emphasize unconscious processes and the conceptualization and planning of therapy and who stress the personality, spontaneity, and personal involvement of the therapist. Consistent with this description of ther- apists along such a general dimension is Patterson's (1967) conclusion that the results of these factor analytic studies offer support for his earlier proposal that the therapy process can be described by a single dimension or dichotomy, Rational-Affective. The Rational therapy process tends to be planned, objective, and impersonal, while the Affective therapy process is emphasized as being warm, personal, and spontaneous. Moreover, Wolff (1956) had earlier predicted the existence of two basically different kinds of psycho- therapy: one based upon a preconceived set of notions in which therapeutic change is accomplished by means of inter— pretations, and a second based upon an evolving personal relationship in which the relationship itself is the vehicle by which change is accomplished. The kind of therapeutic relationship therapists consider appropriate, or "good," constitutes a major com- ponent of such a general Analytic-Experiential factor. The emergence of this factor indicates that there are two basic, quite different conceptualizations among therapists 4 regarding what constitutes a good client—therapist relation- ship and that therapists of diverse orientations can be compared and differentiated by the way in which they concep- tualize a good relationship. One conceptualization portrays the therapist as being restrained, aloof, and impersonal in relation to the client and as analyzing and interpreting the client's feelings and behavior in a rational, objective manner in order to facilitate the client's achievement of insight. The parent-child relationship is the prototype for this Analytic kind of relationship (Hobbs, 1962). The con- ceptualization at the opposite pole portrays the good rela- tionship as mutually spontaneous, open, and intimate and as being the basic therapeutic agent of client change. A second series of recent factor analytic studies also have provided evidence of two different kinds of rela- tionships similar to those reflected by the Analytic- Experiential factor. The studies also provided additional empirical support for the importance generally accorded to the therapeutic relationship. Among the factors which emerged from factor analysis of therapist and client descriptions of actual therapy sessions were several dis- tinct factors specifically reflecting the relationship. Orlinsky and Howard (1967a, 1967b) found that productive or good and nonproductive or bad therapy sessions, as judged by both the therapist and the client, were differentiated by a factor, Mutual Personal Openness. This factor depicted the quality of the client-therapist relationship in a manner 5 similar to the Analytic-Experiential factor. This factor characterized good therapy sessions, and thus presumably also a good client-therapist relationship, as involving a person-to-person encounter with recognition and expression of one's own feelings in the immediate here-and—now situ- ation by both the client and therapist. Bad therapy hours, on the other hand, were characterized by both participants feeling neutral, detached and withdrawn from one another. In addition, Howard, Orlinsky, and Hill (1969a) found a dis- tinct content factor related to the relationship, Therapy and Therapist, which reflected dialogue consisting of eval— uation of the therapist or therapy and exploration of client feelings about the therapist. And Howard, Orlinsky and Hill (1968) reported a factor, Toying with Therapist, which reflected the client's phenomenological experience during therapy sessions of feeling playful, affectionate, superior, and flirtatious toward the therapist. The results of the studies by Orlinsky and his asso— ciates indicate that therapists translate their beliefs about psychotherapy into actual practice since relationships similar to those derived from self-professed attitudes, Analytic and Experiential, were reported by both therapists and clients to exist in therapy sessions. There is a sug— gestion, however, that an Experiential type of relationship is experienced as being more satisfactory since this type of relationship was characteristic of therapy sessions described as good whereas the analytic type of relationship 6 was typical of therapy hours described as unproductive or bad. Hobbs (1962) and Wallach and Strupp (1964) have iden- tified the essence of this basic difference or Analytic- Experiential dichotomy in the client-therapist relationship as consisting of the way in which therapists handle the client's feelings about the therapist or his transference manifestations, including both those based on reality and classical psychoanalytic transferred reactions. There are two basic, quite different ways, corresponding to the Analytic and Experiential types of relationships, in which therapists respond to client feelings about the therapist. Therapists who establish an Analytic type of rational, objective, and personally aloof relationship consider the successful resolution of the transference neurosis to be the ultimate goal of therapy. To achieve this goal transference manifestations are analyzed and interpreted to the client in order to facilitate his gaining insight into the origin of his transference feelings (Wolman, 1967). In contrast, therapists who establish an Experiential type of intimate, personally involved relationship consider the relationship to be curative in and of itself and to provide an oppor- tunity for the client to directly and immediately experience the impact his transference strategems have on another and to experience that these strategems are neither necessary nor appropriate; consequently, the therapist reacts openly and spontaneously to the client's transference feelings. Thus, an essential difference among therapists has been 7 delineated as the kind of response therapists make to client's feelings regarding the therapist, that is, whether the therapist analyzes and interprets these feelings to the client or whether he responds or reacts to these feelings in the context of an open, intimate relationship. The therapist's response to the client's therapist- relevant feelings has also been noted by various other authors as constituting a central component of the therapy process within different theoretical orientations (Hobbs, 1962; Shapiro, 1961; Shoben, 1949; Strupp, 1958; Yulis & Kiesler, 1968). For example, in formulating the common sources of gain in diverse kinds of psychotherapy, Hobbs (1962) prOposed that one such gain accrues from the thera— pist's being alert to and reinforcing any reaching out towards himself by the client so that the client will be increasingly able to express his feelings in the presence of another and "even to go so far as to dare to include the therapist as an object of these feelings" (Hobbs, 1962, p. 743). Shoben (1949) has noted that most therapists, regardless of theoretical orientation, consider those feel- ings and attitudes that clients have about the therapist to be intimately related to the success or failure of therapy and, in addition, noted the agreement among therapists that these feelings, in particular, must not be ignored or rejected by therapists. Moreover, the desirability of the therapist's encouragement and reinforcement of client expression of feelings about the therapist has also been 8 stressed from a learning theory framework on the basis that such client expressions provide an opportunity for the reduction or extinction of anxiety associated with inter- personal relationships (Hobbs, 1962; Murray, 1956, 1962; Shoben, 1949). In addition, numerous authors have prOposed that the therapist's discussion and sharing with the client of his own feelings regarding the client can be a most effective element in psychotherapy (Fromm-Reichmann, 1948, 1949, 1950; Sullivan, 1949; Berman, 1949; Winecott, 1949; Little, 1951; Heiman, 1950; Tauber, 1954; Spitz, 1956; Whitaker, Felder, Malone, & Warkentin, 1962). In summary, then, the type of relationship thera- pists consider to be most appropriate and, more specif- ically, the manner in which therapists respond to the client's thoughts and feelings regarding the therapist, constitute a major dimension, if not the_major dimension, upon which therapists of different orientations and training can be compared and differentiated. Thus, the evidence cited has brought into direct focus the critical role occupied by that dimension of the therapy process consisting of the client's feelings regarding the therapist and the therapist's response to these feelings, regardless of the particular manner or theoretical framework within which the therapist responds to these client feelings, i.e., regard- less of whether the therapist analyzes and interprets or spontaneously reacts to these feelings. The therapist who conducts a rational, objective, and personally aloof 9 Analytic type of therapy considers the analysis and reso- lution of the transference to be the primary objective of therapy and therefore analyzes and interprets the client's feelings about the therapist: the therapist who conducts a mutually open, intimate, Experiential type of therapy considers the effective element of therapy to be the client- therapist relationship itself and therefore reacts sponta- neously to the client's feelings about the therapist. Thus, although the two types of therapists reSpond in different ways to the client's therapist-relevant expressions, both place equally strong emphasis on the integral, critical role that the client's feelings about the therapist and the therapist's response to these feelings play in the outcome of therapy. Consequently, both types of therapists would be expected to strongly encourage and reinforce client expression of feelings related to the therapist and to devote a major effort in therapy to exploration and dis- cussion of these feelings and to focus extensively on the therapeutic relationship. In View of this emphasis upon the critical impor— tance to client improvement of the client's feelings about the therapist and the therapist's response to these client feelings, it is noteworthy that relatively so few studies have appeared in the psychotherapy literature that directly and specifically deal with client verbalizations of therapist-relevant feelings and therapist responses to these client verbalizations. This paucity of studies has 10 also been noted by Shapiro (1961), who specifically called attention to the neglect of both client and therapist ver— balizations regarding one another and the therapeutic rela- tionship in investigations of psychotherapy. He also considers this aspect of therapy to be a major dimension characterizing the client-therapist relationship, as well as all social interaction, which can be readily measured and should be investigated. His admonition for investigation of this dimension of the therapeutic relationship has however, for the most part, gone unheeded. A review of the literature revealed several trends in the relatively few studies in which the client's expression of feelings related to the therapist and/or the therapist's response to such client expressions was specif— ically investigated. Four early studies focused on the relationship between client expression of feelings regarding the therapist and the success or failure of therapy, but no attempt was made to investigate the therapist's response to these client eXpressions (Braaten, 1961; Gendlin, Jenney & Shlien, 1960; Lipkin, 1954; Seeman, 1954). Two other studies investigated the relative frequency of relationship- relevant statements and statements unrelated to the rela- tionship by both clients and therapists over the course of therapy (Karl & Abeles, 1969; Murray, 1956). And one experimental study investigated the relative frequency with which high and low anxious therapists explicitly interpreted ll client statements in relation to the therapist (Yulis & Kiesler, 1968). Conflicting findings were obtained within the group of related studies of Client-Centered therapy which investi- gated the relationship between the extent to which clients increasingly focused on the relationship over the course of therapy and the success or failure of therapy. No attempt was made to investigate therapist responses. Two studies found no relationship between therapist's ratings of client focus on the relationship and therapist's judgment of client outcome (Gendlin 25 31., 1960; Seeman, 1954) while another study reported a positive relationship with the therapist's judgment of outcome but no relationship with either the client's own perception of his improvement or a diagnos- tician's assessment of outcome using the TAT (Braaten, 1961). In a departure from the above studies, Lipkin (1954) measured relationship focus as the amount of client dis- cussion regarding the therapist by objective ratings of actual therapy sessions and found a negative relationship with outcome ratings. In an attempt to integrate these contradictory findings, Gendlin gt 31., (1960) designed a study in which the therapist as a frequent topic of client discussion was differentiated from the client's use of the relationship for significant experiencing and found only the latter to be related to client success. They concluded that client success was related to the manner in which the client focused upon the relationship, i.e., whether he simply 12 talked about the relationship or whether he used the rela- tionship for achieving new experiences and insights. Moreover, the therapeutic relationship was often found to be a momentary instance of a problem and, in the case of successful clients, to provide a new experience constituting the first step in overcoming the problem. One very recent study appeared in which the fre- quency distributions of all relationship-relevant verbali- zations of both clients and therapists were investigated throughout individual sessions (Karl and Abeles, 1969). Client verbalizations pertinent to the therapeutic rela- tionship were found to be evenly distributed throughout single sessions, but therapist statements related to the relationship occurred least frequently during the first 10-minute segment of the hour and then increased and stabi- lized for the remainder of the session. Information regarding therapist reactions to client feelings about the therapist is provided by several studies, most of which suggest that therapists may not recognize or may be reluctant to respond to such feelings. For example, when a group of therapists, heterogeneous with respect to experience, orientation and sex, and their female clients reported the topics discussed during therapy sessions that were judged to be productive by the therapists, clients and therapists were in agreement on all topics except one: clients reported that they talked about the therapist or the relationship, while "the therapists, 13 perhaps out of modesty, seemed not to have noticed" (Orlinsky & Howard, 1967b, p. 624). And in a study of the feelings experienced by therapists during therapy sessions, therapists reported having no feelings whatsoever in response to the client's relating to them in a playful, affectionate, superior, and flirtatious manner, although the client perceived the therapist as experiencing an uneasy intimacy and as feeling ineffective (Howard, Orlinsky,& Hill, 1969b). Several studies suggest that therapists may avoid client expressions of specific kinds of feelings, particu- larly hostile feelings. For example, Bandura, Lipsher, and Miller (1960) found that therapists were much more likely to avoid client eXpressions of hostility which were directed toward the therapist than those directed toward an object other than the therapist. This finding was later replicated for experienced therapists but not for interns who were involved in group supervision which emphasized the client- therapist relationship and the therapist's ability to respond to the client's feelings about the therapist (Varble, 1968). Two other studies have also suggested that therapist avoidance responses to client expressions of dependency, particularly those directed toward the thera- pist, may result in the client's premature termination (Alexander & Abeles, 1968; Winder, Ahmad, Bandura, & Rau, 1962), although other studies failed to confirm this finding (Caracena, 1965; Schuldt, 1966). 14 A recent experimental studylnrYulis and Kiesler (1968) investigated the extent to which therapists explic- itly verbalized to the client that the therapist himself was the object of some of the feelings the client was expressing as a function of the therapist's anxiety level. In this well designed study high and low anxious therapists listened to three client tape recordings characterized, respectively, by sexual, aggressive, and neutral content which bore fairly obvious relevance to the therapist. At ten choice points in each tape the therapists were asked to select one of a pair of interpretive responses which differed only in whether or not the response explicitly verbalized the implication the client's statement had in relation to the therapist. While the low anxious therapists chose more responses interpreting client statements in relation to the therapist than did the high anxious therapists, regardless of the specific content of the client's statement, only slightly more than half, 52%, of the responses chosen by the low anxious therapists and 38% chosen by the high anxious therapists verbalized the relevance the client statements bore to the therapist. And finally, Murray (1956) found that although a group of therapists heterogeneous with respect to experience and orientation made more active responses, i.e., responses that encouraged client continuation of the current dis- cussion, to client statements related to the therapist than to those unrelated to the therapist, they made active responses to only 15% of client therapist-relevant 15 statements. However, client statements related to the ther- apist were found to increase significantly over the entire course of therapy from 5.3% to 10.6% of the total number of client statements, and the rate of increase was positively related to the extent to which the therapist responded in an active rather than a passive way to client statements of all kinds, regardless of whether or not the client statements were related to the therapist. In discussing these findings, Murray (1956) states that the client initially brings into therapy a large part of his feelings about the therapist, but direct expression of these feelings is inhibited by anxiety. Consequently, at the beginning of therapy these feelings are expressed as displacements, i.e., the client expresses his feelings about the therapist in a displaced, covert, or indirect way when he is talking about something not manifestly related to the therapist. The therapist's active responses at these points of indirect references to the therapist may serve as a reinforcement and decrease the client's anxiety about direct or overt expression of these feelings. An example presented by Murray provides an illus- tration of a client's indirect expression of feelings related to the therapist when the manifest content is unrelated to the therapist. Thus, for example in Case D, the patient spent the first few hours in abstract intellectual discussion. In the content analysis this was scored intellectual defense and properly so. But there was an undercurrent in this intellectual discussion. The patient seemed to be telling the therapist, "Look, I'm clever and I've read a lot of books. I'm worthy of your.interest and 16 respect." Later, the patient spent a good deal of time with physical complaints. Here he seemed to be saying, "I'm frightened and helpless. I need your help." (Murray, 1956, p. 22). The fact that the client may be expressing feelings about the therapist even though his statements are not mani- festly related to the therapist has been recently noted and discussed by several other authors (Beier, 1966; Kell & Mueller, 1966; Searles, 1965; Wiener & Mehrabian, 1968). Searles (1965) has presented the most encompassing formu- lation of this phenomenon in his statement that everything a client says contains references, whether or not the client is aware of these references, to the immediate psychotherapy situation. Furthermore, he believes thattjns phenomenon occurs to a very great extent in the everyday conversation of ordinary individuals. He states the case for this phenomenon in the following way. Whenever A is expressing to B a felt attitude about an absent third person, C, then A is simultaneously revealing that he holds the same atti- tude, in some degree, toward B. A is usually unaware of his communication to B of his own (A's) attitude toward B, and B is unaware of this communication regarding A's attitude about himself unless he has a practiced ear for this kind of communication. Kell and Mueller (1966) have described this phenomenon in their statement that the manifest content of the client's statements mirrors the therapeutic relation- ship. They report that the choice of content of therapy sessions reflects the nature of the therapeutic relationship 17 and that, particularly at times of stress, the client talks about past experiences that symbolically communicates to the therapist some of his feelings about the therapist. Kell and Mueller (1966), Murray (1956), and Searles (1965) have suggested that the therapist's ability to recog- nize and respond to client feelings about the therapist, including both manifest and covert feelings, is a critical component of therapist effectiveness and the success or failure of therapy. Moreover, Murray (1956) has specif- ically stated the need for investigation of the relationship between client improvement and the amount and quality of therapist responses to manifest and indirect therapist- relevant client statements. However, to this writer's knowledge, no such investigation has been conducted to date. Accordingly, this study is an investigation of the relationship between client outcome and the extent to which therapists focus on the client—therapist relationship in therapy by verbalizing the client's feelings, both manifest and covert, which are related to the therapist. More spe- cifically, it is an investigation of the frequency with which therapists respond to implicit and explicit client references to the therapist by either acknowledging and verbalizing or by avoiding and ignoring the client refer- ences and the implication or relevance they bear to the therapist or the client-therapist relationship. It is hypothesized that there will be a significant positive relationship between client improvement and the extent to 18 which the therapist explicitly relates client statements to himself and thus focuses directly on the client's feelings regarding the therapist and on the client-therapist relationship. The extent to which therapists verbalize or inter— pret client statements in relation to themselves or therapy will be measured by a scale, the Immediate Relationship Scale (IRS), specifically constructed for this purpose by this writer (see Appendix A). The underlying assumption of the scale is that all client statements, regardless of the manifest content of those statements, contain some reference to the therapist or the therapeutic situation: this refer- ence may be direct and explicit, indirect and implicit, or opaque so that it is not readily or clearly discernible. The scale quantifies the therapist's response to these client references. It consists of six stages which are conceived as lying along a continuum reflecting the extent to which the therapist attempts to focus directly on the immediate client-therapist relationship by verbalizing or interpreting the implications the client statements have in relation to the therapist or the therapeutic situation. At the lowest stage of the scale, stage 1, the client makes a direct and explicit reference to the therapist, thereby providing the therapist with maximal Opportunity to acknowl- edge or interpret the client statement to himself and to focus on the therapeutic relationship, but the therapist responds by completely ignoring the direct reference to 19 himself. In contrast, at the highest stage, stage 6, the therapist responds by clearly and directly focusing on the immediate relationship by making explicit the client refer- ence to the therapeutic relationship, regardless of the manifest content of the client statements or the degree of directness of the reference, i.e., the Opportunity provided by the client. Thus, at the lowest stage the therapist actually avoids personal involvement and focus on the relationship even though a clear invitation to do so is presented by the client, while at the highest stage the therapist directly and explicitly relates or connects the client reference to himself and focuses on the relationship, regardless of the opportunity afforded by the client's statements. For all six stages of the scale, each succeed- ing higher stage indicates that the therapist has advanced a step away from out-right refusal to discuss the client- therapist relationship and advanced a step closer toward a direct focus on the relationship. The essential elements of the IRS scale are summa— rized in Table l. The scale is used to rate two distinct kinds or classes of responses which the therapist makes in response to these client references: responses which ignore the reference and make no attempt to relate the reference to himself and to focus on the relationship, and responses which attempt to relate the reference to himself and to focus on the relationship. Therapist avoidance responses in which an attempt to focus on the relationship is absent 20 mdqmmo Ammmapumwmmv no .uomufip somouma< waouum Bdefixmz IIII somouad< uomufla IcH .uomufin o mavmao Ammwapumwmmv no .uomufip somouaa< mumumpoz wumumpoz IIII :omoudd< booufich IGH .uomuHQ m Ammmapumwmmv oncmdo no nomouam< nae: ummmq IIII Lemonad< popSMIcmao uooufich q Aucmmb< paw: IIII ummmq mocmpfio>< muoawH movmdo m Anammnav moamwfio>< mumumpoz IIII mumumpoz mocwpfio>< mnogmH nomufipaH N Augmmp< wcouum IIII anafixmz moampfio>< muocwH uomufin H mmwaapmom mmaoamom ucmfiao moampfio>< moamummmm moamummmm umfinmumnH m.uwammumse mp pmpfi>oum no somouaa< ou uncommom udmwau «0 mo mmmauomuwn kufiaauuoamo m.umwamum:9 umaamnmnH mmmauomuwn mmmum manmaowumflmm mmoomfia ou omMum 50mm um poumm umaamumny mo mmmcfipmmm Hmofiumnuoaxm mo cowumaaaumumm momaodmom umfiamumnH pan uGMfiHo mo GOHuaauommn .mmH Sb pounmmmz noncommmm uamHHo van umfidemcH .H manna 21 are rated at stages 1, 2, or 3; therapist responses in which such attempts are present are rated at stages 4, 5, or 6. Whether the therapist's avoidance response is rated at stage 1, 2, or 3 is determined by the degree of directness or overtness of the reference contained in the client state- ments. Whether the therapist's approach response is rated at stage 4, 5, or 6 is determined by the degree of direct- ness or explicitness of the therapist's response, regardless of the directness Of the client's reference. Thus, ratings at stages 1, 2, and 3, respectively, indicate the absence Of a therapist attempt to focus on the relationship in response to a client reference which is progressively less overt or direct: at stage 1 the client reference is overt and explicit; at stage 2 the reference is indirect and implicit but clearly conveys a message regard- ing the therapeutic situation; and at stage 3, although an implicit reference is assumed, that reference is unclear and opaque and cannot be readily determined or deciphered. In contrast, ratings at stages 4, 5, and 6, respectively, indi- cate a therapist attempt to focus on the relationship which is progressively more direct and explicit, regardless of the degree of overtness or directness of the client's reference: at stage 4 the therapist's response does not relate the client's reference to any specific person or situation, including the therapeutic situation, but rather is struc- tured in an Open-ended fashion which provides an Opportunity or tends to increase the probability for the client to 22 generalize, apply, or translate the reference to the thera- pist or their relationship; at stage 5 the therapist relates the client's reference to the relationship in an indirect, tentative, hesitant, or cautious manner; and at stage 6 the therapist directly and eXplicitly relates the client's reference to the therapeutic relationship and clearly attempts to focus discussion on the relationship. The scale may be conceived as providing an index or measure of the strength of the therapist's attempts or lack of attempts to relate client references to himself and to focus on the immediate therapeutic relationship or, in other words, the strength Of the therapist's approach or avoidance of focusing on the relationship. Stages 1, 2, and 3,respec- tively, reflect avoidance response of decreasing magnitude. At each Of these three stages a therapist avoidance response is indicated by the failure of the therapist to attempt to relate client references to the therapeutic situation, and the strength of that avoidance reSponse is indicated by the overtness of the client reference, i.e., by the Opportunity which the client statements seemingly provide the therapist to focus on the therapeutic situation. A strong avoidance response is reflected at stage 1 where the client reference is explicit and overt and thus provides maximal Opportunity for responding to it, a moderate avoidance response is reflected at stage 2 where the client reference is indirect and implicit yet clearly discernible, and a mild avoidance response is reflected at stage 3 where the reference is 23 assumed but is not clearly discernible. Stages 4, 5, and 6, respectively, reflect therapist approach responses Of increasing magnitude. An approach response is indicated at each of these three stages by the presence of therapist attempts to relate client references to the therapeutic situation and to focus on the relationship, and the strength of that approach is indicated by the directness or explicit- ness with which the therapist attempts to relate the refer- ence to himself and to engage the client in discussion Of the immediate relationship. A mild approach response is reflected at stage 4 where the therapist's response is Open— ended and does not attempt to specifically relate the client reference to the therapeutic situation but, rather, prepares the way for the client to make the application tO the thera— peutic situation; a moderate approach response is reflected at stage 5 where the therapist makes a very tentative or indirect attempt to relate the reference to himself and to focus on the immediate relationship; and a strong approach response is reflected at stage 6 where the therapist makes a direct and explicit attempt to focus on the immediate relationship. Thus, the IRS scale provides a measure Of whether the therapist avoids or approaches the client's references to the therapist or to the therapeutic situation. When the therapist's response is one of avoidance, the strength or degree of the therapist's reluctance or unwillingness to deal directly with the reference to the therapeutic relationship 24 is measured by the directness or overtness of the client's reference. When the therapist's response is one Of approach, the strength of the therapist's attempts to focus on the immediate relationship is measured by the directness and explicitness of the therapist's response. Recent reviews and discussions of psychotherapy research have advocated the use Of multidimensional models in which the independent variable is investigated or described in relation to relevant therapist, client, and time variables (Kiesler, 1966; Paul, 1967; Strupp & Bergin, 1969). In accordance with this suggested approach, the present research investigates the IRS variable not only in relation to client outcome, but also in relation to certain relevant therapist and client variables as well as at dif- ferent periods of time within the psychotherapy process. The IRS variable is investigated in relation to three thera— pist variables--the core conditions, orientation, and experience level--and two client variables--depth of self- exploration and inpatient or outpatient status. These variables, with the exception of the relatively global client status variable, have been subjected to a great deal of empirical research and are among those prOposed by Strupp and Bergin (1969) as showing the greatest promise for unrav- eling the psychotherapy process. 