A COMPARISON OF DIFFERENT APPROACHES TO THE MEASUREMENT OF COUNSELOR EMPATHY IN PERSONAL COUNSELING Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY ROBERT R. KURTZ 1970 IIIIII IIIILIII II II III I II III—III I f . LIBRA R Y Idichigan Stet!) ' University This is to certify that the thesis entitled A Comparison of Different Approaches to the Measurement of Counselor Empathy in Personal Counseling presented by Robert R. Kurtz has been accepted towards fulfillment -. of the requirements for ph.d Counseling, Per. Ser. 8 Ed. Psy. ' degree in Qumflyfl A l/MAAAAJA Date __8_/_]3_/ I 0-169 ‘7 airtime av 3“ "MG & SUNS' BOOK BINDERY IND. Li»: ARY BINDERS ’ HIClllliAIi Leg: .‘ :—:«Jlll i 7 7 ‘ ‘h—vm “—17fi‘77,‘ , I L. _v- .¢ ABSTRACT A COMPARISON OF DIFFERENT APPROACHES TO THE MEASUREMENT OF COUNSELOR EMPATHY IN PERSONAL COUNSELING BY Robert R. Kurtz The present investigation was concerned with the counselor's empathic ability as a variable influencing the process and outcome of personal counseling. Using the same group of 31 counselors and clients, six different ways of measuring counselor empathy were compared with each other. Then, each measure was related to the processes and outcomes of personal counseling. The six different empathy measures were samples of operational definitions and procedures used in previous counseling research. Each counselor in the study was assigned a research client who took a series of personality inventories before and after counseling. The data were collected as a part of a larger research study conducted with the counselors and clients from the Michigan State University Counseling Center during the 1968-1969 school years. Robert R. Kurtz The empathy measures used in the study were: (a) counselor predictions' of client self-descriptions on the Leary Interpersonal Check List administered after the third interview. (b) counselor predictions of client self- descriptions based on constructs derived from a modified Kelly Rep. Test taken after the third interview. These two measures were called "predictive empathy." (c) Counselor scores on Kagan's et a1., Affective Sensitivity Scale (1967), which uses videotaped segments of actual counseling interviews as stimuli for the 89 multiple choice items on the questionnaire. This measure was called the "situation test" of empathy. (d) Ratings of counselor empathy by two judges using Carkhuff's Empathic Understanding in Inter- personal Processes Scale (1967). This measure was labeled "judged" or "rated" empathy. (e) Client perceptions of the counselor empathy using the Barrett-Lennard Relationship Inventory (1962), administered after the third interview and at termination. (f) The counselor perceptions of his own empathy based on the Relationship Inventory given after the third interview. These measures were called "perceived empathy." The six counselor empathy measures were correlated with each other, and no correlation reached the 5 per cent level of confidence, indicating that the six measures were not related. The client's perceptions of the counselor's empathy after the third interview did correlate .31 (p i 10) Robert R. Kurtz with the Carkhuff rating of judged empathy present in the counseling interviews, but the P values of the correlations between all other measures were larger than .10, and several of these were negative correlations. Three process measures were used to assess counseling processes in the study. The measures were: (a) ratings of client self exploration by two judges based on Carkhuff‘s Self Exploration in Interpersonal Processes Scale (1967), (b) ratings of client experiencing based on Gendlin and Tomlinson's Experiencing Scale (cited in Rogers' g£_gl., 1967) by one judge on 19 clients, and (c) counselor and client affective attitudes toward each other based on Snyder's (1961) Client Affect and Therapist Affect Scales which are 200 item true and false questionnaires. Correlations calculated between the six empathy measures and the process measures yielded significant corre- lations between rated counselor empathy and client self ex- ploration (r = .47); between client perceptions' of counse- lor empathy and client positive affect (r = .83); and between client perceived empathy and counselor positive affect (r = .45). The only process measure which con- sistently related to the outcome indices was the Client Affect Scale. Six measures of the client change were used as out- come variables. These were: (a) difference scores on two scales of the Tennessee Self Concept Scale (Fitt's, 1965), Robert R. Kurtz (b) judges ratings of pre- and post-MMPI Inventory scales for amount of improvement, (c) counselor ratings of success of counseling, (d) client ratings of how helpful counseling had been, and (e) a combined score of all of the indices. Correlation coefficients calculated between each of the empathy measures and outcome variables resulted in sig- nificant correlations (t = < .05) between the client per- ception empathy measure (Relationship Inventory) and four outcome variables. Positive correlations were found between the judge ratings of counselor empathy and outcome variables although only one correlation was high enough to reach a .05 level of significance. Multiple regression equations calculated for each outcome variable revealed that client perceptions of counse- lor empathy was the best predictor of counseling outcome, accounting for 44 per cent of the variance on the MMPI and 30 per cent of the variance on the combined outcome measures. Combining the different empathy measures did not account for more of the outcome variance except that judge ratings and client perceptions together accounted for more of the outcome variance and the Tennessee total positive score (a measure of self esteem). Conclusions The major conclusion emerging from this investigation is that the six different measures of counselor empathy, which sampled the operational definitions and procedures Robert R. Kurtz used in previous counseling research, were unrelated. This conclusion is of considerable importance to the interpre- tation of previous counseling research. Another more tentative conclusion is that client per- ceptions' of their counselors' empathy is the best pre- dictor of counseling outcome on a college student population undergoing personal counseling. This conclusion is in agreement with Rogerian theory, but it must be regarded as tentative since several other studies have demonstrated that judged empathy is the best predictor of counseling outcome. Client perceptions' of counselor empathy appear to be relatively stable over counseling and are highly related (r = .83) to the client's positive affect toward his counselor. Serious questions about the usefulness of the pre— dictive approach to the measurement of counselor empathy were raised by the study. The counselor judgement about his own empathy was concluded to be a less valid assessment of counselor empathy than assessments by the client or by unbiased judges. Another tentative conclusion was that one promising situational measure of counselor affective sensitivity was not related to counselor empathy in this counselor population. Finally, it was concluded that the rated counseling process measures in this study were of questionable validity. A COMPARISON OF DIFFERENT APPROACHES TO THE MEASUREMENT OF COUNSELOR EMPATHY IN PERSONAL COUNSELING . BY <- 6 Robert RQUKurtz A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services, and Educational Psychology 1970 DEDICATION To my wife Barbara, whose understanding and support throughout this dissertation helped me to personally realize how im- portant empathy and support are to a person. ii ACKNOWLEDGEMENTS The writer wishes to express sincere appreciation to Dr. Donald L. Grumman, dissertation chairman, for his original ideas which stimulated this research, and for his help in completing the study. The writer also wishes to thank Dr. Norman Kagan, guidance committee chairman, Dr. William Mehrens, and Dr. Bill L. Kell for their assistance on the doctoral committee. A special expression of gratitude goes to Dr. Bill L. Kell, a truly empathic counselor, for his support and understanding during the writer's professional training. This research would not have been possible without the assistance of the Michigan State University Counseling Center, and the many persons who worked on the research of which these data form a part. Dr. Normal Abeles, Dr. Donald Grummon, and Dr. William Mueller planned and organized the collection of the raw data. Dr. Forest Erlandson, Mr. Robert Harris, Dr. Richard Pierce, Dr. Samual Plyler, and Dr. Paul Schauble served as expert judges in various capacities for this and related studies; and Mrs. Judith Davids made the Kelly Rep test data available. iii Finally, the writer is indebted to the thirty-one counselors and clients who generously agreed to participate in the research. iv TABLE OF CONTENTS Chapter I. THEORETICAL PERSPECTIVES AND OVERVIEW OF THE STIJDY O O O O O O 0 Purpose . . . . . . . . Organization of Study . . . II. REVIEW OF RELATED LITERATURE . Research on Empathy in Counseling, and Psychotherapy . . . . Empathy and Therapist Experience Empathy and Counseling Process and Outcomes . . . . . Positive Results . . . Negative Results . . . Summary of Research . . . Definitions of Empathy . . . The Measurement of Empathy . Classification of Empathy Measures Classification and Definitions Used in the Present Study . Early Research . . . . . Predictive Approaches . . Situation Approaches . . . The Rating Approach . . . Perceived Empathy Approaches Previous Comparative Studies Summary . . . . . . . . III. PROCEDURES AND STATEMENT OF RESEARCH QUESTIONS 0 o o o o o o 0 sample 0 o o o o o o o Counselors . . . . . . Clients 0 O O O O O O Page 21 24 24 26 26 35 40 44 47 48 51 51 51 52 IV. Collection of Data . . . . . . . Measurement Instruments . . . . . Measures of Empathy . . . . . . The Situation Test. . . . . . The Predictive Tests . . . . . Tape Ratings or Judged Empathy. . Perceived Empathy . . . . . . Process Measures . o o . . o 0 Self Exploration . . . . . . Experiencing . . . . . . . Therapist and Clien Affect . . Outcome Measures . . . . . . . Tennessee Self Concept Scale . . Minnesota Multiphasic Personality Inventory . . . . . . Counselor Evaluation Form . Client Evaluation Form . . Combined Outcome Score . . Research Questions . . . . . . . Relationships Between Empathy Measures . . . . . . . . .- Relationship Between Counselor Empathy and Counseling Processes . Relationship Between Process and Outcome . . . . . . . . . . Relationship Between Counselor Empathy and Outcome . . . . . . RESULTS AND DISCUSSION . . . . . . Results . . . . . . . . . . Relationship between Empathy Measures . . . . . . . . . Relationship between Counselor Empathy and Counseling Processes . Relationship between Counseling Processes and Outcome . . . . . Relationship between Counselor Empathy and Counseling Outcomes . . vi Page 52 54 54 54 56 57 59 60 60 62 64 66 66 68 69 69 70 70 7O 73 74 74 76 76 76 80 83 85 Discussion . . . . . . . Concept . . . . Predictive Empathy Situational Empathy Judged Empathy . . Perceived Empathy . Counselor Empathy as a Measured Counselor Perceptions . . Client Perceptions . . . Empathy and Rated Counseling Processes . . . . . . Counseling Outcomes . . . Limitations of the Study . V. GENERAL SUMMARY AND CONCLUSIONS REFERENCES APPENDICES APPENDIX A. Letters to Research Clients and Instruc- General Summary . . . . . Conclusions . . . . . . tions to Intake Counselors . . B. Number of Interviews per Client C. Empathy Measures Used in Study . D. Raw Scores Obtained by Bach Counselor on the Six Empathy Measures . . E. Calculation of Interjudge Reliability for Carkhuff Empathic Understanding in Interpersonal Processes Scale . F. Counselors and Clients Evaluation of Counseling Scores . . . . . G. Means, Standard Deviations, and Range of Empathy Scales . . . . . . H. Process Measures Used in Study . I. Scores on Process Measures . . J. Ranks and Scores of the Low Scoring Counselors (Lowest Seven) on the Affec- tive Sensitivity Scale Compared to Ranks and Scores on Other Empathy Measures . . . . . . . . vii Page 89 89 91 92 94 95 95 95 97 98 99 101 101 105 107 115 115 121 122 134 140 141 142 143 148 150 Raw Score Differences on Pre- and Post- Testing on Tennessee Self Concept Scale (Fitts) . . . . . . . . Instructions to Judges for Pattern Analysis of MMPI Profiles and Calculations of Inter-Judge Relia— bility . . . . . . . . . . MMPI Profile Analysis . . . . . Calculation of Interjudge Reliability for Carkhuff Self Exploration in Interpersonal Processes Scale . . . Conversions of Outcome Scores to Standard Scores for Combined Outcome Score . . . . . . . . . . Least Squares Addition for Prediction of Outcome Analysis of Variance Tables for Overall Regression for all Outcome Measures . . . . . . . . viii Page 151 152 155 156 157 158 Table 1. LIST OF TABLES Summary of Therapy Outcome Studies Using Accurate Empathy Scale . . . . . Correlations Between Empathy Measures Correlations Between Empathy Measures and Process Measures . . . . . Correlations Between Process Measures Correlations Between Process Measures and Outcome Measures . . . . . Correlations Between Empathy Measures and Outcome Measures . . . . . Least Squares Multiple Regression for Prediction of Outcome From Empathy Measures 0 O O O I O O O 0 Correlations Between Outcome Measures Number of Interviews per Client . . Raw Scores Obtained by Each Counselor on the Six Empathy Measures . . . Inter-Judge Reliability for Carkhuff Tape Ratings . . . . . . . . Counselors and Clients Evaluation of Counseling Scores . . . . . . Means, Standard Deviations, and Range of Empathy Scales . . . . . . Scores on Process Measures . . . . Ranks and Scores on the Low Scoring Counselors (Lowest Seven) on the Affective Sensitivity Scale Com- pared to Ranks and Scores on Other Empathy Measures . . . . . . . ix Page 16 77 81 83 86 88 89 121 134 140 141 142 148 150 Table Tennessee Self Concept Difference Scores . . . . . . . . . . Inter-Judge Reliability MMPI Data . MMPI Profile Analysis--Average Ratings Inter-Judge Reliability for Self Exploration Tape Ratings . . . . Post t Scores--for Combined Outcome Score . . . . . . . . . . ANOVA for Overall Regression Between Relationship Inventory Client Per- ceptions and Tennessee Self Concept Total "P" Differences Scale . . . ANOVA for Overall Regression Between Carkhuff Ratings, Relationship Inventory Client Perceptions and Tennessee Self Concept Total "P" Difference Scale . . . . . . ANOVA for Overall Regression Between Relationship Inventory Client Per- ceptions and Tennessee Self Concept, Number of Deviant Signs Differences ANOVA for Overall Regression Between Relationship Inventory Client Per- ceptions and MMPI Ratings . . . ANOVA for Overall Regression Between Relationship Inventory Client Per- ceptions and Counselor Evaluations ANOVA for Overall Regression Between Relationship Inventory Client Per- ceptions and Combined Outcome . . Page 151 153 155 156 157 158 158 159 159 160 160 CHAPTER I THEORETICAL PERSPECTIVES AND OVERVIEW OF THE STUDY Orlo Strunk (1957, p. 53) concluded his review of the literature on empathy by saying: "There is nearly ’\// complete agreement on the part of psychologists and social psychologists that empathy is an important aspect of both personality study and social intercourse and as such de- serves priority in terms of research." Empathy is important I in counseling and psychotherapy as well, and many authors of diverse theoretical orientations stress the counselor's or therapist's empathic ability as one of the necessary prerequisites to effective counseling and therapy (e.g., May, 1939; Fromm-Reichman, 1950; Rogers, 1957; Truax & Carkhuff, 1967; Tyler, 1969) . Despite this agreement on the importance of the concept, little agreement has been reached about the operational definitions and measurement of empathy. The research literature reveals several different ways of conceptualizing and measuring empathy, and it is not altogether clear whether these studies deal with one variable or several different variables employing the same label. Few studies have compared these alternative approaches to the measurement of the concept. Buchhiemer (1963, p. 61) concluded that "empathy is an intervening variable. It is both an inferential and intuitive process. Consequently, the nature of the operational definitions, the kinds of metaphore employed, and the criteria used be- come very important." The present investigation is concerned with the counselor's empathic ability as a variable influencing the process and outcome of personal counseling. Using the same group of counselors and clients, this study compares six different ways of measuring counselor empathy, and examines how each is related to the process and the outcomes of personal counseling. The six different methods of measuring empathy sample the operational definitions and procedures employed in previous counseling research. The study is thus a replication of previous research relating counselor empa- thy to counseling process and outcome, but more importantly, it provides one basis for clarifying the concept of empathy and for determining which operational measures are most useful in counseling research. Purpose More specifically, the purpose of this study is: (a) to examine the relationships between six different measures of empathy which have been used in previous re- search on counseling and psychotherapy, (b) to examine the relationship between each of the empathy measures and counseling processes, (c) to examine the relationship be- tween client change over personal counseling and each measure of empathy and, (d) to determine which empathy measure or combination of measures is the best predictor of client change. By investigating the interrelationships between the different empathy measures and their relationships to counseling process and outcome, it is hoped that the construct of empathy as it applies to counseling and psy- chotherapy will be clarified. Organization of Study The following chapter will include a review of the literature related to the study. Specific issues of direct relevance to the study emerging from the review will be discussed. Chapter III will contain a report on-the pro- cedures used in the study and a statement of the research questions. The results and discussion will be reported in Chapter IV. Chapter V will contain a summary of the study and conclusions drawn from the results. Since the present investigation is a comparative study, examining relationships between operational measures rather than testing specific hypotheses, the results and discussion will be organized around general and specific research questions rather than around statements of hy- potheses to be tested. CHAPTER II REVIEW OF RELATED LITERATURE The literature on empathy is briefly reviewed in this chapter with special emphasis on different ways of defining and measuring counselor empathy as a variable in- fluencing counseling process and outcome. An attempt is made to clarify the specific issues of direct relevance to the present study emerging from the review. Research on_§mpathy in Counseling, and Psychotherapy Although many theorists have written about the im- portance of empathy in counseling and psychotherapy, Rogers (1957. 1959) formulated the most explicit theoretical statement about the relationship between empathy and counseling process and outcome. He specifies those aspects of the therapist's attitudes and behavior, one of which is empathy, which have a direct influence upon therapeutic process and outcome. The greater the counselor's empathy, the more likely it is that the counseling processes will get started, which in turn, produces more client change. The Roger's position, plus the writings of other theorists have stimulated considerable research upon the relationship between empathy and effective counseling. Empathy and Therapist Experience Fiedler (1950) compared the therapeutic relation- ship created by 10 experts and 10 non-experts from three schools of therapeutic thought by means of rating recorded interviews. Three trained judges rated these tapes on a variety of relationship dimensions. 