A STUEPY 10F (SOUNSELO.R-SUPERV§SOR AND COUNSELOR-SUE?” DYAEIC RELATIONSI‘HPS Thesis for the Degree of Ph. {3. MiflHIGAN STATE UEWERSE‘I‘Y LOEEISE PETERS‘GN FORSUZFF 296? 0-169 This is to certify that the thesis entitled A Study of Counselor -Supervisor and Counselor -Client Dyadic Relationship 3 presented by Louise Peterson Forsleff has been accepted towards fulfillment of the requirements for E;— degree in _fi*_e__ College of Education Counseling and’Guidance. ) ,/ ,i 11 ‘1' Major brofesscx ‘QI‘ - mm“ A“. -I‘iii:"\ I 1 _-_‘-‘ ,. . 1.9: *5 . L I B RA R Y 5: Michigan Stem ‘ Universi'fy' .M 7 /‘ ‘l ,- ' ’i‘aL'um l/(L/l v" \../ ABSTRACT A STUDY OF COUNSELOR-SUPERVISOR AND COUNSELOR-CLIENT DYADIC RELATIONSHIPS by Louise Peterson Forsleff This is a study of the consistency of counselor feeling-verbalization in counseling and supervision situations. More specifically, it is an examination of the degree of consistency of feeling-verbalization exhibited by counselors ~in -training in counselor-client and counselor-supervisor relation ships 5. The hypothesis is derived from the theoretical assumption that per- sonality characteristics tend to be consistent from one situation to another and, further, that such consistency is particularly evident in patterns of interpersonal behavior. The specific hypothesis tested is the following: Consistency of feeling-Jverbalization exists between the responses made by the counselor in the counselor- supervision and counselor-client dyadic relationships . The review of related literature suggests that (a) "outcome" studies, so prevalent in counseling research have provided information about the effectiveness of therapy, but, in treating therapy as an homogene- ous treatment variable, have largely failed to probe the complexities of interpersonal behavior which need to be included in an understanding of the nature of therapy; (b) that some degree of counselor consistency exists Louise Peterson Forsleff 2 across counseling and non -counseling situations; and (c) that feeling-verba- lization is an important and empirically sound research construct. The study was conducted during a summer National Defense Educa- tion Act Counseling and Guidance Institute. Data for the hypothesis -testing portion of the study were collected by tape recording all counseling and supervisory sessions held during the Institute. Typescripts of counselor responses were made from tape samples. The counselor responses were rated by three judges as follows: (F) if the response included expression of feeling or affect, and (C) if it was cognitive in content. The proportion of F responses determined a counsel- ing rank order and a non-counseling rank order for the subjects. For the counseling data, percentages of agreement on F responses for pairs of judges were as follows: Judge A with B, 81 .3%; B with C, 77. 1%; A with C, 62.4%. For the supervision data, percentages of agree- ment on F responses for pairs of judges were as follows: Judge A with B, 76.3%; B with C, 69.6%; A with C, 65.6%. The coefficient of correlation (Kendall's tau) between the counseling and supervisory rankings was +0.23. This value of tau is significant be- yond the .05 level of confidence. Therefore, the null hypothesis of no consistency of feeling verbalization was rejected. The research hypothesis that consistency of feeling verbalization exists between the responses made by the counselor in the counselor-supervisor and counselor -client dyadic Louise Peterson Forsleff relationships was accepted. To further explore possible non-«hypothesized relationships between supervision and comiseling, each supervisor described himself and the counselors with whom he worked on the Leary Interpersgmal 91139.15 List. Counselors described themselves, their supervisors, and their clients; clients described themselves and their counselors on the 19.1.1; Analysis of 191.: data demonstrates apparently significant differences between clients' self descriptions and descriptions of them by their counse— lors. Examination of the lg: data revealed several interesting factors which could be implemented by further research. In this study, a significant correlation was observed between feeling- verbalization in supervision and in counseling. One of the ultimate possible implications of this kind of consistency is that some understanding of the counselor-client relationships which will be established by a counsel or in his counseling practice may be obtained by observing his other relationships. A STUDY OF COUNSELOR-SUPERVISOR AND COUNSELOR-CLIENT DYADIC RELATIONSHIPS BY Louise Peterson Forsleff A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Education Counseling and Guidance 1967 I ll lllllll‘f Ill-I'll!!! 454:3: 3/07 ACKNOWLEDGMENTS The author wishes to express her sincere appreciation to the following individuals who gave most generously of their time and experience during the course of this research. To Dr. Norman Kagan, major adviser and chairman of the thesis committee for the many hours spent in consultation, and the encourage- ment, kindness and understanding with which he acted in this capacity. To Dr. Donald L. Grummon for his assistance with experimental design, his critical reading of the manuscript, and his kindness while serving as a member of my guidance committee. To Drs. Lee Dean, Richard Rank and James Costar for their inter— est in the project; and for their willingness to serve as members of my guidance committee. _ To Dr. Albert Elwell for his excellent help as a statistical con- sultant. To Drs. Donald A. Davis, Kenneth B. Engle, and Robert L. Betz who gave generously of their time to serve as judges. To the supervising counselors, enrollee counselors, and clients who participated in the National Defense Education Act Institute and who acted as subjects for this research. To my family. Without their help this task would have been impossible. ii TABLE OF CONTENTS CHAPTER Page I. THE PROBLEM OF THE RESEARCH . . . . . . I Introduction . . 1 Statement of the Problem . 3 Definition of Terms. 3 Delimitations 5 Basic Assumptions . 6 Underlying Rationale and hypothesis. 7 Need for the Study and Theoretical Bases 9 11,. THE RELATED LITERATURE. . . . . . . . 12 Client-Therapist Interaction . . . . . . . 12 Counselor Consistency . . . . . . . . 25 Feeling Verbalization . . . . . . . . 30 Summary . . . . . . . . . . . . . 40 III. METHODOLOGY. . . . . . . . . . . . 41 Overview. . . . . . . . . . . . . 41 The Setting . . . . . . . . . . 41 The Research Subjects. . . . . . . 42 Assignment of Subjects to Supervisors . .. . -. 44 Tape Sampling and Typescripts . . . . _ ._ _ 47 Evaluation of Typescripts by Judges . . . , . . 49 Interjudge Agreement . . . . . . . . 50 Collection of the Data . . . . . . . . . 52 Treatment of the Data . . . . . . . . . 53 The Interpersonal Check List . . . . . . . 55 IV. TREATMENT OF THE DATA . . . . . . . . 59 Results of the Statistical Analysis . . . . . . 60 Non-Hypothesized Observations . '. . . . . 65 iii V. SUMMARY, CONCLUSIONS, AND DISCUSSION . Summary . Conclusions Discussions . . . . . . Implications for Further Research BIBLIOGRAPHY APPENDICES . iv 67 67 70 71 77 79 87 LIST OF TABLES TABLE ' Page 3.1 Summary Chart of Information about the Research Subjects. . . . . . . . . . . . . 45 3.2 Rank Order of Subjects from High to Low Dominance. . 46 3.3 Chart of Assignment of Subjects to Supervisors . . . 48 3.4 Percentages of Agreement on F (Feeling) Responses by Pairs of Judges for Counseling, Supervision, and CombinedData.............51 3.5 Number and Totals of Responses for Subjects in Counseling and Supervisory Situations . . . . . . 54 4. l Ranks of Subjects for Counseling and Supervision Data . 61 4.2 Presentation of Ranked Data Showing S-component for Computation of Kendall's Tau . . . . . . . . 62 4.3 Octant Summary Data for Clients’ Self Descriptions and Counselors' Descriptions of Clients . . . . . 64 4.4 Chi Square Table Testing Significance of Difference Between Clients’ Self Descriptions and Counselors' Descriptions of Clients . . . . . . . . . . 65 LIST OF FIGURES FIGURE Page 3. 1 Illustration of Interpersonal Behaviors Classifications . ; . . . . . . . . . . . 56 3. 2 Circular Grid for use with Interpersonal Check List . 57 Vi APPENDIX LIST OF APPENDICES Selection of Clients . The Judges: Biographical Material Instructions to the Judges . A Sample Type script Page for Judges Rating of Subjects Responses. . The Interpersonal Check List -- Form IV Diagnostic Grids, Illustrative Diagrams, and Charts of Linear Relationships Based on Leary _I___C_I_.Data........ Page 87 89 91 93 95 98 CHAPTER I THE PROBLEM OF THE RESEARCH Introduction As recently as thirty years ago, the literature of psychology was dominated by concepts oriented toward nonsocial aspects of " human behavior. Emphasis on the individual, his instinctual past, his racial past, his symptoms, his intelligence, his achievement, his learning processes or his temperament has given way to a new way of looking at man as a uniquely social being, involved in constant and crucial interactions with other social beings. In the last two decades, the social psychologists and other Systems theorists have put their weight behind the movement toward an interpersonal approach to behavior. Researchers in the field of counseling and psycho- therapy have found this social, interpersonal view of behavior to be an in- valuable contribution to their thinking. Out of this view has come a vast literature of studies concerned with interpersonal influence, conformity, suggestibility, and similar behavior functions. 2 Consistency emerges as an important and relatively unstudied dimen- sion of interpersonal behavior. Practices in psychotherapy imply that per- sonal characteristics remain relatively consistent from one interpersonal situation to another. Theorists representing widely spread schools of thought have referred to personal characteristics which they assumed to be consistent from one life setting to another. Freud's idea of "transfer" in psychotherapy of certain interpersonal relationships (Freud, 1949), Adler's "style of life" concept (Hall and Lindzey, 1958), the learning theorists' explanation of "generalization" across similar situations (Dollard and Miller, 1950), and ROgers' term "congruency" (Rogers, 1959) all describe a kind of consistency. The goal of counselor consistency in and outside of the therapy setting has been a central tenet of the client~centered theorists. Rogers' (1959) insistence that the counselor be congruent in all his relation- ships and that he have "unconditional positive regar " for his client and for others, is a clear statement of expected consistency. All of the above-mentioned investigators and numerous others have indicated that the therapist’s personality, feelings, emotional blind spots, and past interpersonal experiences influence his counseling procedure and effectiveness. In the literature of therapy, there is a paucity of research related to consistency of counselor characteristics such as sensitivity, feeling- verbalization, hostility, or inhibition. Feeling-'verbalization has been identified by numerous researchers as a sound empirical construct, but the only systematic attempt to evaluate consistency of feeling -verbalization across counseling and non-counseling situations which appears in the litera- ture is Ellsworth's study (1962). Statement of the Problem This study is designed to investigate the consistency of counselor feeling-verbalization in counseling and non-counseling situations . More specifically, it is an examination of the degree of consistency of feeling- Verbalization shown by counselors-in -training in counselor-client and counselor-supervisor relationships. Definition of Terms Counselor refers, in this study, to a National Defense Education Act Counseling and Guidance Institute enrollee. Client is understood to mean a high school age counselee who engages in a counseling relationship with one of the Institute enrollee counselors. Supervisor is understood to mean a National Defense Education Act Counseling and Guidance Institute staff member who supervises the counsel— ing practicum experience of enrollee counselors. Supervisory session refers to a period of time approximately one hour in length once a week during which a supervisor and counselor interact. Counselingvsesvsiown refers to a 50-minute period of time each week 4 during which a counselor and client interact. Counselor response is defined as each separate response or proposi- tion made by the counselor during a supervisory session or counseling ses- sion. Each counselor response will be separated from another counselor reaponse either by a client or supervisor response or by a period of silence. Consistency is defined as the presence in the counselor-client inter- action of the same proportion of feeling-verbalization on the part of the counselor as is found in the counselor-supervisor interaction for a given counselor. Feelingjrerbalization is operationally defined as the judgements made by experienced counselors (judges) about the presence or absence of feeling~verbalization in responses made by the counselors in counselor- client and counselor-supervisor interactions. Each counselor response is rated by the judges as (F) or (C). (F) is the rating used by judges to describe a feeling response char- acterized by the expression of affect. (C) is the rating used by judges to describe a content response which does not contain any expression of affect. Feelingor Affect is further defined in this study to mean the expres- sion of present experience of emotion, including expressed, current feelings about self or another individual (i. e., client, supervisor, or other), re - statement of a specific emotional experience of the client or supervisor, and feelings underlying feelings. Feeling or affect as defined in this study does 5 not include vague "How do you feel?" questions, complaints to the counselor, or verbal behavior which is active, cognitive resistance even though these expressions may be couched in "feeling" language. (For further detail, see instructions to the judges in Appendix C.) Deliminations The sample for this study is limited to enrollee-counselors in the National Defense Education Act Counseling and Guidance Institute held at Western Michigan University during the summer of 1966. The sample in- cludes twenty~six males and four females, between the ages of 25 and 51 . Since the selection of participants was based on certain National Defense Education Act requirements, the present sample may be representative of the total population of National Defense Education Act Counseling and Guid- ance Institutes in the United States. Generalization of the research findings is limited to the extent to which the counselors included in the sample are Similar to and representative of other counselors -in~training. No assumptions or inferences are made about the emotional states underlying the verbalization of affect. The design of this study does not in- clude dynamic interpretation or causality involved in the feeling-verbaliza~ tions of the counselors. Although some of the findings of the present research may contribute to an understanding of counselor effectiveness, that is not the primary con- cern of this study. 6 Loosely structured verbal data such as those found in the typescripts used for analysis in this study present problems in precise statistical sam- pling. However, every effort was made to obtain responses representative of all responses made during the time over which the inVestigation extended. Generalization of the findings is limited to the extent to which the data were sampled in a random manner. Any extension or generalization of the research findings should be in accordance with the delimitations outlined above. Basic Assumptions It is assumed that the supervisory sessions and the counseling ses- sions from which data for this study were drawn are not unique or signifi- cantly different from other counseling and supervisory sessions in programs of counselor education. It is further assumed that the behavior of the subjects during the time of the study is representative of their behavior in other similar situa- tions and is not unlike their future behavior. It is assumed that verbal statements of affect proceed from under- lying experiences of affect. More specifically, the assumption is made that such verbalization can be identified and classified. Some research evidence has been produced in support of this assumption (see chapter two, especial- 1y Anderson, 1956). HoweVer, it still seems appropriate to state the matter as an assumption . 