25 Therapeutic Core Conditions Certain characteristics of therapist behavior, the core conditions of empathy, genuineness, and nonpossessive warmth, cut across virtually all theories of psychotherapy and are common elements in a wide variety of approaches to psychotherapy (Truax & Carkhuff, 1967; Bordin, 1955; Rogers, 1957; Shoben, 1953). Operational definitions of each of the three components of the therapeutic core conditions are provided by the research scales constructed by Truax which concretely specify along quantified dimensions these three central ingredients (Truax & Carkhuff, 1967). Therapist Genuineness is described as the thera- pist's openness to experiencing himself and his feelings during the therapy encounter. A high level of Genuineness does not mean that the therapist must overtly express his feelings but only that he does not deny them. A therapist who is highly genuine is integrated, authenic, nondefensive and presents himself to the client as a real person, without phoniness and without hiding behind a professional facade or role. The Truax (1962c) scale Of therapist genuineness descriptively specifies five stages along a continuum beginning at a very low level where the therapist presents a facade and defends or denies feelings and continuing to a high level of self-congruence where the therapist is freely and deeply himself. At the lowest level the scale includes such descriptions as "there is explicit evidence of a very considerable discrepancy between his experiencing 26 and his current verbalizations,"'the therapist makes strik- ing contradictions in his statements," and "the therapist may contradict the content with voice qualities." At inter- mediate stages "the therapist responds in a professional rather than a personal manner," and "there is a somewhat contrived or rehearsed quality." At the highest stages "there is neither implicit nor explicit evidence Of defen- siveness or the presence Of a facade," and "there is an openness to experiences and feeling by the therapist of all types, both pleasant and hurtful, without traces of defen- siveness or retreat into professionalism." Therapist Warmth is described as the therapist's interest in and respect for the client and is communicated by his nondominating, nonjudgmental attitude toward the client. The therapist who communicates a high degree of warmth conveys a deep interest, concern, and reSpect for the client, does not approach the client in a moralistic manner, but rather, accepts the client for what he is and not for what he should be. The Truax (1962a) scale of therapist Nonpossessive Warmth consists Of five stages ranging from a high level where the therapist warmly accepts all Of the client's experience as part of the client without imposing conditions, to a low level where the therapist evaluates the client or his feelings, expresses dislike or disapproval, or expresses warmth in an evaluative way. The lowest level of the scale includes such descriptions as "the therapist acts in such a way as to make himself the focus Of 27 evaluation, he may be telling the client what would be 'best' for him or may be in other ways actively trying to control his behavior," and "the therapist responds mechan- ically to the client and thus indicates little positive warmth," and "the therapist ignores the client where an unconditionally warm response would be expected--complete passivity that communicates a lack of warmth." At higher levels the therapist "clearly communicates a very deep interest and concern for the welfare of the client. Attempts to dominate or control the client are for the most part absent, except that it is important that the client be more mature or that the therapist himself is accepted and liked." At the highest level "the client is free to be himself even if this means that he is temporarily regressing, being defensive, or even disliking or rejecting the therapist himself." Therapist Empathy is defined as the therapist's ability to perceive the client's feelings, including those which are partially hidden as well as those which are Obvious, and to communicate his perception in words attuned to the client's feelings in order to clarify and expand what the client has hinted at by voice, posture, or cues. The Truax Accurate Empathy scale (1961) is a 9-stage scale which measures on a continuum the therapist's perception and com— munication of the sum total of the client's feelings. The lower stages of the scale include such descriptions as the therapist "seems completely unaware of even the most 28 conspicuous of the client's feelings. His responses are not appropriate to the mood and content of the client's statement and there is no determinable quality of empathy." At intermediate stages "he often responds accurately to more exposed feelings. He also displays concern for more hidden feelings which he seems to sense must be present, though he does not understand their nature." At the highest stages the therapist accurately interprets all of the client's present, acknowledged feelings. He "moves into feelings and experiences that are only hinted at and does so with sensi— tivity and accuracy. He Offers additions to the client's understanding so that not only are underlying emotions pointed to, but they are specifically talked about." In summary, then, the essence of non-possessive warmth is to preserve the client's self respect as a person and a human being and to provide a trusting, safe atmosphere; the purpose Of genuineness is to pro- vide an honest nondefensive relationship which allows us to point to unpleasant truths about the relationship and about the client rather than to hide behind a facade; accurate empathic understanding serves as the work of the therapeutic relationship (Truax & Mitchell, 1970, pp. 34-35). Research evidence to date suggests that for a therapist to have effective therapeutic impact, he must not be decidedly low on any of the three component core conditions and must be moderately high on at least any two Of them (Truax & Mitchell, 1970). Carkhuff and his associates have condensed and revised the Truax Accurate Empathy, Nonpossessive Warmth, and Genuineness scales (Truax, 1961, 1962a, 1962c) in an 29 attempt to reduce the ambiguity and increase the reliability of those scales. The revised Empathy scale (Berenson, Carkhuff, & Southworth, 1964) collapses the 9-stage Truax scale into five stages, and the revised Warmth scale (Carkhuff, Southworth, & Berenson, 1964) is essentially a simplified restatement of the Truax scale, Similarly, the revised Genuineness scale (Carkhuff, 1964a) is a rewording of the Truax scale although the scoring of the therapist's negative reactions to the client are made more explicit at the lower stages Of the scale. Both the Truax scales for the measurement of the core conditions and the revised scales by Carkhuff and his associates have been used in a great number of studies and have been demonstrated to be reliable (Carkhuff & Berenson, 1967; Truax & Mitchell, 1970). A great deal of converging research has produced extensive evidence indicating the critical importance of these three central ingredients to the psychotherapy process for changing client behavior (Bergin, 1967; Carkhuff, 1969b, 1969c; Carkhuff & Berenson, 1967; Rogers, Gendlin, Kiesler, & Truax, 1967; Strupp & Bergin, 1969; Truax & Carkhuff, 1967; Truax & Mitchell, 1968, 1970). Moreover, these ingredients have been found to be Of major importance in a wide variety of situations in addition to psychotherapy which involve human relationships, including teacher-student relationships (Aspy, 1965; Aspy & Hadlock, 1966; Dickinson & Truax, 1966; Truax & Tatum, 1966; Wagner & Mitchell, 1969» re ct CC Mu 19‘ Gez tez all enc be Tru p511" (Ca *1: J. 30 rehabilitation counseling (Truax & Mitchell, 1970), parent- child relationships (Shapiro, Krauss, & Truax, 1969), verbal conditioning (Truax, 1966; Vitalo, 1970), educational- personal counseling by untrained housewives (Stoffer, 1968) and by college dormitory counselors (Wyrick & Mitchell, 1970), peer relationships among college students (Shapiro g; 31., 1969; Shapiro & Voog, 1969), as well as therapists' personal and social interpersonal relationships involving spouse, colleagues, and friends (Collingwood, Hefele, Muehlberg, & Drasgow, 1970; Hefele, Collingwood, & Drasgow, 1970). Indeed, the core conditions of Empathy, Warmth, and Genuineness have been conceptualized as personality charac— teristics or attributes not only of psychotherapists but of all individuals, and as playing a critical role in all human encounters which are intended to change human behavior or to be helpful (Carkhuff, 1967b; Carkhuff & Berenson, 1967; Truax & Carkhuff, 1967). The critical role of the core conditions in the psychotherapy process has been empirically demonstrated by repeated findings indicating that therapeutic progress varies as a function of therapists' characteristics Of Empathy, Warmth, and Genuineness. This extensive research has been widely reviewed (Bergin, 1967; Strupp & Bergin, 1969; Truax, 1967), and summaries are published periodically which include current research on the core conditions (Carkhuff & Berenson, 1967; Truax & Carkhuff, 1967; Truax & Mitchell, 1968, 1970). The research converging on the 31 therapeutic core conditions in psychotherapy lead to the following major conclusions. The level of therapist func- tioning on the core conditions is related to client change: most improved clients receive higher levels of the core conditions than least improved clients, and clients of therapists with relatively high levels on the core condi— tions show more improvement than either clients of thera- pists with relatively low levels on the core conditions or clients in no-therapy control groups. Moreover, not only are high levels of therapist-Offered core conditions related to and predictive of client improvement, but low levels are related to no improvement or even deterioration, i.e., clients Of low functioning therapists are no better Off or actually worse off after therapy than before therapy. Con- sequently, therapy can be "for better or worse." These conclusions indicating that therapists who are empathic, warm, and genuine are indeed effective, are based on the research of numerous researchers and appear to hold for a variety of therapists and counselors regardless of their training and orientation and with a variety of clients, including college underachievers, college counselees, juvenile delinquents, hospitalized schiZOphrenics, mild to severe outpatient clients, and hospitalized clients with mixed diagnoses. Moreover, the findings appear to hold up across diverse therapeutic settings including hospitals, outpatient clinics, college counseling centers, rehabili- tation centers, and private practice, and in both individual 32 and group psychotherapy. Truax and Mitchell (1970) have recently compiled and presented in tabular form the results of studies which have related the Truax E, W, and G scales to client outcome measures. The findings that client improvement or deterio- ration is related to the high or low level of therapist functioning on the core conditions strongly indicates that all therapists cannot be assumed to constitute a homogeneous group and to be relatively equally helpful and effective. Kiesler (1966) has specifically cautioned against the assumption, which he designates the "therapist uniformity myth," that therapists constitute a homogeneous group. Since differences among therapists on the core conditions have been demonstrated to be associated with important and significant differences in client outcome, it would appear that failure to take into account the therapists' level of functioning on the core conditions implies subscription to the Therapist Uniformity Myth. Such studies run the risk Of obtaining misleading results in which diverse trends in the data mask or cancel each other. Moreover, studies using different or unspecified compositions of therapists with respect to the core conditions may produce findings which cannot be compared across studies. Indeed, Truax and Mitchell (1970) have even suggested that previous research on therapist characteristics which failed to take account of the core condition variable needs to be re-done. 33 Consequently, this study investigates the relation— ship between the IRS and core condition variables and attempts to demonstrate that the IRS variable is related to the core conditions but, at the same time, is distinct and can be differentiated from the core condition variable. That is, it is predicted that although the two variables are related, they are not identical and do not simply measure the same thing. In addition to the attempt to control for the therapist uniformity myth there is a second, equally compelling, reason for the inclusion of the core condition variable in this study. To date, no other therapist var- iables have been subjected to a comparable programmatic research effort and have been as consistently related to client therapeutic progress and outcome. Two specific hypotheses are tested regarding the relationship between the IRS and core condition variables. The first hypothesis states that there is a significant positive correlation between the extent to which therapists focus directly upon the immediate client—therapist relation- ship and their level of functioning on the core conditions. The second hypothesis which deals with the assessment of both the utility of the IRS variable in predicting client outcome and the distinctness of the IRS variable from the core condition variable states that the IRS variable signif- icantly predicts measures of client outcome independently Of the core condition variable, i.e., with the effect of the core conditions on outcome measures controlled statistically 34 Therapist Orientation and Experience Level The attitudes of therapists regarding psychotherapy processes and techniques have been investigated in a number of studies as a joint function of therapist experience level and orientation or training. In a frequently cited series Of studies Fiedler (1950a, 1950b, 1951) concluded that the nature of the therapeutic relationship is a function Of the therapist's experience and not a function Of orientation or school. In contrast, studies by McNair and Lorr (1964), Sundland and Barker (1962), and Wallach and Strupp (1964) indicated that therapist attitudes differed as a function of orientation and not as a function of experience level. Attempts have been made to explain the contradictory find— ings regarding the relative influence of orientation and experience level as determinants of therapist attitudes in terms of differences in the kind of therapist attitudes sampled (Gardner, 1964; Sundland & Barker, 1962). Sundland and Barker (1962) suggested that when therapists are com- pared on attitudes with which most therapists tend to agree, such as those regarding empathy and understanding as Fiedler did, then therapists differ as a function of experience, but when therapists are compared on controversial attitudes and preferences for diverse therapy techniques, as was done in the three later studies, then therapists differ as a function of orientation. 35 There is more recent evidence, however, to suggest that contrary to Fiedler's results, orientation has a greater effect than experience in determining therapist attitudes and preferences. Spilken and Jacobs (1968), using a much larger sample of therapists than Fiedler used, exam— ined self-professed attitudes of therapists regarding ten therapist variables, including empathy, warmth, sincerity or genuineness, and respect, which were similar to the atti- tudes examined by Fiedler. Therapists were found to differ as a function of orientation and training, but differences occurred as a function of experience level on only one variable, directiveness. Quite similar results regarding therapist attitudes and preferences for therapy techniques as a function of orientation were obtained in the three studies by McNair and Lorr (1964), Sundland and Barker (1962), and Wallach and Strupp (1964). When therapists were classified into three orientations--Freudian, Sullivanian, and Rogerian--the three groups differed on nine Of the 16 scales used to measure attitudes, with the Sullivanians consistently being in the middle position (Sundland & Barker, 1962). In comparison with the Rogerians, the Freudians differed on all scales specifically measuring attitudes regarding the nature of the therapeutic relationship and preferred a relationship in which the therapist analyzes and conceptualizes the nature of the client's relationship with the therapist rather than responding spontaneously to the client's behavior as it 36 occurs in therapy. When therapists were classified into four orientations--Orthodox Freudian, Psychoanalytic-GeneraL Sullivanian, and Rogerian--and compared on their preference for a client-therapist relationship characterized by the therapist's maintenance of personal distance, the Orthodox Freudians were highest in their preference for this kind Of distant, personally aloof relationship, the Psychoanalytic- General group was next highest, and the remaining two, similar to one another, least preferred this kind of rela— tionship (Wallach & Strupp, 1964). These findings were also consistent with those obtained by previous studies (Fey, 1958; Fiedler, 1950b, Strupp, 1955) on those aspects in which the studies were comparable. And in contrast to Fiedler's (1950b) Often quoted conclusion that therapists of different orientations do not differ in how they actually behave in therapy, a study of the actual therapy behavior of highly experienced Client-Centered and Analytic therapists by Cartwright (1966) found extreme differences between ther— apists of the two schools in their proportional use Of various therapeutic responses such as encouragement, reflection, clarification, and interpretation. Thus, the evidence cited suggests that therapist attitudes and views regarding the client-therapist rela- tionship differ as a function of orientation and not as a function of experience level, and this difference consists of the degree to which therapists of diverse orientations prefer an impersonal, aloof relationship with the client. 37 As noted earlier, since the importance of the client's feel- ings regarding the therapist is universally recognized and accepted by therapists, including those who prefer an imper— sonal, aloof relationship as well as by those who prefer a personal, intimate relationship, therapists would not be eXpected to differ on the extent to which they respond to therapist-relevant client statements as a function of their orientation or the degree of personal distance they prefer to maintain in the relationship. However, it has been suggested that the therapist may prefer an impersonal, distant relationship because Of his own neurotic conflicts about intimate contact with others. Bugental (1964), for example, has proposed that one of the major neurotic conflicts of therapists consists of a great need for and a concomitant fear of intimacy with others. The therapist with this neurotic conflict strives to establish a one-way intimacy with the client in which he holds himself and his feelings aloof from the client and thus, to some extent, gratifies his need for intimate con- tact, and at the same time, avoids the anxiety that would result from his permitting the client to become close to or intimate with him. Since the therapist cannot tolerate the anxiety associated with the closeness of another, efforts made by the client to reach out toward the therapist are met with anxiety and some form of rejection, avoidance, or resistance from the therapist. One manifestation of this resistance may be indicated by Hobb's (1962) suggestion that 38 interpreting, as opposed to reacting to, the client's trans- ference behavior may be tantamount to negative reinforcement Of the client's attempts to reach out toward the therapist. Marcondes (1960) has emphatically proposed that this type Of neurotic one-way gratification which prevents the therapist from responding positively and adequately to the client's feelings about the therapist is exemplified by the imper- sonal, objective type of relationship typically established by psychoanalytic therapists. Therefore, in view of the evidence suggesting that the therapist's orientation, but not his experience level, is related to the degree to which the therapist prefers a therapeutic relationship characterized by his maintenance Of personal distance from the client, and the proposals by Bugental (1964), Hobbs (1962), and Marcondes (1960) sugges- tive of a relationship between a preference for maintaining personal distance and the therapist's inability or reluc- tance to accept and deal with the client's feelings about the therapist, the following predictions are made in this study regarding the relationshp between the extent to which the therapist responds to therapist-relevant client state— ments and his orientation and experience level. 1. Therapists are ordered on the basis Of their IRS scores according to orientation as follows--Re1ation- ship, Eclectic, Client-Centered, and Analytic. 39 2. Relationship therapists have significantly higher IRS scores than Eclectic, Client-Centered, or Analyt- ically oriented therapists. 3. No significant differences in IRS scores occur between high and low experience level therapists within any of the four orientation groups of therapists. Client Depth pf Self-Exploration A mass Of research has accumulated which indicates that the client's self-exploratory behavior and attempts to understand and define his own beliefs, values, motives, and actions play an important role in the outcome of psycho- therapy (Truax & Carkhuff, 1967). Much of this research has been based on the 9-stage Depth of Self—Exploration Scale (DX) developed by Truax (1962b) and subsequently condensed into five stages by Carkhuff (Carkhuff, 1964). These DX scales define along a continuum the extent to which clients engage in self-exploration, ranging from no demonstrable intrapersonal exploration to a very high level of self- probing and exploration. At the lowest level of the scale there is no discussion of personally relevant material by the client and no opportunity for it to be discussed. Per- sonally relevant material includes self-descriptions by the client intended to reveal his innermost feelings and thoughts to the therapist, and communications Of his per— sonal values, perceptions Of his relationships to others, his personal role and self—worth in life, as well as re d1 15 st al ti. (3e Ca 4O communications indicating emotional turmoil and expressions Of specific feelings such as anger, affection, etc. At the middle level the client introduces discussion of personally relevant material but does so in a mechanical manner, without spontaneity or emotional proximity and without an inward probing to newly discover feelings and experiences. At the highest level the client actively and spontaneously engages in an inward probing to discover or rediscover feel- ings or experiences about himself, his relations with others, and his world. A number of studies using the DX scale with a vari- ety of client populations and outcome measures have reported that successful clients show more self-exploration during psychotherapy than do unsuccessful clients (Truax, 1966; Truax & Wargo, 1966; Truax, Wargo, & Carkhuff, 1966) and that this outcome was predictable even in the initial stages of therapy (Truax & Carkhuff, 1963, 1965). There is also a great deal of evidence indicating that the level of the therapeutic core conditions is positively related to the degree of client self-exploration (Truax, 1966; Truax & Carkhuff, 1967; Carkhuff & Berenson, 1967; Carkhuff, 1969c). Cross cultural validating evidence from West Germany has recently become available for both the Truax DX and Accurate Empathy scales. In a replication of an experi- mental manipulation study by Truax and Carkhuff (1965), Sander, Tausch, Bastine and Nagel (1968) reported that when therapist level of empathy was lowered there was a 41 corresponding lowering of client self—exploration and when therapist level of empathy was raised there was a concom— itant rise in client self-exploration. In addition, Tausch, Eppel, Fittkau and Minsel (1969) related therapist level Of empathy to both client depth of self-exploration and client improvement and Obtained results similar to those reported by Truax and his associates (Truax & Carkhuff, 1967; Truax & Mitchell, 1968, 1970): the higher the therapist level of empathy, the higher the client level Of self—exploration and the greater the degree of client improvement. In contrast to the bulk of evidence which indicates that client depth of self-exploration is positively related to constructive client change and improvement, Truax and Wargo (1966) found that depth of self-exploration was unrelated to some and negatively related to other measures of client outcome in a large sample of male juvenile delin- quents in group psychotherapy. In explanation of this unexpected finding, the authors suggested that engaging in high levels of self-exploration may constitute "unmanly" behavior for juvenile delinquents. In addition to the evidence derived from the body of research based upon the DX scale, similar measures of similar constructs have received extensive support in the literature of counseling and psychotherapy (Blau, 1953; Braaten, 1961; Kirtner & Cartwright, 1958; Peres, 1947: Seeman, 1949; Tomlinson & Hart, 1962; van der Veen, 1967). 42 Thus, the depth of self-exploration variable has been empirically demonstrated to be a reliable measure of client in-therapy behavior and to differentiate among clients who improve and those who fail to improve in therapy. Moreover, the self-exploration variable appears to be relevant to the IRS variable in View of the fact that high levels of self-exploration include the client's dis- covery and experience of new aspects of himself and new feelings in his relationships with others. The therapist's responses to therapist-relevant client statements in which the therapist attempts to focus directly on the client's feelings about the therapist and their immediate relation- ship would have the anticipated effect Of enhancing the client's exploration of himself in relation to the thera— pist and, further, frequently in relation to other signif— icant persons. As Gendlin g; 31. (1960) have observed, the client's feelings about the therapist and their relationship is frequently representative of one of the client's central problems, particularly problems in his relationships with others, and the client's exploration of his feelings in relation to the therapist is Often a new experience consti- tuting the initial step in his overcoming of the problem. Consequently, in this study prediction of a significant positive relationship between the IRS and DX variables is made. 43 Inpatient and Outpatient Status In an attempt to reduce the heterogeneity within the sample, the IRS variable is investigated for inpatient and outpatient therapists separately. Since the IRS var- iable is conceptualized as being determined primarily by the therapist, it is predicted that the IRS variable does not differentiate between therapists Of inpatients and thera— pists of outpatients, and no alternative predictions are made regarding client status. Investigation of the IRS variable in relation to the foregoing specified variables is carried out by means Of two separate psychotherapy studies. The first is a process study in which the IRS variable is investigated in relation to the therapists' core conditions, orientation and experi- ence level as well as the clients' depth of self-exploration and status as inpatient or outpatient. The second is an outcome study in which the IRS variable is again investi- gated in relation to the therapeutic core conditions and in relation to measures of client improvement or outcome. The following specific hypotheses of each study are to be tested. Study Ie-Process Study Hypothesis 1. There is a significant positive relationship between the therapists' core condition and IRS scores . 44 Hypothesis II. The group of outpatient therapists and the group of inpatient therapists do not differ sig- nificantly on mean IRS scores. Hypothesis III. There is a significant positive relationship between the therapists' IRS scores and the clients' DX scores. Hypothesis IV, Three specific hypotheses are used to test the general hypothesis that IRS responses signifi- cantly differentiate therapists on the orientation variable but do not significantly differentiate on the experience level variable. A. Orientation to psychotherapy is ordered on the basis of IRS scores from highest to lowest in the following way: Relationship, Eclectic, Client-Centered, and Analytic. B. Relationship therapists have significantly higher IRS scores than Eclectic, Client-Centered, or Analytically oriented therapists. C. NO significant differences in IRS scores occur between the group of high experience level therapists and the group of low experience level therapists within any of the four orientation groups. Study Elf-Outcome Study Hypothesis I. There is a significant positive relationship between the therapists' core condition and IRS scores . 45 Hypothesis 1;. The higher the therapists' IRS scores the greater the improvement (outcome score indicating better adjustment) the clients show, and the lower the therapists' IRS scores the less improvement the clients show on each of five different outcome measures, regardless Of the therapists' core condition scores, i.e., with the effect of the core conditions on the outcome measures controlled statistically. METHOD Study I—-Process Study Tape recordings of the first psychotherapy session Of 56 different therapist-client dyads were used in this study. The tape recordings were borrowed on the basis of availability from various researchers at institutions across the country, including the Universities of Kentucky, Massachusetts, Maryland, Arkansas, and Wisconsin. The therapists represented a variety of settings, disciplines, theoretical orientations, and a wide sample of psychotherapy experience ranging from advanced level grad— uate students in Clinical and Counseling Psychology, to therapists with more than 15 years of post-doctoral psychotherapy experience. Clients ranged from minimally disturbed college students to hospitalized chronic schizophrenics. Of the 56 clients, 23 were seen in university counseling centers and 33 were seen in hospital settings: 21 were female, and 35 were male. Formal diagnoses were not available for the students, but on the basis of the tape recordings, most appeared to be mildly to moderately disturbed although several seemed more severely disturbed and would best be 46 47 described as ranging from character disorders to ambulatory schizophrenics. The hospitalized clients were either acute or chronic schiZOphrenics with the exception of approxi- mately five who were diagnosed at the time of hospitali- zation as sociopathic personalities with alcohol or drug addictions as primary symptoms. Table 2 provides demographic information describing the therapists by discipline, orientation, setting, experience, sex, and type of client. Information regarding the therapists' psychotherapy experience was Obtained from therapists' self-reports and was available for only 39 Of the 56 therapists. The crite- ria employed by Strupp (1960) was used to classify therapists as less (0-5 years) and more (6-15 years) experienced and resulted in dichotomization of therapists into graduate students and post-graduate therapists. The more experienced group thus consisted Of the 22 therapists who had completed their training, with an average Of 7.5 years of post-graduate psychotherapy experience. The less experienced group consisted of the 17 therapists who were graduate students in Clinical or Counseling Psychology, with an average of 2.5 years of supervised psychotherapy experience. Therapists' orientation was also available for these 39 therapists and was determined by two colleagues who were well-acquainted with the way in which the therapists con— ceptualized and conducted psychotherapy. Four 48 Table 2. Summary of Therapists' Characteristics Discipline p Clinical Psychology 14 Counseling Psychology Psychiatry 12 Social Work Clinical Psychology Students Counseling Psychology Students 10 O I a Orientation Analytically-Oriented Client-Centered Eclectic 11 Relationship 12 Setting Counseling Center 23 Hospital . a Experience Low Experience Level 17 High Experience Level 22 Sex Female 6 Male 50 aAvailable for only 39 therapists. 49 classifications of orientations were used: Relationship, Eclectic, Analytic, and Client-Centered. Therapists were classified as Analytic on the basis of their relative emphasis on Freudian concepts such as unconscious processes, Oedipal conflict, transference neurosis, resistance, and use of the techniques of free association and interpre- tation; as Client-Centered on the basis of relative emphasis on Rogerian concepts such as client self-actualization, communication of therapist congruence, empathic understand- ing, and an attitude of unconditional acceptance toward the client, and the use of non-directive techniques such as reflection and restatement; as Relationship on the basis of emphasis on conceptualization Of client maladaptive behavior in terms of problems in relating to others, emphasis on the client-therapist relationship as a major focal point Of therapy, and use of extensive exploration Of the therapeutic relationship to delimit and work through client conflicts and problems in relationships with others in addition to the therapist; and as Eclectic on the basis of utilization Of a combination of concepts and techniques derived from a variety Of theories. The Immediate Relationship, Depth of Self- Exploration (Carkhuff, 1964), Empathy (Berenson, Carkhuff, & Southworth, 1964), Warmth (Carkhuff, Southworth, & Berenson, 1964), and Genuineness (Carkhuff, 1964a) scales were each rated on the same five 3-minute segments selected randomly with the restriction that no segments overlapped. 