'This was accomplished by arranging 75 statements along a Q-sort continuum for each therapist. It was found that regardless of which therapeutic school was represented, that items such as the following placed high in the descriptions of relationships created by experienced therapists: The therapist is well able to understand the patient's feelings. The therapist is never in any doubt about what the patient means. The therapist's remarks fit in just right with the patient's mood and content. ' The therapist's tone of voice conveys the complete ability to share the patient's feelings. Although Fiedler doesn't label this dimension as empathy, it seems that one of the important characteristics which differentiates experts from non-experts in the thera- peutic relationship is this ability to understand and com- municate with his client. Fiedler's descriptions have been used in other re- search studying counselor understanding and outcomes of counseling (Lesser, 1958; Gonyea, 1963). Empathy and Counseling Process and Outcomes Positive Results A number of research studies found positive relationships between counselor empathy and counseling and psychotherapy processes and outcomes. Cartwright and Lerner (1963) explored the relation- ship between the therapist's empathic ability, the clients need to change, and improvement in psychotherapy. One of the primary hypotheses in this study was: the empathic understanding of the patient is directly related to the degree of improvement in the patient with psychotherapy. Empathic understanding was measured early in counseling and later by having counselors predict what their clients' responses would be on a scale developed from the Kelly Rep. Test. The constructs derived from the Rep. were arranged on a polar continuum. The clients would check closest to the construct they felt described them. Improvement was measured by ratings on four components: (1) patient's in- tegration, (2) defensiveness vs. openness, (3) present life adjustment, and (4) therapist's post ratings of therapy outcome. Cartwright and Lerner found no relationship between client change and early therapist prediction of client responses. Predictions made a second time, later in therapy, were significantly related to improvement in therapy. The investigators concluded that later empathy was more crucial for client change than empathy early in therapy. Dombrow (1966) investigated the relationship be- tween the therapist's ability to predict their clients' Q-sort self descriptions and changes in client self concepts as a result of therapy. Differences between pre- and post- Q-sort descrepancies between present and ideal self were used as indexes of change. Twenty-nine mothers in treatment at a child guidance clinic and their 14 social work thera- pists were the subjects used in the study. Results showed a significant relationship between the therapist predictive ability and client change in self concept. The actual changes observed were reported to be small after six to 10 months of therapy. "Assumed simi- larity" between the therapist and client had no Significant relationship to client change. This result differed with the findings of some previous research. Significant relationships were found between counselor empathy (plus respect and genuineness) on client self exploration by Hountras and Anderson (1969). The 27 male and 27 female undergraduates seeking counseling were assigned to one of three appropriate problem categories (educational, vocational, or personal-social). Each client then saw one of nine counselors on the staff. Counselor empathy was rated by three judges using the Carkhuff Empathic Understanding in Interpersonal Processes Scale. Client self exploration was rated by three judges using the Truax Depth of Self Exploration S2313, Significant correlations were found between counse- lor empathy and client self exploration in all problem categories. The study provided strong support for the hypothe- sis that counselor empathy affects client's depth of self exploration on a college client population. Unfortunately, the relationship between depth of self exploration and resolution of the client's problems was not investigated. Dickenson and Truax (1966) investigated the effect of counselor empathy (as part of a combined measure) and improvement of academic achievement for college under- achievers. The 24 students who received group counseling twice a week for 12 weeks showed greater improvement in grade-point average than 24 matched, non-counseled control subjects. Further, those counseled students who received the highest therapeutic conditions tended to show the greatest improvement. Counselor empathy was measured by ratings on the Accurate Empathy Scale (Truax, 1961) by two judges. It was difficult to determine the effects of counselor empathy per se on outcome in this study, since the Accurate Empathy ratings were combined with other counselor variables (genuineness and regard) in the analysis. Hansen, Moore, and Carkhuff (1968) rated tapes of nine multiple counseling groups conducted by high school counselors with the Accurate Empathy Scale. The clients were 70 high school students from different high schools. They found that ratings on the scale correlated with client change based on pre- and post-self concept Q-sort de- scriptions (Baymuir & Patterson, 1954) at r = .82. Barrett-Lennard (1962) was also interested in exploring the relationship between the "necessary and suf- ficient conditions" proposed by Rogers and outcome. This investigator approached the problem from a different per- spective, however. The general postulate was that the clients' experience of his therapists' response was the primary locus of therapeutic influence in their relationship. In order to measure these perceptions, Barrett- Lennard developed the Relationship Inventory which was de-. signed to provide scales for five therapist variables, one of which was empathic understanding. The inventory con- tained statements which could be responded to with three grades of "yes" and three grades of "no" responses which were identified as +1, +2, +3, and -l, -2, -3; respectively by the respondent. Both therapist and client forms were prepared. Relationship data was gathered from clients and from therapists after the first five therapy interviews, after 15, 25, and at termination of therapy. The total sample 10 consisted of 42 clients at the Counseling Center of the University of Chicago and their 21 separate therapists. The degree of change in the client during therapy was assessed by therapist rating measures, scores on Q-sort- adjustment, the Taylor Manifest Anxiety, and Minnesota Multiphasic Personality Inventory Depression scales. Two composite indices of change were used--one derived from therapist ratings and the other from the client self- descriptive data. The 16 cases falling in the upper category of change on both component indices perceived their therapist as having greater empathic understanding (M = 29.3) than those falling in the lower category (M = 16.4). This relation- ship also held when the therapists perceptions were em- ployed, but it was found that the client's perceptions were a better predictor of outcome than those of the therapists. Barrett-Lennard also confirmed Fiedler's findings that the more experienced therapists were perceived as being more empathic by themselves and by their clients. A number of studies have used data from the research project on psychotherapy with 16 hospitalized schizophrenic patients at the University of Wisconsin under the leadership of Rogers (1967). The findings of these studies have con- tributed a great deal to the research on therapist empathy as it effects therapy processes and outcome. The design, results, and conclusions of this very complicated and 11 carefully designed research program are too extensive to report in this review, however. Therefore, only the con- clusions and findings of direct relevance to this study will be summarized. The complete research report is presented in Roger's ep_pl., The Therapeutic Relationship and Its Impact, 1967. Therapist empathy was measured in the project in three ways. First, judges rated therapists on the Accurate Empathy Scale (Truax, 1961). It was also measured by client's perceptions of their therapist's empathy on the Barrett-Lennard Relationship Inventory. Finally, therapists rated their own empathy on the Relationship Inventory. The major process measure used in the project was cast in terms of the experiencing dimension proposed by Gendlin (1962). It was operationalized in the form of a seven—point rating scale. At the lowest levels of experi- encing the patient makes no personal reference or little or no expression of feeling. At the higher levels, the patient experiences his feelings, understands them, and in- tegrates them into a meaningful framework. A great number of outcome indeces were used in the project. These included: the Minnesota Multiphasic Person- alipy Inventory, the Butler-Haigh Q-sort, the Wittenborn Psychiatric Rating Scales, Therapist Rating Scales, Hospi- talization Status, the Thematic Apperception Test, and others. 12 In general, it was found that: (l) judge ratings of therapist empathy (Accurate Empathy Scale--AE) were significantly correlated to the level of process the pa- patient engaged in; (2) patient perceptions of therapist empathy were moderately associated with process level; (3) judge ratings (AE) and patient perceptions were re- lated (Initial, r = .71; Terminal, r = .38); (4) therapist perceptions of their empathy were negatively correlated with judge ratings (Initial, r = -.52; Terminal, r = -.12), and non-significantly with patient perceptions (Initial, r -.53; Terminal, r = .02); (5) therapist perceptions of their empathy was negatively correlated with level of process (Initial, r = -.47; Terminal, r = -.37) the patient was engaged in; (6) patient perceptions of therapist empathy tended to remain stable over therapy; (7) the process level exhibited by patients was positively asso- ciated with many measures of outcome, although patients did not move on the process dimension as a group (to greater experiencing); (8) the patients receiving the highest level of accurate empathy (judge rating) in their relationship showed the greatest reduction in schizophrenic tendencies as measured by the Minnesota Multiphasic Person- ality Inventogy. Some of the findings regarding the relationship between therapist empathy as measured by the Accurate Empathy Scale were reported in another source (Truax & Carkhuff, 1967). 13 Truax and Carkhuff (1967) reported a series of studies designed to explore the relationship between therapist empathy and therapy outcome. Therapist empathy was measured by the Accurate Emppthy Scale (Truax, 1961). One of the first studies compared the levels of accurate empathy provided to four hospitalized patients who showed clear improvement on a variety of personality tests with four who showed personality deterioration after six months of therapy. Rating of the therapy sessions was done on the Accurate EmpathyScale (1961). A total of 384 two-minute samples were selected from the middle third of the recorded sessions and randomly assigned to raters. The findings suggested that the psychotherapists whose patients improved on the personality tests rated consistently higher on accurate empathy than those with test-deterioration scores (p < .01). The level of accurate empathy also tended to remain at a consistent level throughout the six months of psychotherapy. This initial finding relating accurate empathy to case outcome was extended by later studies (Truax, 1963) involving 14 schizophrenic patients who were in psycho? therapy for periods ranging from six months to four years. Five raters (undergraduate students) who were em- Ployed on the scale had no knowledge of the therapist, the Patients, the case outcomes, or the order in which the saHHPIes were given to them for rating. 14 A comparison was made between the mean level of accurate empathy offered by the therapist in each case and personality and behavioral change in the patient. The final outcome criterion included psychological test change data, diagnostic evaluations of personality change, and a measure of time actually spent in the hospital since the initiation of therapy. This criterion correlated with accurate empathy ratings at r = .77. A second outcome cri- terion based upon blind evaluations of change in personality functioning, interpreted from pre- and post-Rorschach and MMPI tests by two psychologists, correlated with the level of accurate empathy offered by the therapist at r = .48. In an effort to extend to an outpatient population the findings obtained with hospitalized schizophrenics, Truax (1963) obtained an additional 14 cases (seven rela- tive successes and seven relative failures) from Stanford University and the University of Chicago. Using samples obtained from early and later interviews with the 14 hospi- talized and the 14 outpatient cases, he found that the level of accurate empathy offered in therapy was significantly higher for successful cases than for failures. This posi- tive relationship between accurate empathy and outcome of therapy held for both hospitalized schizophrenics and out- Patients seen in counseling. In particular, the failure cases had a high frequency of therapist responses character- ized by inaccuracy in responding to "preconscious" material. 15 Table 1 presents a summary of the studies done on therapy outcome using the Accurate Empathy Scale as the independent variable. As can be seen from the table, many of the studies support the hypothesis of a relationship be- tween accurate empathy ratings and therapy outcome. Negative Results Not all investigations have found a significant relationship between empathy and counseling outcome. Katz (1962) investigated the relationship between two different measures of empathy and client change in short term counsel- ing. Using 21 high school students and 21 counselors in- training as subjects, Katz took measures of the counselor's predictive empathy and behavioral empathy. Predictive empathy was measured by the counselor's ability to predict how his client would arrange the Self-Concept Q-sort. Be- haviorally, empathy was measured by judges ratings of counselor behavior within the interview on the Rating Scale of Empathic Behavior. Two measures of client change were employed. The first was the difference in scores on the Self-Concept Q-sort which each client took before and after counseling. The second was the difference in judges ratings of client behavior within the initial and terminal interviews of the counseling sequence on the Rating Scale of Counselee Be- havior. Clients in the study had at least six interviews. 16 nwma .m INwm pom mdaaomcooo m>wuoomwm pum3oa "ommowso .xmsua .m .aamH .ocHamHHnmm mcHaHa .mmsfiumo .m a .oofluomnm can mcficflmnu .0 Beam cmxou magmas wNmmumouono ow ucmnvcwama o H as HmuHmmom macho AammHo segue Hommav mwoaxnou o o om ucmwummuoo moouw pom comma one xmsua Ammmav o H om ucosvcaamo macho omuoz can xmsua Ammmav o H oma Hmuwmmom moonu omumz can stua “momHv :msoox o o ov Hmuammom macaw can mmosxumo .xmsue laach macaw ace .nmmz .owummlcnmom .mHuumm .umnsH o N ov ucmwummuoo .ocH .xcmum .omumz .xonue o H aH HmHHmmom .aaH lmomHv xmsua o H «H ucmHummuso .aaH HmmmHv xmsua :oHumon>m owumocmofio .nomsomnom .Hmzz .mumma o m «H Hmqumom .ocH “momHv swans aoHumaHnsoo o H m HmuHmmom .caH lemHV xasus mmo. mvm MMO.m av m . m umcwm 4 . m cfluo> SOHO monommmz monomowz 1mm mousmmoz z ucowwu MW .psHV mooum mo mm a mEoouoo vocab meoouso pecan mo 0 B ucmsvomua 1800 Hamum>o IEOU Haoum>o «.mamom anummEm wumusoom moans mmwooum mEoouso ammuosb mo mumEESmII.H mqmde 17 Katz found no significant relationship between the counselor's predictive or behavioral empathy and client change in short term counseling. The author concluded that the study raised questions about whether empathic ability of the counselor is related to client change at least in short term counseling. Lesser (1958) explored the relationship between empathic understanding and progress in counseling. The counselor's empathic understanding was measured by means of an Empathic Understanding Scale. The 12 statements of the scale, were rated by both counselors and clients on a seven point continuum from most to least. These statements were obtained from Fiedler's (1950) statements of therapeutic relationships. The investigator also had the counselors predict what the client's initial Q-sort would be as a measure of the counselors ability to predict his client's self-concept. Differences in pre- and post-Q-sorts were used as an index of client change. In general, Lesser found no sig- nificant relationship between empathy as measured by the Empathic Understanding Scale, or the counselor's prediction of his client's Q-sort and outcome. Gonyea (1963) found no significant relationships be— tnveen counseling outcome and the quality of therapeutic re- .lationship provided by eight counseling interns. Counsel- ling'supervisors were asked to sort statements (Q-sort) 18 describing the therapeutic relationship established by each of the interns they supervised. These statements were the same used by Fiedler (1950) in his study of experienced and inexperienced therapists. As reported earlier, the state- ments which differentiated experienced from inexperienced therapists centered around the therapist's ability to under- stand his client and communicate that understanding. The resulting Q-sorts were then correlated with Fiedler's de- scriptions of experienced therapists to provide a measure of the therapeutic relationships provided by each counselor. A measure of improvement of self-concept, work and study adjustment, and social reactions (Self Description F255, Gregg and Kelley) was used as a measure of counseling outcome. The 208 clients seen by the eight interns in the study took the measure before and after counseling. Posi- tive changes on the scale were considered improvement. These outcomes were correlated with the supervisor's rat- ings of the counselors. No significant relationships were found between the ratings of the counselors and the counseling outcome. It was impossible to isolate the counselor empathy variable from others influencing the supervisors ratings in this study. Fiedler's statements, while containing many references to counselor understanding, include state- ments of other variables as well. 19 Summary of Research The review of the research on the relationship be- tween counselor empathy and counseling process and outcome revealed mixed results. Some studies show a clear relation- ship between counselor empathy and counseling process and outcome (Cartwright & Lerner, 1963; Dombrow, 1966; Hountras & Anderson, 1969; Hansen, Moore,£§Carkhuff, 1968; Barrett- Lennard, 1962; Truax&Carkhuff, 1967), while others (Katz, 1962; Lesser, 1958; Gonyea, 1963) found no significant relationship. A number of alternative explanations might be sug- gested to explain the mixed results of these research studies. First, the different studies sampled quite differ- ent populations. Truax used hospitalized patients, Cart- wright, Lerner, and Lesser used University Counseling Center populations, whereas Katz used high school students. Indi- cations are that the counselor competence in the various studies was widely divergent as well. Another difference in the studies was the outcome criteria used as an indica- tion of counseling progress. Some studies used combinations <3f personality test scores and rating scales (e.g., Truax, 1967; Barrett-Lennard, 1962). Others used changes in client Q-sort self concept (e.g., Katz,l962; Dombrow, 1966; Lesser, .1969) and some process measures (e.g., Rogers, 1967; Ikauntras & Anderson, 1969). 