7 The further assumption is made that the three expert counselors who served as judges for this investigation are competent and qualified by educa- tion and experience to identify feeling-Verbalization and rate the counselorst responses with appropriate validity for the purposes of this research. Underlying Rationale and Hypothesis The theoretical assumption has been made that personality charac- teristics tend to be consistent from one situation to another and, further, that such consistency is particularly evident in patterns of interpersonal behavior. Sullivan’s definition of personality as "the relatively enduring pat- tern of recurrent interpersonal situations which characterize a human life" (1953, pp. llO-—lll) implies that understanding an individual and his behav- ior requires knowledge of the interpersonal techniques that he employs to minimize or avoid anxiety and of the consistent patterns of relationships that he forms as a result of these techniques. Leary's (1955) general probability principle, which states that sub- jects tend to pull from others (with a probability significantly greater than chance) interpersonal responses which lead to a repetition of the subjects' own favored interpersonal security operations, is derived from Sullivanian theory. In emphasizing interpersonal aspects of behavior, Sullivanian theory assumes that the essence of human success and failure, happiness and de- 8 spair, is determined in large part by the way in which an individual _w- ently sees, symbolizes, and communicates with others. The consistency which is hypmhesized for human beings, generally, is, in this reserach, being hypothesized for counselors in specific elements of their interpersonal behavior. This study is designed to evaluate assumptions related to previous research findings and, specifically, to examine the degree of consistency of feeling-Verbalization exhibited by counselors in counseling and supervisory sessions. The hypothesis -testing portion of this study will analyze ratings of counselor responses for support or rejection of thefollowing hypothesis:l Consistency of feeling-verbalization exists between the res- ponses made by the counselor in the cormselor-supervisor and cormselor-client dyadic relationships. The hypothesis -generating portion of this study will be concerned with relationships which appear to exist between personality variables (as meas- cognitive dimension of verbal behavior in counseling and supervision. The hypothesis stated above was formulated from the purpose and ra- tionale of the study. The hypothesis also reflects the writer's observations in counseling and supervising experiences, and from listening to the tape- recorded interviews of others. 1The hypothesis is restated in testable null form in chapter four. Need for the Study and Theoretical Bases The literature devoted to counselor consistency contains an abun~ dance of theory which assumes consistency, but empirical demonstrations of consistency are conspicuously missing. The demands of science permit neither acceptance nor refutation of assumptions or postulates in the ab- sence of empirical verification. Thus, verification of an assumption upon which we seek to build more theory is an important and timely step in coun- seling research. The consistency of the counselor’s interpersonal behavior outside of the counseling situation is believed to be positively related to his counseling effectiveness (Lubo r s ky, ,1952, Pa rl off, 1956, Abel e s, 1962). Increasing evidence is also being offered in support of the hypothesis that a positive relationship exists between a therapist's ability to communicate with clients on a feeling level and his effectiveness as a therapist (Fiedler, 1950, Rogers, 1959, Strupp, 1963, Abeles, 1961, 1962). In the light of findings from these inVestigations it would seem that better understanding of counselor consistency, generally, and consistency with respect to feeling- verbalization, specifically, could make a real contribution to programs of counselor training. The theoretical formulations and research findings of a number of individuals have contributed to the framework for this study. The impetus for the present study is f01md within the theoretical structure presented by 10 Kell and Mueller (1966) who suggest "that a counselor may often attempt to create a relationship with his supervisor which is similar in some signifi- cant dimensions to the relationships he has with his clients'.‘ (1966, p.100). Also, they submit the idea that, in some respects, "the supervisory rela- tionship parallels the counseling relationship . . . . . The supervisory re- lationship itself is a human interaction process and as such the dynamics of the supervisee and supervisor are reciprocally affecting and changing each other" (p. 129). Rogers' (1959) observations about counselor congruency and consist- ency lend support to the hypothesis underlying the present investigation as do other research results which indicate that counselors appear to have in- dividual "modal approaches" or "styles" which can be seen as consistent with past emotional experiences (Dipboye, 1954, Peterson e_t_a1_. , 1958, Bandura, 1960). Few studies have been concerned specifically with the interpersonal behavior of counselors outside of the counseling relationship. In a study of counselor sensitivity and affective complexity, Abeles (1961) has stated the importance of counselor consistency for counselor effectiveness. In order to be therapeutically sensitive, not only must one be able to perceive and to respond to complex and varied feelings, but there must also be consistency in the responses, not. only toward the client, but OUTSIDE of the therapist -c1ient relationship. Otherwise, the conflicts aroused by the inconsistency of one ’8 response to non-clients might be expected to reduce therapeutic sensitivity. (1961, p. 3). InAbeles' own research on therapeutic sensitivity and self-consist- ency (1961), his findings suggest that the beginning counselor who has great- ll er sensitivity also has a widerrange of feeling-Verbalization available to him and demonstrates this greater capacity for feeling-verbalization on pro- jective tests and in simulated therapy settings . Ellsworth's study (1962) lends support to the consistency hypothesis with respect to counseling and group case conference situations. More research in the area of counselor consistency is needed. A portion of the present study is designed as an extension of Ellsworth's (1962) investigation of consistency of feeling-'verbalization. CHAPTER II THE RELATED LITERATURE The review of literature related to the present study will be present- ed under three headings: (a) client -therapist interaction, (b) counselor con- sistency, and (c) feeling-Verbalization. The literature dealing with client- therapist interaction and counselor consistency provides the background for the development of the present study. Feeling-Verbalization is included be- cause it is to be conceptualized as an experimental construct in this investi— gation. The review of literature pertinent to that construct will include a survey of methodological approaches to the study of such a concept. Client-Therapist Interaction A number of significant empirical studies have been devoted to inves- tigation of the process of client~therapist interactions. Many of these studies could be described as "outcome" research, They have focused largely on the question of effectiVeness of therapy and have attempted to define compo- nents of successful therapy. 12 13 Arbuckle’s (1956) study of client perception of counselor personality is a classic example of the systematic definition of therapist characteristics associated with favorable and unfavorable therapeutic prognosis . Linked with favorable prognosis are such therapist qualities as scientific curiosity, self ~confidence, low M scores, patience with clients, warmth, interest, tolerance, highly developed persuasive and literary skills, and genuineness . Arbuckle found the following therapist characteristics to be associated with poor prognosis: moralizing, lack of understanding, disinterest, authoritar- ian or aggressive attitudes, and behavior which conveys a superior attitude. Aronson (1953), in contrast to Arbuckle and most authors of similar studies, evaluated dependency as prognostically favorable and intellectual curiosity and exploration as unfavorable. Operating from the point of view of the client-centered approach, he also describes adequate emotional con- trol and social adjustment with co-workers as therapist qualities associated with therapeutic success. Rubenstein and Lorr (1957), using self and peer personality ratings of psychotherapists, describe effective therapists as motivated for long- term goals, having a sense of personal adequacy and emotional understand—- ing, conscientious and orderly. Therapist qualities which they labeled as associated with poor client prognosis were anxiety, tension, suspiciousness, and rebellion toward authority. Holt and Luborsky (1958) in their lengthy two ~volume report, which is undoubtedly the most extensive discussion of therapeutic prognosis grow- 14 ing out of research, present findings based on a sample of 228 Fellows in Psychiatry at the Menninger School of Psychiatry. These beginning psychi- atrists were studied almost microscopically during several years of resi- dency in an attempt to devise criteria which would effectively discriminate between the "good" and "poor" ones. In outline form, under six general headings, the therapist qualities associated with a good therapeutic progno- sis, according to Holt and Luborsky are as follows: I. Motivation for psychological work 1. Freedom from material preoccupations 2. Freedom from status-mindedness 3. Need to help2 4. Psychological ~mindednessl 5. Interest in Psychia II. Self-adjustment l . Self~confidence1 2. Self-objectivity2 3. Capacity for personal growth1 4. 5 . 1 Adequate emotional controlz Adequate handling of hostility2 6. Emotional appropriatenes III. Orientation toward clients 1. Objective attitude to authority figures1 2. Objective toward dependent figures1 3. Empaihy (sensitivity)2 4. Conscious of social injustice:Z 5. Freedom from etlmocentrism1 IV. Social adjustment 1. Mature heterosexual adjustment2 2. Social adjustment with co-workers2 3. Feelings of security1 V; Intellectual effectiveness 1. Clarity of thought1 2 . Cultural wealth1 3 . Quality of vierbalizationl 1Significant at .05 level 2Significant at .01 level 15 VI . Honesty 1 . Genuineness 2 Some aspects of intellectual effectiveness which Holt and Luborsky had hy- pothesized as being important to therapy (1. e., perceptual sensitivity and freedom from stereotypy of thought) were not supported as significant by their investigation. A number of researchers and theorists have included honesty or its synonyms in their writings. For Holt and Luborsky it was by far the most significant single variable which differentiated between good and poor therapists. In much of the literature of research and speculation about counselor effectiveness, concern with what might be referred to as "orientation toward clients" continually presents itself as a foremost consideration. While this observation is not surprising, it does support the notion that the interperson- al relationship between the therapist and client is the most important aSpect of psychotherapy- Axelrod (1952) found counselor -client similarity on certain Rorschach variables, particularly those reflecting intellectual functioning, to be condu- cive to therapeutic progress . T‘uma and Gustad (1957) hypothesized that systematic relationships exist between the amount of self -learning by clients in a counseling situation and the degree of client-counselor similarity on selected personality traits. The dependent variable, self ~learning, was assessed by the Self-Knowledge 2Significant at .01 level 16 Inventory. Ten measures were used to assess similarities and differences between counselors and clients . These measures were the Taylor Manifest Anxiety Scale, California F Scale, (Authoritarianism), and the Dominance, Social Participation, Social Presence, Tolerance, Flexibility, Impulsivity, Self~Acceptance, and Good Impression scales of the _C_al__ifgr_n_i_a Personality Inventory. As the indices were set up, better criterion performance was associated with smaller differences between clients and counselors . Only three of the scales attained significant results. They were Dominance, Social Presence, and Social Participation. The Anxiety and Self-Acceptance scales approached significance. On all five of these scales, it was found that the greater the similarity between counselor and client, the better the client's criterion performance. In a similar study with somewhat different results, Gerler (1958) used the Ewing 29219331... Rating wand found that some medium degree of similarity was more conducive to a favorable outcome of therapy than was a high degree of similarity. Snyder's ambitious research project which he reports in Thfinggip- _d}_e__rapy Relationship (1961) was designed to study the interpersonal relation- ship between client and therapist, to discover whether mutual reciprocity exists in the client's and therapist's interpersonal affect, and to identify personality variables which may affect the therapeutic relationship. Exten~ sive use was made of the Edwards Personal Preference Schedule in the in- v vvv—v vestigation. TWO affect scales, the Therapist’s Affect Scale and the Client's l7 Affect Scale, were constructed by Snyder and validated by use in the study. Mhmespta Multiphasic Personality Inventory scores, Leary classifications, biographical material and diagnostic ratings were included among the data. At the conclusion of therapy, each client was ranked by the therapist on the following six characteristics: hostility, amount of rapport, amount liked by the therapist, dependency, success of case, and guardedness. Dependency was the only characteristic that produced a significant correlation. Clients rated low on dependency by the therapist were the clients whose PPS Factor IV (independent, status ~seeking) scores most nearly correlated with those of the therapist. No marked trend appeared in PPS score patterns of clients’ need structures toward approximation of the therapist's need structures. Clients who changed in the direction of the therapist's need structure showed a tendency to increase in needs for achievement and autonomy and to de- crease in needs for succorance and abasement. Clients whose need struc- tures changed in a direction away from that of the therapist showed in- creases in need for succorance and decreases in endurance, introception, and need for change. Clients who were more like the therapist in PPS need structure at the start of therapy were among the most unsuccessful and dif- ficult cases according to the therapist's rating. The four clients who were least like the therapist in TEES need structure at the beginning of therapy were ranked as the most successful of cases and possessed the best rapport with the therapist during the course of therapy. The latter two findings might be seen as a corroboration of Leary's (1957) theory that persons with 18 certain characteristics "pull" predicted kinds of responses from others whose patterns of response are least like their own. Snyder factor -analyzed client w scores and extracted four factors (Factor I: uninhibited, manic, hysterical; Factor II: hypochondriacal, schizoid; Factor III: obsessive, schizoid; Factor IV: uninhibited, depressive, hysteric) which he correlated with therapist rankings and with mean ranks on the Therepist’e Affeet Scale. Clients toward whom the therapist felt the most favorable affect were in the Factor I and Factor [II clusters . A factor which limits generalization of Snyder’s findings is the fact that only one therapist (Snyder, himself) was involved in the research. He also worked with a relatively small, homoge- neous sample -- twenty graduate students in psychology. The study is de- scribed by Snyder as "an exploratory one, designed more for the purpose of hypothesis generating than that of hypothesis testing" (1961, p. 19). Basing identification of counselor effectiveness on judgements by fellow'enrollees in a National Defense Education Act Counseling and Guid- ance Institute, Stefflre, _e_t el. (1962) identified three variables which were associated with being chosen as effective counselors. The variables were (a) higher academic performance, (b) somewhat more appropriate Strong scores, (c) less dogmatism as measured by the Rokeach Degmatism Scale. Demos and Zuwaylif (1966), using (a) the Allport-Vernon-Lindzey Study of Values, (b) the Kuder Preference Record, and (c) the Edwardsfier- sonal Preference Schedule in an attempt to differentiate between most and least effective counselors as rated by supervisors, found only the EPPS to 19 show sensitivity to differences between the two groups. The most effective counselors showed significantly higher scores on need for affiliation and nurturance. The least effective counselors had high scores on abasement, autonomy, and aggression scales. Truax and Carkhuff (1965) present research findings which show a significant positive relationship between therapist transparency or self-con- gruence and the client's level of self ~disclosure or self -exploration. Kahn (1957) examined the question of whether the counselor's choice of therapeutic approaches has influence on the therapy relationship. His findings indicate that the passive-active qualities of a counselor are related to the accuracy of his perception of the client in therapy of a leading type. He found that the therapist who has more negative feelings toward one client than another, perceives the behavior and dynamics of the less favored client more accurately in leading therapy than in reflective therapy. Heilbrun (1961) in his analysis of personality correlates of early ter- mination in counseling suggested that analysis of sex differences or sex matching would improve predictive power. Mendelsohn and Geller (1963) disclaimed the usefulness of that approach. They found no significant sex differences in their research. Using the Myere—Briggs Type Indicator, they rated clients and counselors on the four dimensions of the instrument (Judgment-Perception, Thinking-Feeling, Sensation-Intuition, and Extraver- sion-Introversion). Summing the absolute differences between the scores of the counselor and client for each client-counselor pair on each of the four 20 dimensions resulted in total difference scores. The difference score was a measure of client~counselor dissimilarity, with greater dissimilarity repre- sented by high scores. The number of counseling sessions before termina- tion was the dependent variable. The findings indicated that the greater the client-counselor difference score on each dimension, the fewer the number of sessions. The data consistently indicate that similarity between counse- lor and client is associated with a greater number of counseling sessions. If, as Mendelsohn and Geller assume, the number of sessions reflects to some extent success in counseling, their results can be seen to be similar to those of Tuma and Gustad (1957) and Axelrod (1952). The value of studies which will provide us with information about the effectiveness of therapy cannot be denied. Such an approach has not been abandoned because the effectiveness of therapy has been demonstrated une- quivocally, but because researchers have recognized a difficult problem that arises as a result of looking upon therapy as an homogeneous treatment var- iable in the "outcome" designs. There is some evidence that therapeutic effectiveness rests upon characteristics brought to the situation by each member of the dyad, and upon the nature of the interaction of the members rather than upon differ- ences in theoretical orientation. Lennard and Bernstein (1960) borrowed ex- tensively from the work of Parsons, Bales, and others in their examination of therapist ~client pairs. They describe their project as "research designed to develop and test a methodology and to uncover hypotheses -- not sample 21 statistics to be generalized, but parameters of our group of eight therapist- patient pairs" (p. 35). They also stated "though process is considered in relation to change in structure, we are not primarily concerned with thera- peutic results" (p. 3). Their study is multi-dimensional and multi -level. Their concept of primary system references seems especially noteworthy. The primary system references are concerned with therapist ~client role relations and revolve around the activities and obligations of the therapist and client vis ~a?ViS each other (discussion of the purposes of therapy, com- munication between therapist and client regarding their respective roles, etc.). The other reference categories which they use are the secondary system (reference to therapist and client in other than primary roles -- transference, etc.), the tertiary system (reference to family and other so- cial systems), and the self system (references to self that do not fit in any other category). They also identify "evaluative propositions" which give or ask for appraisals or statements of value, and "affective propositions" which express feeling or emotion. An increase was found in similarity of thera- pist and client behavior over time. The sample size, as in Snyder's re- search, is too small for generalization of findings. Bohn (1965) sought to clarify the relationship of each of three vari- ables (counselor dominance, counselor experience, and client type) to coun- selor directiveness.. Experienced and inexperienced counselors matched on California Psychological Inventory Dominance scale scores anddivided into groups on the basis of high and low dominance, were asked to respond to 22 recordings of first interviews. Three client types (typical, dependent, and hostile) were presented. The three types of clients elicited significantly different responses from the same therapists. The dependent client elicited the most directiveness. The directiveness scores of the high dominance groups were no: significantly different from those of the low dominance groups. Bohn expressed dissatisfaction with the measure of dominance used in his study and suggested that, even though his hypothesis linking counselor dominance and counselor behaviorwas not supported, personality variables such as dominance deserve further attention. Aronson (1953), using a measure of directiveness in the therapy setting and counselor peer ratings of six personality ratings, obtained re- sults which suggest a positive relationship between counselor dominance and directiveness in therapy. Heller, Myers, and Kline (1963) successfully predicted counselor behavior as a function of client type. Their hypothesis that dominant client behavior would evoke passivity from counselors was supported. Tuma (1955) found that dissimilarity in dominance between counselor and client, regardless of the direction of the difference, correlated highly with rated improvement in a counseling situation. Hypothesizing that the productivity of a pair depends upon the degree to which each member can utilize his habitual patterns of interpersonal behavior, T‘uma predicted a rank order of achievement for various combinations of dominant and sub- missive individuals placed in dominant, submissive, and neutral role as‘- 23 Sigiments. He found that assigned roles which were contrary to the pre— ferred modes of the individuals involved aroused anxiety and disrupted achievement. In his study, Tuma obtained a rank order correlation of ..78 between the predicted and empirical rank orders of achievement. The most productive group was made up of pairs in which the dominant subject was assigned the dominant role and the submissive partner the submissive role. The least productive was the group with roles reversed. Paired dominant subjects were more successful than paired submissive subjects. Tuma also observed that the degree of dominance attributed to others appeared to be a function of the subject's personal dominance. The degree of dominance ascribed to the self and partner was observed to vary in the direction of the assigned role. Pallone and Grande (1965) report significant relationships betvveen such counselor personality characteristics as lack of rigidity, spontaneity, ego strength, liberalism, and rapport with the client. Concepts originally derived from research on small groups have been found to be useful tools for examination of psychotherapeutic dyads . Such studies as those by Barrett-Lennard (1962), Brams (1961), Cahoon (1963), Canon (1964), Heller _e_t 3:1 . (1963), Izard (1960), Mendelsohn (1963 and 1966), Meuller (1963), and van der Veen (1965) attempt to define inter- personal variables and assess their effect on relationship. Frank (1961) comments as follows: 24 . . . all research that deals with the dynamics of any one-to.one interpersonal relationship, such as with hypnosis, non~clinical studies of attitude change, etc. are relevant for an understanding of psychotherapy. One can reason that all these are but special instances of the dyadic rela- tionship, therefore, the phenomena pertinent to one should be pertinent to the others (pp. 89-90). The theoretical structuresof several authors suggest ways of con- ceptualizing interpersonal structure. Bales' (1950) interacting procees analysis provides a valuable technique for analysis and for construction of small group models. His observations and laboratory records have led Bales to postulate certain uniformities in group interaction which, for heu- ristic purposes, he presents as a closed system. Sullivan (1953) hypothesizes that an individual's ways of relating to others are functional, enabling him to keep anxiety at a minimum. He takes the theoretical position that interpersonal situations which permit or encour- age the use of a salient interpersonal technique such as dominance or sub- mission give rise to less anxiety than situations that do not. If, as he assumes, the experiencing of anxiety disrupts cognitiVe function and leads to less effective performance in a given task situation, then greater achieve- ment and/ or satisfaction should occur when individuals are encouraged or permitted to assume their habitual modes of relating to others. This hy- pothesis is borne out by I‘uma's (1955) research results reported earlier. Despite the theoretical importance and the explicitness of this recip- rocal aspect of interpersonal theory, very few experimental studies have been concerned with it. 25 Leary has built, on a Sullivanian foundation, an elaborate multilevel interpersonal system which is both theoretical and methodological. Discus- sion of some of the methodology of the interpersonal system appears in chapter three. Leary describes his approach as a "dynamic behaviorism. '.' Two dynamic attributes are contained in his system; (a) the impact an indi- vidual has in interaction with others, and (b) the interaction of psychological pressures among the different levels of personality which he describes. The use which is made of Leary's system in the present study is limited to the former (i.e., the impact of an individual in interaction with others). Over 5, 000 cases, including psychiatric and medical patients and normal controls, have been included in the research for development and validation of the interpersonal system; The literature in client ~therapist interaction summarized here suggests that "outccme" studies have provided information about the effec- tiveness of therapy, but, in looking at therapy as an homogeneous treatment variable, have largely failed to examine the relationship between the coun- selor as a person in a variety of relationships and the counselor as a func- tioning part of the counselor -client dyad. Counselor Consistency In comparison to the number of studies which deal with some aspect of client-therapist interaction, there is a paucity of research which is di- 26 rectly concerned with counselor consistency. Carl Rogers (1959) presents theoretical formulations and empirical evidence which are related to counselor consistency. His concept of coun- selor "congruency" requires the counselor to understand himself and be "deeply himself" in relationships with clients. Genuineness in the thera- peutic encounter includes behaviors which are characteristic of the counse- lor and which have their sources in his personality dynamics . If he is "genuine" in counseling situations and in non-cmmseling situations, it fol- lows that consistencywill exist across both settings. In his attempt to define the ideal therapeutic relationship, Fiedler (1950) examined counselor and client responses to questions and to given therapeutic situations. He concluded that the counselor's past experience was the most important factor in his provision of empathic, communicative, therapeutic relationships with clients. He suggests that the experienced counselor's behavior in the counseling situation is influenced only slightly by his professional affiliation or theoretical background. Strupp (1955), ina study of interaction patterns of counselors from different theoretical orientations and varying levels of experience, found no significant difference in the process of therapy as practiced by the counse- lors included in his sample, Strupp made a series of similar studies, and concluded that the counselor’s own personality, past interpersonal experi- ences, feelings, and emotional blind spots are the prime determiners of his behavior in the therapy situation. Strupp (1960) also indicates that the ther- 27 apist’s feelings toward and perception of a client reflect the therapist's own personality dynamics. He adds that only if a counselor's personality encom~ passes "integrity, honesty, and dedication" (1960, p. 219), can he achieve positive therapeutic results. Strupp found a significant difference in the counseling behavior of therapists who had undergone personal psychoanaly- sis and those who had not. Wrenn (1960), using a modified version of Bales (1950) interaction process analysis categories, attempted to complement the experimental weaknesses of Fiedler's study (1950). Wrenn sought to provide data which would emphasize the theoretical differences between counselors instead of avoiding them. He looked for similarity of counseling behavior in the inter- Views of therapists subscribing to different theoretical orientations. His findings concur with those of Fiedler and Strupp. He reported finding no significant procedural differences among the therapists he studied which could be accounted for by their theoretical orientations . The differences between counselors, as studied by Fiedler (1950), Strupp (1955, 1957, 1960), Wrenn (1960), and Peterson, Snyder, Guthrie, and Ray (1959) appear to be personal differences rather than differences of technique or theory. The findings of such studies strongly suggest that the personality characteristics of the counselor greatly influence his comseling behavior. Other writers support these findings . In a study of counselor sub- roles, Danskin (1955) and Hoffman (1959) found individual counselor prefer- 28 ences a factor which determined the sub~roles played by the therapist in the counseling situation. Dipboye (1954) stated his belief that there is a general similarity in "counselor style" when units of style are considered together. Grigg and Goodstein (1957) reported that counselors in their study were found to show identifiable individual differences in their counseling behavior or "approach, " and that the counselors showed consistency in the use of such "modal approaches. " Bandura's findings (1960) demonstrate that the counse- lor who expresses hostility during the counseling interview also permits and encourages the expression of hostility by his client. Kell and Mueller (1966) suggest that if a counselor avoids affective struggle with his client, he will also avoid it with his supervisor. Counselor consistency across counseling and non-counseling situa- tions is implied by a number of authors. Some treat consistency in a more specific manner. For example, Bordin (1955) makes the statement that The emotional tone of the counselor'8 relationship to the client prob- ably reflects most fully his natural ways of interacting with others, his per- sonality conflicts and integrations, his life style.