50 Copies of the Carkhuff scales appear in Appendices B through E. Ratings on the E, W, G, and DX scales were made in connection with previous studies (Berenson, Mitchell, & Laney, 1968; Berneson, Mitchell, & Moravec, 1968). Two experienced clinicians with eight and four years of post- doctoral psychotherapy experience, respectively, made independent ratings on the E, W, and G scales. Ratings on the DX scale were made independently by two colleagues of the senior researchers who had previously had considerable experience in rating this scale. Ratings on the IRS scale were made independently by two advanced level graduate students in Clinical Psychology and by this writer, who had previously trained the two raters in the use Of the IRS scale on tape recordings Of therapy sessions not used in the present research. Each rater subsequently made a second set of ratings on approximately one-third or 19 of the 56 tapes after an interval of at least six weeks. Stud II--Outcome Stud ____1 __ X The data for this study were based upon the tape recorded psychotherapy sessions Of 40 clients seen in individual psychotherapy at the Henry Phipps Psychiatric Outpatient Department of The Johns Hopkins Hospital by four therapists who were randomly assigned ten clients each. The recordings were initially collected in connection with studies carried out at Johns Hopkins University 51 (Hoehn-Saric, Frank, Imber, Nash, Stone, & Battle, 1965; Nash, Hoehn-Saric, Battle, Stone, Imber, & Frank, 1965) and were subsequently used in studies by Truax (Truax, Wargo, Frank, Imber, Battle, Hoehn-Saric, Nash, & Stone, 1966a, 1966b). The initial study for which the tape recordings were collected (Hoehn-Saric ep_al., 1965) was an investigation Of the effect of client instruction regarding psychotherapy processes and typical therapist and client in-therapy behavior, designated as Role Induction, on client outcome following brief psychotherapy which lasted four months. The design consisted of a screening interview conducted by one Of two research psychiatrists followed by weekly psycho- therapy sessions over a period of four months. At the time Of the screening interview a personal history was taken on each client, information relevant to outcome measures was obtained, clients were randomly assigned to either a Role Induction (RI) or a No Role Induction (NRI) group, and those clients assigned to the RI group were given the Role Induction instructions based on the Anticipatory Sociali- zation Interview Of Orne (Orne & Wender, 1968). This instruction, emphasizing the analytic model of therapy, covered four aspects: 1) a general exposition Of psycho- therapy, 2) a description and explanation Of expected client and therapist behavior, 3) preparation for certain typical phenomena in the course of therapy such as resistance, and 4) the induction of a realistic expectation for improvement 52 within four months of therapy. Clients in the NRI group did not receive this instruction. Immediately following the screening interview the research psychiatrist designated each client as being either an Attractive (ATT) or Unat- tractive (UATT) candidate for psychotherapy. This desig- nation was a global rating based on client characteristics Of age, education, general appearance, psychOpathology, warmth, and ability to relate easily with others. An equal assignment of ATT and UATT therapy clients was made to the four therapists and to the RI and NRI groups. Within these restrictions clients were assigned randomly. Thus, each therapist saw ten clients: three ATT RI, three ATT NRI, two UATT RI, and two UATT NRI. Additional client outcome data was obtained four months after the beginning of treat- ment or at the time of termination for clients who discon- tinued treatment earlier. No statistically significant differences existed between clients in the RI and NRI groups with respect to age, education, sex, race, or ratings by the research psychiatrists of severity of illness, prognosis, and difficulty in establishing a therapeutic relationship. The sample of 40 neurotic clients excluded those with a history of alcoholism, brain damage, or mental deficiency and those having prior psychotherapy or theo- retical knowledge of the therapeutic process. A total Of 58 clients was given a screening interview; Of the 12 who were excluded from the sample, eight were excluded because of prior psychotherapy experience or theoretical knowledge 53 of the therapeutic process, two were pregnant and would have delivered within the projected four-month treatment period, one was psychotic, and one refused therapy. Of these 46 clients, six clients began treatment but dropped out before the third therapy session. The remaining 40 clients, com- posed Of 17 males and 23 females ranging between the ages Of 18 and 55 and having a mean education of 11 years, constituted the research sample. The four therapists, three males and one female, were psychiatric residents at The Johns HOpkins University and were predominantly Analytically oriented. Three were in their second year of residency and one in his fourth. The therapists were unaware of the design and nature of the research and were told only that the researchers were inter- ested in the effects of brief psychotherapy and preferred that treatment be terminated after four months. Therapists were also given the client's case history notes. Therapists met with clients at least once a week for one-hour sessions. Specific therapeutic techniques, scheduling of sessions, and further treatment following the four-month period were left to the therapists' discretion. Five predictor measures obtained on each client as part of the initial study conducted at The Johns Hopkins University (Hoehn-Saric 3; al,, 1965) were used in the present study in the analyses Of the five outcome measures: 1) The total number Of therapy sessions received by each client (NO. Sess.). 2) The Initial Adjustment score (IAS), 54 with higher scores representing a greater degree of malad- justment. The IAS was the score obtained on the Discomfort Scale administered prior to therapy. The Discomfort outcome measure, in distinction, was the change score on the Dis— comfort Scale from pre-therapy to post-therapy adminis- tration of the scale (see Appendix J). 3) A Relationship (REL) score which consisted of the post-therapy rating made by the therapist on a five point scale indicating the degree Of difficulty encountered in establishing and maintaining a satisfactory therapeutic relationship with the client. The degree of difficulty was rated in terms of extreme, marked, moderate, slight, and not at all difficult, with a higher rating indicating a greater degree of difficulty. 4) The Patient Attractiveness condition (PA) which consisted of the client's pre-therapy designation by a research psychiatrist as either an Attractive (ATT) or an Unattractive (UATT) client for psychotherapy. 5) The.Role Induction condition which consisted of the client's pre—therapy random assign- ment to either the Role Induction (RI) or the NO Role Induction (NRI) group. Five measures of client outcome were available from the initial study conducted at The Johns Hopkins University (Hoehn-Saric pp al., 1965). Two were global measures of overall improvement: Patient Statement and Therapist Statement. Three were more specific measures Of client outcome developed by Frank and his associates and used in a series of studies: Target Symptom (Battle, Imber, 55 Hoehn-Saric, Stone, Nash, & Frank, 1966), Discomfort (Frank, Gliedman, Imber, Nash, & Stone, 1957; Parloff, Kelman, & Frank, 1954; Stone, Frank, Nash, & Imber, 1961), and Social Ineffectiveness (Frank 33 31., 1957; Imber, Frank, Nash, Stone, & Gliedman, 1957; Parloff g; 31., 1954; Stone §E_al., 1961). The Discomfort outcome measure consisted of the change score on the Discomfort Scale, which was filled out by each client just prior to the screening interview and again at the time of termination. The Discomfort Scale consists of 50 items describing symptoms of anxiety, depression, and somatic complaints. The client completed the scale by indicating the extent to which he had been bothered during the previous seven days by each of the 50 symptoms by checking one of the following alternatives: 0) not at all, 1) just a little, 2) pretty much, 3) very much. These alternative responses were assigned numerical values from zero through three, respectively, and were summed over the 50 symptoms to obtain the Discomfort Scale score. The algebraic change in this score between the pre- and post-therapy administrations Of the scale constituted the Discomfort Outcome score (see Appendix J). The Patient Statement of global improvement con— sisted Of the client's post-therapy rating of the degree of his global or overall improvement in therapy on a 5-point scale. The five points of the scale were 1) worse, 2) same, 56 3) slightly better, 4) some better, 5) a lot better (see Appendix L). The Therapist Statement of client global improve- ment consisted of the therapist's post-therapy rating of the degree of the client's global improvement in therapy on a 5-point scale. The five points Of the scale were 1) worse, 2) no change, 3) slight, 4) moderate, and 5) marked (see Appendix M). The Target Symptom outcome measure consisted of the client's rating of the amount Of improvement in the three complaints or symptoms that he initially most wanted changed by therapy. Each client was asked to state three target symptoms during the screening interview. At the time of termination the client rated the amount of improvement of each target symptom on a 5-point scale. The five points were 1) a lot better, 2) some better, 3) slightly better, 4) the same, or 5) worse. The average Of the ratings of improvement for the three symptoms constituted the Target Symptom outcome score (see Appendix K). Each client was rated on the Social Ineffectiveness Scale by a member of the research staff who had no knowledge of the client's group assignments or performance during therapy. The ratings were based on a structured interview which focused on the client's day-to—day relationships with each significant individual in his life. The frequency and degree of the client's ineffective behavior with each sig- nificant individual was rated on a 6-point scale in each of 57 15 areas of social and interpersonal relations: overly- independent, overly-dependent, superficially-sociable, withdrawn, extrapunitive, intrapunitive, officious, irresponsible, impulsive, over-cautious, hyper-reactive, constrained, overly systematic, unsystematic, sexual malad- justment. A single 6-point rating was then assigned to each category by using the rating made in connection with each significant person as a guide and by taking into consider- ation the relative importance to the client of the persons with whom the ineffective behavior was shown as well as the number of persons to whom it was shown. Thus, ratings of ineffective behavior in each category were made on the basis of the frequency, degree, and importance to the client of that behavior. The Social Ineffectiveness score consisted of the sum of the numerical ratings given to each of the 15 categories (see Appendix N). Ratings of E, W, G, and DX were subsequently made on the recorded psychotherapy sessions of these 40 clients by Truax £3 a1. (1966) in a study investigating the rela- tionship between the therapeutic core conditions and the five measures of client outcome and were used in the present study. From the tape recorded therapy sessions of each of the 40 clients six 3-minute segments were excerpted for study, two segments from the first session, two from the tenth session, and two from the fifth interview before the final one. In each case, one segment was taken from the 58 middle third and one from the final third of the session in question. These segments were rerecorded onto small tape spools and randomly assigned code numbers. Each of these segments were rated on the E (Truax, 1961), W (Truax, 1962a» G (Truax, 1962c), DX (Truax, 1962b), and IRS scales (see Appendices F through I). Undergraduate college students who were naive with respect to psychotherapy theory were trained in the use Of the E, W, G, and DX scales. A total of four different raters independently rated each of the scales after training. The coded segments were presented to each set of raters in a different sequence. Two Master's level research technicians employed at the Arkansas Rehabil- itation Research and Training Center who had no prior familiarity with the IRS scale rated this scale after training in its use. RESULTS Study I—-Process Study Reliabilities of the E, W, G, and DX scales were determined in a previous study (Berenson pp 31., 1968). Pearson intercorrelations between two independent raters were as follows: E, .96; W, .96; G, .80; DX, .76. Pearson r rate-rerate reliabilities for two raters were as follows: E, .90, .88; W, .92, .89; G, .90, .85; DX, .90, .95. Pearson intercorrelations between three independent raters on the IRS were .83, .86, and .87. Pearson rate-rerate reliabilities for the three raters were .85, .88, and .91. Using Fisher's Z transformation for Pearson's r, the mean reliability between raters was .855, and the mean rate- rerate reliability was .880. The following procedure was followed in Obtaining a core condition score for each therapist. First, separate scores of E, W, and G were obtained for each therapist by computing the mean of the five ratings made on each of the three respective scales. Following this, a core condition score was determined for each therapist by computing the average of his E, W, and G scores. An IRS score was deter— mined for each therapist by computing the average Of the 59 60 five IRS ratings. Similarly, a DX score was determined for each therapist by computing the average of the five DX ratings. Whenever the raters disagreed on the rating to be given a particular segment on any of the scales, the average of the discrepant ratings was used. Table 3 presents the Pearson intercorrelations among the E, W, G, and core condition scores for the sample of 56 therapists. All r's reached at least the .001 level of significance.1 Since core condition scores were highly and significantly related to each of the separate E, W, and G scores, and since the r's among the E, W, and G scores were also high and significant, the core condition score, i.e., the mean of the E, W, and G scores, was subsequently utilized throughout the process study to represent each therapist's ratings on the E, W, and G scales. Table 3. r's among E, W, G, and Core Condition Scores E w c w .74a G .71a .76a Core .76a .79a .78a ap < .001 Core condition scores were used to classify thera- pists into a high and a low core condition group. Those 28 1All tests Of statistical significance are tw0m tailed tests. 61 therapists whose core condition score fell above the median score of 1.65 for the entire sample were designated as high core condition therapists, and those 28 therapists whose score fell below the median were designated as low core con- dition therapists. This classification of therapists is the referent for subsequent references to high and low core condition, or simply core, groups Of therapists. Table 4 presents the means and standard deviations of the E, W, G, and core condition scores for the high and low core condition groups and the entire sample. Both the means and standard deviations of the E, W, G, and core con- dition scores were quite similar within the high and low core groups and the total sample, further indicating the appropriateness Of the use of the single core condition score to represent E, W, and G scores. A t-test for inde- pendent measures indicated that the mean core condition score was significantly higher for the high core group than the low core group of therapists (t = 9.17, p < .01). Thus, this high and low core condition classification resulted in two groups of therapists which differed significantly on the relevant variable. Hypothesis I There is a significant positive relationship between the therapists' core condition scores and IRS scores. Hypothesis I was tested directly by computing a Pearson r between IRS scores, i.e., the mean of the five IRS 62 mm.mloo.H mm. hm.a mm. mm.H em. mm.a Hm. vm.a Houoe mm H z mm.HIoo.H ha. om.H ma. mm.H mm. mm.a mm. Hm.H muoo 30A mm H z mm.thm.H mu. mm.~ om. mm.m an. mm.m mo. nm.m muoo nmflm mmcmm .O.m x .O.m x .O.m x .O.m x msouw GOHDHOCOU w 3 m ouou mumflmmnmge OHOU BOA paw smflm MOM mouoom aoflpflpcoo OHOU can .0 .3 .m mo m..a.m paw m.m .e manna 63 ratings per therapist, and core condition scores of the entire sample. A significant positive relationship was obtained (r = .65,p < .001» thus confirming Hypothesis I. r's were also computed between IRS and core con- dition scores for the high and the low core groups and for the 56 therapists divided into quartiles on the basis of their core condition scores. A significant relationship was obtained within the high core group (r = .60, p < .001) but not within the low core group (r = .15). A significant positive relationship was also obtained within the first core quartile group of therapists (r = .82, p < .001), but no significant relationship was found within the second (r = .21), the third (r = .02), or the fourth (r = .03) quartiles. Thus, although IRS and core condition scores were significantly and positively related fOr the total sample, additional analyses indicated that this finding was primarily accounted for by the strong relationship between IRS and core scores which existed within the first quartile group. Table 5 presents the means and standard deviations of the IRS scores, i.e., the average of the five IRS ratings per therapist, and core condition scores and the range Of the core scores for the high and low core groups of thera- pists and for the therapists divided into core condition quartiles. A t-test for independent measures indicated that the high core group had a significantly higher mean IRS score than the low core group (t = 3.76, p < .01). A 64 Table 5. X's and S.D.'s of IRS and Core Scores for High and Low Core and Core Quartile Therapists Core Condition IRS Scores Scores Group Mean S.D. Y S.D. Range Core Ql 3.52 1.25 3.27 .44 2.42-3.96 Core 02 2.76 .78 1.98 .21 1.67-2.33 Core Q3 2.39 .58 1.46 .09 1.30-1.62 Core Q4 2.21 .49 1.15 .07 1.00-1.29 High Core 3.14 1.11 2.63 .73 1.67-3.96 Low Core 2.30 .54 1.30 .17 1.00-1.62 Total 2.72 .97 1.97 .85 1.00-3.96 65 one-way analysis of variance, used to evaluate differences in IRS scores among the quartile groups, revealed a signif- icant difference among the quartile groups (F = 4.96, p < .01), and Duncan's New Multiple Range test indicated that the first quartile had a significantly higher mean IRS score than each Of the three lower quartiles (p < .05), while the three lower quartiles did not differ significantly among themselves. In addition to the analyses used in testing Hypothesis I, the frequency with which the IRS scores, the average of the five IRS ratings per therapist, of the 56 therapists occurred at each Of the six stages of the IRS scale was also investigated. These frequencies are pre- sented in Table 6 as the percent of IRS scores at each IRS stage within each Of the core quartile groups Of therapists. These percents could not be evaluated statistically because of the many cells which contained a zero. Inspection of this table, however, suggested that the second, third, and fourth quartiles were quite similar in the distribution of IRS scores among the stages and, at the same time, quite different from the distribution in the first quartile. Therefore, the three lower quartiles were combined, and the test for the difference between two independent proportions was used to evaluate the differences in IRS scores at each IRS stage between these 42 therapists and the 14 therapists in the first quartile. Significant differences were obtained at both stages 2 and 4, indicating that a 66 Table 6. Percent Of IRS Scores at Each IRS Stage for Core Quartile Therapists IRS Stages Group 1 2 3 4 5 6 Core Ql 7.14 7.14 42.86 21.43 14.29 7.14 Core Q2 0 50.00 42.86 0 7.14 0 Core Q3 7.14 50.00 42.86 0 0 0 Core Q4 7.14 57.14 35.71 0 0 0 core Q1 7.14 7.14 42.86 21.43 14.29 7.14 N=l4 Core Q1, Q3, & Q4 4.76 52.38 40.48 0 2.38 0 N=42 Total 5.36 41.07 41.07 5.36 5.36 1.79 67 significantly greater proportion of first quartile thera- pists had an IRS score at stage 4 (Z = 2.40, p < .01) while significantly fewer had an IRS score at stage 2 (Z = 2.68, p < .01) than the remaining group of 42 therapists. Table 6 shows that the IRS scores of 82.14% Of the entire sample of therapists were divided evenly between stages 2 and 3, while the remaining scores were fairly evenly divided among the remaining stages. When the dis- tribution Of IRS scores among the stages within the first and within the combined three lower quartiles was considered separately, striking differences between the two groups were apparent. The IRS scores of approximately 41.00% of the therapists in both groups fell at stage 3. However, for the remaining therapists within the first quartile, three times as many of their IRS scores fell above stage 3 as fell below stage 3; in contrast, almost all of the IRS scores of the remaining therapists in the three lower quartiles fell at stage 2, and 92.86% of all the therapists within this group had IRS scores at either stage 2 or 3. Since the IRS scores, the average of the five IRS ratings per therapist, tended to rather severely reduce the variability of the individual IRS ratings so that a con- sequent piling up of IRS scores occurred near the middle stages of the scale, the five individual IRS ratings made on each therapist were also investigated. The frequency Of these ratings made at each IRS stage was tallied for the 68 high and low core groups, for each Of the core quartile groups, and for the 56 therapists as a group (see Table 7). There is no appropriate statistical test with which to evaluate the statistical significance of the difference in the frequency or prOportion of ratings obtained by the groups of therapists at each stage Of the IRS scale. This rating data violates the assumption of the independence Of observations which is made for the statistical tests with which the rating data might otherwise be analyzed, e.g., chi square, the test for the difference between two inde- pendent proportions. The IRS ratings were not independent; i.e., there were five repeated ratings for each therapist and, moreover, it was possible for each therapist to have more than one rating at any particular stage. The only strictly appropriate way of analyzing the rating data would involve obtaining a single measure or score per therapist to represent his five IRS ratings similar to the procedure above where t-tests were used to evaluate group differences in IRS scores. 'However, since the distribution or pattern of ratings among the stages of the scale was also very much of interest, the use of such a single measure or score was not entirely satisfactory for all of the purposes of this study. Since even a somewhat crude or approximate statis- tical evaluation of group differences in the proportion Of ratings at each stage was considered preferable to reliance on inspection or intuition so long as the assumption 69 violation and the consequent approximate nature of the statistical analysis is fully noted, the test for the dif- ference between two independent prOportions, corrected for discontinuity (Edwards, 1962, pp. 51-57), was selected as the most apprOpriate and direct test with which to analyze these differences. A chi square test proved to be unfeasible because the IRS rating data consisted of too many cells which contained a zero or an expected frequency less than five. Thus, with the realization that this use of the proportion test should be considered to provide only an approximate statistical evaluation of the data and could produce errors of unknown magnitude and kind, the test for the difference between two independent prOportions was sub- sequently used:h18tudies I and II for evaluating group differences in the prOportion of IRS ratings which occurred at individual IRS stages. However, some justification for this use may be drawn from the theoretical statistical analyses regarding the problem of the statistical independ- ence of Observations done by Chassan and Bellak (1966, pp. 493-496). On the basis of their work, Chassan and Bellak have concluded that even with relatively high degrees of statistical dependence of Observations standard tests such as the t-test, which is highly related to the test for the difference between two prOportions, can be used as a reasonably accurate tool. In addition, throughout Studies I and II the results of all tests for the difference between two proportions of 70 IRS ratings were reported whenever they reached at least the .05 level of significance so the reader could evaluate these differences for himself; however, because of the large number of such tests computed, only those differences reach- ing at least the .01 level were used as a basis for drawing conclusions. With the foregoing in mind, then, the following significant differences in the proportion Of IRS ratings at the individual IRS stages were obtained between the high and low core groups of therapists: the high core group gave fewer stage 2 responses (Z = 2.88, p < .004), more stage 6 responses (Z = 4.48, p < .001), and more stage 4 responses (Z = 3.49, p < .006) than did the low core group. The response pattern or the distribution of ratings among the IRS stages for the high and low core therapists displayed some striking dissimilarities. While both groups of therapists gave approximately the same number of stage 3 responses, the group of low core therapists showed much less variability than the high core therapists in the kinds Of IRS responses they made, with 82.86% of all responses in the low core group occurring at either stages 2 or 3. The following significant differences in IRS ratings were obtained between the core quartile groups of therapists. At stage 2, fourth quartile therapists had a greater pro- portion of IRS ratings than the first quartile therapists (Z = 3.30, p < .001). At stage 6, the first quartile gave a greater prOportion of responses than the second (Z = 3.17, Table 7. Percent Of IRS Ratings at Each IRS Stage for High and Low Core and Core Quartile Therapists IRS Stages Group 1 2 3 4 5 6 High Core 12.14 26.43 30.00 12.86 2.14 16.43 Low Core 15.00 43.57 39.29 1.43 0 .71 Core Ql 12.86 22.86 21.43 14.29 1.43 27.14 Core Q2 11.43 30.00 38.57 11.43 2.86 5.71 Core Q3 15.71 35.71 45.71 1.43 0 1.43 Core Q4 14.29 51.43 32.86 1.43 0 0 Total 13.57 35.00 34.64 7.14 1.07 8.60 72 p < .0001), or the fourth quartiles (Z = 4.39, p < .0001). At stage 3, the first quartile gave fewer reSponses than either the second (Z = 2.03, p < .04) or the third (Z = 2.86, p < .004) quartiles. Thus, the first quartile made signifi- cantly fewer stage 2 responses than the fourth quartile, significantly fewer stage 3 responses than the third quartile, and significantly more responses at stage 6 than each of the three lower quartiles. These results suggested that the differences found at stages 2, 3, and 6 were pri- marily a function of the first quartile group of therapists differing significantly from one or more of the lower quartile groups. Description of the characteristic manner of respond- ing on the IRS by therapists grouped into the core quartiles can be seen from Table 7. In comparison with the first quartile, second quartile therapists made fewer responses at stage 6 while simultaneously making more responses at both stages 2 and 3. In comparison with the second quar- tile, the third quartile made fewer responses at both stages 4 and 6 while making a greater number at stagesZ and 3. In contrast to the third quartile, the fourth quartile simul- taneously made fewer stage 3 and more stage 2 responses. Thus, the pattern Of IRS responses within each of the core quartiles displayed a general trend progressing from the first through the fourth quartiles: within each successive quartile there occurred both a decrease in the variability of responses and a decrease in the number Of reponses made 73 above stage 3 with a concomitant increase in the number Of responses made at stage 2. In summary, a significant positive relationshp was found between core condition scores and IRS scores, i.e., the mean Of the five IRS ratings per therapist, for the entire sample of therapists, and Hypothesis I was therefore confirmed. A significant positive relationship was also found within the high core group, but no significant rela— tionship was found within the low core group. Moreover, in comparison with the low core group the high core group had a significantly higher mean IRS score and gave a signifi- cantly greater number of responses at both stages 6 and 4 and a significantly fewer number Of responses at stage 2. Analyses of the core quartile groups indicated that within the first quartile IRS scores were significantly and posi- tively related to core scores and, in addition, were sig— nificantly higher than the IRS scores Of each of the three lower quartiles. Moreover, significant differences between the quartile groups in the proportion Of IRS ratings at stages 2, 3, and 6 tended to differentiate the first quar- tile from the remaining quartiles. Thus, two distinct groups which differed on both IRS ratings and mean IRS scores as well as on the relationship between IRS and core condition scores emerged from the data: the group Of 14 therapists having the highest core scores and the remaining 42 therapists having the lowest core scores. 74 Hyppthesis I; The group of outpatient therapists and the group of inpatient therapists do not differ significantly on mean IRS scores. Hypothesis II was tested directly by evaluating the difference between the mean IRS scores, the average of the five IRS ratings per therapist, of the 33 therapists of the inpatients and the 23 therapists of the outpatients with a t-test for independent measures. The Obtained t indicated that the inpatient and outpatient therapists did not differ significantly on mean IRS scores (t = .38); therefore, Hypothesis II was confirmed. Means and standard deviations of IRS scores of inpatient and outpatient therapists are shown in Table 8. 1 On the basis of their IRS scores the 56 therapists were divided into a high and a low IRS group, using the median IRS score of 2.56, and subdivided into IRS quartile groups. t-tests for independent measures were used to evaluate the differences in mean IRS scores between the inpatient and outpatient therapists within each of these six IRS groups. The results indicated that mean IRS scores Of the high IRS group did not differ significantly between inpatient and outpatient therapists (t = .63), but within the group of low IRS therapists those who saw outpatients had a significantly higher mean IRS score than those who saw inpatients (t = 2.86, p < .05). NO significant differences were found between the inpatient and outpatient therapists oo.cloo.a mm. on. HH.H Nn.N mm.N wo.~ HmuOH 75 mm.~uoo.a am. as. me. so.~ RN.N ma.H wmuz man 36a oo.euoe.~ 3a. 6a. so.H am.m 3N.m a3.m mmuz mmH swam ma.~-oo.a mm. 50. mm. 65.H ao.~ a6.H 30 man mm.~-o~.m on. on. on. 6m.m em.m am.~ mo mmH oo.m-oe.~ an. ON. om. 3N.N 3a.~ 3R.N No maH oo.onoo.m mm. as. so.H 00.3 ma.m H~.e no maH emaaaaoo mumwmwumsH mumfiamuosh muwfiamuwnH mumaamuwne mumaamuoca mumfiamumnH a u ufimfiumm ufiUHUNQUDO UGOHUGQHH quHqu ufiUHUNQuDO muGQHUmQGH DO U .650 s .6H -650 s .6H ommwm .o.m cmmz mumwmmumce oaauumno mmH pan uamfiumquso pcm IGH pom mouoom mmH mo m..n.m pom O.M .w memH 76 within any of the IRS quartile groups (t's = 1.53, 0.00, .76, and 1.38, respectively) (see Table 8). The distribution of the five IRS ratings per thera- pist at each IRS stage was also investigated for the groups of inpatient and outpatient therapists as well as for the inpatient and outpatient therapists within both the high and the low IRS therapist groups. Tests for the difference between two independent proportions indicated that the only significant differences occurred at stage 3 where the group of outpatient therapists had a greater proportion Of responses than the inpatient group (Z = 3.20, p < .001), and within the group of low IRS therapists outpatient therapists had a significantly greater proportion of responses than inpatient therapists (Z = 2.83, p < .005) (see Table 9). To further clarify the differential IRS responses made by inpatient and outpatient therapists at the indi- vidual IRS stages, the percent of the total number of IRS ratings at each stage for the entire sample occurring in the inpatient and outpatient groups of therapists was calculated. Table 10 shows that inpatient therapists gave 75.00% of the total number of stage 6 responses given by the entire sample, 73.68% of the total number of stage 1 responses, and 64.29% of the total number of stage 2 responses. Thus, more variable and extreme IRS responses were given by the inpatient therapists than by the outpatient therapists. 