20 Finally, many of the studies used different opera- tional definitions and measures of counselor empathy. Katz, Cartwright and Lerner, and Dombrow asked their counselors to predict their clients' responses on certain personality measures as an indicator of empathy. Gonyea used supervisor ratings of counselor understanding. Barrett-Lennard used a post interview questionnaire to ask clients how empathic their counselor had been. Hountras and Anderson; Hansen, Moore, and Carkhuff, and Truax used judge's ratings of counselor empathy based on the Accurate Empathy Scale. Katz also used a rating scale which was different from Truax's scale and based on the concept of confluence. The different operational procedures and measures used in the different studies were purporting to measure empathy or some similar concept as an indepedent variable. If indeed, the measures were measuring the same variable, then the differences in the operational procedures between them are of little importance. If, on the other hand, the different measures are not related, then the differences in the reported results could be due to the fact that the same independent variable was not being measured in the differ- ent studies. This hypothesis is one of the concerns of the ‘present investigation. Often, the different operational procedures used in research reflect disagreement about the basic conceptions and definitions of empathy. 21 Definitions of Empathy Attempts to define empathy have been numerous and varied. A standard dictionary (English & English, 1958) of psychological terms defines empathy as:‘ Apprehension of the state of mine of another person without feeling (as in sympathy) what the other feels. While the empathic process is primarily intellectual, emotion is not precluded, but it is not the same emotion as that of the person with whom one empathizes. The parent may empathize with the child's puny rage, feeling pity or amusement, whereas in sympathy he would feel rage along with the child. The attitude in empathy is one of acceptance and understanding of an implicite "I see how you feel." The above definition emphasizes the intellectual, objective aspects of empathy. This objective, detached, although con- cerned attitude is what differentiates empathy from sympathy according to the definition. Some psychoanalytic defini- tions also stress these elements of the process. Fenichel (1954, p. 511) states that empathy " . . . consists of two acts: (a) an identification with the other person, and (b) an awareness of one's own feelings after the identification, and in this way an awareness of the objects feelings." Some current analytic writers still accept this definition (Chessick, 1965). Dymond (1948) stressed the role-taking aspect of the empathic process in her definition. She defines it as " . . . the imaginative transposing of oneself into the 'thinking, feeling, and acting of another and so structuring 'bhe world as he does." 22 Some definitions stress the interactive component of the empathic process. Sperroff (1953) defined it as " . . . the ability to put yourself in the other person's position, establish rapport, anticipate his feelings, reactions and behavior . . . empathy and role reversal are mutually com- plementary." This definition differs from Dymond's in that it not only requires the ability to put oneself in another position, but adds the dimension of interaction. Roger's (1951, p. 29) included a similar interaction dimension in his definition of empathy. "It is the counselor's function to assume in so far as he is able, the internal frame of reference of the client . . . and to com- municate something of this empathic understanding to the client." This definition stresses the importance of the communication of understanding or empathy as an important aspect of the process. In an attempt to contrive a more generic paraphrasing and, at the same time, to avoid the semantic confusion in- herent in general definitions of empathy, Kagan, Krathwohl and Farquhar (1965) defined a more restricted trait of "affective sensitivity" as "the ability to detect and de- scribe the immediate affective state of another, or in terms of communication theory, the ability to receive and decode affective communication" (cited in Kagan, et_21,, 1967, p. 463). This ability as defined, is a component of the larger concept of empathy. Other writers have also 23 identified different components of empathy in an attempt to clarify the broad concept (Cronbach, 1955; Smith, 1966). These efforts , while useful, have not clarified the concept sufficiently to bring an agreement about its definition and methods for measurement. Some investigators have employed different terms for what appears to be a concept similar to empathy. One such term often used is "sensitivity" or "social sensitivity." Bender and Hastorf (1952, p. 574) defined social sensitivity as the "perceptions of the thoughts and feelings of other people. . . ." Smith (1955, p. 3) defined sensitivity as "the ability to predict what another will feel, say and do about you, himself and others." Nortcutt and Silva (1951) in a study of the ability to predict the personality characteristics and needs of others, described this ability as insight, i.e., "ability to predict must depend on insight into the meaning and pat- tern of the subject's behavior, not merely on detailed knowledge of his habits or memory for small facts [p. 32]." Hogan (1969) had 14 non-psychologists and nine graduate students in psychology arrange 50-Q-sort behavioral descriptions of how they saw an "empathic man." He found that the descriptions correlated at r = .53. The investi- gator concluded that the group had common behavioral refer- ents for the concept. 24 These descriptions also correlated with Q-sort de- scriptions of the "good man" and the "well adjusted" man. Therefore, there seems to be some overlay in terms of be- havioral referents to these concepts. The general idea of the ability to put yourself into another person's place seems to run throughout the definitions of empathy. Such global agreement leaves much to be desired, however. Some definitions tend to stress the predictive element of the concept where others stress the communicative element. Disagreement seems to run highest over differences in the shades of meaning and in differentiating empathy from other concepts. Again, since empathy must be inferred, operational definitions become crucial to understanding what a particu- 1ar investigator means by empathy. This leads to the problem of suitable measuring instruments about which there is considerable disagreement. The Measurement of Empathy Classification of Empathy Measures Several authors have attempted to differentiate and classify different empathy measures. Astin (1957) dis- tinguished between the "Predictive Empathy Test" and the "Situation Test of Empathy." The predictive test is one which involves the prediction of the behavior of another. The judge is either acquainted with the subject or is given 25 data about him. The judge's task is to predict the subject's performance on test items or his responses to personality, or attitude inventories. In contrast, the Situation Test of Empathy provides the subject with some standardized situation to which he is to respond, em- pathically. More specifically, subjects are presented with audio tape recordings, films, video tapes, role played sessions, and/or typescript materials. These materials provide the standardized situation to which the subject re- sponds. Kagan and Krathwhohl's e£_pl. (1967) review of the literature on empathy uses the same classification of empathy tests as Astin. Chessick (1965) defined the "retrospective approach" to the measurement of empathy as one where the counselor is asked to rate his empathy in a particular interview after the completion of that interview. In a similar way, the client can be asked to report how empathic his counselor was during a particular interview. Barrett-Lennard (1962) labeled this procedure "experienced" or "perceived" empathy and deve10ped a questionnaire for both client and therapist to rate the therapist's empathy. Katz (1962) distinguished between the predictive ap- proach (as defined earlier) and what he termed "behavioral empathyfl' The latter was measured by having trained judges listen to or observe interactions between counselor and client. On the basis of certain defined criteria, the 26 judges rated the amount of empathy displayed by the counselor in the interaction. Classificatipn and Definitions Used in the Present Study This study will use some of the same definitions and classifications of empathy measures described by the previ- ous authors. The "Predictive Approach" will refer to those measures which involve prediction of another person's re- sponses. "Predictive Empathy" refers to the variables measured by the predictive approach. The "Situation Ap- proach" will refer to measures which create a standardized situation to which the counselor responds. "Perceived Ap- proaches" will refer to the procedure of having clients and counselors rate how empathic the counselor had been during the previous interview. "Perceived Empathy" refers to the perceptions of counselors and clients measured by the "Per- ceived Approach." The "Rating Approach" refers to the pro- cedure of having trained judges rate counselor empathy. The results of this procedure will be called "Judged Empathy." Early Research Early studies concerned with the measurement of empathy dealt with the question of recognition of emotions and characteristics of an individual. Often, the subject was presented with photographs of peOple and was asked to 27 indicate what emotion was being expressed or was asked to judge some other characteristics of the person on the photograph (Busby, 1924; Kanner, 1931). By 1938, Estes concluded that the validity of these judgements based on first impression static poses were al- most worthless. In an effort to correct this, Estes con- ducted a series of six experiments using silent motion picture segments of subjects engaged in various activities (e.g., building a "card house"). The 323 judges were to estimate personality characteristics of the subjects on film. The judgments were validated against elaborate cri- teria provided on the filmed subjects as a result of co- operative research investigation of their personalities. Estes concluded that the accuracy of judgments varied with: (1) the judge, (2) the subject, and (3) the aspects of per- sonality being judged. It is likely that the activities which the filmed subjects engaged in did not provide a relevant sample of behavior to judge from. Garden (1934) reported a device for demonstrating empathy. It consisted of four photographs of a Mexican figure which had one arm raised. The negatives of the four images were reversed and printed so that a total of eight photographs were shown to subjects. Subjects were asked to indicate which arm, right or left, was raised. They were then observed to detect whether they made any overt physi— cal responses in an attempt to mimic the images in the 28 photographs. These mimicking responses were taken to mean that the subjects were attempting to "feel" with the photo- graphed images. Rather than being a measurement of empathy, the device was used more to demonstrate the concept. Steinmetz (1945) proposed using personality tests a different way. He discussed the possibilities of having one subject answering the questions on the Guilford-Martin Per- sonality Inventory as if he were another person as a way of measuring the psychological understanding ability of the subject. This proposal may have been a forerunner to the early predictive studies of empathy. Predictive Approaches Predictive approaches have been subdivided into those involving predictions based on a "specific other person" and those based on a "generalized other" such as predictions about how the average member of a particular group will respond. The majority of studies asked the judge to make predictions about people he had seen and talked to but some studies have used written descriptions, filmed and taped interviews (Smith, 1966). The responses which are to be predicted are usually test responses on some type of self report inventory but may be predictions of other behavior as well. In an early attempt to develop an empathy measure utilizing this approach, Dymond (1949) divided 53 subjects 29 into five groups. After they had met three times for a class project, the investigator had the subjects rate each other on a number of traits including: self confidence, superiority-inferiority, selfishness, friendliness, leader- ship, and a sense of humor. The following procedure was used: (1) A rates himself on the traits, (2) A rates B as he sees him, (3) A rates B as he thinks B would rate him- self, and (4) A rates himself as he thinks B would rate him. This procedure was followed by all subjects, rating the other members in their group. Empathy was defined as the ability of the subject to predict how the others in his group would rate themselves and rate him. Many studies followed Dymond's example by using simi- lar procedures (Bender & Hastdorf, 1950; Cowden, 1955; Helpern, 1951, 1955). Some investigators began to point out problems associated with this approach, however, and a number of questions were raised about its adequacy. Hastorf and Bender (1959) and Cowden (1955) questioned whether the empathy scores obtained by this method were not confounded by variables such as projection. In other words, is the subject predicting the responses of another or is he just describing his own responses and pro- jecting them on to another? Lindgren and Robinson (1953) question whether the procedure measures empathic ability or the ability to re- spond to the cultural norm. They found that accurate 30 predictions were highly correlated with the average re- sponses given by the group. Halpern (1955) investigated the relationship between similarity and predictive empathic ability. In other words, this investigator hypothesized that subjects would be better able to predict the responses of people who were similar to them or who answered the self report inventory measures much like they did. His hypothesis was supported in a study involving 38 student nurses. Two other studies investigating the same hypothesis (Inglis, 1966; Lesser, 1958) did not find evidence to support this hypothesis, however. Because of these criticisms, the predictive approach has not been used as extensively as other approaches in the current research (Hobart, 1965). The approach has some appeal, however, because it requires the subject to assume another person's perspective and to respond like the other person would. This ability appears to be what many defini— tions of empathy describe. According to Wittich (1955) and Taft (1955), the ability to predict the responses of another may be regarded as a trait. This conclusion rests on the assumption that persons will demonstrate strong generality in that behavior across similar situations (Mischell, 1968). In other words, it is assumed that a counselor showing high predictive ability in one situation will also be high in this ability 31 in other predictive situations. How valid is this conclusion? Assuming that this ability to predict the responses of another is a trait, then what is the relationship of this ability to other measures of empathy? Is high pre- dictive ability related to the way a counselor interacts with his client from moment to moment in an interview? Or, is this predictive ability separate and unrelated to the abilities measured in other ways? A number of studies (Astin, 1957; Katz, 1962; Lesser, 1958) suggest that pre- dictive ability is unrelated to other measures of empathy. If predictive ability is unrelated to other measures of empathy and also unrelated to process and outcomes in counseling, then this method of assessing empathy has little relevance to counseling research. A conclusion such as this would be important since much of the research on empathy has utilized the predictive approach. In order to investigate the above questions, two different predictive measures of empathy will be given to the counselors in this study. These measures will be cor- related with each other and with process and outcome measures of counseling. The generalized other predictive measure.--The pro- cedures for this type of empathy test are similar to the ones used in the "specific other" predictive measure. The primary difference being that the perceivers are required 32 to predict the responses of a specified population rather than predict those of an individual. For example, the per- ceiver is asked to predict the interests of the average college student (Livensparger, 1965, cited in Smith, 1966). Kerr and Speroff (1954) developed a group paper and pencil measure called The Empathy Test utilizing this ap- proach. Subjects were required to rank as the "average man" would rank: (1) preferences of different musical forms, (2) the popularity of different magazines, and (3) the annoyance magnitude of different experiences found by most people to be annoying. In general, the test appeared to be poorly put to- gether and has questionable validity. Wallace B. Hall (Buros, 1965, p. 215) criticized the test's normative data, manual, forms, references, scoring keys and concludes his review with this comment: In view of these negative features and the implication that the test is more of a measure of general informa- tion and prediction of opinions than of interpersonal empathy, there appears little to recommend this test for the purposes stated by the authors. Hall and Bell (1953) duplicated Dymond's techniques using leaderless discussion groups in order to give subjects a chance to interact with each other and to develop "empathy toward one another." Fifteen groups of five members and three groups of four members met for 30 minutes. Dymond's procedures were then administered. Kerr's and Speroff's test was administered one week later. The results 33 of the two tests were correlated for each group and an average correlation of r = .02 was estimated between the tests. The authors concluded that no empirical relationship exists between Dymond's ("Specific Other Predictive Test") and Kerr and Speroff's test ("Generalized Other Predictive Test") even though they reported to be measuring the same construct. Patterson (1962) correlated the results of the Kerr and Speroff Empathy Test with the MMPI K scale, the Edward's Personal Preference Scale on affiliation, intraception and nuturance with 300 males in rehabilitation counseling. Cor- relation coefficients ranged from r = -.13 to r = .04. The author concluded that no support for a common factor for empathy or for the construct validity of the Kerr-Speroff Empathy Test was found. In the Sixth Mental Measurements Yearbook (Buros, 1965) an instrument named the Diplomacy Test of Empathy, constructed by one of the authors of The Empathy Test, William Kerr was reviewed. The test was proposed to extend the basic idea of the earlier test measuring a number of dis- tinct empathy factors. The latter test purports to measure understanding of other's feelings by having the examinee predict the self descriptions of a large variety of general- ized others. Thorndike (Buros, 1959) comments that he does not think that the purported factors are true factors and that the results from the test are meaningless. 