ABSTRACT EFFECTIVE LEADERSHIP IN SMALL GROUPS DESIGNED TO TEACH INTERPERSONAL COMMUNICATION SKILLS By David Lloyd Runyan This study's goal was to ascertain the nature of effective leadership in small groups for acquiring interpersonal communication skills. Fifteen groups, with about 15 members each, led by eight different leaders, comprised the subject population. These groups (Interpersonal Process Laboratories) are the initial undergraduate experience (Education 200) in the Education Department at Michigan State University. Empathy, Respect, Genuineness, Self-Disclosure, Concreteness, Immediacy, Confrontation, Risk Taking, Non-verbal Behavior, Group Climate, Group Structure, and Trust were the variables hypothesized as integral to effective leadership and participant benefit. Constituted of item quintets, the Leader Behavior Question- naire (LBQPAR) was designed to assess each of these variables. It was hypothesized that satisfied group members would perceive the leader as high on each predictor variable in comparison to dissatis- fied group members. Member benefit was measured by the Group Member Benefit Questionnaire (GMBQPAR). Participants completed these two questionnaires post-group. David Lloyd Runyan The Interpersonal Checklist (ICL), a personality inventory, was administered to all participants pre- and post-group. It was hypothesized, but not substantiated, that those who benefited from the group would score higher on the post-test scales for Love and Dominance on the ICL. Each group was visited on five different occasions by at least three different non-participant observers. On each visit observers assessed leader functioning using the LBQOBS. The LBQOBS is identical'UJthe LBQPAR. The first seven predictor variables were also rated by the observers on a separate measure, the Index of Accu- rate Communication (IAC). IAC scores were used to assess leader functioning, both prior to and during the group (PREIAC, IACOBS). Observers also rated each group member's Self-Exploration (SE), and participant benefits using an instrument much like the GMBQPAR (GMBQOBS). ' Another measure of participant benefit was taken by having a friend of the participant from outside the group evaluate the group member's post-group benefit based on his/her behavior outside of the group setting (GMBQFR). Stepwise regression equations using the LBQPAR, LBQOBS, and IACOBS with all measures of participant benefit showed Respect to be the single best predictor of participant benefit. Group Structure, Group Climate, Genuineness, Immediacy, and Empathy, respectively, added significantly to some regression equations. The criterion measures were examined for reliability and validity. Cronbach's alpha reliability coefficient was .86 for the David Lloyd Runyan total GMBQPAR and .85 for the total GMBQOBS. Typal analyses of the GMBQPAR and GMBQOBS yielded clusters which related to the predictor variables in a manner quite similar to the total benefit scores. This suggests that GMBQ total benefit scores are a valid measure of participant benefit. The predictor measures were examined for convergent and divergent validity using Campbell and Fiske's multitrait-multimethod matrix. Sufficient convergent and marginal divergent validity was demonstrated for observer-rated measures (LBQOBS, IACOBS). Identical leader behavior questionnaires (LBQOBS and LBQPAR) completed by observers after each group session and by participants at the group's end showed the least evidence of validity. The results support the work of Carkhuff who suggested Respect, Empathy, Genuineness, and Immediacy as core conditions necessary for effecting helping relationships and the work of Hemphill concerning the importance of Consideration to effective leadership in industry. The results affirm the importance of Caring to effective encounter group leadership as demonstrated by Yalom, Lieberman, and Miles. Finally, they confirm the research of Hurley and Force that participant ratings of leader Acceptance/Rejection of Others (ARO) is highly correlated with reports of encounter group gains made by participants, observers, or intimates. EFFECTIVE LEADERSHIP IN SMALL GROUPS DESIGNED TO TEACH INTERPERSONAL COMMUNICATION SKILLS By David Lloyd Runyan A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1977 To my mother, Mrs. Eloise Runyan ii [UIU‘IU .IIUIII All . 