77 Table 9. Percent of IRS Ratings at Each Stage for In- and Out- Patient and High and Low IRS Therapists IRS Stages Group 1 3 4 6 33 Inpatient Therapists 16.97 38.18 26.67 5.45 1.82 10.91 23 Outpatient Therapists 8.70 30.43 46.09 9.57 0 5.22 High IRS 16 Inpatient Therapists 7.50 16.25 41.25 8.75 3.75 22.50 12 Outpatient Therapists 5.00 11.67 56.67 16.67 0 10.00 Low IRS 17 Inpatient Therapists 25.88 58.82 12.94 2.35 O 0 ll Outpatient Therapists 12.73 50.91 34.55 1.82 0 0 Total 13.57 35.00 34.64 7.14 1.07 8.60 78 Table 10. Percent of Total IRS Ratings at Each Stage for Inpatient and Outpatient Therapists IRS Stages Group 1 2 3 4 5 6 33 Inpatient Therapists 73.68 64.29 45.36 45.00 100.00 75.00 23 Outpatient Therapists 26.32 35.71 54.64 55.00 0 25.00 Total 100 100 100 100 100 100 In summary, there was no significant difference in mean IRS scores between the 33 inpatient therapists and the 23 outpatient therapists, and Hypothesis II was therefore confirmed. However, the group of outpatient therapists gave a significantly greater proportion Of stage 3 responses than inpatient therapists. Additional analyses indicated that this difference at stage 3 was primarily attributable to the group Of low IRS therapists who saw outpatients and who had both a significantly higher mean IRS score and a signifi- cantly greater prOportion of stage 3 responses than the low IRS therapists who saw inpatients. High IRS therapists who saw inpatients and those who saw outpatients, on the other hand, were not differentiated by either mean IRS scores or by the proportion of IRS ratings at any of the individual stages. 79 Hypothesis III There is a significant positive relationship between the therapists' IRS scores and their clients' DX scores. Hypothesis III was tested by computing a Pearson r between the IRS and DX scores of the entire sample of 56 therapists and their respective clients. A significant positive relationship was obtained (r = .31, p < .05); therefore, Hypothesis III was confirmed. Further delineation Of the relationship between DX and IRS scores was achieved by classifying therapists on the basis of their client's DX scores into a high and a low DX group, using the median DX score Of 1.88, and into UK quartile groups. This classification Of therapists and their respective clients thus made possible analysis Of therapists' IRS scores and the prOportion Of IRS ratings at each IRS stage in relation to their clients' DX scores. See Table 11 for means and standard deviations of IRS and DX scores for these six DX groups. Analyses of the IRS scores indicated that therapists in the high DX group had significantly higher (t = 3.04, p < .01) and significantly more variable (F = 5.03, p < .01) IRS scores than therapists in the low DX group. Duncan's New Multiple Range test, used to evaluate the differences in mean IRS scores among the DX quartile groups, further indi- cated that therapists of clients in both the first and second DX quartiles had significantly higher IRS scores than those in the fourth DX quartile (p < .05). 80 Table 11. X's and S.D.'s Of IRS and DX Scores for High and Low DX and DX Quartile Groups IRS Scores DX Scores Group _ _ X S.D. X S.D. Range DX Ql 3.05 1.04 2.89 .33 2.50 - 3.25 DX Q2 3.14 1.34 2.06 .17 1.88 - 2.38 DX Q3 2.41 .60 1.71 .10 1.62 - 1.88 DX Q4 2.30 .46 1.42 .14 1.12 - 1.50 High DX 3.10 1.18 2.48 .50 1.88 - 3.25 Low DX 2.36 .53 1.56 .20 1.12 — 1.88 Total 2.72 .97 2.02 .59 1.12 - 3.25 Analyses of the frequency of IRS ratings at each IRS stage for therapists who saw high DX clients and those who saw low DX clients indicated that therapists Of high DX clients gave a significantly greater number Of responses at both stage 4 (Z = 3.06, p < .002) and stage 6 (Z = 3.63, p < .0003) and, at the same time, gave significantly fewer responses at stage 2 (Z = 3.40, p < .0007) than therapists of low DX clients. With therapists divided into the DX quartile groups, differences were Obtained at the same stages, stages 2, 4, and 6, at which differences were obtained between the high and low DX groups. At stage 6, both the first and second DX quartiles had a greater pro- portion of responses than the third (Z = 2.10, p < .04; 81 Z = 2.31, p < .02, respectively) or the fourth DX quartile (Z = 2.50, p < .01; Z = 2.70, p < .007, respectively). At stage 4, the first quartile had a greater proportion Of responses than either the third (Z = 2.10, p < .04) or the fourth quartile (Z = 2.50, p < .01). Finally, at stage 2, the first quartile had fewer responses than either the third (Z = 1.99, p < .05) or the fourth quartile (Z = 3.13, p < .002) (see Table 12). Table 12. Percent of IRS Ratings at Each Stage for High and Low DX and DX Quartile Groups IRS Stages Group 1 2 3 4 5 6 High DX 13.57 25.00 32.14 12.14 2.14 15.00 Low DX 13.57 45.00 37.14 2.14 0.00 2.14 DX Ql 17.14 22.86 30.00 14.29 1.43 14.29 DX 02 10.00 27.14 34.29 10.00 2.86 15.71 DX Q3 14.29 40.00 40.00 2.86 0 2.86 DX Q4 12.86 50.00 34.29 1.43 0 1.43 Total 13.57 35.00 34.64 7.14 1.07 8.60 In summary, a significant positive relationship was found between IRS and DX scores within the entire sample, and Hypothesis III was consequently confirmed. With therapists divided into DX groups on the basis of their client's DX scores, IRS scores were significantly higher and 82 more variable for therapists of high DX clients than for therapists of low DX clients and, in addition, therapists in the first and second DX quartiles each had significantly higher mean IRS scores than therapists in the fourth DX quartile. Therapists whose clients had high DX scores were also found to give a significantly greater number of responses at both stages 4 and 6 while giving significantly fewer responses at stage 2 than therapists whose clients had low DX scores. Hypothesis IE Three specific hypotheses are used to test the general hypothesis that IRS responses significantly differ- entiate on the variable of orientation but do not signifi- cantly differentiate on the variable of experience level. Hypothesis IVA Orientations to psychotherapy are ordered on the basis of IRS scores from highest to lowest in the following way: Relationship, Eclectic, Client-Centered, and Analytic. Hypothesis IVB Relationship therapists have significantly higher IRS scores than Eclectic, Client-Centered, or Analytically oriented therapists. Hypothesis IVC No significant differences in IRS scores occur between the group of high experience level therapists and 83 the group of low experience level therapists within any of the four orientation groups. Since information regarding orientation and experi- ence level was available for only 39 of the 56 therapists, the following analyses were based on only those 39 thera- pists. Table 13 shows the number of inpatients and out- patients as well as the average number years of experience in conducting psychotherapy for the 39 therapists classified according to orientation and experience level. The IRS scores, i.e., the average of the five IRS ratings per therapist, formed a 2 x 4 factorial design with classification of therapist experience level into high (above five years) and low (five years or less) and classi- fication of therapist orientation into Relationship (RE), Eclectic (EC), Client-Centered (CC), and Analytic (AN). A square root transformation, using the formula X + .5, was performed on the IRS scores because of the simultaneous occurrence of heterogeneity of variance and unequal n's in the cells. Table 14 presents the means and standard devia- tions of both the transformed and nontransformed IRS scores for the four orientation groups divided into high and low experience levels. Table 14 shows that the means of the transformed IRS scores for the RE, EC, CC, and AN groups were 2.05, 1.77, 1.75, and 1.72, respectively. Thus, the means of the trans- formed IRS scores for the orientation groups demonstrated the predicted order; therefore, Hypothesis IVA was confirmed. 84 mm m m m.m m ma m.h Hmuoe m H H m m m n.m oaumamce n m m m.m H m m.n commucmo pcmflau Ha m m h.H o m n UHUUmHom NH m N m.H m v n mflcmcoflmeom mymflmmnmza pcmflumm ucmflumm .mxm .mnw ucmflumm yamflumm .mxm .mum Hmuoa luoo IGH M Iuoo ncH m moonw mocmflnmmxm 30A mocmwnmmxm swam mpmfimmnmna Hm>mq wocmflummxm paw coflumucmfluo ou mcflpnoood pmamawmmao How mucmfipmmpso pom mucmfiummcH mo Honesz pom mocmfluwmxm mo mummy m .mH magma 85 vo.a wm.m Hm. om.m om. nm.m Ho.H an.m vm.a m>.m Hmpos hm.a mH.m em. vm.H mm. nw.m mm.H mm.m ov.H vv.v mocmwummxm Boa mo. mh.m we. ww.m mm. mv.m mm. Hm.m om. mm.m mocmflummxm swam coflumEH0mmcmua ow Hoflum mmuoom mmH hm. vm.H om. Nb.a 0H. mh.H mm. hh.a hm. mo.m Hmuoe mm. mm.a om. mm.a 0H. mn.H em. Hm.a em. om.m mocmfluomxm 30A 5H. om.H ea. nn.a 0H. H5.H «a. mn.H em. mm.a mocmflummxm swam mmuoom mmH UmEHowmcmue .Q.m Gmmz .D.m Cmmz .Q.m :mwz .Q.m cmwz .Q.m ammz msouw mumwmmumne oaymfimnfl Umnmucmu UHMUmHom mflnmcoflumamm mm Hopoa uswflau Hw>mq mocmflummxm pom coflumpcwfluo 0p mcflpuoooé pmwmammmau mumfimmnmse Mom coH0wEuom Imcmue ou Hoanm mmuoom mmH paw mmuoom mmH meHommcmua nuom mo m..o.m paw m.m .va magma 86 The 2 x 4 analysis of variance, performed on the transformed IRS scores and using the unweighted means method for unequal n's, indicated there was no significant differ- ence in mean IRS scores between the high and low experience therapists and no significant Experience Level x Orientation interaction effect. This failure to obtain a significant interaction effect thus indicated that high and low experi— ence therapists within each of the orientation groups did not differ significantly on mean IRS scores. Both of these findings were in agreement with Hypothesis IVC; therefore, Hypothesis IVC was confirmed. However, a significant orientation main effect (F = 4.68, p < .05) was obtained from the analysis of variance. Duncan's New Multiple Range test, used to evaluate the difference in mean IRS scores among the four orientation groups, indicated that the RE group had a significantly higher mean IRS score than the AN group (p < .05) but only tended to have a higher score than the CC (p < .10) and EC group (p < .10). Therefore, Hypothesis IVB was rejected. Table 15 summarizes the analysis of variance. The frequency of the five IRS ratings per therapist at each IRS stage was also investigated for the high and low experience level groups of therapists, for the four orientation groups, and for the high and low experience therapists within each orientation. Table 16 presents the percent of IRS ratings at each stage for the high and low experience therapists. Tests for 87 Table 15. Summary of Orientation X Experience Level AOV df SS MS F Experience Level 1 .041 .041 —- Orientation 3 .785 262 4.68 Experience X Orientation 3 .207 .069 1.23 N.S. Error 31 1.73 .056 Total 38 ap < .05 Table 16. Percent of IRS Ratings at Each Stage for Experi— ence Level and Orientation Groups IRS Stages Group 1 2 3 4 5 6 High Exp. 8.18 31.82 41.82 11.82 0 6.36 Low Exp. 16.47 24.71 17.06 8.24 3.53 20.00 RE 3.33 25.00 21.67 18.33 1.67 30.00 EC 12.73 40.00 29.10 7.27 1.82 9.10 CC 14.29 40.00 31.43 8.57 2.86 2.86 AN 20.00 11.11 64.44 4.44 0 0 Total 11.79 28.72 35.38 10.26 1.54 12.31 88 the difference between two proportions indicated the follow- ing. At stage 6, low experience therapists gave a signifi- cantly greater proportion of responses than high experience therapists (Z = 2.65, p < .008). At stage 3, high experi- ence therapists gave a greater proportion of responses than low experience therapists (Z = 1.99, p < .05). Thus, although no significant differences were found between the two experience groups on mean IRS scores, significant differences were obtained in the analyses of IRS ratings of these two groups. Table 16 also shows the percent of IRS ratings at each stage for therapists in the four orientation groups. Tests for the difference between two proportions indicated the following. At stage 1, the AN group gave more responses than the RE group (Z = 2.43, p < .02). At stage 2, the AN group gave significantly fewer responses than either the CC (Z = 2.75, p < .006) or EC group (Z = 3.03, p < .002). At stage 3, the AN group gave significantly more responses than the CC (Z = 2.69, p < .007), the EC (Z = 3.34, p < .001), or the RE group (Z = 4.22, p < .0001). At stage 6, the RE group gave significantly more responses than the EC group (Z = 2.57, p < .01), the CC (Z = 2.93, p < .003), or the AN group (Z = 3.78, p < .0002). These findings are summarized in Table 17. The major differences, then, between the orientation groups in the proportion of responses at indi- vidual IRS stages occurred between the RE and AN group where the RE therapists gave a significantly fewer number of 89 pmumpcmo #cmflao p oauomaom u o oaumamcd u Q mflnmcowumamm u m 0N0. V hm omOO V Q ..mxm nmflm mm ..oxm 30A z¢ v .oxm 304 mm v .mxm roam z¢ .moo. v a .mxm .mo. v o ..mxm swam v .mxm Bog 30A v .mxm swam oNO. V m ..mxm 30g mm .mooo. v m .z< v mm .Hooo. v o .mm v 24 v .mxm nmflm mm .moo. v o .oo v mm .Hoo v m .om v 24 .moo. v m .oom A z« .Ho. v m .om v mm .500 v o .oo v 24 .moo. v o .poo A 24 .mo. v m .mmm v m m N a mmmmum mmH mmsouw Hm>mq monmflummxm pom coaumucmflno Mom mmepm mmH up mmcwumm mmH mo :owuuomoum CH mmocmummmwa pcmoflMHcmHm .na manna 90 responses at both stages 1 and 3 and, simultaneously, gave a significantly greater number at stage 6. The EC and CC groups were quite similar to one another in the number of responses at each stage, and both groups differed signifi— cantly from the AN group at stages 2 and 3 and from the RE group at stage 6. Comparison of the distribution of IRS responses among the stages within the AN and RE groups revealed striking differences between these two groups. Ninety—five percent of all responses given by the RE thera- pists were approximately evenly divided among stages 2, 3, 4, and 6. In contrast, approximately 65.00% of all responses given by the AN therapists were given at stage 3 alone, 20.00% at stage 1, and 11.00% at stage 2. Thus, the response pattern of the RE therapists was more variable and consisted of more responses at the extreme stages of the scale than that of the AN therapists. The response pattern of the EC and CC therapists fell about midway between the more extreme patterns of the RE and AN therapists. Finally, the percent of ratings at each IRS stage was tabulated for the high and low experience therapists within each of the four orientations (see Table 18). Tests for the difference between two proportions were computed at each stage between the two experience groups within each orientation and indicated the following. High experience RE therapists gave more stage 2 (Z = 2.26, p < .02) and fewer stage 6 responses (Z = 2.28, p < .02) than low expe— rience RE therapists. Within the AN orientation, high 91 Table 18. Percent of IRS Ratings at Each Stage for Experience Levels Within Orientation Groups IRS Stages Group 1 2 3 4 6 Relationship High 2.86 37.14 17.14 25.71 17.14 Low 4.00 8.00 28.00 8.00 .00 48.00 Eclectic High 8.00 48.00 36.00 4.00 4.00 Low 16.67 33.33 23.33 10.00 .33 13.33 Client Centered High 6.67 53.33 33.33 6.67. 0 Low 20.00 30.00 30.00 10.00 .00 5.00 Analytic High 14.29 5.71 74.29 5.71 0 Low 40.00 30.00 30.00 0 0 Total 11.79 28.72 35.38 10.26 .54 12.31 92 experience therapists gave more stage 3 responses (Z = 2.20, p < .03) than low experience therapists. All significant differences obtained in the proportion of ratings at each stage by therapists classified according to orientation and experience level are summarized in Table 17. In summary, the four orientation groups were ordered from highest to lowest on the basis of mean IRS scores in the following manner: RE, EC, CC, and AN. Thus, Hypothesis IVA was confirmed. The RE therapists had a significantly higher mean IRS score than the AN therapists but only tended to have a higher score than the CC or EC groups, and Hypothesis IVB was therefore rejected. In addition, com- parisons between the orientation groups on the total number of ratings at each IRS stage indicated that the RE thera- pists gave a significantly greater proportion of responses at stage 6 than each of the three other orientation groups and gave significantly fewer responses at both stages 1 and 3 than the AN therapists. The 22 high and the 17 low experience therapists did not differ significantly on mean IRS scores nor did the high and low experience therapists within each of the four orien- tation groups differ significantly on mean IRS scores, and Hypothesis IVC was confirmed. However, the low experience therapists gave a significantly greater number of stage 6 responses and tended to give fewer stage 3 responses than the high experience therapists. Comparisons between high and low experience therapists within each of the 93 orientations indicated that this difference found between the two experience groups at stage 6 was primarily accounted for by the RE therapists, while the difference found at stage 3 was primarily due to the AN therapists. Study II-—Outcome Study Separate scores were computed for the IRS, E, W, G, and DX scales for each of the 40 cases and for each of the four therapists. The design on the study called for six ratings to be made on each scale for each of the 40 cases: for each case the six ratings per scale consisted of one segment selected from the middle portion and an additional segment from the last third portions of the first, tenth, and fifth from final therapy sessions. The score on each scale for each of the 40 cases was obtained by computing the average of the six ratings per scale; the score on each scale for each of the therapists was obtained by computing the average of the scores on the respective scales of the ten cases randomly assigned to each therapist. In the cases where not all six ratings were available because of missing segments, the score consisted of the average of the avail— able ratings. In addition, two measures of the core conditions, a ZEWG score and a EEG score, were determined for each case and for each therapist. Each of the distri- butions of the 40 E, W, and G scores was converted into a standard Z distribution with a mean of zero and a standard deviation of one. The first core condition score, ZEWG, was 94 then obtained by summing the E, W, and G Z-scores, and a second core condition score, EEG, was obtained by summing the E and G Z-scores for each case and for each therapist. The reliability of the E, W, G, and DX scales was determined in the previous study by Truax pp 31. (1966). The reliability as measured by intraclass correlation for the combined four raters on the mean ratings per case was E, rkk = .63; W, rkk = .59; G, rkk = .60; DX, rkk = reliability of the mean IRS ratings per case for the two .71. The raters as measured by a Pearson correlation was .85. Table 19 presents the means, standard deviations, and intercorrelations for the 12 predictor variables in the present study for the 40 cases. The Pearson intercorrelations among the E, W, and G scores for the 40 cases were as follows: E and G (r =.60, p < .001), E and W (r = .07), and W and G (r = -.11). Thus, for the entire sample of cases E and G scores were signifi- cantly related, while W scores were not significantly related to either E or G scores, and moreover, a negative r was obtained between W and G scores. Truax (Truax e3 31., 1966) has suggested that whenever one of the three compo- nents of the core conditions is negatively related to the other two, client outcome is best predicted by the two con- ditions which are most highly related. For this reason, the EEG scores were included as an additional predictor variable in this study. 95 . a Hoo v p Ho v an Emuwoum wouaaaoo so .xao>fiuommmmu .q a n ma pmumamfimmpm No. v as Hammond umuamaou so .m~m>fiuomammu .m w m mm pmumcmwmmpm mo. v mm Ho.H oo.o oo. Hmz\HH no. mm.q wH.I mo.| xa oo. no.m 5a.: HH. mo. mmH Hm. om.m 50.: oo. oo.: om.| 3 mm. om.m mo. om. mo.l Ho. HH.I 0 mm. m~.¢ Ho. «0.: ma.1 mm. 50. com. m mm.H om.N mmm. om. NH.| mo. mm. No.| mo. 4mm cm.q mo.~a «N.I mm.| Ha. 00.: um¢.| no. mo.l mo.| .mmmm oz mo.~ me. no. No. mH.| no. ooq. noq. pmo. «a. mmm.| 03mm ow.a mo. «0. mo. HH.I NH. No.| ppm. too. No. mo. two. wmw mH.mH ow.m~ ma. HH., no.1 mo.1 ma. HN. mmm. oa.l no.1 one. men. m “anodemum NH 96 Hypothesis I There is a significant positive relationship between the core condition scores and IRS scores within the entire sample of 40 cases. Hypothesis I was tested directly by computing Pearson r's between the EEWG scores and the IRS scores, i.eq the average of all IRS ratings per case, and between the IRS and EEG scores of the entire sample of 40 cases. The results indicated that no significant relationship existed between IRS scores and either EEWG scores (r = .08) or EEG scores (r = .18); therefore, Hypothesis I was rejected. The 40 cases were dichotomized into high and low core condition groups of therapists, each group consisting of two therapists and their 20 clients. Each of the sepa- rate EEWG, EEG, E, and G scores of the four therapists resulted in identical dichotomies of two high core and two low core condition therapists. t-tests for independent measures indicated that the high core group of therapists had significantly higher mean scores on E (t = 3.82, p < .01), G (t = 3.44, p < .01), EEG (t = 4.26, p < .01), and IRS (t = 2.33, p < .02) than the low core group. Table 20 presents the means and standard deviations of the E, W, G, EEWG, EEG, and IRS scores for each therapist and for the high and the low core condition groups of therapists. Pearson intercorrelations were computed among the IRS, E, W, G, EEWG, and EEG scores for the high and low core groups of therapists and the entire sample and are 97 ov N Z OO. mO.m OO.H mO. mo.m OO. mm. Om.m Hm. OO.m hm. ON.O Hmuoe ON n z m a H om. OO.N Om.H mO.Hn OH.N OO.| ON. mm.m Rm. 50.m Hm. O0.0 mumHmmuwse muoo 30H ON n z e O N OO. ON.m Om.H mm. OH.H mm. OH. RO.m mm. em.m mm. O0.0 mpmHmmumae mnoo anm OH n 2 en. mm.m ON.H ON.H am. am. OH. OO.m Hm. OO.m ON. mm.O e umHmmumce OH u 2 me. OO.m mm. Om.m- HO.N mm.u mH. RO.m RH. Om.m mm. Hm.m m HmHmmumze OH n 2 mm. RH.m OH.H HO. Rm.H Om. ON. OO.m mm. OO.m mm. m0.0 m umHmmumne OH H 2 we. mO.m mm. mm. Om.H mp. OO. me.m mm. m0.0 OO. nm.e H umHmmnge .o.m m .o.m m .o.m m .o.m m .o.m m .o.m m mmH omw ozmw mumflmmnmna mnou 304 039 com swam 039 How Ocm umHmmHmne somm How mmnoom mmH qu omw .ozmw .o .3 .m m0 m..o.m cam m.x .Om oHnme 98 shown in Table 21. For the sample of 40 cases IRS scores tended to be positively related to E scores (r .27, p < .10) and negatively related to W scores (r = -.26, p < .11) but unrelated to G scores (r = .01), EEWG scores (r = .08), or EEG scores (r = .18). The IRS scores had strikingly different patterns of relationships with E and G scores within the high and low core groups of therapists. In the high core group IRS scores had a significant positive relationship to E (r = .45, p < .05) and EEG scores (r = .47, p < .05), while in the low core group IRS scores had a significant negative relationship to G (r = -.55, p < .05) and EEG scores (r = -.50, p < .05). Consequently, the failure to find IRS scores significantly related to E, G, or EEG scores within the entire sample of 40 cases reflected these opposing relationships within the high and low core therapist groups which tended to mask or cancel one another in the entire sample. The lack of a significant relationship between IRS and EEWG scores within the entire sample, however, reflected primarily the positive relation- ship of IRS scores with E scores, the negative relationship with W scores, and the lack of relationship with G scores within the total sample. The frequency distribution of IRS scores, i.e., the average of all IRS ratings per case, at each stage of the IRS scale was tallied separately for the two high core therapists and the two low core therapists with their respective sets of 20 clients. Tests for the difference 99 Table 21. r's Among IRS, E, W, G, EEWG, and EEG Scores for High and Low Core Therapists w G EWG EG IRS High .06 .28 .57b .65 .45b g Low .40a .56C 30 .89 .30 Total .07 .60C 27a .84 .27a High —.20 .43a —.33 .25 3 Low .19 65C .35 .13 Total -.12 44° -.11 .26 . C High .59 .78 .10 9 Low .17 .87 .55b Total 39b .87 .01 High .62 .13 EEWG Low .46 .12 Total .55 .08 High .47b b EEG Low .50 Total .18 a < .10 bp < .05 Cp < .01 100 between two independent proportions revealed no significant differences between the high and low core therapists in the proportion of IRS scores at any of the individual stages. Reference to Table 22 shows that 70.00% of the total number of IRS scores in both the high and low core groups fell at stage 3; however, in the high core group the remaining 30.00% fell at stages above stage 3 while in the low core group 20.00% fell at stage 2. Table 22. Percent of IRS Scores at Each IRS Stage for the High and Low Core Therapists IRS Stages Group 1 2 3 4 5 6 High Core N = 20 0 0 70.00 25.00 5.00 0 Low Core N = 20 0 20.00 70.00 10.00 0 0 Total N = 40 0 10.00 70.00 17.50 2.50 0 In addition to the above frequency distribution of IRS scores, the frequency of the IRS ratings at each stage of the IRS scale was also tallied for the high and low core therapists. Tests for the difference between two inde- pendent proportions indicated that the greatest difference between the two groups occurred at stage 6, where the high core therapists had a higher prOportion of responses than the low core therapists (Z = 2.03, p < .04). The ratings . q-a . .viv 0!: V '3» 0 cm pv-I'" . n I, C Cr “‘V V‘ n'. 2" : O‘H‘y I (I) r?- 101 made at stages 4, 5, and 6 were then combined: 26.53% of the responses of the high core therapists and 8.33% of the responses of the low core therapists occurred at the com— bined stages 4, 5, and 6. This difference was significant (Z = 3.00, p < .003), thus indicating that the high core therapists made a significantly greater proportion of responses in which attempts were made to refer client state- ments to the immediate therapeutic relationship than did the low core therapists. Table 23 shows the percent of ratings at each stage for the high and low core therapists and for the entire sample. Table 23. Percent of IRS Ratings at Each IRS Stage for the High and Low Core Therapists IRS Stages Group 1 2 3 4 5 6 High Core N = 98 2.04 16.33 55.10 13.27 2.04 11.22 Ratings Low Core N = 84 5.96 20.24 65.48 4.76 1.19 2.38 Ratings Total N = 182 3.85 18.13 59.89 9.34 1.65 7.14 Ratings In summary, IRS scores were not significantly related to either measure of the core conditions, EEWG or EEG scores, within the total sample of cases, and Hypothesis I was therefore rejected. The nonsignificant relationships 102 between IRS scores and the two core condition measures reflected the opposing patterns of relationships between IRS scores and the individual E, W, and G scores within the high and low core groups and within the total group of cases. However, high core therapists had a significantly higher mean IRS score and gave significantly more IRS responses at the combined higher stages of the scale, stages 4, 5, and 6, than low core therapists. IRS Responses pp Middle and Last Segments Selected from the First, Tenth, and Fifth from Final Therapy‘Sessions Although no specific hypotheses were proposed regarding differential IRS scores as a function of time within an individual therapy session or as a function of time across the entire course of therapy, this information was of considerable interest and was consequently investi— gated insofar as was possible on the basis of the available data. Many of the middle and last segments designated for analysis in the design of the study, particularly those from the fifth from final session, were not available because of inaudible tape recordings, missed sessions, or fewer than six sessions. For the first therapy session, middle and last segments were available for 38 clients, 18 seen by low core and 20 by high core therapists. Segments for 20 clients, seven seen by low core and 13 by high core thera- pists, were available from the tenth therapy session, where the tenth session occurred prior to the fifth from final session: thu" stifled. In :3 the tenth zents were a‘7 grist to the :5 core and fifth from fl :hrczological. final session session 12.5. ;ri3r to the lEth session. 251‘; umber c raged from 1 In 0: im, the IR ‘lemiddle a Set ‘ $108 Were 103 session, thus making the tenth session the second one to be sampled. In all but one case this reduction from the first to the tenth session in the number of clients for whom seg- ments were available was a result of termination of therapy prior to the 14th session. Segments for 13 clients, three low core and ten high core cases, were available from the fifth from final session where the fifth from final session chronologically followed the tenth session. This fifth from final session for these 13 clients was, on the average, session 12.5. Altogether, a total of 19 clients terminated prior to the 14th session, and 30 terminated prior to the 16th session. Within the total sample of 40 clients the mean number of sessions was 12.7 and the number of sessions ranged from four to 19. In order to make maximum use of all the available data, the IRS ratings obtained on the segments taken from the middle and last third portions of the first therapy session were analyzed for the group of 20 clients seen by the high core therapists and the group of 18 clients seen by the low core therapists. A 2 x 2 factorial design was used. The high and low core groups constituted one factor, and the middle and last segments constituted the second factor for which there were repeated measures. Results of the analysis revealed no significant differences in IRS ratings between either the high and low core therapists or between the middle and last segments and no significant Therapist x Segment interaction. The analysis of variance . ‘3.’ -u'r. *5 5321:5071 the :21 th fir 212 x 2 fa high and 10"- :e two rema first and te :itained frc Espent x Se far repeatec :m‘: p «.10.. CCRSI 104 is summarized in Table 24, and the means and standard deviations of the IRS ratings are shown in Table 25. IRS ratings were then analyzed for those 20 clients, seven in the low core group and 13 in the high core group, for whom the middle and last segments were available from both the first and tenth sessions. The data formed a 2 x 2 x 2 factorial design with one factor consisting of high and low core therapists and with repeated measures on the two remaining factors, middle and last segments and first and tenth sessions. The only significant difference obtained from the analysis of variance consisted of a Segment x Session interaction (F = 8.30, p < .05). t-tests for repeated measures indicated that the significant inter— action consisted of a significantly higher mean IRS rating on the last segment of the tenth session than on either the middle segment of the tenth session (t = 2.10, p < .05) or the last segment of the first session (t = 2.73, p < .02). The analysis of variance is summarized in Table 26. IRS means and standard deviations are shown in Table 25. A t-test for repeated measures indicated that no significant difference existed between mean IRS ratings on the middle and last segments selected from the fifth from final session for those 13 clients, ten seen by high core and three by low core therapists, for whom segments on all three sessions were available (t = 1.01). Therefore, these IRS ratings on the middle and last segments were averaged and analysis of variance for repeated measures was used to 28224. Su La f0 Source \ :‘I‘Jn‘een 85 N ~ High VS. Lc Core There Error 105 Table 24. Summary of 2 x 2 AOV on IRS Ratings of Middle and Last Segments Selected from First Therapy Session for High and Low Core Therapists Source df SS MS F Between §§ 37 19.23 High vs. Low Core Therapists l .61 .61 1.17 N.S. Error 36 18.62 .52 Within Sp 38 Middle vs. Last Segments 1 .03 .03 -- Therapists x Segments 1 1.16 1.16 4.14 N.S. Error 36 9.93 .28 Total 75 |lll|l\|lll‘ uth oHoon HmoH oHosz COHmwwm >Qmu0£8 SuCOB :Cflmmom xuoumnfi umhwk m a. E H 138.82%. SHUU 33.4 3.2% Ca. w : HS .H PU f- a... .U My Uh a... flu -—C.H h— H»: a UK. n. 1.1 «U. at .v c pb—Zquv n... pnhhv m,ur.~.fl,hlv..u.eu. Tnfi-u sW.H-¥€.H u-I.>< Lav no... a\.~:.... ~...-. Fun .7. mu C. 0 fl mm. H... 5 .5 3 d 2 6.8L. 106 1.1.:lilllliull. : l‘flE'u.’ .mGOHmmmm gummy 0cm uwnfim cpon Eoum oHanHm>m 0Ho3 mucoamom pmma pom mappHE 80:3 How mucoHHo om n zo .manmaflm>m ouo3 cofimmom mmmuosp umuflw Eonm mucofimom “mmH ppm mappHE 8033 How mucofiao mm H zm hH.H ov.m mm.H mh.m mm. on.m on. om.m namuoa om. gm.m mm. mo.m mo. vo.m on. om.m h u z oHOO 30H OH.H Hm.m mm.a mm.m mu. mn.m mm. NH.m ma n z ouoo non mm. Hm.m mm. mm.m mm H z mampoe an. gm.m om. mn.~ mH u z onou 30H mm. mm.m mm. mH.m om n z ouoo sOHm .o.m M .o.m m .o.m M .o.m m ummq matte: ummq matte: msouw cowmmom ammumna Spams coflmmom wmmnoza umuflm mumflmmnmne muoo 30A pom swam How mcoflmmmm spams pom “when Eoum pmuooamm mucofimmm so mmuoom mmH wannawm>< mo m..o.m 0cm m.M .mN OHQMB file 26. V 1‘ 63 F) (I? / \ > v‘,_..'—' _, Between 55 Eigh vs. Lo Core Thera; 1 H 1301' ..- I‘ ‘ i. J '16” (I) u “Lot \ ‘ VH; ....1e Vs. Segments First vs. '1 Sessions m,I . ”9331315 ts H :e» theflts Tnerapists SESSiOns Sassions 107 Table 26. Summary of 2 x 2 x 2 AOV on Available IRS Ratings on Segments Selected from the Middle and Last Thirds of the First and Tenth Therapy Sessions for High and Low Core Therapists Source df SS MS F Between Sp 19 35.93 High vs. Low Core Therapists 1 6.14 6.14 3.70 Error 18 29.79 1.66 Within S5 60 47.94 Middle vs. Last Segments 1 1.13 1.13 1.20 First vs. Tenth Sessions 1 1.38 1.38 1.82 Therapists x Segments 1 .03 .03 .07 Therapists x Sessions 1 .96 .96 1.26 Segments x Sessions 1 3.82 3.82 8.30 Therapists x Segments x Sessions 1 1.75 1.75 3.80 Errorl 18 16.91 94 Error2 18 13.69 .76 Error3 18 8.27 .46 Total 79 a p < .05 p ~ 3 827 V soon #5? ‘, .‘vo- v‘ I In J» 'A‘nl .