34 Mahoney (1960) developed a Literature Test of Empathy. This paper and pencil instrument contained four selections from literature portraying markedly different personalities. After each section was read the subjects were to answer 20 multiple choice and incomplete sentence items. The scoring key was developed by 23 psychologists who were selected by peers as being "empathic individuals." The answers represented 78 per cent agreement by the psy- chologists. Split-half and test-retest reliability ranged from r = .66 to r = .92. Low positive correlations were obtained between the test and intelligence and academic aptitude measures. The test also correlated at r = .44 with measures of reading ability. No validity studies were de- veloped for this test, however. In summary, both of the tests published utilizing the "generalized other" predictive approach have received poor reviews in the Mental Measurements Yearbook (Buros, 1959, 1965). There is a good reason to believe that they are measuring general information and the prediction of opinions. There is evidence that they have no relationship to specific other predictions (Hall & Bell, 1953). In general, the approach does not seem to be particularly applicable to counseling or interpersonal situations since these require interpersonal interaction and sensitivity to a specific other. The Mahoney Test seems to be highly 35 correlated with reading ability and has not reported any validity studies. Therefore, this type of test will not be sampled in this study. Situation Approaches Reid and Snyder (1947) in an early attempt at de- veloping a situation type empathy measure, phonographically reproduced excerpts from counseling sessions to play to counselors. The fifteen counselors were to identify the feelings being communicated on the recordings. The feelings recognized by the subjects varied considerably. Fifty per cent of the subjects agreed 50 per cent of the time with the modal designation of feelings. The modal designation was considered the "correct" response, although it was ap— parent that more than one feeling was being communicated per excerpt. Subjects were ranked in order of skill by their pro- fessor in a practicum type of course. A correlation of .70 was obtained between the professor's ranking and the frequency with which each counselor agreed with the modal designations of feeling on the recorded excerpts and .49 with the typescript materials which were also used. Astin (1957) developed a situation test of empathy by having a professional actor portray various emotion laden statements on to audio tape recordings. The state- ments were taken from Wicas' (1955) client typescripts of eight different clients undergoing therapy. The statements 36 were chosen on the basis of their being strongly emotional- ized, verbal expressions. Items were also selected on the basis of havingyelicited similar responses by the majority of 22 "expert" counselors functioning in the role of counselor trainers, professors of guidance, and directors of counseling and testing. Subjects responded to the statements verbally by talking to a tape recorder. These responses were then rated by seven judges. The basis for rating was how much each response communicated as understanding of the essential feeling and content being expressed. The inter judge agree- ment index was .82. This situational test was able to discriminate counselors from non-counselors. Arbuckle and Wicas (1957) developed a test similar to Astin's using audio tape recordings of counseling inter- views. Using the tapes as stimuli, they developed rating systems for evaluating counselor responses. Little follow up has been reported on this measure since its original description, however. Strupp (1960) used a sound film of an actual inter- view from a therapy session as a stimulus for reaponses. The film stopped in 28 places and audience therapists were asked, "What would you do?" The free responses were then rated by two independent raters on a gross plus-to-minus rating. Minus was given to "cold," "distant," "unper- ceiving," "unfeeling" responses and plus ratings were given 37 to "warm, "perceiving," "close," responses. Whereas a zero was given when there was insufficient evidence to rate the response or the tone was emotionally neutral. Strupp was able to discriminate experienced thera- pists from inexperienced ones by means of this test. He also used it for discriminating between therapists who had been through analyses themselves and those who had not. The major problem with the study was that many of the therapist responses were rated "zero" and the ratings seemed to be too global. O'Hern (1964) recorded 30 different client problems onto audio tape, using actors for "clients" or empathy stimuli. Empathizers were selected for three degrees of expertness in counseling (novice, trainee, expert). A panel of clinicians then rated the correctness of judge's responses to each of the 30 scenes. O'Hern then identified those items which best discriminated empathic skill as de- fined by training and experience. The test showed no sig- nificant correlation between the class grades and a judge's evaluation when presented to 212 counseling students. Kagan and Krathwohl, ep_gl, (1967) developed a situ- ation test using videotape segments of actual counseling interviews. The test, while not measuring empathy, pur- ports to measure an important component of empathy, af- fective sensitivity. Affective sensitivity is the ability to perceive and identify the affective states of others. 38 The test, called the Affective Sensitivitnycale, contains 89 multiple choice items. The subject chooses the state- ment which best describes the last feeling expressed by the client on videotape. Correct answers for items were derived from judges and from client responses during recall sessions while watching themselves on the tape. The reliability co- efficients obtained from a number of studies conducted on the scale ranged from r = .53 to .77. The scale was re- vised a number of times to increase its discrimination and to increase reliability. A number of concurrent and predictive validity studies were carried out with the scale (Kagan, ep_§l., 1967). One study calculated a therapist's ranking of the members of three counseling groups in terms of their sensi— tivity. These ranks were correlated with the ranked scores obtained on the Affective Sensitivity Scale. The average rho correlation coefficient was .53. Another study, using similar rankings on doctoral candidates in counseling by their supervisors found a rho correlation of .32 with the scores obtained on the scale. Form B of the scale was ad- ministered to three groups towards the end of their academic year NDEA institute experience. At the same time, peer and staff ratings of counselor effectiveness were collected. An average correlation coefficient of r = .30 was obtained between the scale and peer ratings whereas the correlation coefficient between staff ratings and the scale on the groups was r = .24. 39 Evidence of the predictive and concurrent validity of the scale is significant but this validity is far from being adequately established. The major advantage of the situatiOn approach is that the stimulus situation is held as nearly uniform as possible for all subjects. Variations in counselor empathy therefore, cannot be attributed to the effects of different stimulus input such as different clients. Because of this, norms can be developed on the instrument. The situation test is related to the predictive ap- proach in that counselor empathy is regarded as a trait. The counselor's performance on these tests is expected to be reflected in other situations such as in a counseling interview. Presumably, the closer the standardized situa- tion approximates the situations to which it predicts, the better the measure. I In general, the validity of this approach rests on how predictive the measured trait is to behaviors reflecting the trait in other situations. In other words, counselors receiving a low score on this type of test would be expected to show low empathy in other situations. The generalizability of one situation test will be investigated in this study by correlating it with other measures of empathy and with process and outcome measures of counseling. 40 The Rating Approach This method of measuring the concept of empathy con- sists of having trained judges to listen to or observe interactions between persons. On the basis of certain de- fined criterion, the judges rate the amount of empathy dis- played by the subject being judged in the interaction. Buchhiemer (1961) developed a rating scale of empathic behavior based on the concept of confluence. Confluence denoted the flowing together or moving along of the counse- lor with the counselee within the counseling interview. The rating scale was developed to measure confluence and from these ratings empathy was inferred. The interjudge re- liability correlations reported in Katz's (1962) study using this scale ranged from r = .80 to r = .91. The investi- gator found no relationship between the ratings on this scale and counseling outcomes with 20 different student trainees working with short term counseling. Truax (1961) developed the Accurate Empathy Scale which consists of nine levels of empathic behavior from "inaccurate to obvious feelings" to "always accurate to- ward obvious feelings, unerringly accurate in both content and depth." The interjudge reliability on this scale seems to center around the sensitivity of the judges. They must be sensitive enough to be able to discriminate differences along the nine dimensions or the range of rating will be 41 too narrow. More research has been done with this scale A than with any other rating scale. Carkhuff (1967) developed an adaption of Truax's Accurate Empathy Scale. This revision, called the Empathic Understanding in Interpersonal Processes Scale, collapses the nine stages on the Accurate Empathy Scale to five levels. It was written to reduce ambiguity and to stress the additive, subtractive and interchangeable aspects of the communication of understanding. At the lowest level, the counselor detracts significantly from impressions of the cli- ent in that he communicates le§§_of the clients feelings than the client did. At the highest level, counselor responses add significantly to the feelings and meaningsmH Ho. Hm unmoHMHsmHmas .Hm>mH Hoo. um HGMOHMHsmHms Hm om mH om n z loam “mamas mGOHummonm ustHU «ems. «mm. ao.- am. snoucm>cH mHsmcoHHmHmm msoHumoouom HOHomssoo «.om. am. mo.- oo.- snouam>aH mHsmcoHumHmm umws mo.- mH.n mo. mm. .mmm SHHmmum>HuoHomum umHH xomso no.1 SH. mH. HH.| HmsomummuwusHim>Hu0HomHm mo. mm. om.- .«sa. mcHumm mama mmnsxuao mHmom mo. mo. mH. mm. SHH>HuHmcmm m>Huommmm uommmd mmsaumm mosaumm Hummus HcmHHo . . Honmssou mcHocwHHmmxm coHuoHonxm MHmm haummEm mmoooum . mOHDmMOE mmmooum pom monommme mausmEm cmmzuwn mGOHHMHmHHOUII.m mHmea bei 1m la 82 An inspection of Table 3 reveals that no one measure of counselor empathy was consistently related to the counseling process measures. Judged empathy was sig- nificantly correlated with Self Exploration ratings (r = .47) but negatively with Experiencing ratings (r = -.30). Client perceived empathy was significantly correlated with counse- lor affect (r = .45) and with client affect (r = .83). Counselor perceived empathy was significantly correlated with counselor affect (r = .50) but not with the other process measures. Specific Question: 1. What is the relationship between client perceived empathy based on the Relationship Inventory and scores on the Client Affect Scale? Table 3 reveals that there was a significant re- lationship between these measures (r = .83). vThis high re- lationship suggests that client perceptions of counselor empathy and positive affect toward the counselor are highly related. 2. Are the three process measures used in this study related? The correlations between the process measures are presented in Table 4. The correlations in Table 4 suggest that the process measures were not related. The only sig- nificant correlation reported is between the Counselor 0‘. Q ‘HH P M. . re: 83 TABLE 4.--Correlations between process measures. 1 2 3 4 1. Self Exploration Ratings 2. Experiencing Ratings .16 3. Client Affect .04 -.13 4. Counselor Affect .28 .22 .38* N = 30 19 30 31 *Significant at .05 level. Affect Scale and the Client Affect Scale. This finding suggests that the counselor's positive attitudes toward his client are related to the client's positive attitudes to- ward him. Relationship between Counseling Process and Outcome General Question: Are the process measures related to counseling outcomes as measured in this study? Specific Question: Which process measure is most related to counseling outcome? The correlations between the process measures and the outcome measures are reported in Table 5. An examination of Table 5 indicates that only one process measure is consistently related to the outcome 84 .Hm>mH mo. Hm ucmonHcmHmee .Hm>mH Ha. Ha HaaonHcmHme Hm u 2 am. Hm. «om. oo. mo. so. Hummus Honmasoo om u 2 tom. RH. mm. “mm. smm. «om. Hummus ucmHHo mH u z mmcHumm mm.. mm. RH. mm.- .00.- .«oa.u maHocoHuomxm on u z mmcHuam SH. RH. mm. mH.- mo.- mo. coHuaHoanm HHmm mamHm msHuHmom mmusmomz coHumsHm>m :oHumsHm>m mmcHumm usmH>mo Hmuoe omcHHsoo HamHHo nonmasoo Has: no Henssz mommmssma wommmssme .mouomome mEoouoo tam mondmmoe mmmooum :mm3umn mGOHDMHmHHooII.m mqmda “.4123: C5 ~"C L1 85 measures. The Client Affect Scale is significantly corre- lated with four of the six outcome measures including the combined outcome score. The experiencing ratings were negatively correlated with two of the outcome measures. The results suggest that the rated process measures were not related to counseling outcome. The counselor affect measure was also not related to outcome. The client affect measure was the only process measure which was sig— nificantly correlated with the outcome criteria. Relationship between Counselor Empathy and Counseling Outcomes General Question: Is counselor empathy signifi- cantly related to counseling Outcomes? The correlations between the different counselor empathy measures and the outcome indices are reported in Table 6. An inspection Of Table 6 discloses that counselor empathy is significantly correlated tO counseling outcomes. In general, there are few significant relationships between most Of the counselor empathy measures and the outcome measures, but the client perception measure appears to be consistently related to outcome. Significant correlations between the scores on the Tennessee Self Concept Scale-- Total Positive, MMPI ratings, Counselor Evaluation, and Combined Outcome Score and the Client Perception Measure were found. llllllllul. t‘,ll F IIIII v‘c',\ll IQ ...N G - CAN. 86 .Hm>mH mo. Hm unmoHHHcmHm..e .Hm>mH Ho. um HaaoHanmHmee .HmsmH Hoo. um HcmoHHHamHm. Houm Hmummv msOHumOO esmm. em. «tsmm. «mo. em. «stew. IHOm MQOHHUIWHOH 1sm>cH HomCOHuMHmm msoHummoumm Ho. mo. mH. HH.1 mH. 0H.1 HOHmmssOUI>HOH 1co>sH mHnmsOHuonm . puma .mmm mH.1 mH.1 mm.1 mm.1 no. mo. sHHmelm>HuoHamnm umHH xoonu HmsOmnmm HH.1 «H.1 vH.1 NH. om.1 mH.1 IHOHGH1O>HHOHOOHm mcHumm om. mo. NH. Hm. mm. .«ema. mama musexumo OHoOm >MH>HH nH.1 no.1 HH. mo.1 om.1 mm.1 IHmsmm O>Huommm4 mwusmmo mcmHm O>HuHmom 2 sOHumsHm>m sOHumsHm>m mmsHuom HGMH>OQ HmuOB OOQHQEOU ucmHHu HOHOmSHOU Hmzz mo Hwnfidz commences ommmmscwa woummfim Oeoouoo .mOHSmmoe osoouso was monsmmoe woummEm consume msOHuMHOHHOU11.m mqmde u v a :u i . p y «L . CC A.» ~ rant Y. C . ~‘b (Nil. 1 R\U 1 S Rd .PIU 87 Specific Questions: 1. Which Empathy measure is the best predictor of counseling outcome? 2. Which combination of Empathy measures are the best predictors of counseling outcome? Table 7 presents the results of a series Of least squares multiple regression equations, one for each outcome measure. Each empathy measure was fitted into the regres- sion equation to check whether it predicted the outcome measure at a significant level of probability. As can be seen in the Table, the Barrett-Lennard Relationship Inven- tpry_(C1ient Perceptions after the third interview) was the best predictor of the outcome measures. The Carkhuff ratings and the Client Perceptions together predicted the changes in Total "P" scale of the Tennessee Self Concept §EE$S at a .02 level of probability and this cOmbination accounted for 30 per cent of the variance. None of the other empathy measures predicted the outcome measures at a .10 level of probability or less. The correlations between the outcome measures are presented in Table 8. An examination Of the correlations in Table 8 indi- cates that the outcome measures in the study were related. The only outcome measure which consistently did not corre- late with the other measures was the Counselor Evaluation of Counseling Measure. 1 PJ..re1~ ~H U1 ~n...J~C i 1 1 . .. IAN ,. H. .~dn: mueUHH-naud hue-mu U .rflnueflleI I 1 A .n.‘«4..~<.~. l H i IvaP-FUIJJH—PU Tun”- n-h~hvfi.l~U4.N7zULHO~ in thu :flvlflfikmonVINNVavlfl 1.v N fl 0 1N \A f- 1... ‘11 5.7 e Cuban 9 1 . 88 .mo. u H as msHHmmssoo mo GOHumsHm>m uGOHHO on» £HH3 OOHMHOHHOO Humomv msOHumooumm usOHHU1>H0usO>cH mHnmcoHuoHom« Acum Hmumev msOHummOHOm ucmHHU NN moo. om.m om. em. monsoon: OmsHQEOO anousm>sH mHsmsoHuMHOm «ucmum 11 11 11 11 11 soHumsHm>m usmHHo 1:00 ucmOHmHsmHm Oz loam “mouse coHum mcoHummouom usmHHO mN mo. mo.v NH. mm. 15Hm>m HOHmmssou >HOHCO>GH mHanOHOMHOm Apum “cause msowumoouom usmHHU Hm Hoo. om.aH as. am. mmcHumm Has: snoucm>cH oHamcoHumHmm mcmHm loam “mouse uGMH>OQ mo msOHamooumm usOHHO mN mo. NH.m NH. vm. HonEdz mommmssoe wuousm>sH QwsmsOHuoHOm O>HuHmom msHumm were mmscxuou mm No. oa.a am. am. Hence mmmmmccme mst mcoHuamoumm HamHHo Hpum HmumHuHmom mcowummoumm usmHHU mm amo. om.m om. as. Hence mmmmmccme snoucm>aH mHnmcoHumHmm m Mom usOHOHmumOO mo mo mocoo m wwwmwmwwd conowOHHOO mEOouso HOHOHOOHm 1HHHcmHm no assumes mHmHuHsz . mmufimmme mauomfim EOHM oEoousO m0 coHuOHOOHm How conmmHmmH OHQHHHDE mmuosvm ummmall.h mqmda 89 TABLE 8.