(p. 170). Patterson (1959) voiced his view that the counselor's behavior in therapy must be consistent with his own personality structure; consistent with his feelings and attitudes. Recognizing situational variance, he added that the counselor's personality may appear to be different at different times or with others, but stressed the common elements which he believes exist in inter- actional situations. Fiedler (1950) described a good relationship in cormsel- ing as very much like any good interpersonal relationship, and supported his 29 view with research findings. He also found that lay raters described a good therapeutic relationship in a way that Very closely approximated the ratings by experienced therapists . Luborsky (1952) reported that beginning psychiatrists who received better ratings as therapists were also rated better in their personal relation- ships with research staff, supervisors, ward personnel, and fellow students. Concurring with Luborsky's findings, Parloff (1956) indicated that of two therapists, the one that was rated highest for the quality of therapeutic inter- action with members of therapy groups with whom both therapists worked, was also rated higher for other social relationships. Little investigation has been made of consistency in terms of speci- fic components of counselor behavior such as feeling-Verbalization, hostili- ty, sensitivity, or inhibition. Abeles (1961), in his study of therapeutic sen- sitivity and self consistency attempted to define therapist response to client affect. He described counselor consistency as an essential ingredient in gemiine sensitivity. He also postulated that counselor sensitivity includes the consistent ability to perceive affect and to respond to affect which varies in complexity. Using Henry and Schlien’s (1958) affective complexity score, Abeles found significant correlations between affective complexity scores on a projective test and in a simulated therapy situation. Indicating that the most frequent activity of the counselor is understanding and/ or respond- ing to client affect, Abeles put particular emphasis on feeling-verbalization. Ellsworth (1962) reported significant correlations between feeling- 30 Verbalization (one dimension of behavior) in counseling sessions and in group case conferences. Such findings as those reported above lend convincing support to the hypotheses that (a) the personality characteristics of the counselor are an integral part of his counseling bdiavior, and that (b) as a result, consistency canbe expected to occur in interpersonal relationships which the counselor forms in and outside of the therapy setting. Feeling-Verbalization The literature which is concerned with measurement, identification, or classification of feelingdverbalization in counseling reports two basic approaches to the problem: (a) methods of rating by judges, and (b) methods using semantic instruments such as word lists, grammatical ratios, or physiological correlates. There is some overlap or combination of the two methods in some studies, making discreet categorization difficult at times. The client ~centered movement with its strong emphasis on affect, led to much interest in the development of methodology for making content analyses of verbal material from counseling interviews. A number of inter- action variables, feeling-verbalization prominent among them, were sub- jected to empirical scrutiny. Among the early studies in this group, a study by Porter (1943) ap- pears to mark the first reference to feeling-verbalization as an empirical 31 construct. He reports a study in which judges rated counselor behavior (in- cluding identification of feeling-Verbalization) in interview situations. Find- ing an interjudge reliability coefficient near .95, Porter concluded that ". . . judges can agree with ahigh degree of reliability as to the procedures which the counselor follows. " I Earlier, Baldwin (1942) in approaching a similar problem revealed both the fruitful possibilities and the tedious process involved in a quantita- tive approach to verbatim interview material . Snyder (1945) found during his investigation into the nature of non- directive therapy that he was able to classify eighty per cent of content and feeling items from forty-eight interviews in the same manner after a time lapse of one month. An untrained judge was able to categorize sixty per cent of the items in exactly the same way and seventy-four per cent in the same general category groups. These findings led Snyder to conclude that material from therapy interviews can be analyzed by the categorization method and that measurable data can be obtained in that manner. Later, Reid and Snyder (1948) attempted to determine the accuracy of therapist recognition of client feeling-Verbalization in the therapy inter- view. They computed a rank correlation of .70 between what "good" coun~ selors thought was productive feeling-'verbalization and what supervisors identified as such when they listened to taped recordings of therapy inter- vie'WS. They also reported that counselors appeared to have consistent personal preferences. for designation of certain feelings, that "good" coun- 32 selors showed this tendency toward consistency more clearly than those counselors who were rated as less effective, and that the "good" counselors were able to recognize a broader range of feelings than the less effective counselors . Seeman (1.949), also investigating process in non-directive therapy, presents evidence which supports the validity and reliability of the feeling- Verbalization construct. He reported concurrence by four judges eighty‘one per cent of the time when affective units were identified and ninety-five per cent of the time when cognitive content units were identified. In a study of self-reference in therapy interviews, Raimy (1948) re- ported that four judges, rating positive, negative, and ambivalent client feelings about self and others, achieved 81. 8 per cent agreement in the initial rating and 80 per cent agreement in test-retest ratings over a period of six months. Raimy referred to the work of Baldwin (1942), Curran (1945), and Rogers (1951) to support his claim of validity. He also stated, "The reliability studies indicate that the method can be applied to verbatim res- ponses . . . with considerable hope of gaining an objective picture of changes taking place in Verbalized self-reference" (p. 158). In an attempt to measure tension, Dollard and Mowrer (1947) estab- lished a Discomfort-Relief Quotient (DRQ) . They instructed judges, using written case nOtes as material, to list all the words which they identified as indicating discomfort (unhappiness, suffering, pain, tension, etc.), and all the words which they saw as indicating comfort or relief (satisfaction, enjoy- 33 ment, comfort, etc.). Sentences and clauses were also rated. Instructions to the judges about rating sentences were: "Does the sentence (a) disturb, (b) relieve, or (c) fail to affect you either way? If you react with tension to the sentence, i.e., if it makes you feel annoyed, excited, appetitive, or apprehensive . . . " rate it as a discomfort sentence. If, conversely, ". . . the sentence gives you a sense of well-being, relaxation, or satis- fl faction . . . rate it as a relief sentence. Independent clauses or "thought units" were also rated in the same manner. Interjudge reliability for eight judges of individual word analysis was .80. lnterjudge reliability for sen- tence analysis, again with eight judges, was .81. An interjudge reliability coefficient of .88 was found when ten judges rated the clauses or thought units. Positive inter-correlations were found to exist between the graphic representations of the three rating methods. Clause and sentence curves showed a correlation coefficient of .90, as did word and sentence curves. Clause and word curves correlated positively at .93. The clause scoring method produced greater parallelism and coincidence between curves than either of the other methods. Dollard and Mowrer state, "It seems indubi- table that something ’real’ is being scored when results so similar are ob- tained by three somewhat varying methods" (1947, p. 13). Their study and subsequent investigations using the DRQ indicate that it is a valid and reli- able measure of tension change which can be applied to case records. Such success with the DRQ implies not only that gross feeling-Verbalization can be scored reliably and used as a valid empirical construct, but also that 34 more specific kinds of feeling-verbalization can be identified reliably. Murray (1956) stated his belief that all of the observable events in any thermieutic encounter can be meaningfully sorted into such categories as verbal, physiological, and gross behavioral. Selecting the area which be labeled "verbal" for investigation, he sought, by means of content analysis procedures, to identify "meaning content" of the language used. He used a "meaning phrase" construct which was similar to the DRQ thought-clause. Reliability coefficients for identification of the meaning phrases ranged from .88 to . 94. TWO judges classified the meaning phrases into such categories as affection, independence, dependence, sex, etc. with reliability coeffici- ents ranging from .58 to .98. Interjudge reliability for the two judges was .86 for classification of affection, .77 for classification of anxiety, and .78 for classification of frustration. Murray also explored expressions of client feeling about the therapist and found interjudge reliabilities (again, for two judges) ranging from .58 to .97 for eight feeling categories and a reliability coefficient of . 97 when all eight categories were pooled. Murray concluded that although the validity of his approach cannot be ideally estab- lished, some validity is inferred by the usefulness of his data. . Anderson (1956) in a study designed to link phenomenological and physiological operationalism concerning feeling-verbalization obtained re- sults which convincingly establish validity for judges’ ratings of feeling- Verbalization. Anderson studied the same client over a period of ten thera- py sessions. The client was "wired up" to demonstrate changes in heart 35 rate and other physiological changes which are known to be related to emo- tional states. He obtained a correlation of .69 (tau coefficient) between the physiological changes measured during the sessions and the therapist's judgements about intensity of feelings. He also obtained the following reli- ability coefficients for two judges rating four categories: (a) person refer- ents, . 82, (b) time referents, .75, (c) feeling-vepqiression, .73, and (d) intensity of affect or feeling expressed, .82. After a ten-month interval, a test—retest procedure showed an overall agreement of . 85. In Anderson‘s words, There are concomitant modifications of the emotional and cognitive processes of client behavior during client-centered therapy which are re— 1ated to each other, and they will vary together in a predictable manner. It was assumed that cognitive behavior could be measured from the client’s speech during the interview and the emotional behaviors could be measured from a physiological system known to be responsive to emotional upset . . . As the frequency of client talk units denoting self-reference, nega- tive affect, and present time reference vary together, there should be a concomitant variation in the heart rate . . ." (p. 175). Physiological tension was related to (a) the referent of the client's speech, (b) the affective valuation (positive or negative), and (c) the time reference in the speech. More specifically, self references, negative valuations of experience, and present-time oriented statements were most significantly related to physiological tension elicitations. An example would be "I feel blue, " as opposed to "he felt O.K. " The example illustrates each point respectively as it refers to (a), (b), and 36 (c) above; in other words, "I" as a self-referent rather than "he, " "feel" in the present rather than "felt" in the past, and "blue" as a negative rather than a positive valuation. In addition such statements as "He was six feet tall" or the question "What is your major?" or "How old did you say your mother was?" are illustrations of cognitive intellectual content, as opposed to specific feeling expression. Lending further support to the empirical respectability of the con- struct of feeling~verbalization are the more recent studies by Salzinger (1958, 1960). His theoretical orientation is similar to that of Dollard and Miller (1950) . Salzinger views therapy as a learning situation in which he identifies approach-avoidance conflict about feelings, reduction of anxiety and tension, the decrease of displacement, and the increase in positive feeling expression as the therapy-learning experience progresses. In his research, Salzinger investigated reinforcement of client feeling response by therapists, using both schizophrenic and normal client populations. He defined feeling or affect responses as follows: The response class of affect used in this experiment was defined as any statement by the patient describing or evaluating his own state (other than intellectual or physiological). The response class therefore included only those affect statements which begin with pronouns "I" or "we" as the following examples indicate; "I am satisfied, I‘m happy, we enjoyed it, I like him, I'm very close to him, I was mad at him, I feel . . . (followed by other words), I was frightened, we couldn’t take it, I always suffered. " Quotations in which affect was described to the speaker by another individual (e.g. , "my husband said I didn't feel good") although fulfilling all other cri— teria, were excluded on the basis of not being direct expression of the patient's own affect. 37 Statements like "I am happy and excited" were considered as one. affect statement only, because the pronouns "I" or "we" did not precede the second affective word. On the other hand, incomplete statements (in the sense that the object of the affect is not mentioned) like "I love . . . " or "We feared . . . " were viewed as bona fide responses. Certain types of private events or internal states were excluded from the response class of aflect because they referred primarily to intellectual processes, or to actions which are sometimes but not always associated with affect, or to desires which appear to constitute a class of responses different from- the affect class defined here. Some examples are the follow- ing: "I am confused, I am not well, we forgave him, I threaten her constant- ly, I didn't trust them" (1960, p. 242). The interjudge reliability coefficients for two judges as they inde~ endently coded 15 recorded interviews for self—referred affect were com- puted separately for each response class and ranged from .79 to 1.00. The interjudge reliability coefficients were the same in both studies. The high reliability coefficients led Salzinger to conclude that "The aflect response as defined in this experiment can be objectively isolated and counted" (1958, p. 85). "It was further shown that a verbal response class" (like feeling- verbalization) "can be reliably isolated and reacted to" (1958, p. 90). Salzinger also stated his belief that some of the reliability coefficients would have been even higher if the quality of recording had been better. For three judges, Levine (1958) found the same interjudge reliabili- ty (.92) when the amount of. feeling expressed in counseling interviews was rated, when the feelings were further categorized as either positive or nega- tive,,. and when the feelings were rated as mild, moderate, or intense. His study lends further support to the empirical reliability of the judges’ rating method for use with taped material taken from counseling interviews. 38 In a study designed to explore reinforcement in a simulated therapy situation by selective response to content or feeling, Waskow (1962) again demonstrated the experimental reliability and validity of feeling~verbaliza- tion. Seeking to link the client-centered theorists' (Rogers 1942,, Seeman 1954) emphasis of the importance of feeling as opposed to content in client verbalizations with the learning theorists' (Dollard and Miller, 1950) View that previously punished feelings must be expressed in the presence of a non-punitive therapist in order for new learning to occur, Waskow studied 32 subjects in situations where feelings were reflected and in situations where they were not. The aspect of her research which has particular relevance to the present study is the rating of tape recorded interview ma- terial. 'Ihree categories were used for classification of responses, (C) content, (FC) feeling and content, and (F) predominantly feeling. The fol- lowing example illustrates the manner in which the responses were dassi~ fied. Subject: Sometimes my father yells at my mother over nothing. But, boy, can she get mad. And I don't like it one bit when he does it. Therapist: (F): You really get upset about that. (FC)t You feel real upset when your father gets mad at your mother about things that don‘t matter. (C): Sometimes your father gets real mad and yells at your mother about little things. Inter-rater reliability between a therapist and graduate student in clinical psychology was . 81 for the (F), (PC), and (C) classifications. 39 It has been demonStrated that laymen or untrained scorers can make reliable and valid judgements about feelings in counseling and non-counsel- ing siniations. Beldoch (1961) investigated the ability of his subjects (who were not therapists) to identify feeling expression in Vocal, musical, and graphic media. He found that his subjects could reliably identify given feelings from spoken, musical, and abstract art stimuli which were pre- sented to them. He also found significant intercorrelation among the feel- ing identification abilities in the three media. He concluded, "the reliabi- lity coefficients for each instrument ranged from..60 to .70 indicating that the ability to identify the communication of feeling in the separate media is a relatively stable phenomenon" (p. 1274). In a study concerned with awareness of and responsiveness to affect as a function of counselor training and supervision, Abeles (1962) had judges (some of whom were untrained personnel) rate the following fiVe kinds of affect: (a) affect inhibition, (b) monotonic positive affect, (c) mono- tonic negative affect, (d) ambivalent affect, and (e) ambivalent affect with controlled resolutions indicated. He reports an interjudge reliability co- efficient of approximately .70. The literature related to feeling-Verbalization presents evidence that feeling-'verbalization is an important and empirically sound research construct, and reports a number of methodological approaches to the study of feeling-*verhalization. 