‘U. .2388 Th 1 .J l Orl .th S Crap . UH: Mi 0 S A ~v- ‘- uh“ 108 evaluate the differences in mean IRS ratings on the first, tenth, and fifth from final sessions for these 13 clients. Although the mean IRS ratings showed a progressive increase on each successive session, 2.90, 3.13, and 3.38, respec- tively, this trend was not significant (F = 1.50). The analysis of variance is summarized in Table 27. Table 27. Summary of AOV on IRS Scores Obtained on First, Tenth, and Fifth from Final Sessions for Thera- pists of 13 Clients for Whom Data was Available Source df SS MS F Three Therapy Sessions 2 1.50 .75 1.50 Subjects 12 22.72 1.89 3.78 Sessions x Subjects 24 12.00 .50 Total 38 In summary, IRS scores of neither the high nor the low core therapists differed significantly on the first and tenth sessions, nor did IRS scores of the combined four therapists differ significantly on the first, tenth, and fifth from final sessions. There was no significant differ- ence in IRS ratings on middle and last segments selected from the first, the tenth, or the fifth from final sessions. However, the four therapists as a group had significantly higher IRS ratings on the last segment of the tenth session than on either the middle segment of the tenth session or the last segment of the first session. 95 Edti .38 Q.» ‘28; I V F‘- "-4.L re yeti e e m :3. mi iK .9. "v 5: an: QA~ 31b». pl"?! ".338 V c vv 3'“; ~ pf. «A'fi vu'bvv‘“ .mf“ I .ng n .“b .a a .l . S H D. ..... o. v. 71¢ .AH Nan orb In a: i. u < ?. F‘ .4. q.\ n .. . ~ . s 109 Hypothesis 1; The higher the IRS score of the clients' therapists, the better the clients' outcome on each of the five outcome measures, regardless of the therapists' level of core con- ditions, i.e., with the effect of the core conditions on the outcome measures controlled statistically. Hypothesis II was tested and the results reported separately for each of the five different outcome measures. The Pearson intercorrelations between the five out— come measures and the means and standard deviations of each outcome measure are presented in Table 28 for the 34 cases on which all outcome measures were available. On both the Target and Ineffectiveness measures a lower score indicated a greater degree of improvement. All of the outcome measures were significantly interrelated with the exception of the Discomfort measure, which only tended to be related to Ineffectiveness (r = -.32, p < .10) and had no signifi- cant relationship with Therapist Statement (r = .28). A computerized step-wise regression analysis was employed in testing Hypothesis II (Efroymsen, 1960).2 This analysis was selected because it provided a direct and com- prehensive evaluation of the relational prediction stated in Hypothesis II in which independent sources of variance 2BMDX2R--Stepwise Regression—-Version of January 10, 1966 Health Sciences Computing Facility, UCLA Modified for Texas Tech Computer Center, September 10, 1966 110 . m Hoo v 0 HO V dc . 9 mo v Q OH v om vm.m mm. HH.H mN.H mm.mH .o.m mm.ma mm.~ om.m om.m OH.O coo: . mm m me u H B QOO.I pHm.| quaoumvm umfimmuone No.1 mm.| mm. ucoamumum n p p ucoflumm 6mm.u nmg.u mm. nHo. showeoomHo mmoco>wuoommocH Mommas usoEmuMHm ucoEoumum “H0mfioomflo umHooumse HooHpoo mommo vm How monommoz mEoouso 0>Hm Hem m.H Ugo .m..o.m .m.x .mm manme associated w- The extent t1 iicted each < seams of a $1: leasnre. The ii the resul are sumarize azero order predictor val :mm} the h: selected out 3'3 Signific iCrrelatiOnS “I grediCtor Effect Of th -:e Variable 111 associated with each outcome measure could be determined. The extent to which each of the 12 predictor variables pre- dicted each of the five outcome measures was determined by means of a separate run of the program for each outcome measure. The operations of the step—wise regression program and the resulting tests of significance which can be made are summarized as follows. A Pearson r, also referred to as a zero order r, is first computed between each of the 12 predictor variables and the outcome measure. The variable having the highest zero order r with the outcome measure is selected out and entered into a regression equation, and the significance of the zero order r is tested. Partial correlations are then computed between each of the remaining 11 predictor variables and the outcome measure with the effect of the variance of the first selected variable, i.e., the variable which had the highest zero order r with the outcome measure, partialed out. The second variable, the variable having the highest partial correlation with the outcome measure, is then selected out and entered into the multiple regression equation together with the first selected variable, and the partial regression coefficient for each of the two variables is tested for significance with a t-test, with df equal to the number of subjects, minus the number of predictor variables, minus 1. The sig— nificance of the partial r is tested using an F ratio for which the df is reduced by one for each variable held con- stant. A multiple R is computed between the first two :relictor va‘ and the obtai Fixer variat e outcome I.“ L V ,4 selected vari :ance of this the multiple 1 selected and regression co' apation, and selected vari L's-e outcome it related Val-is xltiple reg: SinCe 112 predictor variables selected out and the outcome measure, and the obtained R is tested for significance. The procedure of selecting from the remaining pre- dictor variables the one having the highest partial r with the outcome measure while the effect of all previously selected variables are held constant, testing the signifi— cance of this partial r, entering the selected variable into the multiple regression equation with all those previously selected and testing the significance of the partial regression coefficient for each of the variables in the equation, and finally, computing a multiple R for the selected variable and all those previously selected with the outcome measure, is repeated until all the orthogonally related variables have been selected for inclusion in the multiple regression equation. Since the large number of missing segments made the use of ratings on the individual segments for predicting outcome measures unfeasible, each client's scores, i.e., the average of his ratings, were used in all analyses of the outcome measures. In addition, each of the step-wise regression analyses computed on the total sample of cases employed an n of 40, although four of the outcome measures were available on only 34 clients. Since the six clients who did not return for the post-therapy collection of out- come data could not be assumed to be randomly selected from the entire sample of clients, the mean of the outcome scores of the 34 clients who returned was assigned as the outcome values for t“ Hold the (303 :esulted £1507 :ttcone SCOIE Ht did pemi these six 011 the six clien evaluation, f« :3 were rate :2 inprovemen :lients were 1: having sli Table 55:11 of the ZItCSTie 111933 533388 (IAS) iv-‘IES held 81, fi ‘EHG Sc .: ‘ .34 I I '3 In 113 values for these six clients. This was done in order to avoid the possible biasing of the sample which might have resulted from the exclusion of these cases. Use of the outcome scores did not mathematically alter the outcome data but did permit the use of all the available data related to these six clients, e.g., IRS and core condition scores. Of the six clients who did not return for the post-therapy evaluation, four were seen by low core condition therapists and were rated by their therapists as either having shown no improvement or deterioration, and the remaining two clients were seen by one high core therapist and were rated as having slightly and markedly improved, respectively. Table 29 presents the Pearson zero order r's between each of the 12 predictor variables and each of the five outcome measures for the 40 cases. Discomfort Outcome Measure. Only Initial Adjustment scores (IAS) (r = .50, p < .01), and IRS scores with IAS scores held constant (r = —.31, p < .05), significantly predicted Discomfort scores as each predictor variable was in turn entered into the multiple regression equation. Moreover, the partial r's between IRS and Discomfort scores, with EEWG scores (r = -.31, p < .05) and EEG scores (r = -.34, p < .05) separately held constant, indicated that IRS scores significantly predicted Discomfort scores with the effect of the core condition scores, as measured by either EEWG or EEG scores, controlled statistically. 1. HH. v0.1 mo. emf mu WKH mmmeOKVfiUUCLMUCH UEWHN'H. UCWEQUGUU... UCHWEQUNUMW UHOMEOUWflQ Umufllpwhozpb UEUflUMnM A CG. "- ZV Enfohgnucmiz pro-thin - upsv 8v) u, 1i shut .u 3 3.9.9 ~64: d an F\\/ ,4an .1. ~ 7.9-1...‘ .1! - u av 14 114 m HOO. v HO. v om mo. v an OO. v oo mo. OH. OH.- OH.- OO. HBHO\ee¢ ON. nmm. NN.- OO.- NO. Hm2\Hm NO. NO. OH.- OO.- NO.- xo OH nmm OOH . Om . ON . Hmm OO.- ON.- Oom. OH. NH. .mmom .oz OH. NO. NO. OH. oOO. OOH ON.- ON.- ON.- OO.- HN. 2 NH. NO.- OO. NH. OO.- 0 OO. OH.- OH. omm. mo. m NO. NH.- OH. RN. OO. om“ HN - om I OO Ham HON ozmw OO. HH. eO.- OO. oHO.- OOH mmocm>fluoommosH ummume pcoEoumum ucofimumum unomEoomHQ umHooumze H:6Huoo AO¢ H Zv mmhfimmmz GEOUHDO 0>Hh Sums» meQMMHMNV HO¥UfiUOHm NH m0 WTH .mN wHQmB The I 331.9 the :81 not. sequently, ti“. Zisconfort on therapy 1118186 1.35 score. listed Disco: ride in Hypo ere associa Hypothesis I :easure. See findings rei Lap 3119 leer.) 21* Scores P‘l’edicted P L‘-“‘v9rovenem leasure ' t1 is higher :alS II Wd Tl" \ Stems (r held COnst «a: . “Ctlons c ‘n 5Vr e a C11 115 The relationship of Discomfort with IAS was positive while the relationship with IRS scores was negative. Con- sequently, the more improvement a client showed on the Discomfort outcome measure, the greater his degree of pre- therapy maladjustment on IAS and the lower his therapist's IRS score. Thus, although IRS scores significantly pre- dicted Discomfort scores, in contradiction to the prediction made in Hypothesis II, lower rather than higher IRS scores were associated with client improvement; therefore, Hypothesis II was rejected for the Discomfort outcome measure. See Table 30 for a summary of the significant findings reported for each outcome measure. Patient Statement Outcome Measure. Only Relation- ship (REL) scores (r = -.39, p < .05), and EEWG scores with REL scores held constant (r = .46, p < .01), significantly predicted Patient Statement scores, indicating that the more improvement a client showed on the Patient Statement outcome measure, the less difficult his therapist rated establishing and maintaing a satisfactory therapeutic relationship and the higher his therapist's EEWG score. Therefore, Hypoth— esis II was rejected for this outcome measure. Therapist Statement Outcome Measure. Only REL scores (r = -.70, p < .01), and DX scores with REL scores held constant (r = -.38, p < .05), made significant pre- dictions of Therapist Statement scores, indicating that the more a client improved on the Therapist Statement outcome iEFi n.» F»? GWFLUHUCH N HUME—Ain‘tvalufl.» miujflutwflfl Err-A unwv .1? n paiv OECUuSC 0>HuoNSESU MO I LUCQN Nov .HOUU flcmwhnh mam>fizc< 1,9 01! Iann Alto -, Melvin n.v..l~(-.~ HOUUflUUHQ nhOfiuflIHflVHmvmqvVN i‘..h1‘-§v~.n¥c mMVMud k: \0 3 [AWEV U a n -.v.::- Wu ghzmmmz WEOUUJC RU 'V N: A~ le~t 116 Ho.vm Q mo v on ocoz mmoco>fluoommocH mama OOOH. OOON. omm. oNO.- .mmom .oz ummuma ocoz mmma. mmma. Mom. owm.l OBMN Hmm NNNO. ONOO. OHON. omm.- so ouotouoom oooz OOOO. OOOO. oON. oON.- Hmm umHomHooe Hmm NOOH Ommm oOO oOO ozmw Homeoooom oooz NOOH. NOOH. oOO. oOO.- Hmm HooHooo mNH HONO. OHNO. on. mHm.- OOH QHOmEoomHQ oooz NOON. NOON. oom. oOO. OOH “CmngOU NM CH NM mEOUHSO ®EOUHSO w GHQMHHMNV stmmmz pamm mmmouocH Qufl3 mnouoflpwnm HouONUoum HOQOOUOMm oEooQSO moaQoOHm> o>wumadfidu mo m goo3pom H mmm>HMCN cowmmmumom omwBImoum Eoum pocflmqu oHSmmoz oEOOQDO Qoom m0 wcofluoflpoum “snowmwsmflm mo mumaasm .om mHQma eeasure, the 32.5 maintain ire lower th was rejected Lari P" :05) , am With EEWG sc Caitly predic :‘HCVement E ire higher t] client's numf was relEC IDef \ tie higheSt ECCreS’ but EST-.9 0E the lie SyStEma. ~=~~ .N‘Ve. "'1 J. "18' H_Come Thea The .tCOIne mea “:0 preSen . each Ont ..:Orfie me ‘V‘lctor the . ‘ Si .. 9n 117 measure, the less difficult his therapist rated establishing and maintaining a satisfactory therapeutic relationship and the lower the client's DX score. Therefore, Hypothesis II was rejected for this outcome measure. Target Outcome Measure. Only EEWG scores (r = -.36, p < .05), and the Number of Therapy Sessions (No. Sess.) with EEWG scores held constant (r = -.42, p < .01), signifi- cantly predicted Target scores, indicating that the more improvement shown by clients on the Target outcome measure, the higher the therapist's EEWG score and the greater the client's number of therapy sessions. Therefore, Hypothesis II was rejected for this outcome measure. Ineffectiveness Outcome Measure. Warmth scores had the highest zero order r (r = .27) with Ineffectiveness scores, but this r did not reach significance; moreover, none of the predictions of Ineffectiveness scores made by the systematic selection of each variable reached signifi- cance. Therefore, Hypothesis II was rejected for this outcome measure. The significant findings reported for each of the outcome measures are summarized in Table 30. This table also presents the R for the significant predictor variables of each outcome measure as well as R2, the percent of outcome measure variance accounted for by the best single predictor and the percent accounted for by the combination of the significant predictor variables. The "Increase in 2" R column shows the amount of outcome score variance $9 addition 1.3.1 the if uriame rest ables over it Ear example, ration of IA. . variance of t for 24.87% of addition of J aiéitional 7 In 5' are-dict only I: the predi iteraoists i "Esra-pists ‘. this out iijeeted fc 118 accounted for by the initial or best single predictor and the amount of additional outcome variance accounted for by the addition of a second predictor to the initial predictor, i.e., the increase in the amount of predicted outcome variance resulting from the combination of predictor vari- ables over that predicted by the initial variable alone. For example, reference to Table 30 shows that the combi- nation of IAS and IRS scores accounted for 32.18% of the variance of the Discomfort scores, that IAS alone accounted for 24.87% of the Discomfort score variance, and that the addition of IRS scores to IAS scores accounted for an additional 7.31% of Discomfort score variance. In summary, IRS scores were found to significantly predict only one outcome measure, Discomfort, and contrary to the prediction made in Hypothesis II, clients whose therapists had lpwgp IRS scores rather than those whose therapists had higher IRS scores, showed greater improvement on this outcome measure. Consequently, Hypothesis II was rejected for each of the five outcome measures. Five of the eleven predictor variables, excluding the IRS variable, made significant predictions for either one or two of the outcome measures. While the IRS variable significantly predicted only the one outcome measure, the well-established variable of the core conditions, as reflected in the EEWG scores, significantly predicted only two of the outcome measures, Patient Statement and Target, with higher core condition scores associated with more 01.7 e i at, l U. . a '3" he. .Qf“ K.‘ 1 0 5e 11' #ch“ o QRA .éD‘va ‘ . a sent a-“ no. .wdry v‘l' . 1.1.: ‘A 0‘ 'V .‘ v‘dlo :Ijts 119 improvement on both outcome measures. The Relationship variable also significantly predicted two of the outcome measures, Patient and Therapist Statements, with greater ease in establishing and maintaining a satisfactory rela- tionship associated with more improvement on both measures. In addition, the Initial Adjustment variable predicted the Discomfort outcome measure, with better pre—therapy adjust- ment associated with lesser degrees of improvement. A greater number of therapy sessions was associated with improvement on the Target measure. And contrary to expec— tation, while the DX variable significantly predicted one measure, Therapist Statement, lower rather than higher DX scores were associated with more improvement. Outcome Measures ag a Joint Function pf IRS and EEG Scores Although no hypothesis was stated regarding the effect that the core condition context in which IRS scores occurred may have on the prediction of client outcome from IRS scores, the earlier findings that high and low core therapists differed significantly on IRS scores and ratings suggested that IRS scores in differing core condition con— texts may be differentially related to outcome scores. Consequently, the extent to which IRS responses simulta- neously accompanied by low core conditions and IRS responses accompanied by high core conditions predicted each of the outcome measures was investigated. The 40 cases were divided into 19 high core condition cases and 21 low core 120 condition cases, using the mean EEG score of the 40 cases as the dividing score. Table 31 presents the means and standard deviations of the predictor variables and the outcome measures for the high and low EEG groups, and Table 32 presents the Pearson intercorrelations among the predictor variables within the high and low EEG groups. Step-wise regression analyses were again used to determine the extent to which each of the predictor vari— ables predicted each of the five outcome measures within the high and the low EEG groups separately. In each of these ten regression analyses one additional predictor variable, the Four Individual Therapists, which consisted of the scores on each of the 12 predictor variables obtained by each therapist and his respective set of ten clients, was included along with the previous 12 predictors. Since the high and low EEG groups were classified only on the basis of EEG scores, these groups were not balanced for Role Induction or Patient Attractiveness conditions or for the individual therapists of the clients composing each group. Consequently, by including the scores associated with each of the four therapists as a variable in predicting outcome scores in the step-wise regression analyses, it could be determined whether or not there was a significant difference in the predictions of any of the outcome measures for either the high or low EEG group associated with the individual therapists. The finding of a significant prediction made by the Four Therapist variable would indicate a significant 121 Table 31. X's and S.D.'s of Predictor Variables and Outcome Measures for the High and Low EEG Groups 19 High EEG Cases 21 Low EEG Cases 2 S.D. SE S.D. Discomfort 6.81 17.69 7.43 16.70 Patient Statement 4.05 .97 3.18 1.34 Therapist Statement 3.42 1.12 3.00 1.10 Target 2.22 .88 2.29 1.06 Ineffectiveness 20.84 9.06 18.05 8.02 IAS 33.89 17.54 24.19 15.87 EEG 1.58 1.06 -1.37 1.17 EEWG 1.51 1.56 -.44 1.99 No. Sess. 12.37 4.41 12.90 4.17 REL 2.79 1.44 3.00 1.30 E 4.56 .30 4.03 .22 G 3.53 .16 3.21 .18 W 3.87 .36 3.93 .26 IRS 3.18 .73 2.93 .43 DX 4.95 .79 4.82 .54 4th's 2.63 1.26 2.38 1.02 RI/NRI 6.16 1.01 5.86 1.01 ATT/UATT 3.37 .50 3.43 .51 122 Hmz\Hm m.aoO x9 mmH 3 U m mm. 4mm OH.- OOO.- OOO.- .moom .62 OO. OO. OO. OH.- oOO.- Oaaw NN. HH. OH. Om. OH.- Om. UOO. OOH OO. ONO.- OH. OH. HN.- HO. OOO. NO.- .OOO. OOH Hz 04 Hm GA am 04 H3 04 Hm 4mm .mmmm .02 03mm Qmw mmsouu omN BOA paw Qwfim mQu pom mmHQMHHm> “Ouofipmum waoa< m.u comumom .Nm mHQOH fliruv. HOO. v O OO. v a 123 p Q Ho. v no OH. v om OO. HO.1 HOz\HO OO. OO.- OO.- OH. o.aoO HO. OO.- OO. OO.: OO.- oOO.. HO OH.1 HN.I NH. No. NH. QNm. MN. mo.1 mmH HO. OH.1 OO. NH. ONO.1 OOO.1 OH.1 OO.- OH.1 ON.1 3 mo. HN. Ho.1 Nm. Qom.1 mo. oN.1 mH.1 Hm.1 NH.1 oN. oN.1 0 mo.1 mN. mom.1 mo.1 Nm.1 NH. No.1 mNo.1 MH.1 Nm. mN. oN. m mN. one. mom. oH. HH.1 mm.1 oH.1 oN.1 QHm. NH.1 «H. mm. Hmm NN.1 NN.1 om.1 wH.1 NH.1 moo. mo.1 MN. qu.1 mN. NH.1 omN.1 .mmom .02 NH. NH. mN.1 oN. oH.1 NN.1 No.1 mm.1 Ho. mo.1 Qom. UON. 03mm Ho.1 om. ON.1 MH. Qom.1 oH. HN.1 mqq.1 oN.1 oN. ON. No.1 omN Ho.1 Nm. mo.1 MN. oN.1 oo.1 mm. mm.1 no.1 no.1 oN. oH. wHuummmocH umHamquH ucoHumm manouo omw 30H vow QMHm oQu :HQqu whammoz oaouuoo.:omm Odo oHQmHum> HouoHooum comm aoo3uom w.» umouo oumN .mm OHQOH t:- . \ v 'V i~ ‘_‘ ‘~ in ’\ .6 ‘v 126 Discomfort scores held constant, IAS and EEWG scores, IRS scores also then had a significant partial r with Discomfort scores (r = -.50, p < .05). Thus, although a previous find- ing indicated that IRS scores significantly predicted Discomfort scores within the entire sample of 40 cases, these results of the analyses of the high and low EEG groups indicated that the relationship between IRS and Discomfort scores was much stronger in the low EEG group than in the high EEG group. Within the low core group on the Target outcome measure, IRS scores had a zero order r with Target scores of .37 (p < .10), and with the effect of the best predictor of Target scores held constant, EEWG scores, the IRS scores was the only other variable that significantly pm edicted Target scores (r = .48, p < .05). Within the low core group on the Therapist Statement outcome measure, IRS scores had a significant zero order r with Therapist State- ment scores (r = -.62, p < .01), and with the effect of the best overall predictor held constant, Relationship scores, IRS scores then had a partial r of -.42 (p < .10) with Therapist Statement scores. The significant findings of each predictor variable for each outcome measure obtained on the step-wise regression analyses of the high and low core groups are summarized in Table 34. Thus, IRS scores made a significant prediction of outcome scores for three of the five outcome measures for the low EEG group; moreover, these predictions were also significant when the effect of the core conditions, measured 7. n4 1. OBOO .HOz\HO .OOOO .oz . mmHQmHum> ommmuosH N QOH3 muouoHpmum pom acuoHpoum acuoHOoum . 0>HOOH5850 mo m coo3umm u mmsouo omN 30H Odo Qme onu mo mommHmc< GOHOmouwom omH31moum Baum omaHmqu ouswmoz oaoouao Qomm mo maOHuoHpoum unmonHame mo Opossum .Om OHQOH 128 HO. v ao . a mo v O OH. v a... oaouo msoz OomN. OOmN. omq. Qmo.1 ozmw omw 30H mmmao>HuoowwmaH 03mm oNNH. ONOm. omN. qu. mmH 0coz mmmc. mmmq. OOO. 0OO.1 03mm omw 30H umwume usmumaoo pHom NM aH m mououm maoouso whammoz oaouuao OHQmHum> mmHQmHum> wmmmuoaH N nuH3 muouoHpmum paw HOOUHOOHm u0uoHooum m>HumHsaau mo m o003umm u A.o.uoooV Om OHooe 129 by both EEWG and EEG scores, was controlled statistically. On the Discomfort measure, significant partial r's between IRS and Discomfort scores were obtained with EEWG scores (r = -.57, p < .01) and with EEG scores (r = —.46, p < .05) separately held constant. On the Therapist Statement out- come measure, significant partial r's between IRS and Therapist Statement scores were obtained with EEWG scores (r = -.62, p < .01) and with EEG scores (r = -.65, p < .01) separately held constant. On the Target outcome measure, a significant partial r between IRS and Target scores was obtained with EEWG scores (r = .48, p < .05) held constant but not with EEG scores held constant (r = .29). However, this failure of IRS scores to significantly predict Target scores with EEG scores held constant represented a small and nonsignificant decrease of 5.25% (F = 2.31) (Baggaley, 1964) in the Target score variance predicted by IRS scores; con- sequently, the IRS prediction made with EEG scores held constant did not differ significantly from that made with the effect of EEG scores included. Thus, the significant predictions of the three outcome measures for the low core group made by IRS scores were not attributable to the rela- tionship of IRS scores with core condition scores. These findings indicated that knowledge of the core condition context of IRS scores when that context was high, i.e., an average EEG Z score of 1.58 or raw score of 8.08, was of little or no use in predicting any of the outcome measures, whereas knowledge of the core condition context 130 of IRS scores when that context was low, i.e., an average EEG Z score of -l.37 or raw score of 7.24, was useful in significantly predicting three of the five outcome measures. The relationship between IRS and EEG scores within the high core group was positive (r = .20) while the relationship within the low core group was negative (r = -.29). Since within the low core group higher, rather than lower, IRS scores were associated with both less client improvement and lower therapist EEG scores, this suggested that greater client improvement was associated with lower IRS responses in a lower core condition context. A more precise description of the relationship between the IRS scores in their EEG score context and the three outcome measures for which IRS scores made significant predictions was obtained by investigation of client outcome scores as a joint function of their therapist's IRS and EEG scores. The IRS scores and the EEG scores of the 34 cases for which all of the outcome data was available were each divided into equal thirds, i.e., into high, moderate, and low levels of both IRS and EEG scores, and the Discomfort, Target, and Therapist Statement outcome scores of each client were classified as a joint function of his thera- pist's high, moderate, or low level on both IRS and EEG scores. Table 35 shows the means and standard deviations of the IRS, EEG, and outcome scores for these groups of clients. The EEG scores in Table 35 were transformed into 131 OO. NO.N ON. OO.N OO. OO.N OH. OO.O OOH OO. ON.N ON. NN.N HO. OO.N HO. OH.O OOO ON.H O0.0 OO.H ON.O ON.H O0.0 HN. OO.O .OO .OO OO.H O0.0 OO. ON.O HN.H O0.0 HN. O0.0 .OO .OO O0.0 ON.OH NO.HH ON.OH O0.0 O0.0H O0.0 OO.ON oaHOOOOOoaH OO.H OO.H HO.H OO.N OO. ON.H OO. OO.N OOOOOH HN.OH O0.0 OO.OH ON.O OO.O O0.0 OH.ON OO.ON oooanoOHo mmH mumuwooz ON. N0.0 ON. O0.0 OO. O0.0 NH. O0.0 OOH OO. NO.N ON. HH.N ON. NN.N ON. ON.O OOO NH.H O0.0 OO. OO.N OH.H O0.0 OO.H OO.O .om .oO NH.H O0.0 OO. OO.N OO.H O0.0 HN. O0.0 .OO .OO NO.N H0.0N HO.N O0.0N O0.0H N0.0H N0.0 ON.NN ooHoooOoooH OO. OO.N OO. O0.0 OO.H OO.N OO. OO.N OOOOOO ON.OH NN. H0.0 N0.0 1 H0.0 OO. OH.NN OO.H oOoanoOHO OOH OOHO .O.O N .O.O m .O.O x .o.O m OOO oumuoooz OOO OOHO mmuoom omw paw mmH mo mHo>OH 30H no .muwumpoz .Qme ou waHpuoou< OOHMHmmmHo mmmoo On How mmuoom oaouuoo paw .omN .mmH mo m..o.m Odo m.m .mm mHQmH 132 muGOHHu mo Hman: n zm OO. NO.N NO. OO.N OO. OO.N OO. HH.O OOH OO. OO.N ON. OO.N NO. OO.N NN. ON.O OOO OO.H O0.0 OO.H O0.0 OO.H N0.0 OO.H ON.O .OO .oO OO.H N0.0 OO. O0.0 OO.H OH.O OO. O0.0 .OO .ae OH.OH O0.0N NH.NH ON.ON OO.HH O0.0H N0.0 O0.0H o>HoomOOooH OO.H ON.N NO.H NO.N HO.H HO.H ON. ON.N OOOOOO O0.0H OO.N O0.0H NN.O OO.NH OO.N NO.HN O0.0 oOoanOOHO OO u z HH u OH u 2 OH 1 z Hmooe OO. OO.N OO. OO.N OO. ON.N NO. OO.N OOH NO. OO.N ON. O0.0 OH. OO.N NH. OH.O OOO OO.H O0.0 NO.H O0.0 OO.N ON.O OO.H O0.0 .OO .oO OH.H NN.O NO. O0.0 HN.H ON.N OO. N0.0 .OO .oe NO.NH OH.ON H0.0N ON.ON O0.0H O0.0N NO.HH O0.0H oaHoomOOooH ON.H OH.N OO.H OO.N OO.H OO.N HN.H OO.H OOOOOH NO.NN OO.HH N0.0N O0.0H O0.00 ON.OH O0.0H N0.0 uooanOOHo HH a z O n O u z 11 m n z OOH soH .O.O m .O.O m .O.O m .O.O m HOuoO OOO 33 OOO oomuoooz OOO OOHO N.o.oooov OO mHome 133 raw scores so that the levels of EEG may be more readily interpreted. Since the outcome scores of each client were clas- sified solely on the basis of the joint levels of his therapist's IRS and EEG scores, the resulting groups of clients were only partially balanced for the Patient Attrac- tiveness and Role Induction conditions. Consequently, before attempting to evaluate differences in outcome scores as a function of the joint levels of IRS and EEG scores, it was necessary to determine whether the Role Induction or Patient Attractiveness conditions significantly altered or affected the relationship between either IRS scores or EEG scores and each of the outcome measure scores. Therefore, R's were computed between each of the outcome measure scores and Role Induction conditions and IRS scores. F tests were used to assess whether the correlation of outcome scores with the combination of IRS and Role Induction scores (and the combination of IRS and Patient Attractiveness scores) differed significantly from the correlation of outcome scores with IRS or Role Induction scores alone (and with IRS or Patient Attractiveness scores alone). This same procedure was also carried out with EEG scores. For example, IRS and Discomfort scores had an r of —.31, while Role Induction and Discomfort scores had an r of .07; the R of .33 between IRS, Role Induction, and Discomfort scores reflected a nonsignificant increase of 1.28% (F = .42) in the amount of Discomfort score variance accounted for by the 134 three sets of scores over that accounted for by the IRS scores alone. Thus, the correlation of IRS and Discomfort scores was not significantly affected by the Role Induction/ No Role Induction condition. The results indicated that the r's between IRS and outcome scores were not significantly altered by either the Role Induction conditions or by the Patient Attractiveness conditions, respectively, for any of the outcome measures: Discomfort (F = .42; F = .42), Therapist Statement (F = 0.0% F = 0.00), Patient Statement (F = .07; F = .38), Target (F = 0.00; F = .78), and Ineffectiveness (F = .40; F = .19). Similarly, the r's between EEG and outcome scores were not significantly altered by the Role Induction conditions or by the Patient Attractiveness conditions, reSpectively, on any of the outcome measures: Discomfort (F = .04; F = .05), Therapist Statement (F = .96; F = .77), Patient Statement (F = .40; F 1.23), Target (F = .85; F = .39), and Ineffec- tiveness (F .40; F = .19). Since neither the relationship of IRS scores with each outcome measure nor the relationship of EEG scores with each outcome measure was significantly altered by either the Role Induction or Patient Attractive- ness conditions, analyses of the outcome scores as a joint function of the levels of IRS and EEG scores were carried out even though the cells were not completely balanced for these two conditions. Table 35, based on the 34 cases, reveals a similar pattern of client improvement for each of the three outcome 135 measures: the least amount of improvement, or actual dete- rioration in the instance of the Discomfort measure, occurred within the group of clients that received high levels of IRS scores in the context of low levels of EEG scores. Within the context of low EEG scores, high IRS scores were associated with less improvement than moderate and low IRS scores on all three outcome measures, and this trend reached at least the .10 level of significance on two of the outcome measures: the group of clients whose thera- pists had high IRS scores in a context of low EEG scores had a higher mean Target score (t = 2.04, p < .10), representing less improvement, and a lower mean Therapist Statement score (t = 2.27, p < .05) than the group of clients whose thera- pists had moderate or low IRS scores in the context of low EEG scores. In contrast, the groups of clients who received high, moderate, or low levels of IRS scores in contexts of either high or moderate EEG contexts did not differ signifi— cantly on improvement on any of the outcome measures. It can be seen from Table 35 that high level IRS scores were associated with differing degrees of client improvement depending upon the high, moderate, or low level of the accompanying EEG scores, and to a lesser extent, that the amount of improvement associated with moderate IRS scores was also affected by the EEG context, whereas low level IRS scores were relatively unaffected by the EEG context. Differences in outcome scores between the high, moderate, and low EEG score contexts of high IRS scores were 136 significant for both the Target and Therapist Statement outcome measures: the mean Therapist Statement score was significantly higher for the group of clients whose thera— pists had high levels of IRS scores in the context of high EEG scores than for the clients that had high IRS scores in the context of low EEG scores (t = 3.43, p < .02). Mean Target scores were significantly higher, representing less improvement, for the group of clients whose therapists had high IRS scores in the context of low EEG scores than for the group which had high IRS scores in the contexts of either high (t = 4.78, p < .01) or moderate (t = 3.05, p < .05) EEG scores. Moderate levels of IRS scores in the contexts of high and moderate EEG scores tended to be asso- ciated with differential amounts of improvement on both the Target and Discomfort outcome measures: for moderate levels of IRS scores, the context of moderate EEG scores was asso- ciated with greater improvement on the Target measure (t = 1.94, p < .10) but with less improvement on the Discomfort outcome measure (t = 2.09, p < .10) than the context of high EEG scores. In addition to the three outcome measures for which IRS scores made significant predictions within the low core group, means and standard deviations of the Ineffectiveness and Patient Statement outcome scores of clients classified according to their therapist's levels of IRS and EEG scores are also shown in Table 35 for comparative purposes, although no significant predictions were obtained from IRS 137 scores for either of these two outcome measures. Reference to Table 35 indicates that the pattern of Patient Statement scores was similar to that of the Target and Therapist Statement scores, but the only significant difference in improvement on this measure between the different EEG con- texts of IRS scores consisted of significantly greater improvement for the group of clients who received high IRS scores in a high EEG context rather than in a low EEG con— text (t 2.61, p < .05). On the Ineffectiveness measure one quite interesting trend occurred which differed from that of the other four outcome measures. The group of clients that received high and moderate IRS scores in a moderate EEG context was judged after termination of therapy as being the most effective in interpersonal rela- tionships, while