--Correlations between outcome measures. (N = 25) l 2 3 4 5 1. Tennessee Total "P" Differences 2. Tennessee Number of Deviant Signs Difference .75* 3. MMPI Ratings .67* .75* 4. Counselor Evaluation Of Counseling .008 .02 .14 5. Client Evaluation Of Counseling .08 .32 .47** .36 6. Combined Measures*** .71 .80 .81 .54 .63 *Significant at the .01 level of probability with 23 degrees Of freedom. **Significant at the .02 level of probability with 23 degrees Of freedom. ***Combined measures would be expected to be highly correlated with others because they are not independent Ob- servations. Discussion Counselor Empathyras a .Measured Concept The six different methods of measuring counselor empathy were not related to each other in this research. Table 2 (p. 77) shows low, non-significant correlations between all the different measures sampled. While this finding is curious, considering that all the measures were 11a ,7“ '13 l w 1 C3... ( SEW ccu C1 MM. 90 purporting to measure counselor empathy or some similar variable, it agreed with the findings Of previous research (Astin, 1957; Burstein & Carkhuff, 1968; Hansen, Moore, & Carkhuff, 1966; Katz, 1962; Lesser, 1958; Truax, 1966). Therefore, the results Of the present investigation, plus the previous research, lead to the conclusion that the different measures of counselor empathy are measuring several different variables under the same label. This conclusion has important implications for re- search in counseling since it does make a difference how counsel or empathy is defined and measured because the different approaches are not measuring the same variables. Not only were the different measures unrelated, but the data in Table 6 (p. 86) and Table 7 (p. 88) indicate that only one counselor empathy measure, client per- ception's, was consistently related to counseling outcome indices. The review of the literature concerned with the relationship between counselor empathy and counseling out- comes reported mixed findings with some studies finding positive relationships between counselor empathy and out- come while other research reported no relationship between the two. Since the different studies used different measures Of empathy, the present investigation lends sup- port tO the hypothesis that the differences in the reported results were due to the fact that the same independent variables were not being measured in the different studies. Lu 91 Predictive Empathy According to Wittich (1955) and Taft (1955), the ability to predict the responses of others may be regarded as a trait. If this conclusion was correct, then it was ex- pected that the correlation between the two predictive measures would be high. The data did not support this con- clusion. The correlation coefficient between the two pre- dictive measures was r = .16. Counselor predictive empathy was not constant between the two measures even though the tasks required on the tests were similar, both were taken after the third interview, and both predictions were about the same client. Therefore, Wittich's and Taft's conclusion was not only unsupported, but the evidence suggests that the conclusion is very much Open to question. The predictive empathy measures were not correlated with the other measures of counselor empathy (see Table 3). This finding agreed with a number of other studies (Astin, 1957; Katz, 1962; Lesser, 1958). Thus, the results of the present study, plus previous research, lend strong support to the conclusion that predictive measures are unrelated to other measures of counselor empathy. Finally, predictive empathy was consistently un- related to counseling process and outcome measures in the study. Not one significant correlation was found between the two predictive counselor empathy measures, the three process, and six outcome measures (Tables 3 and 5). These 92 results disagree with the research of Cartwright and Lerner (1958) and Dombrow (1966) but agree with other re- search (Katz, 1962; Lesser, 1958). It must be said, however, that the predictive empathy procedures which replicated those of Cartwright and Lerner (Predictive-Kelly REP Test) differed in an important way. Only one prediction was made in the present study, after the third interview. Cartwright and Lerner had their therapists make two predictions: one early in therapy and one late. They found that the late predictions were sig- nificantly related to client change whereas the early pre- dictions were not. Since the one prediction in this study was made early, then the results do not disagree with the Cartwright and Lerner research. In summary, predictive empathy was not consistent, did not correlate with other counselor or empathy measures, and was not related to counseling process and outcome. Therefore, this study raises serious questions about the usefulness of this type of operational measure of counselor empathy in counseling research. Situational Empathy In general, the validity of a situation test of counselor empathy rests on how predictive the performances lneasured by it are to other situations requiring the ability. In other words, a counselor scoring high on a situation test is expected to show high empathy in counseling CL .5 N 93 interviews. The trait of affective sensitivity as measured by Kagan's Affective Sensitivity Scale did not correlate with counseling process and outcome as measured in this study (see Tables 3, p. 81, and 6, p. 86). Since the scale measured the ability to be sensitive to affective communication, but not necessarily the ability to communicate that understanding, Kagan hypothesized that the scale would be predictive of low counselor empathy on the other measures. In other words, a counselor with high sensitivity may not be able to communicate that sensitivity. But a counselor with low sensitivity, has little sensitivity to communicate. This hypothesis was not confirmed (see Appendix J). The results of the present investigation then, raise questions about the validity of counselor affective sensi- tivity as a trait measured by the Affective Sensitivity Scale. These results were not in agreement with the previ- ous validity studies on the scale (Kagan & Krathwohl, et al., 1967). Since the predictions Of importance from the AET fective SensitivityrScale center around low scoring counse- lors, the counselors used in this study may have been in- appropriate to test Kagan's hypothesis. Appendix C shows that the counselors in the sample obtained a mean score of 57 on the scale with a range of 38 to 72. This is a higher taverage and a more restricted range than was obtained C L. 1 a no. 0 TV 5. f t. K ”We 3C C1 94 previously on the scale (Kagan, et_al., 1967, for 402 subjects; mean = 50, range = 21 to 72 [p. 490]). There- fore, there may have been too few low scoring counselors in the sample to obtain a good test for the hypothesis Of Kagan's test. Judged Empethy Table 6 (p. 86) reports the correlations between judged counselor empathy (Carkhuff Tape Rating) and the counseling outcome measures. Judged empathy, while posi- tively correlated with all outcome measures, is signifi- cantly related to only one outcome index. This result is in conflict with the recent research (Burstein & Carkhuff, 1968; Hansen, Moore, & Carkhuff, 1966; Truax & Carkhuff, 1967) which has suggested that judge rated empathy is more predictive Of outcome than client perceived empathy. I Judged empathy was significantly correlated with client self exploration (Table 3, p. 81). This result agrees with the findings Of Hountras and Anderson (1969). Client self exploration was not related to outcome in the present investigation, however (see Table 5, p. 84). No firm conclusion about the value of judged (empathy can be made. Although judged empathy was less pre- ciictive of counseling outcome than client perceived empathy in this study, previous research has consistently reported 'that judged empathy is predictive Of counseling outcome. In (I) 95 Perceived Empathy Counselor Perceptions Counselor perceived empathy was negatively corre- lated with judge ratings (r = -.24) and only moderately with client perceptions (r = .20). While these relation- ships were non-significant, the results coincided with the findings of previous research (Rogers, et_el., 1967). Counselor perceived empathy was also consistently unrelated to both counseling process and outcome measures (Table 3, p. 81, and Table 5, p. 84). These results are also in agreement with previous research (Rogers, 1967). It appears on the basis of the present research and on previous studies, that the assessment of empathy by counselors and therapists is less satisfactory and probably less valid than the assessment Of counselor empathy by the client or by an unbiased judge. Client Perceptions Client perceived empathy based on the Barrett- Lennard Relationship Inventory was the best predictor of all indices of counseling outcome (Table 7, p. 88). This find- ing is consistent with Rogers' (1957, 1959) theory which says that the client's experience of the therapist's re- sponse is the primary focus Of influence in the relation- ship. This outcome is also in agreement with some previous 1658 I966 the Elm bit It TOE C0' CE (‘1’ n (r J) 96 research (Barrett-Lennard, 1962; Truax, Wargo, et_ei., 1966;1 and Truax, Leslie, et_ei., 19662). Truax and Carkhuff (1967) have suggested that client perceptions are more likely to be related to counse- ling outcome with clients who are not seriously disturbed in their ability to accurately perceive and report perceptions. It is likely that the clients used in this sample had that ability since they were college students. Therefore, the results suggest that the client perception measure Of counselor empathy is most predictive Of counseling outcome with college students. The data in Table 3 (p. 81) reveals that client per— ceptions of counselor empathy and client positive affect toward the counselor are highly related (r = .83). Because this relationship was expressed as a correlation, the question of cause and effect cannot be answered. It may be that clients have more positive affect toward empathic counselors. Or, the client perceptions' of counselor empathy may have been influenced by the more global positive affect felt toward their counselors. The issue of what client perceived empathy actually reflects, therefore, re- mains unresolved. 1Cited in Truax, C. B., & Carkhuff, R. Toward ef- fective couneeling and psychotherapy, Chicago: Aldihe, 1967. P. 137. 21bid. 97 Regardless of the cause and effect Of this re- lationship, client positive affect toward the counselor as measured by the Snyder Client Affect Scale appears to be consistently related to counseling outcOme (see Table 5, p. 84). Client perceptions of counselor empathy appear to remain relatively stable. The correlation between client perceived empathy after the third interview and at termina- tion was r = .66. This finding is in agreement with the re- search conducted in the Wisconsin Project (Rogers, er_ei., 1967). Empathy and Rated Counseling Processes According to Rogers' (1957, 1959) theory, certain therapeutic conditions lead to processes within the thera- peutic interaction, which in turn lead to client behavior change. One of the therapeutic conditions is counselor empathy. Two process measures, self exploration and ex- periencing, were rated which assessed the processes Rogers describes. The question posed and investigated in this study was which measure of counselor empathy relates to these processes? The correlations in Table 3 (p. 81) reveal only one significant correlation between counselor empathy and one of the two rated process measures. Therefore, counselor empathy was not related to rated counseling process in the study. 98 Table 4 (p. 83) shows that the rated process measures were not significantly related (r = .16). This result does not agree with the Rogerian conception of counseling processes since both measures'are supposed to be assessing different aspects of the same overall conception of process. Further, the two rated process measures were unre- lated to outcome (Table 5, p. 84). The experiencing ratings correlate negatively with two Of the outcome measures. These results are not in agreement with an extensive amount of previous research (Rogers, 1967; Truax & Carkhuff, 1967; Carkhuff & Berensen, 1967). Therefore, these findings raise questions about the validity of the rated counseling process measures used in the study. Counseling Outcomes Few researchers agree about the best measures Of counseling and psychotherapy outcome. Previous research has suggested that client changes assessed by different outcome criteria are not related to each other (Zax & Klein, 1960). Therefore, the conclusions drawn in this study must be re- stricted to the outcome measures used, and may not general- ize to other counseling and psychotherapy outcomes. All of the outcome measures in the study with the exception of the counselor evaluation of counseling were based on self-report personality instruments. Zax and 99 Klein (1960) state that the most serious failing with this type of outcome measure is that these criteria have not yet been related to changes in everyday, observable behaviors in the life space of the client. The outcome measures used in this research have been used as acceptable outcome criteria for other research on counseling and psychotherapy however. Two of the counseling outcome measures (the scores from the Tennessee Self Concept Scale) were based on differ- ence scores between pre- and post-tests. Difference scores tend to be unstable (McNemar, 1963) but the data in Table 8 (p. 89) reveals that difference score outcomes were related to the other outcome measures. In general, a wider range of outcome measures would have been preferred, but other outcome criteria of client change was not available. Limitations Of the Study The major limitation of the study was that the number of subjects sampled was small (N = 31) for a corre- lation study of this type. The population of clients were drawn from a Uni- versity setting. These clients tend to be above average in intelligence and were functioning well enough to stay in the University. Perhaps the conclusions cannot be general- ized to other settings where psychotherapy is practiced (e.g., outpatient clinics or hospitals). 100 All of the counselors were either Ph.D.‘s or Ph.D. candidates in clinical and counseling psychology. This may have restricted the range of scores on the empathy measures. According to McNemar (1963), a restricted range tends to reduce correlation coefficients. The range of scores on the different empathy measures are reported in Table G-l (Appendix G). Finally, a wider range of outcome measures, based on criteria other than client self-report would have strengthened the conclusions drawn from the study. The conclusions drawn in the study may not generalize to other counseling outcomes. CHAPTER V GENERAL SUMMARY AND CONCLUSIONS In this, the last chapter, a general summary of the study and the conclusions drawn from the study are presented. General Summary The present investigation was concerned with the counselor's empathic ability as a variable influencing the process and outcome of personal counseling. The purpose of the study was: (a) to examine the relationship between six different measures of counselor empathy which have been used in previous research on counseling and psychotherapy, (b) to examine the relationship between each of the empathy measures and counseling processes, (c) to investigate the relationship between client change over personal counseling and each measure of empathy, and (d) to determine which empathy measure or combination of measures was the best predictor of client change. The study was prompted when a review of literature revealed that several different ways of conceptualizing and measuring counselor empathy had been used in the previous research. It was not altogether clear whether the different 101 102 studies dealt with the same variable or several different variables using the same label. The review Of the research on the relationship be- tween counselor empathy and counseling and psychotherapy outcomes proved inconclusive. Several of the studies found positive relationships between counselor empathy and outcome while other studies found no relationship between the two. One hypothesis for the mixed results found in the research was that different, unrelated measures of counselor empathy were being used in the different studies. The different approaches to the measurement of empathy were reviewed. These approaches were classified into four general categories. The categories included the "predictive" empathy approaches, "situational tests" of empathy, "rated or judged" empathy, and "perceived empathy" approaches. Each approach appeared to have its particular limitations and advantages. Few studies have explored the relationships between the various approaches to empathy measurement. Those studies which have investigated the question, found little to suggest that there was a relationship between the differ- ent measures. Using the same group of 31 counselors and clients from the Michigan State Counseling Center, this study com- pared six different ways of measuring counselor empathy, and investigated how each was related to the process and 103 outcomes of personal counseling. The six different methods of empathy sampled the four general approaches to the measurement Of empathy which had been defined. Each counselor in the study was assigned to a re- search client who had come in for personal counseling and had taken a series of personality inventories before enter- ing counseling and at termination. The empathy measures used in the study were: (a) counselor predictions' of client self-descriptions on the Interpersonal Checklist after the third interview, (b) counselor predictions of client self-descriptions based on constructs derived from a modified Kelly Rep. Test after the third interview. These two measures sampled the "pre- dictive empathy" approaches. (c) Counselor scores on Kagan's, et a1. Affective Sensitivity Scale (1967) , which uses videotaped segments of actual counseling inter- views as stimuli for the 89 multiple choice items on the questionnaire. This measure sampled the "situational ap- proaches" to empathy measurement. (d) Ratings Of counselor empathy by two judges using Carkhuff's Empathic Understand- ing in Interpersonal Processes Scale (1967). This measure sampled the "rating approach" or "judged empathy." (e) Client perceptions' of counselor empathy based on the Barrett-Lennard Relationship Inventory (1962) after the third interview and at termination, and (f) counselor per- ceptions Of his own empathy based on the Barrett-Lennard 104 Relationship Inventory (counselor form) after the third interview. These measures sampled the "perceived empathy" approach. A correlation matrix developed by the IBM 3600 from the scores the counselors Obtained on each Of the six empathy measures revealed no significant correlations be- tween the measures. Three process measures were used to assess counsel- ing processes in the study. The measures were: (a) ratings Of client self exploration by two judges based on Carkhuff's Self Exploration in Interpersonal Processes Scale (1967), (b) ratings Of client experiencing based on Gendlin and Tomlinson's Experiencipg Scale (cited in Rogers' et al., 1967) by one judge on 19 clients, and (c) counselor and client affective attitudes toward each other based on Snyder's (1961) Client Affect and Therapist Affect Scales which are 200 item true and false questionnaires. Correlations calculated between the six empathy measures and the process measures yielded significant corre- lations between rated counselor empathy and client self ' exploration (r = .47); between client perceptions' Of counselor empathy and client positive affect (r = .83) and 'between.client perceived empathy and counselor positive affect (r = .45). The only process measure which con- sistently related to the outcome indices was the Client Af- fect Scale. 105 Six measures of the client change were used as out- come variables. These were: (a) difference scores on two scales of the Tennessee Self Concept Scale (Fitt's, 1965), (b) judges ratings of pre- and post-MMPI Inventory scales for amount of improvement, (c) counselor ratings of success of counseling, (d) client ratings of how helpful counseling had been, and (e) a combined score of all of the indices. Correlation coefficients calculated between each of the empathy measures and outcome variables resulted in significant correlations (t = < .05) between the client perception empathy measure (Relationship Inventory) and four outcome variables. Positive correlations were found between the judge ratings of counselor empathy and outcome variables although only one correlation was high enough to reach a .05 level of significance. Multiple regression equations calculated for each of the outcome variables revealed that client perceptions were the best predictor Of outcome and that the combination of the judge ratings and client perceptions increased pre- dictive ability on one of the outcome indices. Conclusions The major conclusion emerging from this investiga- tion is that the six different measures of counselor empathy which sampled the Operational definitions and procedures in previous counseling research were unrelated. This conclu- sion is of considerable importance to counseling research. 