40 Summary The review of related literature which has beat presented in chapter two suggests (a) that "outcome" studies, so prevalent in counseling research have provided information about the effectiveness of therapy, but, in looking at therapy as an homogeneous treatment variable, have largely failed to examine the relationship between the counselor as a person in a variety of relationships and the counselor as a functioning part of the counselor-client dyad, (b) that counselor consistency has been hypothesized to exist but veri~ fication of this hypothesis remains to be established, and. (c) that feeling- Verbalization is an important and empirically sound research construct. CHAPTER III METHODOLOGY Overview This research was designed to investigate consistency of counselor feeling-verbalization across counseling and supervisory situations. The Setting The study was conducted during a summer National Defense Educa- tion Act Counseling and Guidance Institute held at Western Michigan Univer- sity from June 20, 1966, to August 12, 1966. The structure of the Institute included large group didactic instruction, discussion periods, practicum experience with individual and group supervision, observation of counseling, case conferences, group process interaction, and professional field place- ment in selected community agencies and services. No one~hour laboratory or practicum experiences were scheduled each week for the counselors who were enrolled in the InStitute. These practicum experiences centered around counseling with clients who were 41 42 referred by counselors from secondary schools in the Kalamazoo area.1 The counseling sessions were conducted in rooms fitted with one way glass for observation and were tape recorded for supervision and discussion. In- dividual supervision sessions, held in supervisors' offices, were scheduled once a week and were also tape recorded... The four supervisors on the staff of the NDEA Institute were selected because of their supervisory experience and their recognized ability to work with individuals and small groups. One was a Clinical Staff Psychologist at the Veterans Administration Hospital in Battle Creek, Michigan, and a spe- cialist in counseling theory and process. Two were Assistant Professors in the Counseling Center, Western Michigan University, and specialists in the supervision of individual and group counseling. The fourth was an Assistant Professor in the School of Education, Western Michigan University, and a specialist in research, individual counseling and the supervision of counsel- ing. The Research Subjects The subjects of this study were the enrollee-counselors who partici- pated in the National Defense Education Act Institute at Western Michigan University during the summer of 1966. Thirty enrollees, twenty-four males 1The procedure used in the selection of clients is presented in Appendix A. 43 and six females, were included in the sample. The age range of the enroll- ees was from 24 -to 51 with the mean age being 33.5. Twenty-seven of the thirty held masters' degrees. The other three had earned 28, 27, and 25 semester hours of graduate credit, respectively, and expected to complete the requirements for masters' degrees at the conclusion of the Institute. All of the subjects had been enrolled in at least one counseling prac- ticum course before admission to the Institute and had completed, with a B or better average, not less than 22 and not more than 36 semester hours in graduate courses applicable to a degree in counselor education at Western Michigan University. All of the enrollees were married except two of the women, who were members of religious orders . Twenty-seven of the Institute members had children. The number of children in the 27 families ranged from one to nine with a mean of 2. 9. The enrollees had between one-and twenty-one years of "teaching experience" with a mean of 8.7 years. Included in the years of "teaching experience" for twenty-seven of the enrollees were from one to seventeen years of qu~ or part-time counseling with a mean of 3. 1 years of counseling experience. Three of the Institute enrollees were employed in private schools or institutes. The remaining twentysseven were public school counselors and/or teachers. Twenty-six of the enrollees had been involved in counseling and guidance activities at least half -time in junior or senior high schools prior to enrollment in the Institute. All thirty of the enrollees were assured of em» 44 ployment as counselors in secondary schools at least half time after the In- stitute experience. Enrollees came to the Institute from the following nine states: Cali- fornia, Illinois, Michigan, Minnesota, New Jersey, Ohio, Pennsylvania, Utah, and Wisconsin. Seventeen of the enrollees were from Michigan; thirteen from "out-of-state." A summary chart of information about the research subjects is pre— sented in Table 3. l . Assignment of Subjects to Supervisors The counselor-enrollees were ranked in order of dominance scores obtained by administration of the Leary Interpersonal Check List at. the first session of the Institute.2 The ten enrollees at the high dominance end of the continuum were assigned randomly to the four supervisors.3 The ten coun- selors at the low end of the continuum were assigned randomly to the four supervisors. Random assignment of the middle group was also made. The purpose of this method of assignment was to insure that each supervisor would be working with a group in which a range of dominance could be as- sumed. A chart showing assignment of subjects to supervisors is presented 2Rank order of subjects from high to low dominance is presented in Table 3.2. 3Two supervisors had eight counselors assigned to each of them. The other two supervisors had seven, the difference being a function of other commitments to the Institute by the supervisors. 45 Table 3.1 Summary Chart of Information about the Research Subjects Range Age 24-51 No. of children 1-9 Years of teaching experience 1-21 Years of counseling experience 1-17 Allele Female Number of Enrollees 24 6 No. of Enrollees with Masters Degree No. of Enrollees with counseling practicum experience No. of Enrollees having completed not less than 22 and not more than 36 hours toward graduate degree in counseling No. of Enrollees married No. of Enrollees with children N o. of Enrollees employed as public school counselors and/or teachers No. of Enrollees in private schools or institutes No. of Enrollees in counseling at least half-time prior to enrollment in Institute N o. of Enrollees assured of at least half-time counseling positions following the Instinite State Residence of Enrollees California Illinois Michigan Minnesota New Jersey Ohio Pennsylvania Utah Wisconsin Total Mean “(2)1000 i-‘NOUI Total 30 27 30 30 28 27 26 30 Number i—i—Ntni—i—xtr—r-a 1 (JD 0 46 Table 3.2. Rank Order of Subjects from High to Low Dominance Dominance Ranking Counselor Identification 1 06 2 30 3 29 4 High 18 5 Dominance 08 6 Group 20 7 19 8 25 9 17 10 21 ll 28 12 03 13. 5 07 13 . 5 Average 27 15 Dominance 13 16 Group 24 17 10 18 23 19 05 20 04 21.5 15 21 .5 26 23 22 24 Low 02 25 Dominance ll 26 Group 09 27 14 28 01 29 12 47 in Table 3.3. Tape Sampling and Typescripts All the counseling and supervisory sessions of the Institute were tape recorded. Sample segments from all the tapes were used to derive the research data. Since rate of interaction was not a variable under study in this research, samples were taken in such a way that an equal number of responses from each tape were included. Some researchers (Kiesler, e_t 9:1. .) have suggested that, for most research purposes, tape segments from parallel portions of tapes can be compared more effectively than random segments drawn from different portions of tapes. They also conclude that there is no "best" way to select time segments, but that such a choice de- pends on the focus of a particular study. They indicate that segments from the final portion of the interview are usually superior in discriminatory pow- er to segments from earlier portions . The sample from each tape was taken following thirty minutes of elapsed time from the start of the tape and included the next five counselor responses which followed. If the session ended before five counselor res - ponses were made in the prescribed tape segment, counselor responses made just prior to the thirty minute point were included in the data. In these cases the tape was "back-tracked" to meet the five-response criteri- 011. 48 Table 3.3. Chart of Assignment of fiibjects to Supervisors (showing high, average, and low dominance divisions) Supervises- insures?! 33. - v Cmmselor Rank Counselor Rank 08 5 20 6 21 10 25 8 ---------------------- 17 9 07 13.5 --------------------- 13 15 24 16 23 18 05 19 15 21.5 09 26 12 29 16 30 Supervisor C __ fig Supervisor D - - Counselor Rank Counselor Rank 06 1 30 2 29 3 18 4 19 7 _---..----..--..----..-- ---------------------- 28 11 03 12 27 13.5 04 20 10 17 26 21.5 22 23 11 25 02 24 01 28 14 27 49 Supervisory tapes and counseling tapes were sampled in an identical manner. Typescripts were made containing only the counselor responses from the sample segments of each counselor-supervisor and counselor- client tape. Each tape was heard and converted into typescripts by the investigator. Each response was numbered and groups of responses were coded. A counselor response was defined as each statement separated from another counselor response either by a client or supervisor response or by a period of silence. "Uh huh" or "uh hmm" sounds were not included as responses. Evaluation of Typescripts by judges Three judges4 were selected to evaluate the typescripts. The judges were given typescripts of counselor responses made in counseling and su- pervisory sessions. They were asked to rate each counselor response as (F )if it included expression of feeling or affect or (C) if it was cognitiVe in content. Complete instructions to the judges are reproduced in Appendix C. A sample typescript page for judges’ rating of subjects‘ responses is included in Appendix D . F” v Vvvfivvv v'v‘vv 7+7 fwfi vv f ' ferv—Vfivv 4Biographical material about the three judges is presented in Appendix B . 50 Interjudge Agreements Percentages of agreement between pairs of judges on F (feeling) res- ponses were computed on the basis of the maximum number of possible F- matches between pairs of judges. Chi square assessment of the relative divergence of each pair of judges' actual F ~percentage agreement from P . 50 (chance) agreement was computed. For the counseling data, percentages of agreement on F responses between pairs of judges were as follows: Judge A with Judge B, 81.3 per cent; Judge B with Judge C, 77. 1 per cent; Judge A with Judge C, 62.4 per cent. For the supervision data, percentages of agreement on F responses between pairs of judges were as follows: Judge A with Judge B, 76.3 per cent; Judge B with Judge C, 69.6 per cent; Judge A with Judge C, 65.6 per cent. For all the data (counseling and supervision combined), percentages of agreement on F responses between judges was as follows: Judge A with Judge B, 79.7 per cent; Judge B with Judge C, 74.3 per cent; Judge A with vv fi fvvvrvw V v vv f v—f 7 h ‘w vwv fifi 5Traditionally, computation of inter judge reliability utilizes Ken- dall's concordance (W). The coefficient of concordance is an index of the divergence of actual agreement shown in the data from the maximum possi~ ble (i.e., , perfect) agreement. A more meaningful and less tedious approach to analysis of interjudge agreement for this study was to assess the relative divergence of each pair of judges' actual F-percentage agreement from 10.50 agreement. 51 Judge C, 63.6 per cent. Percentages of interjudge agreement are presented in Table 3.4. Table 3.4. Percentages of Agreement on F (Feeling) Responses by Pairs of Judges for Counseling, Supervision, and Combined Data Pairs of Judges AB BC AC Counseling 81 .3% 77 . 1% 62 .4% Supervision 76 . 3% 69 . 6% 65 . 6% Total Data 79.7% 74.3% 63.6% A chi square test of significance was applied to percentages of inter- judge agreement for the counseling and supervision data combined. The ob- tained value of chi square permitted rejection of the null hypothesis that ratings by the judges are unrelated and could be expected to occur by chance. Applying a chi square test of significance to percentages of inter- judge agreement for comseling and supervision data separately, the obtained value of chi square ( x2: 63.74, P > .001) permitted rejection of the null hypothesis that judges' ratings of counseling and supervision data are unrelated and that the percentages of interjudge agreement could be expected to occur by chance. 52 Rejection of the null hypothesis that the ratings are unrelated may be interpreted as indicating that all the judges applied essentially the same standards in rating all the responses. Ranks for the counseling and supervisory situations were assigned on the basis of agreement on F ratings by two or more judges. The proportion of F responses, as determined by majority agreement of the judges, decided the position for each subject in the counseling and supervisory rankings. Collection of the Data The data for this study were collected during the summer of 1966. The data for the hypothesis-testing part of this research were collected by tape recording all counseling and supervisory sessions which were held dur- ing the NDEA Institute. Sample segments from all the tapes were used to derive the resmrch data. The data for the hypothesis ~generating part of this research were ob- tained in the following manner. Each of the supervisors was asked to de-- scribe himself on the Leary Eerpgrsggal £11351: I:_i§_t before the Institute be- gan. ThelCl. was administered to comiselor-enrollees on the first day of the Institute and to clients when they came for their (second interviews. The second interview was used as the time to obtain client self descriptions at the request of the Institute director who was reluctant to have any kind of testing situation presented at the time of the first interview. 53 After the last supervisory session, each member of a counselor- supervisor dyad was asked to describe the other member of the dyad on the 191:. After the last counseling session each counselor was asked to describe his client on the _I_(_3_._L and each client was asked to describe his counselor on thelglf. Each time the_l_ql.. was administered, the persons involved were asked to describe themselves and/ or the persons they were being asked to describe as honestly as possible. They were assured that the information would be kept confidential and that the instrument was being administered for research purposes only. Early termination by clients and/ or missed supervisory sessions caused a slight variation in the number of responses available from several of the counselors. Table 3.5 shows the number of responses judged per sub- ject in counseling and supervisory situations. Treatment of the Data The proportion of feeling~verbalization in a given counselor's res- ponses was assessed by judges’ ratings. The ratings by judges also deter— mined the rank order of counselors for the counseling and supervisory situ- I ations. The assumption of rankability, which is abasic assumption underly— ing the use of rank correlation methods, is met by the proportions of feeling- Verhalization in two situations available for ranking. Therefore, the degree of consistency of feeling-'verbalization between the counseling and superviso‘ 54 Table 3 .5. Number and Totals of Responses for Subjects in Counseling and Supervisory Situations (EDLHTSEUUDNC1SHTTLATTCXQ SEHHKRNHEIHKY'STFEHXTTCHQ Counselor No. of Responses Counselor No. of Responses "'01 """" fl *“86fiwv"" "61 9 " ' 40' f7 V ' 02 80 02 4O 03 80 03 25 04 80 04 4O 05 8O 05 40 06 80 O6 40 07 80 07 40 08 80 08 40 09 80 09 40 10 80 10 4O 11 8O 11 40 12 80 12 4O 13 80 13 40 14 80 14 40 15 80 15 4O 16 80 16 4O 17 80 17 40 18 80 18 4O 19 80 19 40 20 80 20 40 21 15 21 20 22 80 22 40 23 80 23 40 24 80 24 40 25 8O 25 40 26 8O 26 40 27 8O 27 40 28 75 28 40 29 80 29 40 30 ___~8_Qfi~ 30 __40 Total 2330 Total 1163'“ Grand total (both situations) 3495 Average number of responses, counseling 77 .7 Average mimber of responses, supervisory 38.8 Over~al1 average for both situations 58.2 55 ry situations can be determined by means of a statistical test which results in a rank correlation coefficient. The statistical test which was chosen for analysis of the data in this research is Kendall ‘s tau. The formula for tau given by Kendall (1955) for cases where ties occur in the ranks is the one selected for use in this study and is reproduced here: S tau = ' ' rrfi' ' V172 N '(n-l) itx ' 41/2 N (Nave; Two studies (Anderson 1956, Ellsworth 1962) which have used Ken- dall's tau are reviewed in chapter two. Both deal with assumptions and con- ditions similar to-those of the present research. It appears that the nature of the data and the conditions of this study meet the assumptions for Kendall‘s tau and that this statistical test can be used in a meaningful way in the analysis of data for this study. The Lateraseesnel Cheek 2:18.: The Learylnfrpersonal Check List is a collection of 128 items de- scriptive of interpersonal behavior (see Appendix E for text oflC_L - Form IV). The items are representative of sixteen interpersonal variables each of which has eight items related to it. The sixteen interpersonal variables measured by the 191: can be treated as eight combined octant variables as was done in this study. (see Figure 3.1 for illustration of variables). Oc- tant scores can then be algebraicly weighted, converted into standard scores 56 Illustration of Interpersonal Behaviors Classifications Figure 3.1. .uzr— 5:. 