the group that received low IRS scores in a moderate EEG context was judged as being the least effective; however, the difference between the clients who had therapists with high and moderate IRS scores and those who had therapists with low IRS scores in the moderate EEG context was only suggestive since the obtained t of 1.72 reached only the .11 level of significance. Since the Ineffectiveness measure was the only outcome measure which specifically attempted to assess the quality of the client's interpersonal relationships, this finding very tentatively suggested that dealing with the therapeutic relationship in a manner reflected by the higher stages of the IRS scale may be related to the client's effectiveness in his 138 significant relationships after therapy. Subsequent research, however, is needed to assess the possibility that the existence of such a trend simply reflects a differ- ential pretherapy predisposition for the most effective and least effective groups of clients to explore the therapeutic relationship. DISCUSSION The present research has demonstrated the IRS scale to be a reliable instrument capable of making significant and meaningful discriminations among therapists. Moreover, each of the individual stages of the scale, with the exception of stage 5, proved to be useful in discriminating among therapists on at least one of the following variables: core conditions, orientation, experience, DX, patient status. Discriminations on the greatest number of vari- ables occurred at stages 3 and 6, while the fewest occurred at stage 1. In addition to significant differences obtained in the proportion of IRS responses at individual stages, trends consistent with the predictions frequently occurred. For example, at stage 2 the percent of IRS ratings obtained by the core quartile groups was 22.9, 30.0, 35.7, and 51.4, respectively: only the difference between the first and fourth quartiles reached significance, but the progressive increase through the four quartiles was consistent with Hypothesis I. Since the IRS scale was used for the first time in the present research, detailed description and analyses of differences in IRS ratings at each stage for the high and low IRS therapists and IRS quartiles in each study are 139 140 presented in Appendix 0. Systematic differentiations among therapists in the IRS quartiles were made at each stage, with the exceptions of stage 5 in both studies and stage 1 in Study II. The discriminations at each stage formed a consistent pattern in which successively higher stages of the scale differentiated between therapists with succes- sively higher IRS scores. Although this pattern was more pronounced in Study I where there was a greater variability and a greater number of IRS ratings, it was also clearly evident in Study II. Thus, for example in Study I, stage 6 differentiated the 14 therapists with the highest IRS scores from the remaining therapists, while in direct con- trast, stage 1 differentiated the 14 therapists with the lowest IRS scores from the remaining therapists. The failure of stage 1 to differentiate among dif- ferent levels of IRS scores in Study II appears to be a result of the very few responses, 3.85%, which occurred at this stage. No discriminations were made at stage 5 in either study, and only six of the entire ratings made in both studies occurred at this stage. Either the discrimi- nations required in making a rating at stage 5 were too difficult or the therapists simply did not make stage 5 responses. Since this stage was not useful in either study, it appears that stage 5 should be eliminated from the scale in subsequent research. All hypotheses in Study I were confirmed with the exception of Hypothesis IVB, which predicted differences 141 among Orientations. However, all three findings used to evaluate Hypothesis IVB were in the predicted direction, with one finding reaching the .05 level of significance and the remaining two reaching the .10 level for two-tailed tests. Thus, for the total sample of therapists in Study I IRS scores were: 1) positively related to both core con- dition scores and DX scores; 2) did not differentiate between inpatient and outpatient therapists or between high and low experience level therapists within any of the four orientations, 3) ordered therapists according to Orientation from higher to lower as Relationship, Eclectic, Client- Centered, and Analytic, respectively; and 4) differentiated Relationship therapists from Analytic therapists and tended also to differentiate them from Eclectic and Client-Centered therapists. Since the statistical tests used to evaluate dif- ferences in IRS ratings yielded findings which were approxi— mations, only the results of statistical tests used to evaluate differences in IRS scores were used in testing the hypotheses. However, a great deal of information that was not available from the IRS scores was provided by the IRS ratings, and this additional information was considered to justify their conservative and cautious interpretation. In addition to eliminating the regression toward the mean which occurred for the IRS scores, the IRS ratings were valuable in determining the more precise nature of the differences between therapists who were found to differ on IRS scores by 142 indicating the distribution of responses among all the stages and the particular stages at which therapists differed. For example, although therapists of in- and out- patients were not differentiated by IRS scores, IRS ratings revealed that therapists who saw outpatients gave more stage 3 responses, constituting almost half of all their responses, than inpatient therapists. Moreover, the distribution of IRS ratings indicated that unlike the high IRS therapists who did not respond differentially with in- and out- patients, the 1ow IRS therapists, who almost never focused on the relationship and gave only 2% of their IRS responses at the combined stages 4, 5, and 6, avoided references made to themselves by inpatients to a greater extent than those made by outpatients. Hypothesis I in each study, which predicted a positive relationship between IRS and core condition scores within the total sample of therapists, was confirmed in Study I but rejected in Study II. However, classification of therapists according to relatively high and low function- ing on the core conditions revealed dramatically different relationships between IRS and core scores within these groups in both studies and clarified the failure to find a relationship in Study II. The significant positive rela- tionship between IRS and empathy for the high core thera- pists and the significant negative relationship between IRS and genuineness for the low core therapists was obscured or 143 masked when these two homogeneous groups were combined in the total sample. Moreover, IRS scores as well as the com— bined ratings at stages 4, 5, and 6 were also significantly higher for the high functioning than for the low functioning therapists. Consequently, although hypothesis I was rejected in Study II, a definite pattern of relationship was nevertheless clearly demonstrated in which the two high functioning therapists, who on the average were relatively more empathic and genuine, focused on the relationship to a greater extent in those interactions with clients in which they were more empathic; in contrast, the two low function- ing therapists focused on the relationship more with those clients with whom they were less genuine. Similarly, classification of therapists into rela- tively high and low functioning therapists revealed striking differences in Study I. High functioning therapists focused on the relationship significantly more often and avoided client's direct and indirect references to themselves sig- nificantly less often than the low functioning therapists. Moreover, the more empathic, warm and genuine the high functioning therapists were, the greater the extent to which they focused on the relationship, but within the low func- tioning group of therapists there was no correspondence between their empathy, warmth, and genuineness and the extent to which they focused on or avoided the relationship. Thus, in both studies the high functioning thera- pists related client statements to themselves to a - 144 significantly greater extent and avoided client references to a lesser extent than the low functioning therapists. However, these differences, particularly the extent to which client references were avoided, were greater in Study I than in Study II and reflected the fact that both high and low functioning therapists in Study II gave approximately half as many responses at stages 1 and 2 combined but gave roughly twice as many at stage 3 than those in Study I. In both studies slightly more than a fourth of the responses of the high functioning therapists related client statements to themselves, while fewer than 10% of the responses of the low functioning therapists related client statements to themselves. Moreover, therapists in Study II consistently offered higher and less variable levels of the core condi- tions than those in Study I, but this differential function- ing on the core conditions was more pronounced for the low functioning than for the high functioning therapists. Thus, the difference in therapist-offered conditions between the high and low functioning therapists was considerably greater in Study I than in Study II. In Study I the high function- ing therapists frequently communicated understanding of the client's surface feelings, although they failed to under- stand most of his deeper feelings, and related to the client in an ingenuine or "professional" manner only infrequently, while the low functioning therapists communicated no under- standing of the client's deeper feelings and frequently 145 failed to understand even his surface feelings, and most of the time related to the client in a rigid, nonspontaneous, professional manner. In Study II, in contrast, although the high functioning therapists offered statistically sig- nificantly higher levels of empathy and genuineness than the low functioning therapists, the actual difference between the levels of therapeutic functioning of the two groups of therapists was so small as to be relatively incon- sequential in a practical sense. The two high functioning therapists responded accurately to almost all of the client's more surface feelings as well as to many of his less evident feelings and only rarely related to the client in a professional or phony manner. The two low functioning therapists offered only slightly lower levels of empathy and genuineness to their clients, with differences consist- ing of less than half a stage on the 9-stage Empathy scale and less than a quarter of a stage on the 5-stage Genuine- ness scale. Thus, the fact that the levels of empathy and genuineness offered by the relatively high and low function- ing therapists in Study II did not differ sufficiently to indicate a practical or meaningful difference appears to at least in part account for the smaller overall variability in IRS responses and the fewer differences in IRS responses obtained between the high and low functioning therapists in Study II in comparison with Study I. Not only was the sample of therapists in Study II, which consisted of four low experience, predominantly 146 analytically oriented therapists each of whom saw 10 out— patients, more homogeneous with respect to their IRS responses and their functioning on the core conditions, they were also more homogeneous with respect to Orientation, Experience, and Client Status. Since only four cases in Study I exactly matched this description of the therapists in Study II, such a direct comparison of the IRS responses could not be made. However, when the distributions of IRS ratings of therapists who saw outpatients in each study were compared, the distributions were much more similar, with the differences between the two studies at stages 2 and 3 each reduced to approximately only 13%. The two studies also differed with respect to the scales and raters used to measure E, W, G, and DX. In Study II the original scales devised by Truax were each rated by four different undergraduate college students who were relatively naive regarding psychotherapy theory and processes, and moderate reliabilities were obtained on each of the scales. In Study I the Truax E, W, and G scales revised and condensed by Carkhuff and his associates were rated by two experienced post-doctoral therapists, and the DX scale was rated by two research technicians, and very high reliabilities were obtained on each of the scales. These differences may at least partially account for some of the differences obtained between the two studies. For example, the high interrelationships among E, W, and G in Study I and the lack of relationship between W 147 and both E and G in Study II may be a reflection of one or more of several relevant factors. The ratings on the three scales in Study II may have been more independent than those in Study I as a result of different sets of raters being used to rate each scale. In addition, the scales used in Study I, which had been revised specifically in order to make them more clear, concise, and reliable, were rated by experienced, psychologically sophisticated raters, and higher reliabilities were obtained. However, from listening to the tapes, it appeared that the lack of relationship in Study II between therapists' warmth and their empathy and genuineness was an accurate reflection of the psychotherapy of these therapists: the Analytically oriented psychiatric residents who composed the sample in Study II tended to give a considerable amount of direct advice to clients, and much of this advice-giving, perhaps consistent with the thera- pists' medical training and orientation, appeared at times to constitute an imposition upon the client and thus to communicate a low level of nonpossessive warmth toward the client. The IRS responses of therapists who differed in Orientation were relatively consistent with the prediction that therapists who prefer a more intimate relationship would relate client references to themselves and focus on the relationship to a greater extent than therapists who prefer a more personal, distant relationship. Although the Relationship therapists tended to focus on the relationship 148 to a greater extent than the Eclectic and Client-Centered therapists, they differed dramatically from the Analytic therapists. The Relationship therapists related client statements to themselves in half of their responses, ignored almost no client statements directly related to themselves, and failure to respond to indirect and Opaque references to themselves each accounted for approximately a fourth of their responses. In contrast, in spite of the fact that the cornerstone of an Analytic orientation consists of the working through with the client of his transference feel- ings (Wolman, 1967), Analytic therapists 223$; directly related client statements to themselves or attempted to explore the client's feelings about themselves. They actually ignored direct statements about themselves in 20% of their responses, and in a majority of their responses, 65%, responded literally to the manifest content seemingly unrelated to themselves or the therapy situation. These findings thus indicate that the Relationship therapists actually behaved in therapy in a manner consistent with the tenets-of their orientation, that is, they focused to a considerable extent in therapy on the client-therapist relationship. The findings regarding the Analytic thera- pists, however, seem to be consistent with a more passive approach to therapy in which the therapist is relatively uninvolved personally, quite content oriented, and regard- less of the theoretically deleterious effect on the client, 149 does not encourage the client to explore his feelings about the therapist. Although the high and low experience level thera- pists did not differ on IRS scores, the high experience therapists gave more responses at stage 3 and fewer at stage 6 than the low experience therapists. This difference at stage 3, however, was primarily due to the Analytic therapists. Approximately 75% of the responses of the high experience Analytic therapists consisted of responses to the manifest content of client statements apparently unrelated to the therapist in contrast to 30% of such responses for the low experience Analytic therapists. The difference at stage 6, however, was primarily accounted for by the Rela- tionship therapists. The high experience Relationship therapists focused directly on the relationship less often and, at the same time, more frequently ignored indirect client references to the therapist than did the low expe- rience Relationship therapists. These differences between the high and low experience therapists of the Relationship and Analytic orientations may reflect the tendency for therapists who are still in training to experiment more within the limits of their theoretical orientations and, conversely, for therapists with considerable post-graduate therapeutic experience to become more cautious and conservative in their responses over time. The findings regarding client depth of self- exploration were quite different in Studies I and II. The 150 clients in Study I explored themselves more deeply as their therapists focused to a greater extent on the relationship. More specifically, deeper levels of self-exploration were achieved by clients of those therapists who relatively more frequently interpreted client references to themselves and, at the same time, relatively less frequently ignored their client's indirect references. However, clients' depth of self-exploration in Study II was unrelated to the extent to which their therapists focused on the relationship. Thus, while Hypothesis III of Study I was confirmed, the findings were not replicated in Study II. Although somewhat different scales were used to measure client self-exploration in the two studies, this does not appear to account for the dissimilar relationships between IRS and DX scores obtained in the two studies. The revised DX scale used in Study I is simply a condensed version of the Truax DX scale used in Study II. Moreover, contradictory to the great majority of the evidence cited by Truax and Carkhuff (1967), in the Johns Hopkins study Truax gt a1. (1966) found that clients' level of self— exploration was unrelated to their therapists' level of empathy, genuineness, or warmth and, more surprisingly, was unrelated to any of the measures of the clients' degree of improvement at the time of termination.3 Clearly, the findings regarding client DX generated in Study II are atypical and run counter not only to the 3 . . Personal communication 151 findings in Study I but to the bulk of the research reported by Truax and his associates (Truax & Carkhuff, 1967) and Berenson (1967). Hypothesis II of Study II which predicted greater degrees of improvement for clients whose therapists func- tioned at higher levels on the IRS, independent of the therapist's functioning on the core conditions, was rejected for each of the five outcome measures. However, additional, more precise analyses were suggested by the findings in both studies which indicated that therapists functioning at relatively higher and lower levels on the core conditions responded very differently to client references to them- selves and differed greatly in the extent to which they focused on the relationship. These findings were thus consistent with previous research (Truax & Mitchell, 1970) which has indicated that failure to take into account the therapeutic level of functioning on the core conditions constitutes a critical instance of adherence to the fal- lacious Therapist Uniformity Myth (Kiesler, 1966) and frequently results in misleading findings and conclusions. In other words, analyses of the relationship between IRS and client outcome, without taking into account the fact that the clients in Study II received different levels of the core conditions, resulted in masking the differential relationship of IRS with client outcome. These additional analyses, which took into account the fact that therapists functioning at relatively high and low levels on the core 152 conditions cannot be assumed to constitute a homogeneous group with respect to a variety of process and outcome variables (Truax & Mitchell, 1970), in turn provided further evidence that uniformity or homogeneity of therapists who differ with respect to the core conditions is indeed a myth. The results of these additional analyses were con- sistent with differential conceptualizations of the general nature of the IRS and core condition variables. The core conditions may best be regarded as reflecting rather broad and relatively permanent personal characteristics or attri- butes of the therapist which are not specific to the therapeutic situation or the therapist-client relationship. Indeed, ample evidence has been cited indicating that high levels of the core condition characteristics facilitate growth in areas other than psychotherapy, e.g., teacher— student relations and parent-child relations. In fact, it is quite likely that the core conditions lent themselves readily to early measurement precisely because they reflect broad, easily discerned, and generally effective personal characteristics of helpful persons who are effective in a number of disparate situations. In contrast, the IRS var- iable may be regarded as reflecting a particular class of therapist behavior that is more specific to the psycho- therapy situation but which is not uniformly effective in psychotherapy. That is, a particular therapist response to client references to the therapist, e.g., a therapist response which explicitly relates the client reference to 153 himself or a therapist response which ignores the reference, is not necessarily more effective than another. A partic— ular IRS response can have very different consequences in the hands of different therapists, in this case a therapist who is empathic and genuine with a client in comparison to a therapist who is unempathic and ingenuine with a client. Thus, the effectiveness of a therapist's IRS responses is dependent upon the accompanying level of his empathy and genuineness and, consequently, should be evaluated in con- junction with or in the context of his level of functioning on the core conditions. The major implications of the present research regarding the effectiveness of IRS responses of predomi- nantly Analytically oriented, psychiatric resident thera— pists in relatively short-term therapy with outpatients can be briefly summarized as follows. The extent to which a therapist focuses on the relationship in the context of the therapist's deep understanding and genuine responses to the client is relatively unrelated to client improvement. However, the extent to which a therapist focuses on the relationship is related to client lack of improvement and even deterioration when the context is that of minimal understanding and a lack of spontaneity and genuineness. Furthermore, the more unempathic and ingenuine the thera— peutic context, the more harmful the therapist's attempts to focus on the relationship become in relation to the client, i.e., the less improvement or even deterioration 154 the client shows. More precisely, responses which relate a client's statement to the immediate therapeutic relation- ship when made by an unempathic, ingenuine therapist are adversely related to the client's subsequent improvement; moreover, for such an unempathic, ingenuine therapist either ignoring the client's references to the therapeutic situ— ation or simply responding literally to the client's manifest content may be more appropriate since these thera- pist responses are associated with approximately average levels of client improvement. For relatively highly empathic, genuine therapists whether the therapist relates the client's statements to the immediate relationship or simply responds to the literal content and ignores refer- ences to himself has little relationship to the client's subsequent level of improvement: Unlike the low core con- dition therapist, attempts by an understanding and genuine therapist to relate client statements to himself are not associated with client unimprovement or deterioration. The major significance of the IRS variable in relation to therapeutic effectiveness thus lies in the lack of client improvement associated with an unempathic and ingenuine therapist's attempts to relate client statements to himself or to focus on the immediate therapeutic rela- tionship. The particular way in which an empathic, genuine therapist responds to client references to himself, e.g., ‘whether he avoids or directly approaches and interprets such references, is relatively unrelated to the client's 155 improvement. And more surprisingly, consistently ignoring or avoiding a client's references to the therapist is relatively inconsequential to his improvement, regardless of whether the therapist who avoids these references is relatively highly empathic and genuine or unempathic and ingenuine with the client. Thus, the most important and far-reaching impli— cation of the present research is that the manner in which a therapist responds to his client's references to himself bears relatively little relation to client improvement except in those destructive instances in which a therapist who fails to accurately understand most of the client's feelings and responds to the client in an ingenuine, pro- fessionally stereotyped, defensive manner attempts to inter— pret these references in relation to himself and their immediate relationship. Such instances are destructive in that they are related to client failure to improve or even deterioration. It is not simply that frequent and direct interpretation of client statements in relation to the therapist in and of itself is destructive to the client. The critical element is that these interpretations are made by an unempathic, ingenuine therapist, and the destructive- ness is associated with the inaccurate interpretation of the client's statements and feelings in question and the ingenuine, phony manner in which the therapist offers the interpretation. 156 In speculation of what may actually occur in psycho- therapy with an unempathic, ingenuine therapist that could perhaps account for this destructiveness, two factors seem plausible. A therapist's relatively frequent and consist- ently inaccurate interpretations of the client's statements in relation to himself may be experienced to some degree by the client as attempts by the therapist to impose his own distorted perceptions and reality, and even worse, an unwarranted closeness, on the client, somewhat analogously to the experience of schizophrenics described by Powdermaker (1952), Searles (1965), and Stierlin (1959). Moreover, when a client is encouraged to focus on and to freely and spon- taneously express his innermost feelings about the therapist by a therapist who, at the same time, hides his own feelings from the client behind a professionally stereotyped, defensive facade, the client may very likely experience a double-bind situation somewhat analogous to that proposed by Bateson, Jackson and Weakland (1956). Such a therapist, while seeming to encourage client exploration of feelings about the therapist may in fact be subverting such explor- ation as well as a close, Open therapeutic relationship. Thus, in these respects the destructiveness of this type of therapist behavior appears to resemble that frequently found in mothers of schiZOphrenics. In any event, it seems likely that clients of such therapists would at least feel confused and somewhat distrustful of the therapist's motives and ability to be helpful: these or similar feelings of client 157 dissatisfaction may have been reflected in the fact that the more these unempathic, ingenuine therapists focused on the relationship, the earlier the client terminated therapy. Subsequent outcome research is called for which employs a larger sample of therapists who are more hetero- r~ geneous with respect to level of functioning on the core E conditions, orientation and experience. One of the limita- I tions of the present research is that the relationship between IRS responses and client outcome was based on the therapy of four relatively inexperienced therapists. Inasmuch as the four therapists each saw 10 clients, the specific relationships between therapist IRS responses and diverse measures of client outcome would appear to be reliable. Nevertheless, confirmation of the findings should be attempted with a larger sample of therapists. In addition, the investigation of such complex variables as therapists' level of functioning on the core conditions and their IRS responses and client outcome requires a sample of therapists large and heterogeneous enough to permit detailed study of first and second order interaction effects. SUMMARY The present research was an investigation of the frequency and explicitness with which therapists interpret client statements in relation to themselves or the immediate therapeutic relationship and thereby focus on the client- therapist relationship. All client statements are assumed to be related to the therapist to some extent or with vary- ing degrees of directness or overtness. The Immediate Relationship Scale (IRS) was constructed to measure the explicitness with which therapists respond to client refer- ences to the therapist of varying degrees of overtness. In View of the concensus among therapists regarding the critical role in psychotherapy of the client's feelings about the therapist and the importance attached to the therapist's encouragement of client expression and explora- tions of such feelings, a positive relationship between therapists' IRS scores and client improvement was hypothe— sized. In addition, therapists' IRS scores were hypothesized to be related to their functioning on the core conditions, Orientation but not level of experience, and their clients' depth of self-exploration but not to client status as inpatient or outpatient. 158 159 Two separate studies using tape recordings of actual psychotherapy sessions were conducted. The data for Study I, a process study, was based on tape recordings of first therapy sessions of 56 different therapist—client dyads, which included both inpatient and outpatient clients and therapists heterogeneous with respect to setting, disci- pline, orientation, experience level, and sex. Ratings on the IRS and the E, W, G, and DX scales by Carkhuff and his associates were obtained on five 3—minute segments excerpted from each of the 56 tapes. The results of Study I indicated that for the total sample of therapists, IRS scores were 1) positively related to the therapists' core condition scores and clients' DX scores, 2) ordered therapists according to Orientation from higher to lower as Relationship, Eclectic, Client-Centered, and Analytic, respectively, 3) differentiated Relationship therapists from Analytic therapists and tended to also differentiate Relationship therapists from Eclectic and Client-Centered therapists and 4) did not differentiate between inpatient and outpatient therapists or between high and low experience level therapists within any of the four orientations. All hypotheses in Study I were confirmed with the exception of the hypothesis predicting the Rela- tionship therapists to be differentiated from each of the other orientation groups of therapists. The data in Study II, an outcome study, was based on tape recorded psychotherapy sessions of 40 outpatients seen '1 160 by four psychiatric residents at Johns Hopkins University. Ratings on the IRS and the E, W, and G scales by Truax were obtained on six 3-minute segments excerpted from the record- ings of each client's sessions: one segment was excerpted from the middle and another from the last third portions of the first, tenth, and fifth from final sessions. Five measures of client outcome were used: global improvement ratings by clients and therapists, Discomfort, Target, and Social Ineffectiveness. The results indicated that although for the total sample of therapists IRS scores and core condition scores were unrelated, thus rejecting the hypothesis predicting a positive relationship, IRS scores were positively related to the core condition scores for the high functioning therapists but negatively related for the low functioning therapists. Moreover, the high functioning therapists also functioned at higher levels on the IRS than the low func- tioning therapists. These findings, in addition to the finding that high and low functioning therapists in Study I also responded differentially on the IRS, were interpreted as providing further evidence that therapist uniformity is indeed a myth with respect to high and low functioning therapists and that failure to take into account therapists' level of functioning on the core conditions can result in misleading and even erroneous conclusions. The importance of the therapeutic level of function- ing received further emphasis from the analyses relating IRS 161 scores to client outcome measures. The hypothesis predict- ing a positive relationship between the degree of clients' IRS scores for the entire sample improvement and therapists' was rejected for each of the five outcome measures; indeed, step-wise regression analyses indicated that higher IRS scores were actually predictive of lesser degrees of client 3 W -'_ r. improvement on one outcome measure, Discomfort. However, additional analyses, which took into account the core con- indicated that higher IRS ~ ' L 1 dition context of IRS scores, scores in a context of relatively low levels of therapist offered empathy and genuineness were related to lesser degrees of client improvement on three of the outcome In measures: Discomfort, Therapist Statement, and Target. contrast, IRS scores in a context of relatively high levels of therapist offered empathy and genuineness were unrelated to client improvement on any of the outcome measures. These findings were interpreted as indicating that in the context Of £3 therapist's deep understanding and genuine responses tiOthe client the particular way in which a therapist -reSEx3nds to client statements or references to himself, whether he ignores or explicitly interprets such e.