106 Another more tentative conclusion is that client perceptions' of their counselors' empathy is the best pre— dictor of counseling outcome on a college student population in personal counseling. While this conclusion is in agree- ment with Rogerian theory, it must be regarded as tentative since several other research studies have demonstrated that judged empathy is the best predictor of counseling outcome. Client perceptions' of counselor empathy appear to be relatively stable over counseling and are highly related to the client's positive affect toward his counselor. Serious questions about the usefulness Of the pre- dictive approach to the measurement of counselor empathy were raised by the study. Counselor perceived empathy was concluded to be a less valid assessment of counselor empathy than assessments by the client or by unbiased judges. Another tentative conclusion was that one promising situational measure Of counselor affective sensitivity was not related to counselor empathy in this counselor population. Finally, it was concluded that the rated counsel- ing process measures in this study were of questionable validity. REFERENCES REFERENCES Arbuckle, D. S., & Wicas, E. A. The development of an in- strument for the measurement of counseling per- ceptions. Journal of Counseling Psychology, 1957, 4, 304-312. Ashcraft, C., & Fitts, W. W. Self concept change in psycho- therapy. Psyehotherepy, 1964, 1(3), 115-118. Astin, H. 8. Assessment of empathic ability by means of a situational test. Journal of Counselipg Psychology, 1967, 14, 57-60. Astin, H. A comparative study of the situational and pre- dictive approaches to the measurement of empathy. Unpublished doctoral dissertation, University Of Maryland, 1957. Bachrach, H. Adaptive regression, empathy and psycho- therapy. Psychotherapy: Theory Research and Practice, 1968, 5(417’203-209. Barrett-Lennard, G. T. Dimensions Of therapist response as causal factors in therapeutic change. Psychological Monographs, 1962, 76(43, Whole No. 562). Bender, I. E., & Hastorf, A. H. The perceptions of persons: forecasting another person's responses on three per- sonality scales. Journal of Abnormal and Social Psychology, 1950, 45, 556-561. Bender, I. E., & Hastorf, A. H. On measuring generalized empathic ability (social sensitivity). Journal Of Abnormal and Social Peychology, 1953, 48, 503-506. Berenson, B. G., & Carkhuff, R. R. Sources Of gain in counseling andrpsychotherepy, New York: Holt, Rinehart and Winston, 1967. Berg, J. A. Deviant responses and deviant people: the formulation of the deviation hypothesis. Journal Of Counseling Psychology, 1957, 4, 154-161. 107 108 Borgatta, E. F. The stability of interpersonal judgments in independent situations. Journal Of Abnormal and Social Psychology, 1960, 60, 188-194. Bronfenbrenner, U., Harding, J., & Gallwey, M. The measure- ment of skill in social perception. In D. C. McClelland (Ed.), Talent and society. New York: Van Nostrand, 1958. Pp. 29-108. Buchheimer, A. The development of ideas about empathy. Journal of Counselihg Psychology, 1963, 10, 61-70. Buchheimer, A., & Carter, S. Analysis Of empathic behavior Of counselor trainees in a laboratory practicum. American Peychologist, 13, 1958, 352. Buros, O. K. The sixth mental measurements yearbook. Highland Park, New Jersey: Gryphon, 1965. Burstein, J. W., & Carkhuff, R. R. Objective therapist and client ratings of therapist offered facilitative conditions Of moderate to low functioning thera- pists. Journal of Clinical Psychology, 1968, 24, 240-249. Caracena, P., & Vicory, J. Correlates of phenomenological and judged empathy. Journal Of Counseling Psy- chology, 1969, 16(6), 510-515. Carkhuff, R., & Berenson, B. G. Beyond counseling and therapy. New York: Holt, Rinehart and Winston, 1967. Cartwright, R. D., & Lerner, B. Empathy, need to change, and improvement with psychotherapy. Journal of ConsultingrPsychology, 1963, 27, 138-144. Chessick, R. D. Empathy and love in psychotherapy. Ameri- can Journal Of Psychotherapy, 1965, 19, 205-219. Chiminsky, J., & Rappapart, J. Brief critique of the mean- ing and reliability of "accurate empathy" ratings. Psychological Bulletin, 1970, 73(5), 379-382. Cline, V. B., & Richards, J. M. Accuracy of interpersonal perception: a general trait. Journal of Abnormal and Social Psyehology, 1960, 60, 1-7. Cooper, L. The relationship of empathy to aspects of cognitive control. Dissertation Abstracts, 1967, 27(12b), 4548-4549. 109 Cowden, R. C. Empathy or projection. Journal of Clinical Cronbach, L. J. Essentials ofrpsychological testing. (3rd ed.) New York: Harper and Row, 1970. Cronbach, L. J. Processes affecting scores on "understand- ing of others" and "assumed similarity." Psycho- logical Bulletin, 1955, 52, 177-193. Dickenson, W., & Truax, C. B. 'Group counseling with under- achievers. Personnel and Guidance Journal, 1966, 45, 243-247. Dombrow, R. A. A study Of the relationship between thera- pist's empathy for patients and changes in patient's self concepts during therapy. Dissertation Ab- stracts, 1966, 27(l-B), 301-302. Dymond, R. F. A scale for the measurement of empathic ability. Journal Of Consulting Psychology, 1949, 13, 127-133. Ebel, R. L. Estimation of reliability of ratings. Psycho- metrika, 1951, 16, 407-424. English, N. B., & English, A. C. A comprehensive diction- ary of psychological and_psychoanalytical terms. New York: Longmans, Green and Company, 1958. Estes, S. G. Judging personality from expressive behavior. Journal of Abnormal and Social Psychology, 1938, 33, 217-236} Fenichel, O. The psychoanaiytic theory of neurosis. New York: W. W. Norton, 1945. Fiedler, F. C. A comparison of therapeutic relationships in psychoanalytic, non directive, and Adlerian therapy. gournal of Consulting Psychology, 1950, 14, 436-445. Fitts, W. Tennessee self eencept scale: manual. Nash- ville: Counselor Recordings and Tests, 1965. Fox, R. E., & Golden, P. C. The empathic process in psy- chotherapy: a survey of theory and research. Journal of Nervous and Mental Disorders, 1964, 138, 323. Fromm-Reichman, F. Principles of intensive psychotherapy. Chicago: University of Chicago Press, 1950. 110 Gage, N. L., & Cronbach, L. J. Conceptual and methodolog- ical problems in interpersonal perception. Psycho- logical Review, 1955, 62, 411-421. Gendlin, E. T. Experiencing_and the creation of meanipg, New York: Free Press of Glencoe, 1962. Gonyea, G. G. The "ideal therapeutic relationship" and counseling outcome. Journal of Clinical Psychology, 1963, 19, 481-487. Gorden, K. A device for demonstrating empathy. Seurnal of Experimental Psychology, 1934, 17, 892-893. Guiora, A. Toward a systematic study of empathy. Compre- hensive Psychiatry, 1967, 8(5), 375-385. Hall, H. E., & Bell, G. B. The relationship between two tests of empathy: Dymond's and Derr's. American Psychologist, 1953, 8, 361-362. (Abstract Halpern, H. M. Empathy, similarity, and self-satisfaction. Journal of Consulting Psychology, 1955, 19, 449-452. Hansen, J. C., Moore, G. D., & Carkhuff, R. The differen- tial relationship Of Objective and client per- ceptions Of counseling. Journal of Clinical Psy- chology, 1968, 24, 244-246. Hastorf, A. H., & Bender, I. E. A caution respecting the measurement of empathic ability. Journal of Abnor- mal and Social Psychology, 1952, 47, 574-576. Hawkes, G. R., & Egbert, R. L. Personal values and the empathic response: their inter-relationships. Jourhel of Educational Psychology, 1954, 45, 469-476. Hays, W. Statistics forpsychologists. New York: Holt, Rinehart and Winston, 1963. Hernsdahl, B. Concerning eigafuh lung (empathy): a con- cept analysis Of its origin and early development. Journal of the History of Behavioral Sciences, 1967, 3(2), 180-191. Hobart, C., & Fahlberg, N. The measurement of empathy. American Journal of Sociology, 1965, 70(5), 593-603. Hogan, R. Development of an empathy scale. Journal of Consulting and Clinical Peychology, 1969, 33(3), 307-316. 111 Hountras, P., & Anderson, D. Counselor conditions for self-exploration Of college students. Personal and Guidance Journal, 1969, 48(1), 45-48. Inglis, R. The effects of personality similarity on empathy and interpersonal interaction. Sissertation Abstracts, 1966, 26(10), 6 and 157. Jackson, W., & Carr, A. C. Empathic ability in normals and schizophrenics. Journal of Abnormal and Social Psychology, 1955, 51, 79-82. Kagan, N. Personal Communication, 1969. Kagan, N., & Krathwohl, D., et a1. Studies in human inter- action: interpersonal_process recali‘etimulated by videotape. East Lansing: Education PuBlicatiOn Services, 1967. Kanner, L. Judging emotions from facial expressions Of emotions. Psychological Monograph, 1931, 41(3, Whole NO. 186). Katz, R. Empathy: its nature and uses. New York: Free Press of Glencoe, 1964. Katz, B. Predictive and behavioral empathy and client change in short-term counseling. Unpublished doc- toral dissertation, New York University, 1962. Kerr, W. A., & Speroff, B. J. Validation and evaluation of the empathy test. Journal of General Psychology, 1954, 50, 269-276. La Forge, R., & Suczik, R. The interpersonal dimension of personality: an interpersonal checklist. Journal of Personality, 1955, 24, 94-112. Leary, T. Interpersonel diagnosis in personality. New York: Ronald Press, 1957. Lesser, W. M. The relationship between counseling progress and empathic understanding. Journal Of Counselipg Psychology, 1969, 8, 330-336. Lifton, W. The role of empathy and aesthetic sensitivity in counseling. Journal of Counseling Psychology, 1958, 5, 267-275. Lindgren, H. E., & Robinson, J. Evaluation Of Dymond's test of insight and empathy. Journal of Consultipg Psychology, 1953, 17, 172-176. 112 London, P. The modes and morals ofrpeychotherapy. New York: Holt, Rinehart and Winston, 1964. McNemer, Q. Psychological statistics. New York: John Wiley and Sons, Inc., 1962. Mahoney, S. G. The literature test of empathy. Disserta- tion Abstracts, 1962, 21, 674. Marwell, G. Problems of Operational definitions of "empathy," "identification," and related concepts. Journal Of Social Psychology, 1964, 63, 87-102. May, R. The art of counseling. New York: Abingdon Press, 1939. Mischel, W. Traits and states as constructs. Personality and Assessment. New York: Wiley, 1968. Pp. 41-72. Murstein, B. Some comments on the measurement of projection and empathy. Journal of Consulting Psychology, 1957, 21, 81-82. Norman, R. D., & Ainsworth, P. The relationship among pro- jection, empathy, reality, and adjustment, Opera- tionally defined. Journal Of Consulting Peychol- pgy, 1954, 18, 54-55 and 58. Norman, R. D., & Leiding, W. C. Relationship between measures of individual and mass empathy. Journal of Consulting Psychology, 1956, 20, 79-82. Notcutt, B., & Silva, A. Knowledge of other people. Journal of Abnormal and Social Psychology, 1951, 467*30-37. O'Hern, J. S., & Arbuckle, D. S. Sensitivity: a measure- able concept? She Personnel and Guidance Journal, Patterson, C. H. A note on the construct validity of the concept of empathy. Personnel and Guidance Journal, 1962, 40, 803-806. Pierce, W. D. Anxiety about the act of communicating and perceived empathy. Dissertation Abstracts, 1968, 28(9-B), 3883. Pierce, W., & Mosher, D. Perceived empathy, interviewer behavior and interviewer anxiety. Journal of Con- sulting Psychology, 1967, 31(1), 101. 113 Reid, D. K., & Snyder, W. V. Experiment on "recognition Of feeling" in non-directive psychotherapy. Journal of Clinical Psychology, 1947, 3, 128-135. Remmers, H. H. A quantitative index of social-psycholog- cal empathy. American Journal of Orthopsychiatry, 1950, 20, 161- 165. Rogers, C. R. A theory of therapyl personality, and interpersonal relationships, as developed in the client centered framework. In S. Koch (Ed.), Psychology: a stugy of a science. Vol. III. Formulations of the person and the social context. New York. McGraw-Hill, 1959. Pp. 184-258. Rogers, C. R. Client-centered therapy. Boston: Houghton Mifflin Company, 1951. Rogers, C. R. The necessary and sufficient conditions of therapeutic personality change. Journal of Con- sulting Psychology, 1957, 21, 95-103. Rogers, C. R. (Ed. ). The therepeutic relationship and its impact: a study of psychotherapy with schizo- phrefiics. Madison: University Of Wisconsin Press, 1967. Siegel, A. I. Experimental evaluation of the sensitivity of the empathy test. Journal of Applied Psy- chology,1954, 38, 222- 223. Smith, H. C. Sensitivity to people. New York: McGraw- Hill, 1966. Snyder, W. Theipsychotherapy relationship. New York: Macmillan, 1961. Speroff, B. J. Empathy and role reversal as factors in industrial harmony. Journal Of Social Psychology, 1953, 37, 117-120. ~ Stefflre, B. Client understanding test. Unpublished, Michigan State University, 1962. Stewart, D. A. The psychogenesis of empathy. Psychoana- iytic Review, 1954, 41, 216-228. Strunk, O. Empathy: a review Of theory and research. Peychological Newsletter, 1957, 9, 47-57. Strupp, H. Psychotherepists in action. New York: Grune and Stratton, 1966. 114 Sullivan, H. S. The psychiatric interview. New York: W. W. Norton, 1954. Taft, R. The ability to judge people. Psychological Bul- letin, 1955, 52, 1-23. Tomlinson, T. M., & Hart, J. T. A validation of the process scale. Journal of Consulting Psychology, 1962, 26, 74-78. Toomey, L. C. Factors affecting empathic performance. Dissertation Abstracts, 1962, 22(8), 2880. Truax, C., & Carkhuff, R. R. Toward effective counseling and_psychotherapy. Chicago: Aldine, 1967. Truax, C. B. Influence of patient statements Of judgements of therapist statements during psychotherapy. Journal of Clinical Psychology, 1966, 22, 335-337(a). Truax, C. B. Therapist empathy, warmth, and genuineness and patient personality change in group psycho- therapy: a comparison of unit measures, time, sample measures, patient perception measures. Journal of Clinical Psychology, 1966, 22, 225-229(b). Truax, C. B. Toward a tentative measurement of central therapeutic ingredients. Arkansas Rehabilitation Research and Training Center, University of Arkansas, 1961. Tuchins, A. S. A variational approach to empathy. Journal of Social Psychology, 1957, 45, 11-18. Tyler, L. E. The work of the counselor. (3rd ed.) New York: Appleton-Century-Crofts, 1969. Wittich, J. J. Generality of the prediction of self re- ports. Journal of Consulting Psychology, 1955, 19, 445-448. Zax, M., & Klein, A. Measurement of personality and be- havior changes following psychotherapy. Psycho- logical Bulletin, 1960, 57(5), 435-448. APPENDICES APPENDIX A LETTERS TO RESEARCH CLIENTS AND INSTRUCTIONS TO INTAKE COUNSELORS Dear Student: We at the Counseling Center believe that one of the ways we have of becoming increasingly helpful to students like yourself is through careful study of your problems and of our effectiveness in helping you with them. We ask that you help us in our study Of student problems by participat- ing in a research project that we are currently undertaking at the Counseling Center. If you agree to participate, your commitment would consist Of completing some tests and inventories that we would administer to you before and after counseling, as well as several brief inventories during counseling. We would also want you to permit us to tape record your counseling sessions. ' All of your responses to the test material are strictly confidential and we can assure you that your re- sponses will be used anonymously in our research endeavors. Since our use of the information will be for scientific purposes, your counselor will not see the results of the tests. The pre-testing and the post-testing will each take about four hours of your time. It is important that you complete the pre-testing before you see your counselor for the first time. We recognize that four hours Of testing is a considerable investment of your time, but our experience has indicated that the information from the various tests and inventories is very helpful in understanding student problems. We would like to administer the tests in two two- hour blocks of time as soon after this interview as is possible for us to arrange a mutually agreeable time. Im- mediately after the interview we would like you to complete several brief inventories which are the first Of these tests. 115 116 After you have completed these inventories, you will sign up for the other two testing times in the Testing Office. In summary, your participation consists of your in- vesting four hours of time before and again after counseling as well as a brief session during counseling and in permit- ting us to record your counseling interviews. If you agree to participate, the counselor you are now seeing will direct you to the Testing Office where we ask that you complete brief initial inventories. After you have completed these inventories you will be asked to sign up for the other two two-hour blocks of testing time. We want to emphasize that your willingness to par- ticipate in no way affects your seeing a counselor at the Center. Our most important consideration is to assist you with the resolution of the problems that stimulated you to seek our help. We do want you to know, however, that through the participation of students like yourself in such research projects as this we have an opportunity of studying in a more objective way some Of the kinds of problems students encounter and the ways in which we can help the student to a resolution of his difficulties. Sincerely, William J. Mueller Assistant Director for Research WJM/mlh 117 MEMORANDUM October 18, 1967 TO: Screening Counselors FROM: The Research Group Involved in the Development of a Tape Library SUBJ: Criteria and Instructions for Selecting Research Subjects The group of us who are interested in extending the tape library attempted to draw up some instructions and criteria that would be usable by screening counselors in assisting us to locate cases appropriate for inclusion as research clients in the library. In this memo, we will describe: (1) the method by which we will inform you about our research subject need; (2) the criteria we would like you to use in ascertaining whether a client is a suitable subject for our project; (3) instructions to you regarding how to approach the client about participation; and (4) instructions that we would like you to communicate to the potential research subject regard- ing our project in order to Obtain informed consent from him. . 1. Regarding informing_you about our subject needs: We want a limited number of sfibjects for our research library and a somewhat larger number of subjects for norming purposes. In addition, we want an equal distribution of male and female subjects in the library. In order to inform you about our ongoing needs, we felt that the best procedure would be for Bob Kurtz, our research assistant, to inform you of our subject needs prior to your screening interview time in any given week when we wish your help. 2. Regardingthe criteria you are to use in requesting research participatiOn of potential sub—ects: We are interested onlygin undergraduate male and female subjects. We recognize that counselor intuition and theoretical differences in assessment play a large part in the screener's decision about cases. There- fore, we ask that you use the following two criteria as "rules Of thumb" in asking clients to participate: (a) the subjects we are looking for are clients who specifically request help with personal problems. 118 Further, we are interested only in subjects in whom we will probably invest a considerable amount of treatment time. Reframing this criterion, we could say that if they were motivated to work on them they would involve a considerable investment of time to work through. (b) The potential client may be am- bivalent about therapy. You are not asked to assess the client's motivation to work through his problems to a successful resolution. In other words, we do not ask you to predict how he will resolve his am- bivalence, but only to assess his willingness to try therapy as a possible solution for his problems as you would with any potential client. Inviting therpotential supject'srparticipation: Provided that the client loOks appropriate for re- search with regard to the criteria listed above, we'd like you to inform him about the project. Two things are important to us in this matter. The first is that the client is fully apprised of the fact that his willingness tO participate in the re- search project in no way influences a decision about whether he will be seen at the Counseling Center. By the time a screening counselor invites the client to participate it should have been established that he will be seen at the Counseling Center for treat- ment. The second important item here is that in presenting the research project to the client you do it in such a way that he is fully informed at the outset about what his p rticipation will entail. In addition to your oral dZscription to the client of the research project, we are also asking you to pro- vide him with a mimeographed letter that we have prepared in which we describe the project and on which we receive in writing from him his willingness to participate. The client's commitment--the specifics: Basically, the cIient's participation consists of his willing- ness to take some pre-post tests, as well as two brief inventories in counseling. The pre-post test- ing each will take about four hours of the client's time. The tests are arranged in two two-hour blocks Of time which the client is expected to complete prior to his seeing a counselor for his first inter- view. The other research requirement is that the client permit us to tape record his counseling ses- sions. Perhaps the most important consideration here is your communication to the potential subject that all of the information that is obtained from him both in the pre-post testing and in his relation- ship with the counselor are confidential; the data 119 will be rendered anonymous and will be used in such a way that there is no_possibility that he can be identified from any of that data that he proVides. We think that if the client understands fully the matter Of confidentiality any resistance to partici- pating in the research project will be reduced con- siderably. . 