4:39.! >835)» 31: wsoyufiu no 28.. 94’ «33:35:31: 4. E; ur.0399) is signifi- cant beyond the .05 level of confidence. The null hypothesis of no consist- ency of feeling-verbalization between counselor responses in the counseling and supervisory situations can be rejected. Some support for the research hypothesis is thus indicated. The following hypothesis is accepted: Consistency of feeling-verbalization exists between the re- sponses made by the counselor in the counselor-supervisor and counselor-client dyadic relationships. Using some of the data obtained by administration of the Integerson- §_.l_ Check List analysis was made of the differences between clients' self - descriptions and descriptions of clients by their counselors. After plotting the scores of both descriptive groups on Leary diagnostic grids (See Appen- dix F for diagnostic grids), the octant position for each subject was trans- ferred to a summary sheet having a summary box for each octant. The method routinely employed to establish statistical significance for such data 64 is to divide the diagnostic grid in half by combining four adjacent octants and to use a chi square test of significance. The division used in this research was obtained by grouping octants 1, 2, 7, and 8 and octants 3, 4, 5, and 6. This division is a standard split which combines the idealized "virtues" of strength, responsibility, and conventionality vs. the more "neurotic" fac- tors of dependence, submission, disrrust and anger. Table 4.3 presents the octant data for clients' self ~descriptions and counselors‘ descriptions of clients on the _I__C_L. Table 4.3. Octant Summary Data for Clients' Self Descriptions and Coun- selors‘ Descriptions of Clients "W 7 ' , i Dggfiosgc'bcgani 1 'V - Octant :S:umma' fies Groups ‘fii L2 '-3, {4 5 l 6 7 8 1278 . fi34fi56fi Wfi i 1 ‘ 1 ‘ Clients' Self j. I A ‘ ( ‘ l l Descriptions 1 9 i 5 5 1 l 4 ‘ 0 1 5 l 8‘ 27 1 10 w f'v fivli vv—V v 1* 1 { V fil fii { Descriptions 1 ‘ l , of Clients by j 4 4 7 5 10 l 13 7 21 17 35 Counselors 1 , { in 1 H i - - in A tin The significance of differences between clients' self descriptions and counselors' descriptions of clients was determined by chi square analysis. A X2 value of 14. 99 was obtained which is significant beyond the .001 level of confidence. The self descriptions of the clients were found to be signifi- cantly different from their counselors‘ descriptions of them. 65 In a "healthy," functioning group or dyadic relationship the expect- ancy3 is that the individuals involved will view themselves and each other "accurately, " i.e. , that significant differences will not occur between an individual's self description and the description of him by the other mem- bers of the group. In Table 4.4 the chi square table for testing the significance of the difference between clients' self descriptions and counselors descriptions of clients is presented in tabular form. Table 4.4. Chi Square Table Testing Significance of Difference Between Clients' Self Descriptions and Counselors' Descriptions of Clients 1 v v v 7 'fi 1 T 1 Clients' Self Descriptions . 27 10 I Counselors' Descriptions 3 17 1 35 of Clients 1 fi V ‘ V fl 12 = 14.99 P = .> .001 Non-Hypothesized Observations Although no other statistical analyses were performed on the Leary Interpersonal Check List, an investigation of these data revealed several 3Based on Kaiser Foundation research data (Leary, 1956). 66 interesting factors which could be implemented by further investigation. From examination of diagnostic grids some ordering of supervisors became apparent. Identifying the four supervisors by the numbers 1, 2, 3, and 4, the following orders were observed. Amount of overlap (from most to least) of areas bounded by lines connecting the points marking the enroll- ees' self descriptions and the supervisors' descriptions of enrollees occur- red in the pattern 1, 2, 3, 4. The amount of overlap may be understood as concurrance between supervisors and enrollees about the enrollees' pattern of interpersonal behavior. In the same order for supervisors, an. increase was found in distance between the supervisors' self descriptions and descriptions of supervisors by enrollees. These linear distances may be understood as psychological differences between supervisors' self perceptions and their perceptions of enrollees. In reverse order (4, 3, 2, 1), distance increased between the super- visors' self descriptions and descriptions by supervisors of enrollees. These linear distances may be understood as differences between the super- visors' self perceptions and the enrollees' perceptions of the supervisors. In every case but one, the distance (or difference) between the de- scriptions of each other by client and counselor was found to be greater than the distance (or difference) between their self descriptions. Illustrative diagrams, charts of linear relationships and grids of the Leary data are reproduced in Appendix F. CHAPTER V SUMMARY, CONCLUSIONS, AND DISCUSSION Summary Consistency of personal characteristics from one situation to another has been assumed in and outside of therapy relationships. The personality of the therapist, his feelings, his emotional blind spots and his past interper- sonal experiences are assumed to be determining factors in the nature of the relationship which he develcps with clients. Feeling-verbalization has been identified by numerous researchers as a sound empirical construct. It is one aspect of interpersonal behavior which has been assumed to be consistent across situations, yet little research evidence has been presented which would support this assumption. This study was designed to investigate the consistency of counselor feeling -verbalization in counseling and non-counseling situations. More spe- cifically it was an examination of the degree of consistency of feelingvverbali- zation shown by counselors ~in -training in counselor ~c1ient and counselor- supervisor relationships. 67 68 The following terms were defined in chapter one: (a) counselor, (b) client, (c) supervisor, (d) supervisory session, (e) counseling session, ((1) counselor response, (e) consistency, (f) feeling-verbalization, (g) feeling or affect. Delimitations of the study described in chapter one included (a) limita- tions upon generalization of findings from the study, (b) the absence of as- sumptions about emotional states underlying feeling-verbalization, (c) the limited concern of this particular study with counselor effectiveness as an out- come variable, (d)’limitations in the precise statistical sampling of loosely structured verbal data. The following basic assumptions were stated: (a) the legitimacy of generalizing from the counseling and supervisory sessions sampled in this study to other counseling and supervisory sessions in programs of counselor education, (b) the assumption that verbal statements of affect reflect valid feeling, and (c) the assumption that feeling-Verbalization is a sound empirical construct and that, more specifically, such verbalization can be identified and _ classified. Verbalization of feeling or affect is just one pattern of interpersonal behavior which needs to be evaluated by further research. This study, and others like it, are needed to bridge the gap between an abundance of theory which assumes consistency and a paucity of empirical demonstrations of con- sistency in counselor behavior. 69 Therefore, the following hypothesis was developed: Consistency of feeling-Verbalization can be demonstrated to exist between the responses made by the counselor in the counselor-supervisor and counselor-client dyadic relationships . The review of related literature, presented in chapter two, suggested (a) that "outcome" studies, so prevalent in counseling research, have provid- ed us with information about the effectiveness of therapy, but, in looking at therapy as an homogeneous treatment variable, have largely fail ed to probe the complexities of interpersonal behavior which need to be included in an understanding of the nature of therapy, (b) that some degree of counselor consistency exists across counseling and non -counseling situations, but that further empirical validation of such consistency is needed, and (c) that feeling-Verbalization is an important and empirically sound research con- struct for which high reliability and validity have been found. This study was conducted during a summer National Defense Education Act Counseling and Guidance Institute. The methodology of the research was as follows: Data for the hypothesis-tes ting portion of the study were collected by tape recording all counseling and supervisory sessions which were held during the Institute. Sample segments from all tapes were used to derive the research data. Typescripts containing counselor responses were made from the tape samples. Three experienced counselors serving as independent judges rated the counselor responses as follows: (F) if the response included expression of feeling or affect, (C) if it was cognitive in content. 70 The proportion of F responses, as assessed by judges‘ ratings, deter- mined rank orders of counselors for counseling and supervisory data. For the counseling data, percentages of agreement on F responses for pairs of judges were as follows: Judge A with B, 81.3%; B with C, 77.1%; A with C, 62.4%. For the signervision data, percentages of agreement on F responses for pairs of judges were as follows: Judge A with Judge B, 76.3%; B with C, 69. 6%; A with C, 65. 6%. Percentages of agreement were computed on the basis of maximum possible agreement on F ratings for pairs of judges. Conclusions Kendall’s tau (formula for tied data) was used to test for consistency between counselor rankings. A tau value of + 0.23 was obtained, which is significant beyond the .05 level of confidence. Thus, it was possible to re- ject the null hypothesis of no consistency and accept the research hypothesis. Non-hypothesized observations of the Leary Interpersonal Chick I__._i§_t data led to statistical treatment of an observed difference. A chi square test established difference significant beyond the .001 level of confidence between the octant summaries of self descriptions of clients and their counselors' descriptions of them. An examination of thelg: diagnostic grids revealed supervisor differ- ences. Amount of overlap (from most to least) of areas bounded by lines connecting the summary points of the enrollees' self descriptions and the 71 supervisors' descriptions of enrollees occurred in the pattern 1, 2, 3, 4. In the same order for supervisors, an increase was found in distance between the supervisors' self descriptions and descriptions of supervisors by enrollees . In reverse order (4, 3, 2, 1) distance increased between the supervi- sors' self descriptions and descriptions by superVisors of enroll ees. In every case but one, the distance between the descriptions of each other by client and counselor was found to be greater than the distance be- tween their self -descriptions. Discussion The major finding of the research was that there is a significant posi- tive correlation between the amount of feeling-Verbalization expressed by counselors in counseling and supervisory situations. In other words, the proportions of feeling verbalization in a counselor's counselor-client re- lationships and in his counselor-supervisor relationships are significantly related to each other. Although the statistically significant result indicates some consistency, much of the variance is not accounted for by the postulate of counselor con- sistency. Some possible sources of variance which could have altered the out- come of this Study include: (a) the validity and reliability of judges’ ratings; 72 (b) procedural and sampling methods; (c) additional situational factors such as role-playing, role perception, conformity to authority, and anxiety arising from the practicum situation; and (d) the similarity of the interpersonal rela- tionship in and outside of the therapy setting. Doubt might be raise-:1 that valid feeling can be identified and rated on the basis of feeling-'verbalization, especially from typescripts. This issue r esul ted in the basic assumption made in chapter one concerning verball- zation of feeling. There is also research evidence cited in chapter two(especially Anderson, 1956). It might also be contended that counselors-in-training may be influenced to make feeling-type responses in order to fit the "ideal" coun- selor role as it is seen by them or others. Two forces would appear to miti- gate such pretension of feeling~response. The first would be the professional competence of the judges making the ratings. The second would be the ability of the supervisors to detect such role-playing and explore its dynamics. In any study of content analysis the issue of interjudge reliability is important. Berelson (1952) has stated that reliability among raters can be expected to increase as the task becomes more simple, clear-cut, and spe- cific. The percentages of interjudge agreement in this study ranged from 62.4 per cent to 81.3 per cent and are comparable to percentages of agree- ment found in similar studies reported in the literature (Danskin, 1955; Ellsworth, 1962; and Anderson, 1956). Because there were a large number of data, small samples were taken 73 from each tape and only the counselor responses (not client or supervisor responses) were included. The inclusion in the typescripts of counselor re- sponses only may or may not have influenced the ratings by the judges. Al- though there is no way of knowing, except for future replication of the study, what effect the choice of a particular time segment from the tapes may have had on the outcome of this research, it is clear that sampling techniques utilized in this study are in concordance with accepted techniques reported in the literature ('Keisler, 93.9.1: , 1965) . In order to compensate for a relative- ly small N, Kendall tau statistical technique was employed since its power efficiency rating maximizes sensitivity in assessing probable association be- meen two variables. In a discussion of additional situational factors that could have affected the research, an examination of several theoretical views may be of value. Interpersonal theory, which has perhaps contributed most to the framework of this study, suggests that anxiety is the primary motivating force in human personality. It is plausible to assume that anxieties centering in the practi- cum experience could give rise to behaviors which might be consistent over counseling and supervision or which might be differentiated in terms of the anxiety experienced in the two situations. Counselors ~in ~training, particularly those with a background in class- room teaching, tend to conform to authority. Literature on the personality of the classroom teacher generally agrees that the teacher functions in a rela- 74 tively structured environment in which conformity is rewarded. If the counselor role is perceived in a stereotyped manner or as ex- pectations of an authority figure, such role perception could alter expressive behavior. Goffman (1961) has theorized that an individual, while manifestly participating in one system of role behavior, has some capacity to hold in abeyance his involvement in other patterns, thus sustaining one or more dor- mant roles that are enacted roles on other occasions. Mead (Strauss, 1956) notes the importance of role-taking behavior as a means of facilitating iden- tification, observation and integration of significant "other" roles . It can be assumed that the counselor-in-training will see his role in the counselor-client dyad as different from his role in the counselor-supervi- sion dyad. The correlation between ranks in counseling and supervision could be depressed by the differing role perceptions just described or by certain other possible factors such as session to session changes, develop- ment of a status hierarchy, or task changes. The similarity of the interpersonal relationship in and outside of the therapy setting might also be considered as a possible source of variance on the basis of theory and research evidence. In Mead's words (Strauss, 1956, p. 219). What determines the amount of self that gets into communication is the social experience itself. Of course, a good deal of the self does not need to get expression. We carry on a whole series of different relationships to different people. We are one thing to one man and another thing to another. . . . . It is the social process itself that is responsible for the appearance of the self; it is not there as a self apart from this type of experience. 