§;.' reffielrences, is relatively inconsequential to the client's l'nr'plli‘CDVement. But when a therapist interprets client state- merrtfs ih.relation to himself and thus attempts to focus on time ITelationship in the context of the therapist's failure U) ITEESpond genuinely and to accurately understand the c ' . . llearIt's feelings, the client fails to improve or even 162 deteriorates. Thus, the effectiveness of a therapist's IRS responses is dependent upon the accompanying level of his empathy and genuineness and, consequently, should be evaluated in conjunction with or in the context of his level of functioning on the core conditions. REFERENCES REFERENCES Alexander, J. F., & Abeles, N. Dependency changes in psychotherapy as related to interpersonal relation- ships. Journal g£_Consulting Psychology, 1968, 32, 685-689. ’_MGA_£— I—- Aspy, D. N. A study of three facilitative conditions and their relationships to the achievement of third grade students. Unpublished doctoral dissertation, University of Kentucky, 1965. I'll: {hag-m ‘ Aspy, D. N., & Hadlock, W. The effect of empathy, warmth, and genuineness on elementary students' reading achieve— ment. Unpublished thesis, University of Florida, 1966. Baggaley, A. R. Intermediate correlational methods. New York: Wiley, 1964. Bandura, A., Lipsher, D. H., & Miller, P. E. Psychothera— pist's approach-avoidance reactions to patient's expression of hostility. Journal of Consulting Psycholggy, 1960, 24, 1-8. Battle, C. C., Imber, S. D., Hoehn-Saric, R., Stone, A. R., Nash, C., & Frank, J. D. Target complaints as criteria of improvement. American Journal 9f Psychotherapy, 1966, 29, 184—192. Beier, E. G. The silent language of psychotherapy: Social reinforcement pf unconscious processes. Chicago: Aldine, 1966. Berenson, B. G., Carkhuff, R. R., & Southworth, A. J. Empathic understanding in interpersonal processes--a scale for measurement. Unpublished research scale, University of Massachusetts, 1964. Berenson, B. G., Mitchell, K. M., & Laney, R. C. Level of therapist functioning, types of confrontation, and type of patient. Journal pf Clinical Psychology, 1968, 24, 111-113. 163 164 Berenson, B. G., Mitchell, K. M., & Moravec, J. A. Level of therapist functioning, patient depth of self-exploration and type of confrontation. Journal gf Counseling Psychology, 1968, 15, 136-139. Bergin, A. E. Some implications of psychotherapy research for therapeutic practice. International Journal 9: Psychiatry, 1967, 3, 136-150. Berman, L. Countertransferences and attitudes of the analyst in the therapeutic process. Psychiatry, 1949, 12, 159-166. Blau, B. A. A comparison of more improved with less improved clients treated by client-centered methods. In W. U. Snyder (Ed.), Group report of a program gf research i3 psychotherapy. State College, Pa.: Penn- sylvania State College, Psychotherapy Research Group, 1953. Bordin, E. W. Psychological counseling. New York: Appleton-Century—Crofts, 1955. Braaten, L. J. The movement from non-self to self in client-centered psychotherapy. Journal gf Counseling Psychology, 1961, g, 20-24. Bugental, J. F. The person who is the psychotherapist. Journal 9f Consulting Psychology, 1964, 2S, 272-277. Caracena, P. F. Elicitation of dependency expressions in the initial stage of psychotherapy. Journal 9: Counsel— ing Psychology, 1965, 12, 268-274. Carkhuff, R. R. Facilitative genuineness in interpersonal processes: A scale for measurement. Unpublished research scale, 1964. (a) Carkhuff, R. R. Self—exploration in interpersonal processes: A scale for measurement. Unpublished research scale, 1964. (b) Carkhuff, R. R. The counselor's contribution pg facili- tative processes. Urbana, Ill.: Parkinson, 1967. (a) Carkhuff, R. R. Toward a comprehensive model of facili- tative interpersonal processes. Journal 9: Counseling ngchology: 1967: 13. 67-72. (b) ___.__.__ Carkhuff, R. R. Helper communication as a function of helper affect and content. Journal gf Counseling Psychology, 1969, 16, 126-131. (a) 165 Carkhuff, R. R. Helping and human relations, s primer for lay and professional helpers. Vol. I, Selection and training. New York: Holt, Rinehart and Winston, l969.(b) Carkhuff, R. R. Helping and human relations, s primer for lay and professional helpers. Vol. II, Practice and research. New York: Holt, Rinehart and Winston, l969.(c) Carkhuff, R. R., & Berenson, B. G. Beyond counseling and therapy. New York: Holt, Rinehart and Winston, 1967. Carkhuff, R. R., Southworth, A. J., & Berenson, B. G. Respect or positive regard in interpersonal processes: A scale for measurement. Unpublished research scale, University of Massachusetts, 1964. Cartwright, R. D. A comparison of the response to psycho- analytic and client-centered psychotherapy. In L. A. Gottschalk & A. H. Auerbach (Eds.), Methods gr research 12 psychotherapy. New York: Appleton-Century—Crofts, 1966. Chassan, J. B., & Bellak, L. An introduction to intensive design in the evaluation of drug efficacy during psycho- therapy. In L. A. Gottschalk & A. H. Auerbach (Eds.), Methods 92 research rs psyphotherapy. New York: Appleton-Century-Crofts, 1966. Collingwood, T., Hefele, T., Muehlberg, N., & Drasgow, J. Toward identification of the therapeutically facili- tative factor. Journal 92 Clinical Psychology, 1970, 22, 119-120. Dickenson, W. A., & Truax, C. B. Group counseling with college underachievers: Comparisons with a control group and relationship to empathy, warmth, and genuine— ness. Personnel and Guidance Journal, 1966, 45, 243— 247. Edwards, A. L. Erperimental design 23 psychological research. New York: Holt, Rinehart and Winston, 1962. Efroymsen, M. A. Multiple regression analysis. In A. Ralston & H. S. Wilf (Eds.), Mathematical Methods for Digital Computers, Part V, (17). New York: Wiley, 1960. Fay, W. F. Doctrine and experience: Their influence upon the psychotherapist. Journal 92 Consulting Psycholqu, 1958, 22, 403-409. 166 Fiedler, F. E. A comparison of therapeutic relationships in psychoanalytic, nondirective, and Adlerian therapy. Journal 22 Consultipg PsycholOgYp 1950, 22, 436-445. (a) Fiedler, F. E. The concept of the ideal therapeutic rela— tionship. Journal 92 Consulting Psychology, 1950, 22, 239-245. (b) Fiedler, F. E. Factor analyses of psychoanalytic, nondi- rective,and Adlerian therapeutic relationships. Journal pr Consulting Psychology, 1951, 22, 32—38. Frank, J. D., Gliedman, L. H., Imber, S. D., Nash, E. H., & Stone, A. R. Why patients leave psychotherapy. American Medical Association Archives pr_Neurological Psychiatry, 1957, 22, 283-299. Freud, S. The dynamics of the transference. In Collected papers 22 Sigmund Freud, Vol. 2. London: Hogarth Press, 1959. Fromm-Reichman, F. Notes on the development of treatment of schiZOphrenics by psychoanalytic psychotherapy. Psychiatry, 1948, 22, 263-273. Fromm-Reichman, F. Notes on the personal and professional requirements of a psychotherapist. Psychiatry, 1949, 22, 361-378. Fromm—Reichman, F. The princrples gr intensive psycho- therapy. Chicago: The University of Chicago Press, 1950. Gardner, G. G. The therapeutic relationship. Psychological Bulletin, 1964, 62, 426-437. Gendlin, E. T., Jenney, R. H., & Shlien, J. M. Counselor ratings of process and outcome in client-centered therapy. Journal 92 Clinical Psychology, 1960, 29, 210-213. Hefele, T. J., Collingwood, T., & Drasgow, J. Therapeutic facilitativeness as a dimension of effective living: A factor analytic study. Journal 92 Clinical Psychology, 1970, 2Q, 121-123. Heimann, P. On countertransference. International Journal 92 Psychoanalysis, 1950, 22, 81-84. Hobbs, N. Sources of gain in psychotherapy. American Psychologist, 1962, 22, 741-747. “ISL-5.3. _ 167 Hoehn-Saric, R., Frank, J. D., Imber, S. D., Nash, E. H., Stone, A. R., & Battle, C. C. Systematic preparation of patients for psychotherapy: I. Effects on therapy behavior and outcome. Journal of Psychiatric Research, 1965, 2, 267-281. Howard, K. I., Orlinsky, D. E., & Hill, J. A. The patient's experience of psychotherapy: Some dimensions and determinants. Mpltivariate Behavioral Research, Special Issue, 1968, 55-72. Howard, K. I., Orlinsky, D. E., & Hill, J. A. Content of dialogue in psychotherapy. Journal or Counseling Psycholong 1969, 2g, 396—404. (a) Howard, K. I., Orlinsky, D. E., & Hill, J. A. The thera— pist's feelings in the therapeutic process. Journal or Clinical Psychology, 1969, 2, 83-93. (b) Imber, S. D., Frank, J. D., Nash, E. H., Stone, A. R., & Gliedman, L. H. Improvement and amount of therapeutic contact: An alternative to the use of no-treatment controls in psychotherapy. Journal or Consulting Psychology, 1957, 22, 309-315. Karl, N. J., & Abeles, N. Psychotherapy process as a function of the time segment sampled. Journal 92 Con— sulting and Clinical Psychology, 1969, 22, 207-212. Kell, B. L., & Mueller, W. J. Impact and change: A study of counseling relationshrps. New York: Appleton- Century—Crofts, 1966. Kiesler, D. J. Some myths of psychotherapy research and the search for a paradigm. Psychological Bulletin, 1966, 62, 110-136. Kirtner, W. L., & Cartwright, D. S. Success and failure in client—centered therapy as a function of initial in—therapy behavior. Journal or Consulting Psychology, 1958, 22, 329—333. Lipkin, S. Clients' feelings and attitudes in relation to the outcome of client—centered therapy. Psychological Monographs, 1954, E2, No. l. Little, M. Countertransference and the patient's response to it. International Journal or Psychoanalysis, 1951, 32, 32-40. McNair, D. M., & Lorr, M. An analysis of professed psycho- therapeutic techniques. Journal or Consulting Psychology, 1964, 22, 265-271. 168 Marcondes, D. New aspects of the clinical interview: Countertransference difficulties. Psychosomatic Murray, E. J. A content-analysis method for studying psychotherapy. Psychological Monogrsphs, 1956, 29, (13), Whole No. 420. Murray, E. J. Direct analysis from the viewpoint of learn- ing theory. Journal or Consulting Psychology, 1962, 2Q, 226-231. Nash, E. H., Hoehn-Saric, R, Battle, C. C., Stone, A. R., Imber, S. D., & Frank, J. D. Systematic preparation of patients for short-term psychotherapy. II: Relation to characteristics of patient, therapist and the psycho- therapeutic process. Journal or Nervous and Mental Disease, 1965, 229, 374—383. Orlinsky, D. E., & Howard, K. I. Dimensions of conjoint experiential process in psychotherapy relationships. Proceedings, 75th American Psychological Association Annual Convention, 1967. (a) Orlinsky, D. E., & Howard, K. I. The good therapy hour: Experiential correlates of patients' and therapists' evaluations of therapy sessions. Archives or General Psychiatflr 1967, _l_6, 621—632. (b) Orne, M. T., & Wender, P. H. The use and rationale of anticipatory socialization as a therapeutic technique. American Journal or Psychiatry, 1968, 124, 1202-1212. Parloff, M. B., Kelman, H. C., & Frank, J. D. Comfort, effectiveness, and self-awareness as criteria of improvement in psychotherapy. American Journal or Psychiatry, 1954, 111, 343-351. Patterson, C. H. Divergence and convergence in psycho- therapy. American Journal 2E Psychotherapy, 1967, 22, 4-17. Paul, G. L. Strategy of outcome research in psychotherapy. Journal or Consultipg Psychology, 1967, 2, 109-118. Peres, H. An investigation of non-directive group therapy. Journal or Consulting Psychology, 1947, 22, 159-172. Powdermaker, Florence. Concepts found useful in the treat— ment of schizoid and ambulatory schizophrenic patients. Psychiatry, 1952, 22, 61-71. 169 Rogers, C. R. The necessary and sufficient conditions of therapeutic personality change. Journal or Consulting Psychology, 1957, 22, 95-103. Rogers, C. R., Gendlin, G. T., Kiesler, D. V., & Truax,(L B. The therapeutic relationship and its impsct: is study of psychotherapy with schiZOphrenics. Madison, Wis.: University of Wisconsin Press, 1967. Rosen, J. N. Direct ana2ysis. New York: Grune and Stratton, 1953. Sander, K., Tausch, R., Bastine, R., & Nagel, K. Die auswirkung experimenteller anderungen des psychothera— peutenverhaltens auf klienten in psychotherapeutischen gesprachen. Im manuskript, 1968. Schofield, W. Some general factors in counseling and therapy. In B. G. Berenson & R. R. Carkhuff (Eds.), Sources or gain 23 counseling and psychotherspy. New York: Holt, Rinehart and Winston, 1967. Schuldt, W. J. Psychotherapists' approach-avoidance responses and clients' expressions of dependency. Journal or Counselipg Psychology, 1966, 22, 178-183. Searles, H. F. Collected papers op schizophrenia and related subjects. New York: International Univer- sities Press, 1965. Seeman, J. A study of the process of non-directive therapy. Journal or Consulting Psychology, 1949, 22, 157-168. Seeman, J. Counselor judgments of therapeutic process and outcome. In C. R. Rogers & R. F. Dymond (Eds.), Psycho— therapy and personality change. Chicago: University of Chicago Press, 1954. Shapiro, J. G., Krauss, H. H., & Truax, C. B. Therapeutic conditions and disclosure beyond the therapeutic encounter. Journal or Counseling Psychology, 1969, 22, 290-294. Shapiro, J. G., & Voog, T. Effect of the inherently helpful person on student academic achievement. Journal or Counselipg PsychongY, 1969, 2Q, 505-509. Shapiro, S. B. Some neglected dimensions in psychotherapy. Psychological Reports, 1961, 2, 263-279. Shoben, E. J., Jr. Psychotherapy as a problem in learning theory. Psychological Bulletin, 1949, 46, 366-392. 170 Shoben, E. J. Some observations on psychotherapy and the learning process. In 0. H. Mower (Ed.), Psychotherapy: Theory and research. New York: Ronald Press, 1953. Spilken, A., & Jacobs, M. A. Psychotherapists' preferences for therapeutic styles. Paper presented at 76th American Psychological Association Convention, 1968. Spitz, R. Countertransference: Comments on its varying role in the analytic situation. Journal or the American Psychoanalysis Association, 1956, s, 256-265. Stierlin, H. The adaptation of the "stronger" person's reality: Some aspects of the symbiotic relationship of the schizophrenic. Psychiatry, 1959, 22, 143-152. Stoffer, D. S. An investigation of positive behavioral change as a function of genuineness, non-possessive warmth, and empathic understanding. Unpublished doc- toral dissertation, Ohio State University, 1968. Stone, A. R., Frank J. D., Nash, E. H., & Imber, S. D. An intensive five-year follow-up study of treated psychi- atric outpatients. Journal or Nervous and Mental Disease, 1961, 222, 410-422. Strupp, H. H. An objective comparison of Rogerian and psychoanalytic techniques. Journal 92 Consulting Strupp, H. H. The psychotherapist's contribution to the treatment process. Behavioral Science, 1958, 2, 34-67. Strupp, H. H. Psychotherapists 2p action: Explorations or the therapist's contribution E2 the treatment process. New York: Grune and Stratton, 1960. Strupp, H. H., & Bergin, A. E. Some empirical and con— ceptual bases for coordinated research in psychotherapy: A critical review of issues, trends, and evidence. International Journal or Psychiatry, 1969, 2, 18-90. Sullivan, H. S. The theory of anxiety and the nature of psychotherapy. Psychiatry, 1949, 22, 1-12. Sullivan, H. S. The psychiatric interview. New York: W. W. Norton, 1954. Sundland, D. M., & Barker, E. N. The orientations of psychotherapists. Journal 92 Consulting Psychology, 1962, 22, 201—212. 171 Tauber, E. Exploring the therapeutic use of countertrans- ference data. Psychiatry, 1954, 22, 331-336. Tausch, R., Eppel, H., Fittkau, B., & Minsel, R. Variablem and zusammenhange in der desprachspsychotherapie. Zeitschrift Fur Psychologie, 1969, 176, 93-102. Tomlinson, T. M., & Hart, J. T. A validation study of the process scale. Journal or Consulting Psychology, 1962, 26, 74-78. Truax, C. B. A scale for the measurement of accurate empathy. Psychiatric Institute Bulletin, Wisconsin Psychiatric Institute, University of Wisconsin, 1961, 1, 12. Truax, C. B. A tentative scale for the measurement of unconditional positive regard. Psychiatric Institute Bulletin, Wisconsin Psychiatric Institute, University of Wisconsin, 1962, 2, l. (a) Truax, C. B. A tentative scale for the measurement of depth of intrapersonal exploration (DX). Discussion Papers, Wisconsin Psychiatric Institute, University of Wisconsin, 1962, 29. (b) Truax, C. B. A tentative scale for the measurement of therapist genuineness or self-congruence. Discussion Papers, Wisconsin Psychiatric Institute, University of Wisconsin, 1962, 35. (c) Truax, C. B. Reinforcement and nonreinforcement in Rogerian psychotherapy. Journal or Abnormal Psycholpgy, 1966, 71, 1-9. Truax, C. B. Research findings: Translations and premature translations into practice. International Journal or Truax, C. B., & Carkhuff, R. R. For better or for worse: The process of psychotherapeutic personality change. In B. T. Wigdor (Ed.), Recent advances 2p the study 22 behavior changes. Montreal: McGill University Press, 1963. Truax, C. B., & Carkhuff, R. R. The experimental manipu- lation of therapeutic conditions. Journal or Consulting Psychology, 1965, 22, 119-124. Truax, C. B., & Carkhuff, R. R. Toward effective counseling and psychotherapy: Training and practice. Chicago: Aldine, 1967. 172 Truax, C. B., & Mitchell, K. M. The psychotherapeutic and the psychonoxious: Human encounters that change behavior. In M. Feldman (Ed.), Studies 29 psychotherapy and behavioral changs, Vol. 1, Research in individual psychotherapy. Buffalo: State University of New York Press, 1968. Truax, C. B., & Mitchell, K. M. Research on certain thera— pist skills in relation to process and outcome. In A. E. Bergin & W. E. Garfield, (Eds.), Handbook of psychotherapy and behavior change. —— Truax, C. B., & Tatum, C. R. An extension from the effective psychotherapeutic model to constructive change in pre-school children. Childhood Education, 1966, 92, 456-462. Truax, C. B., & Wargo, D. G. Antecedents to outcome in group psychotherapy with juvenile delinquents: Effects of therapeutic conditions, alternate sessions, vicarious therapy pretraining, and patient self-exploration. Unpublished manuscript. University of Arkansas, 1966. Truax, C. B., Wargo, D. G., & Carkhuff, R. R. Antecedents to outcome in group psychotherapy with outpatients: Effects of therapeutic conditions, alternate sessions, vicarious therapy pretraining and patient self- exploration. Unpublished manuscript, University of Arkansas, 1966. Truax, C. B., Wargo, D. G., Frank, J. D., Imber, S. D., Battle, C. C., Hoehn-Saric, R., Nash, E. H., & Stone, A. R. Therapist empathy, genuineness, and warmth and patient therapeutic outcome. Journal 92 Consulting Psychology, 1966, 22, 395-401. (a) Truax, C. B., Wargo, D. G., Frank, J. D., Imber, S. D., Battle, C. C., Hoehn—Saric, R., Nash, E. H., & Stone, A. R. The therapist's contribution to accurate empathy, non-possessive warmth, and genuineness in psychotherapy. Journal 92 Clinical Psychology, 1966, 22, 331-334. (b) van der Veen, F. Basic elements in the process of psycho— therapy: A research study. Journal 92 Consulting Psychology, 1967, 31, 295-303. Varble, D. L. Relationship between the therapist's approach- avoidance reactions to hostility and client behavior in therapy. Journal 92 Consulting and Clinical Psychology, 1968, 22, 237-242. 173 Vitalo, R. L. Effects of facilitative interpersonal func- tioning in a conditioning paradigm. Journal 92 Counsel- ing Psychology, 1970, 22, 141-144. Wagner, H. M., & Mitchell, K. M. Relationship between per— ceived instructors' accurate empathy, warmth, and genuineness and college achievement. Discussion Papers, Arkansas Rehabilitation Research and Training Center, University of Arkansas, 1969, 22. r ._ Wallach, M. S., & Strupp, H. J. Dimensions of psychothera— ' 1 pists' activity. Journal 92 Consulting Psychology, 1964, 22, 120-125. Whitaker, C. A., Felder, R. E., Malone, T. P., & Warkentin, J. First stage techniques in the experimental psycho— therapy of chronic schizophrenic patients. In J. H. Masserman (Ed.), Current Psychiatric Therapies, Vol. 2. New York: Grune and Stratton, 1962. Whitaker, C. A., & Malone, T. P. The roots 92 psycho— therapy. New York: Blakiston, 1953. Wiener, M., & Mehrabian, A. Language within language: Immediasy, a channel 29 verbal communication. New York: Appleton-Century-Crofts, 1968. Winder, C. L., Ahmad, F. A., Bandura, A., & Rau, L. C. Dependency of patients, psychotherapists responses and aspects of psychotherapy. Journal 92 Consulting Psychology, 1962, 29, 129-134. Winnecott, D. W. Hate in the countertransference. Inter- national Journal 92 Psychoanalysis, 1949, 22, 69-74. Wolff, W. Contemporary,psychotherapists examine themselves. New York: Charles C. Thomas, 1956. Wolman, B. B. The common rationale of the diverse psycho- analytic techniques. In B. B. Wolman (Ed.), Psycho— analytic techniques. New York: Basic Books, 1967. Wyrick, T. J., & Mitchell, K. M. Relationship between resident assistants' accurate empathy and warmth, and judged effectiveness. Journal 92 Student Personnel, 1971, in press. Yulis, S., & Kiesler, D. J. Countertransference response as a function of therapist anxiety and content of patient talk. Journal 92 Consulting and Clinical Esychology, 1968, 32, 413-419. APPENDICES APPENDIX A IMMEDIATE RELATIONSHIP SCALE Stage 2 The client clear2y and explicitly talks about 9r makes a direct, overt reference 29 the therapist or the therapeutic situation, but the therapist completely igno res the erplicit statements related 29 himself or the thera- peutic situation by remaining silent or by responding as the client had said nothing or had referred to some pers or situation other than the therapist or the therapeutic situation. The therapist responds as if he had not hear that part of the client's statements regarding himself ( therapist). Examples: C1: I had to wait for you a long time today. I guess t secretary forgot to tell you that I was here. . . T: Silence, or Mmmmm, or Where shall we begin today? the end of the last session I believe we were talki about your brother. . . Cl: I just can't seem to ask a girl for a date. I get nervous and scared. What do you do in that case? you ever have that trouble? T: Sounds like you're afraid she might turn you down. 174 if on d the he At ng all Did 175 was Implicit in the client's statements is an indirect reference to the therapist or the therapeutic situation. Regardless of the client's specific statements, and in addition to the overt message, the statements contain an indirect or covert reference to the therapist and reflect some of the client's feelings or attitudes about the thera- pist or the therapeutic situation. The therapist ignores the indirect reference 29 himself or the therapeutic situ— ation and makes no attempt to relate the statements to himself. Also rated at this stage is a therapist's response which fails to relate to himself or the therapeutic situ- ation a client's clear and overt statements about physi- cians, psychologists, social workers or other helping personnel, other than the therapist himself. The following are examples of tOpics likely to occur at this stage: authority figures such as doctors, bosses, teachers, ministers, and nosy people, people who just sit there and don't say or do anything, people who feel or act superior to others, peOple who evaluate and judge others, peOple who won't let others be dependent/independent, demanding people, people who are too objective and professional, seductive peOple, people who don't understand others, and people who can read others' minds. Examples: Cl: I've been to three doctors so far and no one has told me what's wrong with me. 176 T: What do you think is wrong with yourself? Cl: I just can't stand people who are always so nosy, always asking questions. . . always trying to mind other peOple's business. I used to have a friend that always wanted to know all about my business but he'd never tell me anything about himself. He always wanted to know what I thought or felt about something. . . . T: It sounds like you really were angry with your friend. 3"} Stage 2 The client's statements contain no clearly dis- cernible reference to the therapist or the therapeutic F a situation, and the therapist makes no attempt whatsoever to relate the client's statements to himself, the therapeutic situation, or to a class of persons, relationships, or situations in which the therapist could be included, e.g., doctors, helping personnel. In other words, 29 the absence 92 discernible client references to the therapist, the therapist respcnds 29 the client's statements quite specif- ically and literally and does not encourage the client to generalize his statements toward the therapeutic relation- ship. Examples: Cl: Well, I've always had problems with guys over me like my boss at the office. T: He really gets to you. . . . is that it? Cl: My brother always seemed to get his own way. T: You must have resented that. . . How about your sister? Did she get her own way, too? 177 ms. The client's statements contain either indirect or indiscernible references, but not overt references, to the therapist or the therapeutic situation, and the therapist's response indicates 92 suggests that the client's statements could 99 related, applied, 92 generalized 29 unspecified persons or situations other than those named by the client. The therapist's response is Open-ended in that it does not refer to a specific person or situation but is structured in such a way that the client could rather readily relate or apply the statement to the therapist or the therapeutic situation. Examples: T: Does that remind you of anyone else besides your boss? T: Who else does that sound like to you? T: Are there any other situations in which you feel that way, or in which something similar happens? T: Why do you bring that up at this particular time? What made you think of that right now? Stage 2_ The client's statements contain direct, indirect, or indiscernible references to the therapist or the therapeutic situation, and the therapist tentative2y, indirect2y, 9r cautious2y attempts 29 relate the client's statements to himself or the therapeutic situation. The therapist's response is more direct than his response at stage 4, but less direct than his response at stage 6. 178 Examples: Cl: I've always found that most doctors kind of think they're up on a pedestal all the time. T: Do you include therapists in that observation? Cl: Well that's just it . . . you go to class for three or four months before you find out how you're doing. She never got the grades out on time. You never knew how you stood with her, how she thought you were doing. T: I wonder . . . Sometimes . . . maybe not all the time but at times maybe you feel a little frustrated about . . . or maybe . . . possibly . . . a little worried about . . . how things are going in here. Stage 2 The client's statements contain direct, indirect, or indiscernible references to the therapist or the thera- peutic situation, and the therapist clearly, directly, and explicitly relates the client's statements 29 himself or to what is going on between them, i.e., to their immediate relationship. Examples: C1: T: C1: T: Christ! I'll be glad to get out of this hospital. And away from me, too! It's always been like that . . . first my mother . . then . . . I don't know . . . I never got . . . Sure, even me, even me. You feel like I'm not really helping you either. Right? Stage Stage Stage Stage Stage Stage 179 Summary 92 Stages Therapist ignores client's direct statements about therapist. Therapist ignores client's indirect references to therapist. Therapist responds literally to client's state- ments, which contain no discernible references to therapist. Therapist's open-ended response sets stage for client to relate to the therapist his statements, which contain indirect or indiscernible references. Therapist tentatively and/or indirectly relates to himself the client's statements, which contain direct, indirect, or indiscernible references. Therapist directly and explicitly relates to himself client's statements, which contain direct, indirect, or indiscernible references. APPENDIX B EMPATHIC UNDERSTANDING IN INTERPERSONAL PROCESSES A Scale for Measurement1 Bernard G. Berenson, Robert R. Carkhuff,J. Alfred Southworth Level 1 The first person appears completely unaware or ignorant of even the most conspicuous surface feelings of the other person(s). Example: The first person may be bored or disinterested or simply operating from a preconceived frame of reference which totally excludes that of the other person(s). In summary, the first person does everything but listen, understand or be sensitive to even the surface feelings of the other person(s). Level 2 The first person responds to the surface feelings of the other person(s) only infrequently. The first person continues to ignore the deeper feelings of the other per- son(s). Example: The first person may respond to some surface feel- ings but tends to assume feelings which are not there. He may have his own ideas of what may be going on in the other person(s) but these do not appear to correspond with those of the other person(s). 180 181 In summary, the first person tends to reSpond to things other than what the other person(s) appearstx>be expressing or indicating. Level 2_ The first person almost always responds with minimal understanding to the surface feelings of the other person(s) but, although making an effort to understand the other person's deeper feelings almost always misses their import. Example: The first person has some understanding of the surface aspects of the messages of the other person(s) but often misinterprets the deeper feelings. In summary, the first person is responding but not aware of who that other person really 2s or of what that other person is really like underneath. Level 3 constitutes the minimal level of facilitative interpersonal functioning. 9221...: The facilitator almost always responds with under- standing to the surface feelings of the other person(s) and sometimes but not often responds with empathic understanding to the deeper feelings. Example: The facilitator makes some tentative efforts to understand the deeper feelings of the other person(s). In summary, the facilitator is responding, however infre— quently, with some degree of empathic understanding of the deeper feelings of the other person(s). 182 Level 5 The facilitator almost always responds with accurate empathic understanding to all of the other person's deeper feelings as well as surface feelings. Example: The facilitator is "together" with the other person(s) or "tuned in" on the other person's wavelength. The facilitator and the other person(s) might proceed together to explore previously unexplored areas of human living and human relationships. The facilitator is responding with full awareness of the other person(s) and a comprehensive and accurate empathic understanding of his most deep feelings. 1The present scale "Empathic Understanding in Inter- personal Processes" has been derived in part from "A scale for the measurement of accurate empathy" by Truax which has been validated in extensive process and outcome research on counseling and psychotherapy. In addition, similar measures of similar constructs have received extensive support in the literature of counseling and therapy. The present scales were written to apply to all interpersonal processes and have already received research support. The present scale represents a systematic attempt to reduce the ambiguity and increase the reliability of the scale. In the process many important delineations and additions have been made. For comparative purposes, Level 1 of the present scale is approximately equal to Stage 1 of the earlier scale. The remaining levels are approximately correspondent: Level 2 and Stages 2 and 3 of the earlier version; Level 3 and Stages 4 and 5; Level 4 and Stages 6 and 7; Level 5 and Stages 8 and 9. APPENDIX C RESPECT OR POSITIVE REGARD IN INTERPERSONAL PROCESSES A Scale for Measurement1 Robert R. Carkhuff, A1fred.J.Southworth, Bernard G. Berenson Level 2 The first person is communicating clear negative regard for the second person. Example: The first person may be actively offering advice or telling the second person what would be "best" for him. In summary, in many ways the first person acts in such a way as to make himself the focus of evaluation and sees himself as responsible for the second person. Level 2 The first person reSponds to the second person in such a way as to communicate little positive regard. Example: The first person responds mechanically or pas— sively or ignores the feelings of the second person. In summary, in many ways the first person displays a lack of concern or interest for the second person. 