5. Specific Procedures: After you have talked with the potential subject regarding the project and he has read and signed the letter, we'd like you to follow these specific procedures. We would hope that the client would have time following your screening in- terview to complete a brief initial inventory in the Testing Office. At this time, he will also be re- quested to indicate when he can arrange a time for the two two-hour blocks. Therefore, whether the client has time to take the initial inventories or not, it will be necessary that he sign up for the testing blocks in the Testing Office. This contact regarding the pre-testing will be made directly be- tween the potential subject and the Testing Office and you need not be concerned about establishing time with'the subject. Finally, you will be re- quested tO fill out a brief First Interview Rating Form. We plan to administer the pre-tests and assign the client to a counselor for a first interview within about two weeks following the screening interview. It is Very im- portant to us that the testing be complete before the client sees the counselor for the first time. In summary, (1) We are asking you to assist us in locating subjects for our tape library during your screening interviews, but only at those times when Bob Kurtz explicitly expresses a need to you before your screening interview time. (2) We are looking for undergraduate clients whose problems are such that long-term treatment would be in order. (3) If the client agrees to participate, he will be asked to complete four hours of testing divided into two two-hour blocks prior to and following his therapy here at the Center. A brief test will be administered immediately after the screening interview if at all possible. (4) The client will also be expected to permit us to tape record his interviews. (5) We feel that it is important that the client's agreement is based upon informed consent on his part regarding the research task. Operationally, this would mean that he would express his willingness to participate, that he is aware of the time involved in testing, and that he knows the fabrics Of confidentiality. The client need not have knowledge of the specific tests that are to be 120 administered, nor about the fact that his responses to the initial inventory may be an additional device upon which some decision is made about his appropriateness for the remainder of the research project. This information could not be evaluated by the client and so would not affect his informed consent. On the other hand, it will be important that the client know that the pre-post testing is Of such a nature that the information gathered from it would be very useful to us in understanding and being able to be Of greater assistance to students who present problems such as his. Since his participation will be based in large measure on his feelings stemming from your contact with him, we hope that you will underline the value of the project. APPENDIX B NUMBER OF INTERVIEWS PER CLIENT TABLE B-1.--Number of interviews per client. Client Number Counselor Number Number of Interviews* 8321 02 3 I 8380 03 16 M 8311 04 24 M 8010 05 24 M 8431 06 16 I 8080 08 9 M 8580 09 6 M 8170 10 20 M 8120 12 12 M 8460 13 9 I 8280 15 13 M 8151 18 7 M 8230 19 22 M 8270 21 4 M 8041 22 5 M 8551 24 6 M 8290 25 6 M 8180 26 17 M 8300 27 9 M 8250 28 7 M 8350 ' 29 5 M 8560 31 6 M 8341 35 7 I 8200 37 ‘5 M 8611 38 22 M 8421 39 27 TR 8590 40 12 M 8480 43 12 M 8450 44 15 M 8471 45 15 M 8490 49 17 M Z = 378 §’= 12.19 *1 Independent Termination Mutual Termination Transfer 3 II II II TR 121 APPENDIX C EMPATHY MEASURES USED IN STUDY AFFECTIVE SENSITIVITY SCALE Instructions You will be viewing short scenes Of actual counseling sessions. You are to identify what feelings the clients have toward themselves and toward the counselors they are working with. ‘ Although in any one scene a client may exhibit a variety of feelings, for the purposes of this instrument you are to concentrate on identifying his last feelings in the scene. On the following pages are multiple choice items consisting Of three responses each. Most scenes have two items, but a few have three. After you view each scene, you are to read the items and ask yourself the following question: If the client were to view this same scene, and if he were completely Open and honest with himself, which Of these three responses would he use to de- scribe his feelings? After you decide which response accurately describes what the client is actually feeling either about himself or the counselor he is with, indicate your choice on the answer sheet. Here is a sample item: CLIENT I Scene 1 Item 1 1. This exploring of my feelings is good. It makes me feel good. 2. I feel very sad and unhappy. 3. I'm groping and confused; I can't bring it all together. After you had viewed Scene 1 for CLIENT I, you would read these three statements (Item 1) and would then decide which one best states what the client would say about his own feelings after viewing the same scene. For example if you decide number two best states what the client is feeling, you would then find the number 1 on your answer sheet and darken in the space for number two. 122 123 l. 122:: 222:: 322:: 42:2: 5:22: We will only make use Of the first three answer spaces fol- lowing each item on your answer sheet. Remember you are to concentrate on the latter part of each scene in determining the most accurate description Of the client's feelings. After you view the appropriate scenes, you will have thirty seconds to answer each of the first twelve items. For each of the remaining items, you will be allowed twenty seconds. CAUTION: The item numbers on your answer sheet go across the page, not down the page as you would usually expect! 124 Scale 1 EMPATHIC UNDERSTANDING IN INTERPERSONAL PROCESSES. II A Scale for Measurement1 Robert R. Carkhuff Level 1 The verbal and behavioral expressions of the first person either do not attend to or detract significantly from the verbal and behavioral expressions of the second person(s) in that they communicate significantly less of the second person's feelings than the second person has communicated himself. Examples: The first person communicates no awareness of even the most Obvious, expressed surface feelings of the second person. The first person may be bored or disinterested or simply operating from a preconceived frame Of reference which totally ex- cludes that Of the other person(s). In summary, the first person does everything but express that he is listening, understanding or being sensitive to even the feelings of the other person in such a way as to detract significantly from the communications of the second person. 1The present scale "Empathic Understanding in Inter- personal Processes" has been derived in part from "A Scale for the Measurement Of Accurate Empathy" by C. B. Truax which has been validated in extensive process and outcome research on counseling and psychotherapy (summarized in Truax and Carkhuff, 1967) and in part from an earlier ver- sion which has been validated in extensive process and out- come research on counseling and psychotherapy (summarized in Carkhuff and Berenson, 1967). In addition, similar measures of similar constructs have received extensive support in the literature of counseling and therapy and education. The present scale was written to apply to all interpersonal processes and represent a systematic attempt to reduce the 125 Level 2 While the first person responds to the expressed feelings of the second person(s), he does so in such a way that he eph- tracts noticeable affect from the communications of the second person. Examples: The first person may communicate some awareness of Obvious surface feelings of the second person but his communications drain off a level Of the affect and distort the level of meaning. The first person may communicate his own.ideas Of what may be going on but these are not congruent with the expressions of the second person. In summary, the first person tends to respond to other than what the second person is expressing or indicating. Level 3 The expressions Of the first person in response to the ex- pressed feelings of the second person(s) are essentially ip- terchangeable with those Of the second person in that they express essentially the same affect and meaning. Example: The first person responds with accurate understand- ing of the surface feelings of the second person but may not respond to or may misinterpret the deeper feelings. The summary, the first person is responding so as to neither subtract from nor add to the expressions Of the second person; but he does not respond accurately to how that person really feels beneath the surface feelings. Level 3 constitutes the minimal level of facilitative interpersonal functioning. ambiguity and increase the reliability of the scale. In the process many important delineations and additions have been made, including in particular the change to a systematic focus upon the additive, subtractive or interchangeable as- pects of the levels Of communication Of understanding. For comparative purposes. Level 1 Of the present scale is ap- proximately equal tO Stage 1 of the Truax scale. The re- maining 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. The levels of the present scale are approx- imately equal to the levels of the earlier version Of this scale. 126 Level 4 The responses of the first person add noticeably to the ex- pressions of the second person(s) in such a way as to ex- press feelings a level deeper than the second person was able to express himself. Example: The facilitator communicates his understanding of the expressions of the second person at a level deeper than they were expressed, and thus enables the second person to experience and/or express feelings which he was unable to express previously. In summary, the facilitator's responses add deeper feeling and meaning to the expressions Of the second person. Level 5 The first person's responses add significantiy to the feel- ing and meaning of the express1ons Of the second person(s) in such a way as to (l) accurately express feelings levels below what the person himself was able to express or (2) in the event Of ongoing deep self-exploration on the second person's part to be fully with him in his deepest moments. Examples: The facilitator responds with accuracy to all of the person's deeper as well as surface feelings. He is "together" with the second person or "tuned in" on his wave;length. The facilitator and the other person might proceed together to explore previously unexplored areas Of human existence. In summary, the facilitator is responding with a full aware- ness of who the other person is and a comprehensive and accurate empathic understanding Of his most deep feelings. 127 INTERPERSONAL CHECK LIST INSTRUCTIONS FOR COUNSELOR1 We would like to have you do the Interpersonal Check List as you think your client filled it out just prior to starting counseling. The client was given the following instructions: ”This is a list of words and phrases which describe the way people behave in relation to one another. "On the separate answer sheet marked SELF, mark the item true if the statement is descriptive of you, in your opinion, at the present time. "Your first impression is generally the best, so go through the list as quickly as you can. However, be certain and answer each item." Use the answer sheet provided. Please note the client was not instructed to supply false answers but only to mark the true answers. 1For Interpersonal Checklist, see Leary, T., Inter- ersonal Diagnosis of Personality, New York: Ronald Press, I957. 128 RATING SCALE Counselor's Code Number Date Client's Code Name The items which you are about to rate are words and phrases which the client used in a previous test to describe himself and important people in his life. Try to fill out the following scale as you feel the client would have de- scribed himself before beginning counseling. Here are the instructions which the client received for this scale: "Place a checkmark above the point on each of the following scales which you feel best describes yourself as you are now (before counseling). A "2" on the scale indi- cates that the word describes how you are fairly closely. A "1" on the scale indicates that the word somewhat de- scribes how you are. A "0" indicates a description which is neither closer to one word nor the other. For example, if the following light hair % i i i i dark hair 2 l 0 l 2 were among the items given, and your hair is platinum blonde, you would check the "2" on the left-hand side Of the scale. If your hair is dark brown, you might check the "1" on the right hand portion Of the scale." Remember to fill out the scale as you think the client would have filled it out. 129 Counselor's Code Number Date Client's Code Name Client's Self-Description Before Counseling .L H + + 1 2 1 0 l 2 2' t 1 _L J— _1 2 1 0 l 2 3' e 1 1 1H J. 2 1 O 1 2 4'7 l 1 _L 1 4 2 1 0 1 2 5' .L 1 + : e 2 l 0 1 2 6' t .L i + 4 2 1 0 1 2 7' 'r ‘r i 1 J. 2 1 0 l 2 8' l l 1 1 n 2 1 0 1 2 9° 1 t t 1 —: 2 1 0 l 2 100 L— J I n J 2 1 0 1 2 130 BARRETT-LENNARD RELATIONSHIP INVENTORY-- COUNSELOR FORM (Please do not write your name On this form. It will be coded anonymously and your answers used for research purposes only.) Below are listed a variety of ways that one person could feel or behave in relation to another person. Please consider each statement with respect to whether you think it is true or not true in your present relationship with your client. Mark each statement in the left margin according to how strongly you feel it is true or not true. Please mark every one. Write in +1, +2, +3; or -1, -2, -3, to stand for the following answers: +1 I feel that it is probably true, or more true than untrue. +2: I feel it is true. +3: I strongly feel that it is true. -1: I feel that it is probably untrue, or more untrue than true. : I feel it is not true. -2 -3: I strongly feel that it is not true. 1. I try to see things through his (her) eyes. 2. I understand his words but not the way he feels. 3. I am interested in knowing what his experiences mean to him. 4. I nearly always know exactly what he means. 5. At times I jump to the conclusion that he feels more strongly or more concerned about something than he actually does. 6. Sometimes I think that he feels a certain way, because I feel that way. 7. I understand him. 10. 11. 12. 13. 14. 15. 16. 131 My own attitudes toward some of the things he says, or does, stop me from really understanding him. I understand what he says, from a detached, Ob- jective point of view. I appreciate what his experiences mean to him. I don't always realize how strongly he feels about some Of the things we discuss. I respond to him mechanically. I usually understand all of what he says to me. When he does not say what he means at all clearly I still understand him. I try to understand him from my own point of view. I can be deeply and fully aware of his most pain- ful feelings without being distressed or burdened by them myself. 132 RELATIONSHIP INVENTORY--CLIENT FORM (Please do not write your name on this form. It will be coded anonymously and your answers used for research purposes only.) ' Below are listed a variety of ways that one person could feel or behave in relation to another person. Please consider each statement with respect to whether you think it is true or not true in your present relationship with your therapist. Mark each statement in the left margin according to how strongly you feel it is true or not true. Please mark every one. Write in +1, +2, +3; or -1, -2, -3, to stand for the following answers: +1: I feel that it is probably true, or more true than untrue. +2: I feel it is true. +3: I strongly feel that it is true. -1: I feel that it is probably untrue, or more untrue than true. -2: I feel it is not true. -3: I strongly feel that it is not true. 1. He tries to see things through my eyes. 2. He understands my words but not the way I feel. 3. He is interested in knowing what my experiences mean to me. 4. He nearly alwaysknows exactly what I mean. 5. At times he jumps to the conclusion that I feel more strongly or more concerned about something than I actually do. 6. Sometimes he thinks that I feel a certain way, because he feels that way. 7. He understands me. 10. ll. 12. 13. 14. 15. 16. 133 His own attitudes toward some of the things I say, or do, stop him from really understanding me. He understands what I say, from a detached, Ob- jective point of view. He appreciates what my experiences feel like to me. He does not realize how strongly I feel about some Of the things we discuss. He responds to me mechanically. He usually understands all of what I say to him. When I do not say what I mean at all clearly he still understands me. He tries to understand me from his own point of view. He can be deeply and fully aware of my most pain- ful feelings without being distressed or burdened by them himself. APPENDIX D RAW SCORES OBTAINED BY EACH COUNSELOR ON THE SIX EMPATHY MEASURES TABLE D-1.--Raw scores obtained by each counselor on the six empathy measures. Counselor Score Counselor Score Scores on the Affective Sensitivity Scale 02 64 25 61 03 67 26 61 04 47 27 57 05 54 28 38 06 52 29 54 08 45 31 56 09 54 35 67 10 52 37 64 12 52 38 66 13 49 39 61 15 72 4o 51 18 58 43 56 19 57 44 57 21 58 45 62 22 55 49 65 24 61 N = 31 §'= 57.19 so = 7.236 Judges Ratings of Counselor Empathy Averate Ratings, Carkuff, Empathic Spderstandingrin Interpersonal Process Scale 02 2.50 25 1.93 03 1.92 26 2.17 04 2.92 27 1.71 05 3.42 28 2.25 06 2.25 29 2.75 08 2.67 31 1.62 09 1.88 35 3.00 10 2.00 37 3.00 134 135 TABLE D-1.--Continued Counselor Score Counselor Score 12 2.25 33 3-21 13 1.71 39 2-25 15 2.17 40 2.50 18 2.13 43 2.00 19 1.58 44 2-50 21 1.79 45 1-92 22 1.58 49 2-13 24 2.50 N = 31 §’= 2.26 SD = .491 Predictions of Client Descriptions on Interpersonal Checklist--Per cent of Correct Predictions 02 .688 25 .762 03 .775 26 .716 O4 .641 27 .662 05 .479 28 .641 06 .559 29 .625 08 .634 31 .604 09 .767 35 .688 10 .612 37 .777 12 .708 38 .574 13 .649 39 .641 15 .664 40 .664 18 .725 43 .649 19 .611 44 .716 21 .650 45 .634 22 .746 49 .611 24 .716 N = 31 'Y = .664 so = .068 136 TABLE D-l.--Continued Client Counselor ' Score Predictions of Self Descriptions on Kelly Role Construct RepertoryiTest--Correct Predictions out of 40 321 02 37 380 03 32 311 04 37 010 05 25 431 06 33 080 08 30 580 09 34 170 10 25 120 12 33 460 13 32 280 15 28 051 18 28 230 19 28 270 21 37 041 22 . 22 551 24 31 290 25 5 32 180 26 34 300 27 30 250 28 31 350 29 31 560 31 28 341 35 25 200 37 29 611 38 24 421 39 35 590 40 29 480 43 34 450 44 23 471 45 26 490 49 27 N = 31 §|= 30 SD = 4.16 137 TABLE D-l.