75 Borgatta and Bales (1953) present evidence that a person will be more consistent, given the same amount of time, when interacting with the same individuals than when interacting with different individuals. They also state that a certain degree of consistency in the interaction pattern may be expect- ed if any common elements (i.e. , the task, subjects, size of groups) exist in the conditions under which the behavior occurs. They conclude (p. 568) that The stability of subjects shown is sufficient to encourage us to believe that the interaction of an individual . . . may tell us something about his personality, in spite of peculiarities due to the fact that he is interacting with particular other persons. . . . supporting the hypothesis that stability or consistency of social behavior may be attributable in some measure to individual personality factors . A plausible explanation for the statistically significant difference found in the Leary LCL data between the self descriptions of clients and their counselors’ descriptions of them may be found in role theory. If Goffman's (1961, p. 85) definition of "role" as consisting "of the activity the incumbent would engage in were he to act solely in terms of the normative demands upon someone in his position" is used, the possibility exists that the counse~ lor perceiving his role in terms of such normative demands creates a "role" for his client which includes less healthy behavior than the client presents . Another related explanation of the same data may be seen in an analo- gy to the medical student who experiences the symptoms about which he is studying. The Institute enrollees were exposed to course work and discus- sions related to psychopathology concurrently with the practicum experience. 76 It seems plausible, therefore, that the counselor-enrollees were acutely aware of deviant behavior patterns and were influenced by their newly ac- quired "diagnostic skills" in their descriptions of clients. Examination of the IQ: data where ordering of supervisors appeared led to a tentative hypothesis that the orderings might be related to an open- close-:1 dimension in the supervisors‘ personalities. In order to obtain consensual validation of an informal ranking of the supervisors according to "openness, " the researcher asked five colleagues in counseling to whom all the supervisors were known personally and profes- sionally to rank the four supervisors from most to least "open. " Four con- curred exactly with the researcher’s ranking. The fifth concurred on the ranking of supervisors 1 and 2, and inverted the rank order of supervisors 3 and 4. It appears that some distortion factors were reflected in the clients' and counselors‘ descriptions of each other, resulting in the greater distance between their descriptions of each other than between their self-descrip- tions. Projection can be eliminated as a possible distortion factor by the findings of the Kaiser Foundation research.1 It is plausible that role ex~ pectations may provide a clue to at least some of the distortion. If Level Hv w w w iviv‘ fifiv—v v vfifiV—v—‘fifi wvv 'va if fiv v—r Vfi— 1On the basis of the normative data collected by the Kaiser founda- tion, two -way projection was hypothesized to result in Level II~Other scores which are closer on the diagnostic grid than the Level II-Self scores for the same individuals. 77 II—Other scores were influenced by role expectations they could be expected to vary from Level II-Self scores in predictable directions. If, conversely, Level II-Self scores are influenced by role-playing or role perceptions, the distortion might be seen to occur as a movement of the Level II~Se1f sum- mary points away from what could be relatively "accurate". Level II-Oth er summary points . Implications for Further Research Areas of further research suggested by the findings of this study are the following. 1 . A parallel study using innovative interpersonal process recall techniques similar to those developed and reported by Kagan, Krathwohl, and Miller (1963). 2. Examination of interpersonal behaviors other than feeling-verbal- ization, such as hostility, warmth, understanding. 3. A parallel study of counselor consistency in other professional environments . 4. Correlated studies of counselor consistency and effectiveness. 5. Extension of the LearylgL findings of this study to include de- sign and implementation of further research with the 191:. 6. Investigation of the validity and effectiveness of pairing therapist 78 and client on the basis of various personality characteristics and/ or role preferences . 7. Perhaps the major implication of this study, not only for future research, but for clinical application and evaluation is the potential meaning of the data for counselor education. In most counselor education programs, judgement about a counselor's potential is withheld until the counselor can- didate is observed in face-to-face relationships with clients. This experi- ence, unfortunately, is not usually enacted until well into the counselor's graduate program. If further research bears out or expands upon our under- standing of how a counselor's behavior with any "other" (in this study, his supervisor) relates to or is predictive of his behavior with clients, we may one day be able to predict an individual’s performance in counseling based on observationof him in other relationships.These other relationships could be observed early in the program, or even "set up" as a screening proce- dure for entrance to a program of counselor education. The method used in this study for quantifying affect could be one method used for such future evaluation . BIBLIOGRAPHY Abeles, N. Therapeutic sensitivity and self consistency: an exploratory study. Paper presented at the 65th meeting of the Michigan Academy of Science, Psychol. Section, March 24, 1961. . Awareness and responsiveness to affect as a function of training and supervision. (Paper presented at the 66th meeting of Michigan Academy of Science, Psychol. Section, March 24, 1962, Ann Arbor, Michigan. Anderson, R. P. Physiological and verbal behavior during client centered counseling. Lvof Counsel. Psychol., 1956, 3, 174-184. Arbuckle, D. S. Client perception of counselor personality. 1. Couns. @311” 1956, 3, 93—96. Aronson, M. A study of the relationships between certain counselor and client characteristics in client -centered therapy. In W.Snyder (Ed.) Group report of a program in research in psychotherapy. University Park, Pennsylvania: Penn.'State Univ., 1953. T T Axelrod, J. An evaluation of the effects on prognosis in therapy of similari- ties and differences between the personalities of patients and their thera- pists. Publication #3604, New York Univer., 1952. Bales, R.F. Interaction process analysis: Avmgthodvforvthe study of small groups. Cambridge, Mass.: Addison-Wesley, 1950. Bandura, A., Lipsher, D. R., and Miller, P. E. Psychotherapists approach avoidance reactions to patients expression of hostility. . Consult. Psychol., 1960, 24, 1-8. Barrett-Lennard, G. T. Dimensions of therapist response as causal factors in therapeutic change. PsychLMonographs, 1962, 76, 1-36. 79 8O Beier, E. S. On supervision in psychotherapy. fsychomgr. Theor. Res. Pract., 1964, 1, 91-95. Berelson, B. , Content-analysis in commuMcadoniresearch. Glencoe, Illinois: The Free Press, 1952. Bohn, Martin J. , Jr. Counselor behavior as a function of counselor domi- nance, counselor experience, and client type. J. Counsel. Psychol., 1965, 12, 346-352. Bordin, E. S. {sychvovlogical counseling. New York: Appleton-Century Crafts, 1955, 152-469. Borgtta, E. F., and Bales, R. F. The consistency of subject behavior and the reliability of scoring in interaction process analysis. American fiSociologvical Review, 1953, 18, 565-569. Brams, J. M. Counselor characteristics and effective communication in cormsel'ing. I. Couns. Psych., 1961, 8, 25-30. Briar, S. , and Bieri, J. A factor analysis and trait inference study of the Leary Interpersonal Check List. JLCfilinLfiPsych. , 1963, 19, 193 -8. Butler, J. M., Rice, Laura N., and Wagstaff, Alice. gugititative naturaliz- tic r;esearch. Englewood Cliffs, N.J.: Prentice Hall, 1963. Cahoon, R. A. Some counselor attitudes and characteristics related to com-- seling relationship. Dissert.vastractvs, 1963, 23, 3473. Campbell, R. E. Counselor personality and background and his interim sub- role behavior. LCouns. VPsyché, 1962, 9, 329-334. Canon, H.J. Personality variables and counselor-client affect. J. Coun__s_. Psych., 1964, 11, 35-41. Carson, R. C., and Heine, R. W. Similarity and success in therapeutic dyads. L Cgsult. Psych... 1962, 26, 38-44. Cottle, W, C., and Lewis, W. W., Jr. Personality characteristics of coun~ selors: 1 male counselor responses to the MMPI and GZTS. J. Couns. Psych., 1954, 1, 27-30. Cronbach, L. J. Processes affecting scores on understanding of others and assumed similarity. Psychol. Bull” 1955, 52, 177-193. -_ .- . j.: l “:w. 7* «tin-arrtr'vtv 7'.” ‘ n.- ”.4 81 Danskin, D. G. Roles played by counselors in their interviews. J. Couns. Psych., 1955, 1, 22~27. Demos, G. D. , and Zuwaylif, F. H . "Characteristics of effective counse- lors." CounseJor Education and Supervision, 1966, 5, 163-165. Dipboye, W. J. Analysis of counselor style by discussion units. J. Couns. Psych., 1954, 1, 21-26. Dittman, A. T. The interpersonal process in psychotherapy: development of a research method. J.fof Abn. and Soc. Psychol., 1952, 47, 236-244. Dollard, J. and Miller, N .E. Personality andpsyghotherapy. New York: McGraw Hill, 1950. , Ellsworth, S. G. The consistency of counselor feeling-verbalization in and : outside of the counselor-client relationship. Unpublished doctoral ii. . dissertation, Michigan State University, 1962. Fiedler, F. E. The concept of an ideal therapeutic relationship. J. Consult. Psychol., 1950, 14, 239-245. - J. A comparison of therapeutic relationships in psychoanalytic, non-directive, and Adlerian therapy. J. Consult. Psychol., 1950, 14, 436-445. ' ' ' ' Frank, G. H. On the history of the objective investigation of the process of psychotherapy. !. Psychol. , 1961, 51, 89-95. Freud, S., An outlinevof psychoanalysis. New York: Norton, 1949. Gerler, W. Outcome of psychotherapy as a function of client- counselor similar arity. Unpublished doctoral dissertation, Univ. of Illinois, 1958. Goffman, E. Encounters. two studies in the sociology of interaction. Indianapolis: Bobbs-Merrill,1961. , The presentation of self in everyday life. New York: Anchor, 1959. ‘" Gonyea, G. G. The “ideal" therapeutic relationship and counseling outcome. J. ClinJJPsychol” 1963, 19, 481-487. 82 Gynther, M.D. Degree of agreement among three "interpersonal system" measures. Abstract, J. Consult; Psych., 1962, 26, 107. W - .. EL. 31: Relations between needs and behavior as measured by EPPS and interpersonal check list. J. Socfi. Psych., 1962, 57, 445-51. Grigg, A. E., and Goodstein, L. D. The use of clients as judges of counse- lor's performance. J.vCounsel. Psychol., 1957, 4, 31-36. . Hall, C. S., and Lindzey, C., Theories of personality. New York: John Wiley and Sons, Inc., 1958. Hansen, J. C., and Barker, E. N., Ehtperiencing and the supervisory 3 relationship. JJCounsel. Psychol., 1964, 11, 107-111. 2» HeiIbrtm, A. , Jr. Personality correlates of early termination in counseling. J. Couns. Psych., 1961, 8, 31-36. L- Heller, K., Myers, R. A., and Kline, Linda V. Interviewer behavior as a function of standardized client roles. !. Consult. Psych .., 1963, 27, 117-122. Henry, W. E., and Shlien, J. M. Affective Complexity and Psychotherapy, LchProjectheJch., 1958, 22, 153-162. Hoffman, A. E., The analysis of counselor sub-roles. J. CouIJSQJPszTchol” 1959, 6, 61-67. Holt, R. , and Luborsky, L. Personality patterns of psychiatrists: a study in selection techniques. New York: Basic Books, 1958. Izard, C. E. Personality similarity, positive affect, and interpersonal attraction. J. Abnormuand Soc. Psych., 1960, 61, 484-485. Kagan, N., Krathwohl, D. R., and Miller, R., "Stimulated recall in therapy using video tape -- a case study." LFOEHISQ: Psych., 1963, 10, 237-243. Kahn, R.K. , Therapist discomfort in two psychotherapies. Unpublished doctoral dissertation. Pennsylvania State University, 1957. Kell, B. , and Mueller, W. J. ImpacJ and'change: a study of counseling re- laJtionshJips. New York: Appleton-Century-Crofts, 1966. 83 Kelz, J. W. , "The development and evaluation of a measure of counselor effectiveness." Pgsonneljipd Guidance Journal, 1966, 44, 511-516. Kendall, M. G. Rank correlation methods. New York: Hafner Publishing Co., 1955. Kiesler, D. J., Klein, Marjorie H., and Mathieu, Philippa L. Sampling from the recorded therapy interview: the problem of segment location. LConqult. Psych. , 1965, 29, 337 -344. Leary, T . , The theory and measurement methodology of interpersonal communication, Psychiatry, 1955, 18, 147-161. . Multilevel measurement of interpersonal behavior: a manual for the use of the interpersonal system of personality. Berkeley. California: Psychological Consultation Service, 1956. . Interpersonal diagnosis of personaligy. New York: The Ronald Press Company, 1957. ., and Coffey, H. S., "Interpersonal diagnosis: some problems of methodology and validation,“ Journal of AbnopmalfiandSocial Psych., 1955, 50, 110-124. . , and Gill, M. The dimensions and measure of the process of psychotherapy: A system for the analysis of the content of clinical evaluations and patient -therapist verbalizations. In E. A. Rubinstein and M. B. Parloff (Eds .), Research in PsLhotherapy. Washington, D.C. American Psychological Association, 1959. Lennard, H. S. , and Bernstein, A. The anatomy of psychotherapy. New York: Columbia University Press, 1960. Levine, G. The effects of two verbal techniques on the expression of feel- ings. 1958,Dissert.Abstr. 1958,29, 363. Luborsky, L. B. The personalities of successful and less successful psychotherapists. Amer. Psychglggifith, 1952, 7, 337. Mead, G. H., Selected Writipgs. Indianapolis: Bobbs-Merrill, 1964. Mendelsohn, G. A., and Geller, M. H. Effects of counselor-client similari- ty on the outcome of counseling. LCoJ1ns.vPsych., 1963, 10, 71-77. 84 __ . Effects of client personality and client-counselor simi- larity on the duration of counseling: a replication and extension. J. ComsLQym. , 1966, 13, 228-234. Mowrer, O. H., Light, B. H., Luria, Z., and Zeleny, M. P., Tension changes during psychotherapy with special reference to resistance, In 0. H. Mowrer, (ed.) PSflhotherapy, Theory and Research. New York: Ronald Press, 1953. GP - V v - - V Mueller, W. J., et _a_l_. The prediction of counselor interview behavior. Pers. andfiGuid.J., 1963, 41, 513-517. Ourth, Lynn, and Lanfield, A. W.,- "Interpersonal meaningfulness and nature of termination in psychotherapy." J. Counsel. Psych. , 1965, 12, 366-371. Pallone, N. J., and Grande, P.P; Counselor verbal mode, problem relevant communication, and client rapport, J. of Cguns. vPvach. , 1965, 12, 357-365. Parloff, M. B. Some factors affecting the quality of therapeutic relation- ships. JLAbn. andeociPsychJol” 1956, 52, 5-10. Parsons, T., and Bales, R. F ., Bamilyi socialigafipn and interactipp rocess. Glencoe, Illinois: The Free Press, 1955. Patterson, C. H. Cgmseling andisychomerapm meoryjnd practice. New York: Harpers, 1959. Peterson, Snyder, Guthrie, and Ray. Explanation of therapeutic bias. V‘v fi' Rice, Laura N. Therapists style of participation and case outcome. J. Cpnsult. Psych., 1965, 29, 155-160. ROgers, C. R. A theory of therapy, personality, and interpersonal rela- tionships as developed in the client-centered framework. In S. Koch (ed.), Psychology: A Study of a Science, Vol. III. Formulations of the Person and the Social Context. New York: McGraw-Hill, 1959, 184-256. Rubenstein, E. A., and Lorr, M. Self and peer personality ratings of psy- chotherapists. J. Clin‘. Psych., 1957, 13, 295-298. 85 Siegel, S., Non-parametric statistics. New York: McGraw Hill, 1956. Snyder, W. U. The psychothergpeutic relationship. New York: Macmillan, 1961. Stefflre, E., King, P., and Leafgren, F. Characteristics of counselors judged effective by their peers. J. Couns. Psych., 1962, 9, 335-340. Strauss, A. , The social psychology of George Herbert Mead. Chicago: .-.. , University at Chicago Press, 1956. Strupp, H. H. Psychotherapeutic technique, professional affiliation and experience levels. J. Consult. Psychol., 1955, 19, 97-102. The effect of psychotherapist's personal analysis upon his techniques. J.JConsult. Psychol. , 1957, 21, 12-20. 