183 184 2221.3. The first person communicates a positive caring for the second person but there is a conditionality to the caring. Example: The first person communicates that certain kinds of actions on the part of the second person will reward or hurt the first person. 1. "1" In summary, the first person communicates that what the second person does or does not do matters to the first person. Level 3 constitutes the minimal level of facili- tative interpersonal functioning. Level 2 The facilitator clearly communicates a very deep interest and concern for the welfare of the second person. Example: The facilitator enables the second person to feel free to be himself and to be valued as an indi- vidual except on occasion in areas of deep personal concern to the facilitator. In summary, the facilitator sees himself as responsible 29 the second person. 927.6212 The facilitator communicates a very deep respect for the second person's worth as a person and his rights as a free individual. Example: The facilitator cares very deeply for the human potentials of the second person. In summary, the facilitator is committed to the value of the other person as a human being. 185 1The present scale, "Respect or Positive Regard in Interpersonal Processes," has been derived in part from "A tentative scale for the measurement of unconditional posi- tive regard" by Truax which has been validated in extensive process and outcome research on counseling and psychotherapy. In addition, similar measures of similar constructs have received extensive support in the literature of counseling and therapy and education. The present scales were written to apply to all interpersonal processes and have already received research support. The present scale represents a systematic attempt to reduce the ambiguity and increase the reliability of the scale. In the process many important delineations and additions have been made. For comparative purposes, the levels of the present scale are approximately equal to the stages of the earlier scale, although the systematic emphasis upon the positive regard rather than upon uncon— ditionality represents a pronounced divergence of emphasis. APPENDIX D FACILITATIVE GENUINENESS IN INTERPERSONAL PROCESSES A Scale for Measurement1 Robert H. Carkhuff Level 2 The first person's verbalizations are clearly unrelated to what he is feeling at the moment, or his only genuine responses are negative in regard to the second person(s) and appear to have a totally destructive effect upon the second person. Example: The first person may be defensive in his inter- action with the second person(s) and this defensiveness may be demonstrated in the content of his words or his voice quality and where he is defensive he does not employ his reaction as a basis for potentially valuable inquiry into the relationship. In summary, there is evidence of a considerable discrepancy between the first person's inner experiencing and his current verbalizations or where there is no discrepancy the first person's reactions are employed solely in a destructive fashion. Level 2 The first person's verbalizations are slightly unrelated to what he is feeling at the moment or when his 186 187 responses are genuine they are negative in regard to the second person and the first person does not appear to know how to employ his negative reactions constructively as a basis for inquiry into the relationship. Example: The first person may respond to the second per- son(s) in a "professional" manner that has a rehearsed quality or a quality concerning the way a helper "should" respond in that situation. In summary, the first person is usually responding according to his prescribed "role" rather than expressing what he personally feels or means, and when he is genuine his responses are negative and he is unable to employ them as a basis for further inquiry. £22912 The first person provides no "negative" cues between what he says and what he feels, but he provides no positive cues to indicate a really genuine response to the second person(s). Example: The first person may listen and follow the second person(s) but commits nothing of himself. In summary, the first person appears to make appropriate responses which do not seem insincere but which do not reflect any real involvement either. Level 3 constitutes the minimal level of facilitative interpersonal functioning. Level 2 The facilitator presents some positive cases indi— cating a genuine response (whether positive or negative) in a non-destructive manner to the second person(s). 188 Example: The facilitator's eXpressions are congruent with his feelings although he may be somewhat hesitant about expressing them fully. In summary, the facilitator responds with many of his own feelings and there is no doubt as to whether he really means what he says and he is able to employ his responses, what- ever their emotional content, as a basis for further inquiry into the relationship. Level 5 The facilitator is freely and deeply himself in a non-exploitative relationship with the second person(s). Example: The facilitator is completely spontaneous in his interaction and open to experiences of all types, both pleasant and fearful, and in the event of hurtful responses, the facilitator's comments are employed constructively to Open further areas of inquiry for both the facilitator and the second person. In summary, the facilitator is clearly being himself and yet employing his own genuine responses constructively. 1The present scale, "Facilitative Genuineness in Interpersonal Processes" has been derived in part from "A tentative scale for the measurement of therapist genuineness or self-congruence" by Truax which has been validated in extensive process and outcome research on counseling and psychotherapy and education. The present scale represents a systematic attempt to reduce the ambiguity and increase the reliability of the scale. In the process, many impor— tant delineations and additions have been made. For com— parative purposes, the levels of the present scale are approximately equal to the stages of the earlier scale, although the systematic emphasis upon the constructive employment of negative reactions represents a pronounced divergence of emphasis. APPENDIX E SELF-EXPLORATION IN INTERPERSONAL PROCESSES A Scale for Measurement1 Robert R. Carkhuff Level 2 The second person does not discuss personally relevant material, either because he has had no Opportunity to do such or because he is actively evading the discussion even when it is introduced by the first person. Example: The second person avoids any self-descriptions or self-exploration or direct expression of feelings that would lead him to reveal himself to the first person. In summary, for a variety of possible reasons, the second person does not give any evidence of self-exploration. Level 2 The second person responds with discussion 29 the introduction 92 personally relevant material 9y the first person but does so in a mechanical manner and without the demonstration of emotional feeling. Example: The second person simply discusses the material without exploring the significance or the meaning of the material or attempting further exploration of that feeling in our effort to uncover related feelings or material. 189 190 In summary, the second person responds mechanically and remotely to the introduction of personally relevant material by the first person. Level 2 The second person voluntari2y introduces discussions of personally relevant material but does so in a mechanical manner and without the demonstration of emotional feeling. Example: The emotional remoteness and mechanical manner of the discussion give the discussion a quality of being rehearsed. In summary, the second person introduces personally relevant material but does so without spontaneity or emotional proximity and without an inward probing to newly discover feelings and experiences. Level 2 The second person voluntarily introduces discussions of personally relevant material with both spontaneiry and emotional proximity. Example: The voice quality and other characteristics of the second person are very much "with" the feel- ings and other personal materials which are being verbalized. In summary, the second person introduces personally relevant discussions with spontaneity and emotional proximity but without a distinct tendency toward inward probing to newly discover feelings and experiences. 191 22215 The second person actively and spontaneously engages in an inward probing to newly discover feelings or experi- ences about himself and his world. Example: The second person is searching to discover new feelings concerning himself and his world even though at the moment he may be doing so, perhaps, fearfully and tentatively. In summary, the second person is fully and actively focusing upon himself and eXploring himself and his world. 1The present scale "Self-exploration in inter- personal processes" has been derived in part from "The measurement of depth of intrapersonal exploration" by Truax which has been validated in extensive process and outcome research on counseling and psychotherapy. In addition, similar measures of similar constructs have received extensive support in the literature of counseling and therapy. The present scale represents a systematic attempt to reduce the ambiguity and increase the reliability of the scale. In the process many important delineations and additions have been made. For comparative purposes, Level 1 of the present scale is approximately equal to Stage 1 of the early scale. The remaining levels are approximately correspondent: Level 2 and Stages 2 and 3; Level 3 and Stages 4 and 5; Level 4 and Stage 6; Level 5 and Stages 7, 8, and 9. APPENDIX F A TENATIVE SCALE FOR THE MEASUREMENT OF ACCURATE EMPATHY Charles B. Truax Stage 2 Therapist seems completely unaware of even the most conspicuous of the client's feelings; his responses are not apprOpriate to the mood and content of the client's state— ments. There is no determinable quality of empathy, and hence no accuracy whatsoever. The therapist may be bored and disinterested or actively offering advice, but he is not communicating an awareness of the client's current feelings. Stage 2 Therapist shows an almost negligible degree of accuracy in his responses, and that only toward the client's most obvious feelings. Any emotions which are not clearly defined he tends to ignore altogether. He may be correctly sensitive to obvious feelings and yet misunderstand much of what the client is really trying to say. By his response he may block off or may misdirect the patient. Stage 2 is distinguishable from Stage 3 in that the therapist ignores feelings rather than displaying an inability to understand them. 192 193 seas; Therapist often responds accurately to client's more exposed feelings. He also displays concern for the deeper, more hidden feelings, which he seems to sense must be present, though he does not understand their nature or sense their meaning to the patient. Stage 2 Therapist usually responds accurately to the client's more Obvious feelings and occasionally recognizes some that are less apparent. In the process of this ten- tative probing, however, he may misinterpret some present feelings and anticipate some which are not current. Sensi- tivity and awareness do exist in the therapist, but he is not entirely "with" the patient in the current situation or experience. The desire and effort to understand are both present, but his accuracy is low. This stage is distin— guishable from Stage 3 in that the therapist does occasion- ally recognize less apparent feelings. He also may seem to have a theory about the patient and may even know how or why the patient feels a particular way, but he is definitely not "with" the patient. In short, the therapist may be diagnostically accurate, but not emphatically accurate in his sensitivity to the patient's current feelings. Stage 2 Therapist accurately responds to all of the client's more readily discernible feelings. He also shows awareness 194 of many less evident feelings and experiences, but he tends to be somewhat inaccurate in his understanding of these. However, when he does not understand completely, this lack of complete understanding is communicated without an antic- ipatory or jarring note. His misunderstandings are not disruptive by their tentative nature. Sometimes in Stage 5 the therapist simply communicates his awareness Of the problem of understanding another person's inner world. This stage is the midpoint of the continuum of accurate empathy. Stage 9 Therapist recognizes most of the client's present feelings, including those which are not readily apparent. Although he understands their content, he sometimes tends to misjudge the intensity of these veiled feelings, so that his responses are not always accurately suited to the exact mood of the client. The therapist does deal directly with feelings the patient is currently experiencing although he may misjudge the intensity of those less apparent. Although sensing the feelings, he often is unable to communicate meaning to them. In contrast to Stage 7, the therapist's statements contain an almost static quality in the sense that he handles those feelings that the patient offers but does not bring new elements to life. He is "with" the client but doesn't encourage exploration. His manner of communicating his understanding is such that he makes of it a finished thing. 195 $2221 Therapist responds accurately to most of the client's present feelings and shows awareness of the precise intensity of most Of the underlying emotions. However, his responses move only slightly beyond the client's own aware- ness, so that feelings may be present which neither the client not therapist recognizes. The therapist initiates moves toward more emotionally laden material, and may com- municate simply that he and the patient are moving towards more emotionally significant material. Stage 7 is distin- guishable from Stage 6 in that often the therapist's response is a kind of precise pointing Of the finger toward emotionally significant material. Stage 2 Therapist accurately interprets all the client's present, acknowledged feelings. He also uncovers the most deeply shrouded of the client's feelings, voicing meanings in the client's experience of which the client is scarcely aware. Since the therapist must necessarily utilize a method of trial and error in the new uncharted areas, there are minor flaws in the accuracy of his understanding, but these inaccuracies are held tentatively. With sensitivity and accuracy he moves into feelings and experiences that the client has only hinted at. The therapist offers specific explanations or additions to the patient's understanding so that underlying emotions are both pointed out and 196 specifically talked about. The content that comes to life may be new but it is not alien. Although the therapist in Stage 8 makes mistakes, these mistakes are not jarring, because they are covered by the tentative character of the response. Also, this thera- pist is sensitive to his mistakes and quickly changes his response in midstream, indicating that he has recognized what is being talked about and what the patient is seeking in his own explorations. The therapist reflects a together- ness with the patient in tentative trial and error explor- ation. His voice tone reflects the seriousness and depth Of his empathic grasp. arc-1922 The therapist in this stage unerringly responds to the client's full range of feelings in their exact intensity. Without hesitation, he recognizes each emotional nuance and communicates an understanding of every deepest feeling. He is completely attuned to the client's shifting emotional content; he senses each of the client's feelings and reflects them in his words and y9299. With sensitive accu- racy, he expands the client's hints into a full-scale (though tentative) elaboration of feeling or experience. He shows precision both in understanding and in communication of this understanding, and expresses and experiences them without hesitancy. APPENDIX G A TENTATIVE SCALE FOR THE MEASUREMENT OF NONPOSSESSIVE WARMTH Charles B. Truax Stage 2 The therapist is actively offering advice or giving clear negative regard. He may be telling the patient what would be "best for him," or in other ways actively approving or disapproving of his behavior. The therapist's actions make himself the locus of evaluation; he sees himself as responsible for the patient. Stage 2 The therapist responds mechanically to the client, indicating little positive regard and hence little nonpos- sessive warmth. He may ignore the patient or his feelings or display a lack of concern or interest. The therapist ignores client at times when a nonpossessively warm response would be expected; he shows a complete passivity that com- municates almost unconditional lack of regard. Stage 2 The therapist indicates a positive caring for the patient or client, but it is a semipossessive caring in the 197 198 sense that he communicates to the client that his behavior matters to him. That is, the therapist communicates such things as "It is not all right if you act immorally." "I want you to get along at work," or "It's important to me that you get along with the ward staff." The therapist sees himself as responsible for the client. Stage 2 The therapist clearly communicates a very deep interest and concern for the welfare of the patient, showing a nonevaluative and unconditional warmth in almost all areas of his functioning. Although there remains some condition- ality in the more personal and private areas, the patient is given freedom to be himself and to be liked as himself. There is little evaluation of thoughts and behaviors. In deeply personal areas, however, the therapist may be con— ditional and communicate the idea that the client may act in any way he wishes--except that it is important to the therapist that he be more mature or not regress in therapy or accept and like the therapist. In all other areas, however, nonpossessive warmth is communicated. The thera- pist sees himself as responsible 29 the client. Stage 2 At stage 5, the therapist communicates warmth without restriction. There is a deep respect for the patient's worth as a person and his rights as a free individual. At this level the patient is free to be himself 199 even if this means that he is regressing, being defensive, or even dislking or rejecting the therapist himself. At this stage the therapist cares deeply for the patient as a person, but it does not matter to him how the patient chooses to behave. He genuinely cares for and deeply prizes the patient for his human potentials, apart from evalua— tions of his behavior or his thoughts. He is willing to share equally the patient's joys and aspirations or depres- sions and failures. The only channeling by the therapist may be the demand that the patient communicate personally relevant material. APPENDIX H A TENTATIVE SCALE FOR THE MEASUREMENT OF THERAPIST GENUINENESS OR SELF-CONGRUENCE Charles B. Truax £22921 The therapist is clearly defensive in the inter- action, and there is explicit evidence of a very con- siderable discrepancy between what he says and what he experiences. There may be striking contradictions in the therapist's statements, the content of his verbalization may contradict the voice qualities or nonverbal cues (i.e., the upset therapist stating in a strained voice that he is "not bothered at all" by the patient's anger). St_asz_e__2_ The therapist responds appropriately but in a pro- fessional rather than a personal manner, giving the impression that his responses are said because they sound good from a distance but do not express what he really feels or means. There is a somewhat contrived or rehearsed quality or air of professionalism present. 200 201 Stage 2 The therapist is implicitly either defensive or professional, although there is no explicit evidence. Stage 2 There is neither implicit nor explicit evidence of defensiveness or the presence Of a facade. The therapist shows no self-incongruence. Stage 2 The therapist is freely and deeply himself in the relationship. He is Open to experiences and feelings of all types--both pleasant and hurtful--without traces of defensiveness or retreat into professionalism. Although there may be contradictory feelings, these are accepted or recognized. The therapist is clearly being himself in all of his responses, whether they are personally meaningful or trite. At stage 5 the therapist need not express per- sonal feelings, but whether he is giving advice, reflecting, interpreting, or sharing experiences, it is clear that he is being very much himself, so that his verbalizations match his inner experiences. APPENDIX I A TENTATIVE SCALE FOR THE MEASUREMENT OF DEPTH OF SELF-EXPLORATION Charles B. Truax Stage 2 No personnaly relevant material and no Opportunity for it to be discussed. Personally relevant material refers to emotionally tinged experiences or feelings, or to feel- ings or experiences of significance to the self. This would include self—descriptions that are intended to reveal the self to the therapist, and communications of personal values, perceptions of one's relationships to others, one's personal role and self-worth in life, as well as communica- tions indicating upsetness, emotional turmoil, or expres— sions of more specific feelings of anger, affection, etc. Stage 2 The patient actively evades personally relevant material (by changing the subject, for instance, refusing to respond at all, etc.). Thus, personally relevant material is not discussed. The patient does not respond to personally relevant material sysp when the therapist speaks of it. 202 203 £2322 The patient does not volunteer personally relevant material but he does not actually evade responding to it when the therapist introduces it to the interpersonal situation. Stage 2 The patient does not himself volunteer to share personally relevant material with the therapist, but he responds to personally relevant material introduced by the therapist. He may agree or disagree with the therapist's remarks and may freely make brief remarks, but he does not add significant new material. Stage 2 Personally relevant material is discussed (volun- teered in part or in whole). Such volunteer discussion is done (1) i a mechanical manner (noticeably lacking in spontaneity or as a "reporter" or "Observer"); and (2) with- out demonstration 92 emotional feelipg. In addition, there is simply discussion without movement by the patient toward further exploring the significance of meaning of the mate— rial or feeling in an effort to uncover related feelings or material. Both the emotional remoteness and the mechanical manner of the patient make his discussion often sound rehearsed. 204 Stage 2 This stage is similar to Stage 4 except that the material is discussed either with feeling indicating emo- tional proximity or with spontaneity, but not both. (Voice quality is the main cue.) Stage 2 In Stage 6 the level of Stage 4 is achieved again, with the additional fact that the personally relevant material is discussed with both spontaneity and feeling. There is clear indication that the patient is speaking with feeling, and his communication is laden with emotion. Stage 2 Tentative probing toward intrapersonal exploration. There is an inward probing to discover feelings or experi- ences anew. The patient is searching for discovery of new feelings which he struggles to reach and hold on to. The individual may speak with many private distinctions or with "personal" meanings to common words. He may recognize the value of this self-exploration but it must be clear that he is trying to explore himself and his world actively even though at the moment he does so perhaps fearfully and tentatively. Stage 2 Active intrapersonal exploration. The patient is following a "connected" chain of thoughts in focusing upon 205 himself and actively exploring himself. He may be discover- ing new feelings, new aspects of himself. He is actively exploring his feelings, his values, his perceptions of others, his relationships, his fears, his turmoil, and his life-choices. 831222 Stage 9 is an extension of the scale to be used in those rare moments when the patient is deeply exploring and being himself, or in those rare moments when he achieves a significant new perceptual base for his View of himself or the world. A rating at this stage is to be used at the judge's discretion. APPENDIX J DISCOMFORT SCALE Listed below are 50 symptoms or problems that people sometimes have. The doctor will read each of the 50 items, one at a time, and you must decide whether you have had the complaint duripg the last seven days including today. For each complaint you have had, the doctor will ask how much it bothered you, that is--not 92 all, jpst a little, pretty much, 92 very much. How Much It Bothered You Symptoms or Complaints Not Just a Pretty Very No Yes at all little much much l. Headaches 2. Pains in the heart or chest 3. Heart pounding or racing 4. Trouble getting your breath 5. Constipation 6. Nausea or upset stomach 7. Loose bowel movements 206 207 ” How Much It Bothered You Symptoms or Complaints Not Just a Pretty Very NO Yes at all little much much 8. Twitching of the face or body 9. Faintness or dizziness 10. Hot or cold spells ll. Itching or hives 12. Frequent urination 13. Pains in the lower part of your back 14. Difficulty in swallowing 15. Skin eruptions or rashes 16. Soreness of your muscles 17. Nervousness and shakiness under pressure 18. Difficulty in falling asleep or staying asleep 19. Sudden fright for no apparent reason 20. Bad dreams 21. Blaming yourself for things you did or failed to do 22. Feeling generally worried or fretful 23. Feeling blue 208 Symptoms or Complaints NO Yes How Much It Bothered You Not at all Just a little Pretty much Very much 24. Being easily moved to tears 25. A need to do things very slowly in order to be sure you were doing them right 26. An uncontrollable need to repeat the same actions, e.g., touch- ing, counting, hand— washing, etc. 27. Unusual fears 28. Objectionable thoughts or impulses which keep pushing themselves into your mind 29. Your "feelings" being easily hurt 30. Feeling that people were watching or talking about you 31. Generally preferring to be alone 32. Feeling lonely 33. Feeling compelled to ask others what you should do 34. People being unsympa- thetic with your need for help 35. Feeling easily annoyed or irritated 209 Symptoms or Complaints ,0 Yes How Much It Bothered You Not at all Just a Little Pretty much Very much 36. Severe temper outbursts 37. Feeling critical of others 38. Frequently took alcohol or medicine to make you feel better 39. Difficulty in speakin when you were excited 40. Feeling you were func4 tioning below your capacities, i.e., feeling blocked or stymied in getting things done 41. Having an impulse to commit a violent or destructive act, for example desire to set a fire, stab, beat or kill someone, mutilate an animal, etc. 42. Feeling shy and uneasy with the opposite sex 43. Unsatisfied with sexual partner 44. Worried about sloppi- ness or carelessness 45. Superstitions 46. Having to check and double check what you do 47. Sex dreams 210 Symptoms or Complaints NO Yes How Much It Bothered You Not at all Just a little Pretty much Very much 48. Seeing anything on a wall that is not hanging straight 49. Difficulty in carrying out normal sex relations 50. Poor appetite APPENDIX K TARGET SYMPTOM SCALE Completed by Patient TARGET OUTCOME MEASURE Clients rated the improvement in each Target Symptom as follows: 1. Targst Symptom 2 ( ) l. A lot Better ( ) 2. Some Better ( ) 3. Slightly Better ( ) 4. The Same ( ) 5. Worse 2. Target Symptom 2, ( ) l. A lot Better ( ) 2. Some Better ( ) 3. Slightly Better ( ) 4. The Same ( ) 5. Worse 211 212 3. Target gymptom 2 ( )l. A lot Better Some Better Slightly Better The Same Worse APPENDIX L PATIENT GLOBAL IMPROVEMENT SCALE Completed by Patient PATIENT STATEMENT OUTCOME MEASURE Clients rated their overall or global improvement as one of the following: (____) l. Worse (____) 2. Same (____) 3. Slightly Better (____) 4. Some Better ( ) 5. A Lot Better 213 APPENDIX M PATIENT GLOBAL IMPROVEMENT SCALE Completed by Therapist THERAPIST STATEMENT OUTCOME MEASURE Therapist rated the overall or global improvement of each client as one of the following: (____) l. Worse (____) 2. No Change (____) 3. Slight Improvement (____) 4. Moderate Improvement ( ) 5. Marked Improvement 214 APPENDIX N SOCIAL INEFFECTIVENESS SCALE Each of the fifteen areas of functioning are scored according to the following procedure. Ratings are made on both the intensity of the behavior (closeness of agreement with the example) and the frequency of the behavior (how typical it is of him with how many peOple or in how many social situations). 1 Behavior pattern is slightly true for him and then it is seldom Behavior pattern is slightly true for him and then it is often, or Behavior pattern is moderately true for him and then it is seldom Behavior pattern is slightly true for him and then it is almost always, or Behavior pattern is moderately true and then it is seldom Behavior pattern is very true for him and then it is seldom, or Behavior pattern is moderately true for him and then it is Often Behavior pattern is very true for him and then it is often, or Behavior pattern is moderately true for him and then it is almost always Behavior pattern is very true for him and almost always 215 216 The following examples of the "Ineffectiveness Scale" items are to be scored. Score (___) Score (___) Score (___) Score (___) Score (___) Score (___) Score ( ) 1. Overly-indspendent: Takes pride in self- sufficiency and competence; makes light of his troubles (as if he should be able to handle them without help), contemptuous of help seeking. N.B. Keep in mind that men are expected culturally to be more inde- pendent than women. Overlyrdependent: Refuses to take initi- ative where he can although it is expected of him. Relies on others for help with problems beyond what observer believes he really needs. N.B. 1) Do not rate pathafis seeking help for problems which are objec- tively too much for him. 2) Keep in mind that women are permitted culturally to be more dependent than men. Superficially-sociable: Seeks many super- ficial acquaintanceships. Breaks off relationships before they become intimate and/or refuses to form close relationships. N.B. Does not apply to relationships forced on patient, e.g., family or neces- sities of occupation. Withdrawn: Has less contacts with others than is culturally expected. This includes own family. N.B. Certain patients may express withdrawal from others in terms of withdrawal from activities. Mere under- activity, unless it implies withdrawal from people, should 992 be scored. Extra-punitive: Tends to disparage or blame others for his difficulties, frustra— tions, or failures. Intra-punitive: Self-critical, tends to blame, criticize and hold himself respon— sible for his difficulties, frustrations, and failures. Over-apologetic. Officious: Volunteers his services where they are neither asked nor needed; meddle- some to others in the use of their skills and capacities; may be overprotective of children or other family members. Score Score Score Score Score Score Score Score 10. 11. 12. 13. 14. 15. 217 Irresponsible: Tends to avoid his respon— sibilities and obligations as an adult; fails to provide culturally expected economic support for self, children, parents, etc. Yields to desires without considering the welfare of those dependent on him; e.g., gambling. N.B. Do not score sexual activity here but under 15 below. Impulsive: Acts or makes decisions without considering consequences; loses interest or becomes impatient if he cannot attain goals immediately. Takes unnecessary risks. Over-cautious: Tends to be timorous, indecisive. Plans too much for the future at the expense of the present. Refuses to take reasonable risks. N.B. Do not score guarded or evasive here. Hyper-reactive: Tends to overreact emotionally; behaves in an uncontrolled manner. Constrained: Tends to be unable to express feelings he experiences to their referent; expression of emotion inhibited. Over2y systematic: Tends to be overly concerned with details and orderliness; needs to recheck his acts. Unsystematic: Tends to be disorganized and unsystematic in daily routine, managing money, etc. N.B. Do not score irrespon- sible, i.e., self-indulgent here. Sexual Maladjustment: Tends to be sexually inadequate or fearful of heterosexual relationships. If married has impotency or frigidity problems or is very restricted sexually. If not married has difficulty or is fearful in dating. Or tends to show overconcern with sexuality and is promis- cuous in hetero or homosexual relation- ships. May have series of sex partners or dates. Sexual adjustment is unsatisfying. APPENDIX 0 Table 36. E's, S.D.'s, and Ranges of IRS Scores and Percent Of IRS Ratings at Each Stage for High and Low IRS Groups and IRS Quartile Groups in Studies I & II IRS STAGE 1 2 3 4 5 6 Hi I 6.43 14.29 47.86 12.14 2.14 17.14 IRS II 2.22 12.22 51.11 16.67 3.33 14.44 LO I 20.71 55.71 21.43 2.14 0.00 0.00 IRS 11 5.43 23.91 68.48 2.17 0.00 0.00 Q I 4.28 10.00 31.43 20.00 4.28 30.00 1 II 2.08 8.33 41.67 18.75 6.25 22.92 1 8.57 18.57 64.29 4.28 0.00 4.28 Q2 11 2.38 16.67 61.90 14.29 0.00 4.76 I 7.14 54.29 34.29 4.28 0.00 0.00 Q3 11 0.00 11.90 85.71 2.38 0.00 0.00 I 34.29 57.14 8.57 0.00 0.00 0.00 Q4 11 10.00 34.00 54.00 2.00 0.00 0.00 T cal I 13.57 35.00 35.64 7.14 1.07 8.60 ° 11 3.85 18.13 59.89 9.34 1.65 7.14 Significant Differences at Each Stage Between High and Low IRS Groups Lo