--Continued # Converted Client Counselor Score Score Barrett-Lennard Relationship Inventory, Counselor Perceptions after 3rd’Interview* 321 02 29 43 380 03 34 48 311 04 14 28 010 05 26 40 431 06 36 so 080 08 11 25 580 09 15 29 170 10 12 26 120 12 1 15 460 13 4 18 280 15 8 22 051 18 28 42 230 19 16 30 270 21 18 32 041 22 8 22 551 24 20 34 290 25 18 32 180 26 26 40 300 27 _ 47 ‘ 61 250 28 29 43 350 29 25 39 560 31 37 51 341 35 7 21 200 37 18 32 611 38 1 15 421 39 7 21 590 40 - 8 6 540 42 No 480 43 42 56 450 44 18 32 471 45 25 39 490 49 26 40 N = 31 i’= 33.29 SD = 12.84 *Added a constant (+14) to avoid negative numbers. 138 TABLE D-1.--Continued # Converted Client Counselor Score Score Barrett-Lennard Relationshingnventory, Client Perceptions after Brailnterv1ew* 321 02 - 4 10 380 03 22 36 311 04 16 30 010 05 23 37 431 06 —11 3 080 08 19 33 580 09 25 39 170 10 - 1 13 120 12 16 30 460 13 — 2 12 280 15 8 22 051 18 17 31 230 19 4 18 270 21 -13 1 041 22 8 22 551 24 20 34 290 25 4 18 180 26 18 32 300 27 , 24 38 250 28 ' 25 39 350 29 32 46 560 31 20 34 341 35 33 47 200 37 26 40 611 38 13 27 421 39 24 38 590 40 8 26 540 42 NO 480 43 34 48 450 44 22 36 471 45 22 36 490 49 25 39 N = 31 i’= 29.52 so = 12.33 *Added a constant (+14) to avoid negative numbers. 139 TABLE D-l.--Continued # Converted Client Counselor Score Score Barrett-Lennard Relationship Inventory, Client Perceptions, Post* 321 02 - 9 5 380 O3 29 43 311 O4 30 44 010 05 29 43 431 06 7 21 080 08 -12 2 580 09 26 40 170 10 4 18 120 12 36 50 460 13 280 15 25 38 051 18 16 30 230 19 - 1 13 270 21 041 22 551 24 16 30 290 25 9 23 180 26 34 48 300 27 26 V 40 250 28 27 41 240 29 30 44 560 31 _ 341 35 ‘ 36 50 200 37 611 38 - 1 13 590 40 7 21 540 42 480 43 27 41 450 44 27 41 471 45 22 36 490 49 40 54 N = 24 5? = 33.16 so = 14.698 *Added a constant (+14) to avoid negative numbers. APPENDIX E CALCULATION OF INTERJUDGE RELIABILITY FOR.CARKHUFF EMPATHIC UNDERSTANDING IN INTERPERSONAL PROCESSES SCALE TABLE E-1.--Inter79udge reliability for Carkhuff tape ratings. (N = 17) Rater l Rater 2 Sum2 z = 219.5 2 = 215.5 (r2+r2) 435 22 = 48180.25 46440.25 94620.5 12027.5 Sum of Squared Ratings = 6024 Product of Sum 435 X 435 and Mean 34 5563'65 Sum of Squares Raters 94620.5 _ ——T7———- 5563.65 - 2.25 Subjects 12027.5 _ _ _——7T—_- 5563.65 - 450.1 Total 6024 - 5563.65 = 460.35 Error 460.35 - 450.0 - 2.25 = 8.00 Mean Square MS Subjects 450.0/16 28.13 Error 8/16 .50 Reliability of 28.13 - .50 27.63 = 965 Ratings 28.13 + .50 28.63 ° Reliability of 28.13 - .50 27.63 = 98 Average Ratings 28.13 28.13 ' l Ebel, R. L., Estimation of the reliability of ratings, Psychometrika, 1951, 16, 407-424. 140 APPENDIX F COUNSELORS AND CLIENTS EVALUATION OF COUNSELING SCORES TABLE F-l.--Counselors and clients evaluation of counseling scores. Counselor Counselor Client Number Rating Rating 2 3 4 3 4 5 4 3 5 5 4 7 6 3 6 8 3 5 9 2 6 10 3 6 12 4 7 13 l - 15 3 7 18 3 - l9 2 5 21 l - 22 4 5 24 2 5 25 4 6 26 4 7 27 4 6 28 4 7 29 3 6 31 3 - 35 4 7 37 l - 38 4 5 39 4 - 4O 3 6 43 3 6 44 3 6 45 4 6 49 4 6 141 APPENDIX G MEANS, STANDARD DEVIATIONS, AND RANGE OF EMPATHY SCALES TABLE G-l.--Means, standard deviations, and range of empathy scales. (N = 31) Mean S.D. Range Affective Sensitivity Scale 57 7.23 38-72 Carkhuff Tape Ratings 2.26 .491 1.53 “3.42 Predictive-Interpersonal Checklist .664 .067 .479- .777 Predictive-Kelly Repertory Construct Test 30 4.16 23-37 Relationship Inventory-- Counselor Perceptions 33* 12.8 6-61 Relationship Inventory-- Client Perceptions (After 3rd) 29* 12.3 1-48 Relationship Inventory-- Client Perceptions (Post) 33 14.7 2-54 *A constant of +14 was added to all of the Barrett- Lennard Relationship Inventories in order to avoid negative scores (see Appendix D for rawidata on empathy measures). 142 APPENDIX H PROCESS MEASURES USED IN STUDY SELF-EXPLORATION IN INTERPERSONAL PROCESSES A SCALE FOR MEASUREMENT1 Robert R. Carkhuff Level 1 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 to the intro- duction of personally relevant material by the first person but does so in a mechanical manner and without the demon- stration 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. In summary, the second person responds mechanically and re- motely to the introduction of personally relevant material by the first person. 1The present scale "Self Exploration in Interpersonal Processes" has been derived in part from "The Measurement of Intrapersonal Exploration" (Truax, 1963) which has been validated in extensive process and outcome research on counseling and psychotherapy (Carkhuff & Truax, 1965, 1965a, 1965b; Rogers, 1962; Truax, 1963; Truax & Carkhuff, 1963, 1964, 1965). In addition, similar measures of similar con- structs have received extensive support in the literature of counseling and therapy (Blau, 1953; Braaten, 1958; Peres, 1947; Seeman, 1949; Steele, 1948; WOlfson, 1949). 143 144 Level 3 The second person voluntarily 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 prox- imity and without an inward probing to newly discover feel- ings and experiences. Level 4 The second person voluntarily introduces discussions of personally relevant material with both spontaneity and emotional proximity. Example: The voice quality and other characteristics of the second person are very much "with" the feelings 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. Level 5 The second person actively and spontaneously engages in an inward probing to newly discover feelings or experiences 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 to perhaps fearfully and tentatively. In summary, the second person is fully and actively focus- ing upon himself and exploring himself and his world. The present 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. 145 A SCALE FOR THE RATING OF EXPERIENCINGl Eugene T. Gendlin and T. M. Tomlinson Revised by Philippa L. Mathieu and Marjorie H. Klein Stages Stage 1 There is simply a narrative of events with no personal referent used. The client may be telling a story that he is connected with in some way but he does not use himself as a reference point--he says nothing about himself, or of his feelings, attitudes, or reactions. The story told is not "his" story. If a personal referent is used, the content is such that the client reveals nothing private or tender about him- self but merely describes the public aspects of his life. The manner of expression would tend to be matter of fact or to have a rehearsed quality. Stage 2 The client establishes the association between the narrative told and himself by the use of personal referents, but he is involved in telling the story and does not go be- yond it. Any comments he offers about the story do not contain personal reference but function only to "get the story across." Any emotions mentioned are described as part of the story, not the client, and are not elaborated beyond the level of pure description. There is no personal "own- ership" of a reaction to the story. The manner of expression at this stage may be less mechanical and more spontaneous than at stage 1. In some cases, however, the client may seem to be emotionally aroused or involved, but the level of this arousal will re- main constant throughout and will not be referred to specifically. lFrom Rogers, et al., The Therapeutic Relationship and Its Impact, Madison, Wisconsin: University 6f Wisconsin Press, 1967, pp. 589-592. 146 Stage 3 The client is primarily involved in telling a story in which personal reference are used. He goes beyond the story at times to make parentietical comments about his reactions and responses, but these associations are based on the external events only. Such comments can be an ac- count of his feelings about the story, his feelings at the time of the events described, or comments about the personal significance of the events to him. These parenthetical comments must contain personal referents. The person's focus is upon telling his story "better" or elaborating upon it, but he does not use the story to show what he is like as a person. Stage 4 The client is now clearly telling something about himself (his feelings, his image of himself), using himself as the referent for his comments. While these comments may be made in the context of a specific story, their function is not to modify the story but to describe the self. In some cases, the client may have great difficulty finding ways to describe himself and the expression of this diffi- culty alone is sufficient basis to rate 4. The client is now aware of his feelings and reactions and is able to express them. He is doing this in order to communicate what he is like; he is not engaged in a struggle to explore himself nor is he using his feelings as the basis for self-understanding. Stage 5 The client is now using his feelings in a struggle to explore himself. This may take several directions. The client may start with his feelings in a given area and work to understand these feelings, to differentiate them, or to understand how and in what situations they arise. The client may also start with some assumption he has about himself and work to understand how this assumption came about or clarify the implications that this assumption has for him. The client at stage 5 is clearly engaged in a process of self-exploration in order to achieve self- understanding; this process may be extremely difficult for the client and may not be maintained throughout the segment. The expression of difficulty in achieving self-understanding 147 is sufficient basis to rate 5 as long as the client is able to express and elaborate his feelings or to present clearly his self-image (as in stage 4). Stage 6 The client is clearly examining the significance of his feelings or self-concept and is able to arrive at con- clusions about them, or to use the results of this self- assessment as the point of departure for further self- exploration. His formulations about himself provide the links between any elaborations of events or expressions of feeling. In stage 6 the client is able to use the results of self-examination in specific areas to arrive at a deeper and more comprehensive self-understanding. Stage 7 The client does not need a narrative as a point of departure. He can travel freely among feelings and under- stands them quickly. The client has no difficulty in tying together what he is saying and presenting a clear picture of himself--what meaning his thoughts, actions, and feelings have for him. He moves easily from one inward reference to another and is able to integrate them into his experiential frame of reference. APPENDIX I SCORES ON PROCESS MEASURES TABLE I-l.--Scores on process measures. Client Counselor Aggggge Client Counselor Agggige Scores on Self Exploration Scale 8321 02 2.41 8290 25 3.10 8380 03 2.54 8180 26 2.50 8311 04 2.83 8300 27 1.46 8010 05 3.08 8250 28 2.08 8431 06 2.17 8350 29 2.63 8080 08 2.10 8560 31 2.00 8580 09 2.22 8341 35 2.75 8170 10 1.92 8200 37 2.58 8120 12 2.25 8611 38 2.33 8460 13 1.83 8421 39 3.08 8280 15 2.17 8590 40 2.00 8251 18 1.96 8480 43 2.25 8230 19 2.00 8450 44 1.41 8270 21 1.88 8471 45 1.96 8041 22 1.71 8490 49 2.83 8551 24 1.75 N=30 if: so= Scores on Experiencing Scale 8321 02 8290 25 2.91 8380 03 2.66 8180 26 2.54 8311 04 2.95 8300 27 8010 05 2.58 8250 28 8431 06 2.50 8350 29 8080 08 2.58 8560 31 8580 09 8341 35 3.00 8170 10 8200 37 8120 12 2.87 8611 38 2.33 8460 13 2.62 8421 39 8280 15 3.16 8590 40 2.66 8051 18 8480 43 2.37 8230 19 2.95 8450 44 2.37 8270 21 8471 45 8041 22 8490 49 2.83 8551 24 2.41 N= Y: so= 148 149 TABLE I-l.--Continued Client (_) Therapist Client Counselor Affect (+) C. A. Affect (+) Scores on Affect Scale 8321 02 64 77 59 8380 03 74 49 69 8311 04 60 67 51 8010 05 68 68 81 8431 O6 48 88 53 8080 08 95 23 34 8580 09 83 42 71 8170 10 54 78 37 8120 12 82 37 57 8460 13 36 114 36 8280 15 55 71 47 8051 18 -- -- 31 8230 19 66 64 44 8270 21 39 105 31 8041 22 57 77 45 8551 24 86 28 48 8290 25 60 70 67 8180 26 81 41 51 8300 27 87 30 , 75 8250 28 82 41 79 8350 29 91 24 69 8560 31 68 69 52 8341 35 90 86 65 8200 37 79 58 39 8611 38 71 57 48 8421 39 59 59 43 8590 40 74 52 17 8480 43 93 33 53 8450 44 85 28 57 8471 45 79 41 71 8490 49 91 26 59 APPENDIX J RANKS AND SCORES OF THE LOW SCORING COUNSELORS (LOWEST SEVEN) ON THE AFFECTIVE SENSITIVITY SCALE COMPARED TO RANKS AND SCORES ON OTHER EMPATHY MEASURES TABLE J-l.--Ranks and scores of the low scoring counselors (lowest seven) on the Affective Sensitivity Scale compared to ranks and scores on other Empathy Measures. Low Scores on Affective Carkhuff Judge Ratings Sensitivity Counselor Score Rank Score Rank 28 38 18 2.25 3 8 45 17 2.67 5 4 47 16 2.92 4 13 49 15 1.71 16 40 51 14 2.50 7 6 52 13 2.25 3 12 52 13 2.25 8 Client Perceptions After 3rd Interview Client Perceptions Post Score Rank Score Rank 25 5 27 6 19 10 -12 16 16 13 -12 4 - 2 18 -- ~- 8 15 7 12 -11 20 7 9 12 16 13 36 2 Number Scor- Number Scor- Number Scor- ing High on ing High on ing High on Judge Client Per- Client Per- Ratings ceptions ceptions (After 3rd) (Post) High Affective 7 2 2 2 Sensitivity Counselors Number Scor- Number Scor- Number Scor- ing Low on ing Low on ing Low on Judge Client Per- Client Per- Ratings ceptions ceptions (After 3rd) (Post) Low Affective 7 l 3 3 Sensitivity Counselors 150 APPENDIX K RAW SCORE DIFFERENCES ON PRE- AND POST-TESTING ON TENNESSEE SELF CONCEPT SCALE (FITTS). TWO SCALES: TOTAL "P" INCREASES AND NUMBER OF DEVIANT SIGNS REDUCTION TABLE K-1.--Tennessee Self Concept difference scores. . Total "P" NDS Cl1ent Counselor Gain Fewer 321 02 36 10 380 03 34 11 311 04 16 2 010 05 44 17 431 O6 - 4 9 080 08 103 40 580 09 53 14 170 10 2 13 120 12 40 2 460 13 280 15 10 - 5 051 18 0 - 4 230 19 -20 -11 270 21 041 22 - 9 -17 551 24 - 9 — 9 290 25 -23 -10 180 26 12 17 300 27 - 3 ' 0 250 28 42 37 350 29 - 2 2 560 31 341 35 34 8 200 37 611 38 421 39 590 40 38 13 480 43 55 33 450 44 68 22 471 45 6 - 7 490 49 60 20 151 APPENDIX L INSTRUCTIONS TO JUDGES FOR PATTERN ANALYSIS OF MMPI PROFILES AND CALCULATIONS OF INTER-JUDGE RELIABILITY INSTRUCTIONS TO JUDGES FOR PATTERN ANALYSIS OF MMPI PROFILES OBJECTIVE: To determine changes on the MMPI as an indica- tion of psychological change. 1. Compare pre-counseling and post-counseling profiled MMPI scores for each subject. Consider the nine common scales (Hs + 5K, D, Hy, Pd + 4K, Mf, Pa, Pt + 1K, Ma + 2K, Sc + 1K). 2. Score the change as follows: 5 satisfactory 4 partly satisfactory 3 - no change N I partly unsatisfactory I-‘ I unsatisfactory 3. In order to establish intra-judge reliability, please score each profile twice; one week apart. 152 153 TABLE L-l.--Inter-judge reliabilityl MMPI data. (N = 45)2 Rater l Rater 2 Rater 3 2c 166.0 137.5 , 156.5 Zcr 460.0 2x2 1,792.75 2(xc)2 70,954.50 2(xc)2/c 70,954.50 / 45 = 1,576.77 2(xr)2 5,239 2(xr)2 r 5,239 / 3 = 1,746.33 (zzcr) 4602 /135 = 1,567.41 Sums of Squares Raters 1,576.77 - 1,567.41 = 9.36 Clients 1,746.33 - 1,567.41 = 178.92 Error 225.34 - 178.92 - 9.36 = 37.06 Total 1,792.75 - 1,567.41 = 225.34 Mean Square df Ms Raters 9.36 (k-l) 2 4.68 Subjects 178.92 (n-l) 44 4.06 Error 37.06 (k-l) (n-l) 88 .42 Total (nk-l) 134 Reliability of Ratings _ 4.06 - .42 _ 3.64 _ rxy ' 4.06 + 2(.42) "'4790 ‘ '74 Reliability of Average Ratings r ave = 4.06 - .42 = 3.64 = 90 xy 4.067 4.06 ’ 154 TABLE L-1.--Continued Reliability of Average Ratings M _ i _ 4.06 _ ‘ Fiducial Limits of Observed rxy = .74 Ft 40 Upper limit 5% 1.59 60 Lower limit 5% 1.64 df FP 40 Upper limit 5% 15.38 60 Lower limit 5% 5.90 df *rxy Upper limit 5% .83 Lower limit 5% .62 lEbel, R. L., Estimation of the reliability of ratings, Psychometrika, 1951, 16, 407-424. 2Calculations not based on same population used. The sample of 31 cases was drawn from the larger N of 45. APPENDIX M MMPI PROFILE ANALYSIS TABLE M-l.--MMPI profile analysis--average ratings. . Average Client Counselor Ratings 321 02 2.17 311 04 1.67 010 05 3.67 431 06 2.33 080 08 5.00 580 09 4.67 170 10 2.17 120 12 3.50 460 13 280 15 4.17 151 18 3.33 230 19 1.67 270 21 041 22 551 24 3.33 290 25 1.33 180 26 5.00 300 27 3.00 250 28 5.00 350 29 3.67 560 31 231 35 5.00 200 37 611 38 421 39 590 40 4.83 480 43 5.00 450 44 5.00 471 45 3.33 490 49 4.33 N = 23 X = 3.61 SD = 1.24 l - Unsatisfactory 2 - Partly Unsatisfactory 3 - No Change 4 - Partly Satisfactory 5 - Satisfactory 155 APPENDIX N CALCULATION OF INTERJUDGE RELIABILITY FOR CARKHUFF SELF EXPLORATION IN INTERPERSONAL PROCESSES SCALE TABLE N-l.--Interjudge reliability for self exploration tape ratings. (N = 17)1 Rater 1 Rater 2 Sum2 Z = 223.5 2 = 229.5 453 12757. 22 = 4995.25 52670.26 102622.5 Sum of squared ratings = 639.15 Product of Sum and Mean 453 _ 205209 453 X T - T 6035.55 Sum of Squares Raters $93§%34§-- 6035.55 1.06 Subjects 12:57 — 6035.55 342.95 Total 6391.5 - 6035.55 355.95 Error 355.95 - 352.95 - 1.06 = 11.94 Mean Square Subjects §£%%2§ 21.43 11.94 Error _T6_— .74 Reliability of 21.43-.74 20.69 _ 93 Ratings 21.43+.74 22.17 ° Reliability of 20.69 = 965 Averate Ratings 21.43 ' 1R. L. Ebel, Estimation of the reliability of ratings, Psychometrika, 1951, 16, 407-424. 156 APPENDIX 0 CONVERSIONS OF OUTCOME SCORES TO STANDARD SCORES FOR COMBINED OUTCOME SCORE TABLE O-l.--Post t scores--for combined outcome score. Client coggie' MMPI Ten "P" Ten Nos Co Ev Cl Ev Ave. 321 02 38 54 50 49 27 44 380 03 -- 53 42 59 39 311 04 34 45 73 49 39 46 010 05 54 61 52 59 64 58 431 06 59 40 53 49 51 50 080 08 61 75 63 49 39 57 580 09 59 60 56 38 51 53 170 10 38 42 47 49 51 45 120 12 49 56 48 59 64 55 460 13 -- -- -— 28 -— 280 15 55 45 43 49 64 51 151 18 48 42 40 49 51 46 230 19 34 40 40 38 39 38 270 21 -- -- -- 28 -— 041 22 -- 38 34 59 39 551 24 48 42 40 38 39 41 290 25 31 34 40 59 41 46 180 26 61 45 59 59 64 58 300 27 45 42 55 59 51 50 250 28 61 56 67 59 64 61 350 29 50 42 53 49 51 49 560 31 -- -- -- 49 -- 341 35 61 54 50 59 64 58 200 37 -- -- -- 28 -- 611 38 -— -— -- 59 39 421 39 —- -- -- 59 ~- 590 40 60 55 56 49 51 54 480 43 61 61 70 49 51 58 450 44 61 65 68 49 51 59 470 45 48 43 42 59 51 49 490 49 66 61 61 59 51 60 157 APPENDIX P LEAST SQUARES ADDITION FOR PREDICTION OF OUTCOME ANALYSIS OF VARIANCE TABLES FOR OVERALL REGRESSION FOR ALL OUTCOME MEASURES TABLE P-1.--ANOVA for overall regression between Relation- ship Inventory Client Perceptions and Tennessee Self Concept total P differences scale. - S.S. df M.S. F Sig ‘ Regression (about mean) 1 3980.73 5.60 .027 Error 23 Total (about mean) 24 N = 25 Multiple Corr. Coeff. Standard Error of Est. r = .4426 r2 = .1959 26.65 TABLE P-2.--ANOVA for overall regression between Carkhuff ratings, Relationship Invento£y_C1ient Perceptions and Tennessee Self Concept total "P" difference scale. SOS. df M.S. F Sig Regression (about mean) 6082.59 2 3041.29 4.69 .020 Error 14240.04 22 647.24 Total (about mean) 20322.63 24 N = 25 Multiple Corr. Coeff. Standard Error of Est. r = .5471 r2 = .2993 25.44 158 159 TABLE P-3.--ANOVA for overall regression between Relation- ship Inventory Client Perceptions and Tennessee Self Concept, number of deviant signs differences. s.s. df M.S. F ' Sig Regression (about mean) 613.58 1 613.58 3.12 .091 Error 4520.41 23 196.53 Total (about mean) 5133.99 24 N = 25 Multiple Corr. Coeff. Standard Error of Est. r = .3457 r2 = .1195 14.01 TABLE P-4.--ANOVA for overall regression between Relation- ship Inventory Client Perceptions and MMPI ratings. ' SIS. df M.S. F Sig Regression (about mean) 148845.66 1 148845.66 16.20 .001 Error 192889.81 21 9185.22 Total (about mean) 341735.47 22 N = 23 Multiple Corr. Coeff. Standard Error of Est. r = .6600 r2 = .4356 95.83 160 TABLE P-5.--ANOVA for overall regression between Relation- ship Inventory client Perceptions and counselor evalua- tions. SOS. df M.S. F Sig Regression (about mean) 3.36 1 3.36 4.05 .054 Error 24.11 29 .83 Total (about mean) 27.48 30 N = 31 Multiple Corr. Coeff. Standard Error of Est. r = .35 r2 = .12 .91 TABLE P-6.--ANOVA for overall regression between Relation- ship Inventory Client Perceptions and combined outcome. s.s. df M.S. F Sig Regression (about mean) 279.02 1 279.02 8.89 .007 Error 658.62 21 31.36 Total (about mean) 937.65 22 N = 23 Multiple Corg. Coeff. Standard Error of Est. r = .54 r = .29 5.60 1010140“