1 The psychotherapist's contribution to the treatment process. Arch. Gen. Psychol., 1960, 3, 219-231. Sullivan, H. S. The interpersonal theory of psychiatiy. New York: Norton, 1953. Taulbee, E. Relationships between certain personality variables and con- tinuation in psychotherapy. JJVConsult. Psych., 1958, 22, 83-89. Tuma, A.H. , An exploration of certain methodological and client-counselor personality characteristics as determinants of learning in the counseling of college students. Dissertation Abstracts, 1955, 15, 2067-2068. - J. , and Gustad, J. The effects of client and counselor person- ality characteristics on client learning in counseling. J. Couns. Psych., 1957, 4, 136-1410 Truax, C. B. , "Effective ingredients in psychotherapy: an approach to unraveling the patient-therapist interaction. " J. Counsel. Psych., 1963, 10, 256-263. ' J - J - J., and Carkhuff, R. R. "Client and therapist transparency in the psychotherapeutic encounter." J.vCounsv. 3,5393“ 1965, 12-3-9. van der Veen, F. Effects of the therapist and the patient on each other’s therapeutic behavior. J. Consult. Psych., 1965, 29, 19-26. 86 Walz, G.G. Practicum supervision: I. theories, self theory. Paper read at APGA Convention, Boston, April, 1963. Waskow, Irene 'E. Selective reinforcement of feelings or content. J. Con- suit. Rsyghol., 1962, 26, 11-19. Wrenn, R. L. Counselor orientation, theoretical or situational? J.__C_o_un: sel. Psychol., 1960, 7, 40-45. APPENDIX A SELECTION OF CLIENTS The high school age clients were selected as follows: Letters were sent by the NDEA Institute director to twenty counselors in area high schools explaining the nature of the Institute and asking for their help in recruiting clients who might benefit from such a counseling experience. A follow-up telephone contact was made to the fifteen counselors who responded to the letter. Arrangements were made with these counselors for the director of the Institute to meet with a group of ten students in each school. The students: selected for inclusion in the groups (a) appeared to be motivated to enter a counseling relationship, (b) would be available for six to eight sessions, and (c) had personal concerns of some depth that might warrant a counseling experience. In talking with the students, the director made some judgements about their levels of motivation for counseling. Following the group meet- ings, the director rank ordered the students according to his estimate of the probability that each would continue in a counseling relationship for the eight- week duration of the NDEA Institute. Students who ranked in the top five from 87 88 each school were selected for participation as clients in the Institute. A sec- ondary list of alternates for clients who might terminate counseling during the eight week period was compiled. Inf-L".- APPENDIX B THE JUDGES: BIOGRAPHICAL MATERIAL The first judge has been admitted to membership in the American Academy of Psychotherapists. He received his A.B. and MLA. degrees in California and his Ph.D. from Michigan State University. He has been Direc- tor of the Counseling Center at Western Michigan University for the past five years. He has held visiting professorships at the University of Dayton, Mich- igan State University, Ohio University, University of Colorado, and the Uni- versity of Wisconsin. He has served and presently is serving as research consultant to school districts and has published numerous research articles on a wide variety of educational problems . The second judge received his B.A. from Western Michigan Universi- ty, his M.A. from the University of Michigan, and his doctorate from Michi- gan State University. He is presently an Associate Professor and Chairman of Guidance and Personnel Services, Western Michigan University. He has been involved as a participant and investigator in depth interviewing and re- search on "underachievers. " 89 90 The third judge received his A.B. degree from Albion College, his M. S. from the University of Illinois, and his Ph .D. from Michigan State University. He is an Associate Professor of Education at Western Michigan University and has served as Director for the National Defense Education Act Summer Institutes for Counseling and Guidance held at Western Michigan m- University in 1965 and 1966. APPENDIX C INSTRUCTIONS TO THE JUDGES The following typescripts contain the responses of counselor trainees in two situations: (1) the counseling interview, and (2) the supervisory inter- action. It may be obvious to which category some responses belong, but the material will be presented in random order and will not be identified as pro- ceeding from one situation or the other. All responses are to be judged in the same manner and according to the criteria which are described below. Verbalization of feeling or affect as opposed to verbalization of intel- lectual content is the behavior which you are being asked to judge. Please rate each response that includes expression of feeling or affect as (F). Res- ponses which are cognitive in content are to be rated as (C) . "Feeling" or “affect" will be further defined to mean the expression of the experience of emotion. Please include in the category of feeling res- ponses (F) only those responses which contain what is evaluated as productive appropriate affect. Productive, appropriate affect for the counselor-enrollee will include expressed, current feelings about self or another individual (i.e. , client, supervisor, or other), restatement of a specific emotional experience of the client or supervisor, and feelings underlying feelings. Productive, appropriate affect for the counselor-enrollee would ESE include vague "How do you feel?" questions, complaints, or verbal behavior which is defined as defensiveness, even though these expressions may be couched in "feeling" language. Cognitive content might be defined as what is left after feeling or af- fect as defined above is removed. Cognitive material has been variously de- fined by different authors as impersonal, intellectual, integration, higher mental process, knowledge, organizing, describing, recognizing, sensing, ordering. 91 92 Please rate a response by content and/or intent. Whether the aim of the response is to clarify, reflect, or interpret, the criteria for rating feel- ing will be the same. In general we might say that usually (though not always) such words as the following include affect: antagonism, annoyed, apprehensive, angry, afraid, bitter, calm, cheerful, close to them, couldn't take it, disturbed, desperate, depressed, detached, embarrassed, excited, emotional, enjoyed, frightened, futile, guilty, happy, hate, it hurt, it was hard for me, irritated, jealous, kind, lonely, loving, longing, nervous, panicky, painful, relieved, relaxed, scared, satisfied, sad, secure, self-pity, shaky, sorry, shy, sulky, timid, tense, terrified, uneasy, unsure, unhappy. Some examples of (F) responses are the following: 1. You really get upset about that. 2 . Where are you now, Jane; I feel you've pushed me clear away from you. 3. It will be hard for me to tell her that. 4. But to me it's expecting her to feel warmly not needing this relationship. 5 . You dislike him a good deal, then. Some examples of (C) responses are the following: 1 . Oh, you haven't? 2 . But most of these kids don’t know what they want to go into, and how can we help them if we don't know either? 3. Well, supposing you don't -- what does that mean? 4. I thought your grades were pretty good. 5. Well, of course, you have to take a lighter course load, because of your work. APPENDIX D A SAMPLE TYPESCRIPT PAGE FOR JUDGES' RATING OF SUBJECTS’ RESPONSES 93 94 132059 1 . Is this -- ah -- do you think this problem has been responsible for low grades? , 2. Now, in taldng a look at this whole thing, you say that the three of you being intermingled with each other and taking each other‘s time, then. 3. What other things do you think we could isolate here? 4. Then, as far as the class is concerned it would be these three guys a... that are bothering you -- I mean these other two fellows in the class including yourself then. 5. Then you feel like this fall when you return to school if these other fellows are in your class, that you want to get away from them. 231060 1 . That‘s what I’m just thinking about here -- trying to figure out what I "“ would tell him. At the end of this what I was thinking about would be that -- well, tell him again when we get to him he tells me a story -- it’s about something else or somebody else ---- 2. Not as bad as he has been, 'cause I was trying to cut that off. But he tells me stories that are contradictory in what he says -- like he doesn't care what people think of him, but, then like you said here, he watches -- he knows very well how they are watching him and that gives him (inaudible) so there are some things like that that contra- dict each other. In fact all these things, as you said, are really denials, or shying away from getting into it. 3. Because he said this three or four times to me last week. . Is it really important or are you telling me a story again? . I feel like he‘s shying away again. He says it‘s not really important, but I think it is. 01“:- 121061 1. By this time I knew I had lost her, but I didn‘t know what to do. I'd have to go back over it and think what I should have done in those cases, because I jump too fast. I‘d never thought of that before. Now this afternoon when she comes, should I go back over all of this? And then I had lost her. But from then on I don‘t think I succeeded with her. MACDN APPENDIX E THE INTERPERSONAL CHECK LIST -- FORM IV This is a list of words and phrases which may be descriptive of you (or the person you are rating). On the separate answer sheet mark the item true if the statement is essentially or most usually descriptive of you (or the other person). Mark it false if it is not essentially or most usually descriptive. Answer fill the items as quickly as you can. 16. 17. 18. 19. 20. Able to criticize self Able to doubt others Able to give orders Able to take care of self Accepts advice readily . Acts important Admires and imitates others Affectionate and understanding Agrees with everyone . Always ashamed of self . Always giving advice . Always pleasant and agreeable Apologetic Appreciative Bi g-hearted and unselfish Bitter Boastful Bossy Businesslike Chn be frank and honest 95 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34'. 35. 36. 37. 38. 39. 40. Can be indifferent to others Can be obedient Can be strict if necessary Can complain if necessary Clinging vine Cold and unfeeling Complaining Considerate Cooperative Critical of others Cruel and unkind Dependent Dictatorial Distrusts everybody Dominating Eager to get along with others Easily embarrassed Easily fooled Easily led Egotistical and conceited 41. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. .68. 69. 70. 71. 72. 73. 74. 75. Encouraging others Enjoys taking care of others Expects everyone to admire him Firm but just Fond of everyone Forceful Forgives anything Frequently angry Frmuently disappointed Friendly Friendly all the time Generous to a fault Gives freely of self Good leader Grateful Hardboiled when necessary Hard -hearted Hardly ever talks back Hard to impress Helpful 96 Impatient with others' mistakes Independent Irritable Jealous Kind and reassuring Lacks self -confidence Lets others make decisions Likes everybody Likes responsibility Likes to be taken care of Likes to compete Loves everyone Makes a good impression Manages others Meek .76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96 . 97 . 98. 99. 100. 101 . 102 . 103 . 104 . 105 . 106. 107 . 108. 109 . 110; Modest Obeys too willingly Often admired Often gloomy Often helped by others Often unfriendly Outspoken Overprotective of others Oversympathetic Passive and unaggressive Proud and self-satisfied Rebels against everything Resentful Resents being bos sed Respected by others Sarcastic Self-confident Selfish Self -punishing Self-reliant and assertive Self -respecting Self -seeking Shrewd and calculating Shy Skeptical Slow to forgive a wrong Sociable and neighborly Somewhat snobbish Spineless Spoils people with kindness Stern but fair Straightforward and direct Stubborn Tender and soft-hearted Thinks only of himself 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121 . 122 . 123 . 124 . 125 . 126 . 127 . 128 . 97 Timid ‘Too easily influenced by friends Too lenient with others Too willing to give to others Touchy and easily hurt Tries to be too successful Tries to comfort everyone Trusting and eager to please Usually gives in Very anxious to be approved of Very respectful to authority Wants everyone's love Wants everyone to like him Wants to be led Warm Well thought of Will believe anyone Will confide in anyone 1:52.- F_—._fi .....'___. _ . .. __ _ .‘n n v -4« ,. . 1 . APPENDIX F DIAGNOSTIC GRIDS, ILLUSTRATIVE DIAGRAMS, AND CHARTS OF LINEAR RELATIONSHIPS BASED ON LEARY g1. DATA key to symbols used on diagnostic grids ® Supervisors' self descriptions 1*- ( ult- 1,2 . . . .n) Description of supervisors by enrollees (number identifies enrollee) D ( Ell-'1, 2 . . . an) Description of enrollees by supervisors ‘1) Enrollees' self descriptions X Clients' self descriptions O Descriptions of clients by enrollees 98 g. 111-314,. a .F Grid Showing £91. Level II-Self and -Other Summary Points for Supervisor A and his Group of Enrollees \ \ 5 \ \ i Q. ‘.'t 0: \ l tut—H. t... 0 I i '/ l 8 VI 2“. . -15 I ¢ £3-15- Bi— )1 a\ ‘11-“qu ‘0 \ v v) til / e l to 20 30 #0 M. _ - 0 7o 80 q -+—*+-l—&-o-H++4-H++-+++T++++-H ”H *‘ ”*1“ ‘.“I,i‘-r-rt--rf++-+« 17 4 1H ii 1 L111 l l t} M ii! i l I t J H mo ; ' " .f 9‘11 4 ‘ U / 13-13 ;> ’l-lrl—J-l l-H 4: o.-- f w H lliltll-Ifillliltl O ., Grid Showing ICL Level II-Self and -Other Summary Points for Supervisor B and his Group~6f Enrollees / 7"— O D- on ’. 3° ® ao-1D ._* $425 “- .ii-‘lr ,s {9 ID 'x \V q, 6;.! “99*!” {3 \ \ U-I? till-Ht: \ \~. \ :10 3o ‘ ++++ F‘”“ 11 -_ use} \ '. /"’ U-Np l o- a" E. J Ht 1ft 4 H“ 1’ \\. '1 .. t/ / ~A \ f w 0 _' -i-tH*'itJt-tlJi-t 0 / /\ ,. ”+1 I/ / +4. -+-.. +1— 0 / I l s {to \_ “3‘ «07:7- 05/." u—RS’; a¥ JI+ ~ so 7.0 so Mile—1H4l-H-HI Ht it Hi Jinan: II!“ If tml tilt—Jil—H‘ 1 Mil mthHlO- 4: AJYI ICP’ " lainLlrrLtJ'L‘LnL rP-H ++t++r 0' lv'fil / O A L.J rh’TfI‘vyrr I w 0 0'9 as 4w? I 25' ¢o is” 3' O - 47 vi? 0‘41). P ‘91 27 Ht t} Hi :1 l H J H -i—+.J+++p ++ +i+t O .7, ‘ O 102 Grid Showing Summary Points of Clients' Self Descriptions and Descriptions of Clients by Enrollees > + '7 ’1 l’ l o (1: / '1 /+ + l' + ‘5 ‘.‘x v . ‘ L1 tH :1:er 44+»! 0 4: rim-H. _ 0 - 14+ X j. 0. O T—IITTIIII1 N L JLJJJulllJlLiluJIDi ’rI—T ' * 'TFTJ T1 I—T—I s n O 1L1 if it _._ i. i‘ In 0 —++ .o 7| JR 1. tl‘t J- o3 Jet-H alt-t— O f—a CD on x "i" ‘I~ 3c (D 20 :30 0 +170 \"I“\‘E ,.-‘Z.o .WWfi—st Hf Wllffifihrfli"- H" 41.1.. .- .l '. ’\ '\ ‘ 6 0 § 2 xx . \ “\ .\ 8 x - X 0 9o t++~+++++3li+~l+lfl149+ 3+ l + _»— t4 +- K .I' Area of Overlap between Supervisor A's Descriptions of Enrollees and Enrollee's Self Descriptions Supervisor‘s descriptbns of enrollees on ICL are represented by outer figure in which each coordinate is the summary Edi-"m for each individual rated. The iimer figure represents surmnary points for enrollees' self de- scriptions on _I_C_I_:. One conceptualization of the data on this page and the following three pages is that the overlap is concurrance between supervisor's perceptions of enrollees and enrollees' self perceptions. gh—o 104 “1"; Area of Overlap Between Supervisor B's Descriptions of Enrollees and Enrollees Self Descriptions Supervisor's descriptions of enrollees on ICL are represented by the smaller figure. The larger figure represents summary points for enrollees' self descriptions on the J91... 8w- 0 105 ‘r ‘_ 11ft” i “.3 Pv:w... .1: . Area of Overlap between Supervisor C's Descriptions of Enrollees and Enrollees' Self Descriptions Supervisor's descriptions of enrollees on ICL are represented by the smaller figure. The larger figure represents summary points for enrollees' self descriptions on the LC}: .' .30 .Ju l 1 106 Area of Overlap between Supervisor D’s Descriptions of Enrollees and Enrollees' Self Descriptions Supervisor's descriptions of enrollees on LCJ. are represented by the figure on the left. The figure on the right represents summary points for enrollees' self descriptions on the ICL. 107 Linear Distances between Self Description Summary Points for Supervisors and Descriptions of Supervisors by Enrollees In each case on this diagram and diagrams on pages 109 and 110 the supervisor's summary point is diagramed as the origin of each line which extends to an enrollee summary point. The linear distance maybe under- stood as psychological distance. 7-0 760 6+0 70 I40 '_ .20 .30 *0 4.- .._. ......._i......._._..._ -_...__...._- -- -f . . ' 108 Linear Distances between Self-Description Summary Points for Supervisors C and D and Descriptions by Supervisors C and D 30 90 £0 70 110 . , I I ll I I II [I in I l I in {I I 1 I11" I. I Iil I i ‘ ll .u... .. ’3..‘Il}tu-!sui 0‘1!" "'llljllljjljJJJJJJJJJlJlEs