1|. III I II. III‘ II III ACKNOWLEDGEMENTS This research could never have been undertaken without the help and cooperation of many people who gave generously of their time and skills. First, I would like to thank the chairman of my dissertation committee, Dr. John Hurley, and the other members of my committee: Dr. Lawrence Messé, Dr. Eugene Jacobson, and Dr. John Lopis. I would also like to thank my excellent observers: Bruce Hillenberg, Mike Farris, Judy Cline, Lew Krash, and Mark Kover. A note of thanks is also due the eight group leaders who allowed me to collect my data from their groups. They are Karen Maroda, Dan Price, Greg Kirsch, Jim Stibbs, John Belaski, Jack Rowlson, Dave Bowyer, and Dan Seik. Vince Cornellier allowed me to use his groups for my pre- study. Alice Semrau was my fine typist. Finally, I would like to thank the l93 students in Education 200 who took the time to answer my questionnaires and allowed my observers to enter their groups. k.I-II‘UII“|‘=[Illl“|‘.|l‘.‘(|ll Il'.llll III [All I I I I II TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES . LIST OF APPENDICES . INTRODUCTION . Chapter I. LITERATURE REVIEW Encounter Group Theory Growth of Group Phenomena Group Process . Group Goals Appropriate Group Behavior Encounter Group Research . . . . . Attaining Group Goals: Positive Results Attaining Group Goals: Negative Results Group Training in Industry: Positive Results : Group Training in Industry: Negative Results . Problems Attenuating Encounter Group Effectiveness in Industry. . Social Psychological Leadership Research Definition of a Group. . . Definition of Leadership. . Leadership, Friendship, Popularity Leadership Traits . . Ohio State Leadership Studies Leader Evaluation . Participative Decision-Making Leadership and Decision- -Making . Encounter Group Leadership Research . Early Studies . . . . Leader As Group Member Leaderless Groups . T-Group Leadership . iv Page viii ix lllIll-‘(lllll‘ l..ll [III {III A [ I'll I I I J I II II I Chapter Interpersonal Sensitivity . Facilitative and Action- Oriented Conditions. Index of Accurate Discrimination Index of Accurate Communication Validity II. FORMULATION 0F HYPOTHESES . Group Leader Interpersonal Sensitivity and Group Member Benefit . . . . . . . Empathy and Respect Self-Disclosure . Genuineness Confrontation . . Concreteness and Immediacy Risk Taking . . Trust . . Non- Verbal Behavior Group Structure and Group Climate in Relationship to Participant Benefit . . Group Structure and Group Climate . III. METHOD, INSTRUMENTS, SUBJECT POPULATION, AND OPERATIONAL DEFINITIONS . Sample Population . Instruments . . . Leader Behavior Questionnaire Group Member Benefit Questionnaire Index of Accurate Communication Interpersonal Checklist Operational Definitions . Instrument Key Hypotheses IV. RESULTS . Introduction . Predictor Variables Criterion Variables . Reliability and Validity Interrater Reliability Predictor Variables Criterion Measures . Correlational Analyses . Participant Data: Leader Behavior and Member Benefit (GMBQPAR) . . . Observer Data: Member Benefits and Leader Behavior (LBQOBS) . Chapter Page Observer Data: Member Benefits and Leader Behavior (IACOBS) . . . . . . . . . . 79 Independent Measures . . . . . . . . . . Bl Interpersonal Checklist . . . . . . . . . 82 Self-Exploration . . . . . . . . . . . 83 Course Grade . . . . . . . . . . . . . 83 Typal Analysis: LBQPAR, LBQOBS, IACOBS . . . . 84 Multiple Regression . . . . . 88 GMBQPAR with Learning and Satisfaction . . . . 90 GMBQOBS with Types 1 to 4 . . . 90 Analysis of Variance for Groups Nested in Leaders . 90 V. DISCUSSION . . . . . . . . . . . . . . . 93 Introduction . . . 93 Reliability and Validity of Predictor Measures . . 94 Reliability and Validity of Criterion-Measures . 98 Hypotheses . . . . . . . . . . . . . lOO Participant Data: Leader Behavior and Member Benefit (LBQPAR). . . . . . lOO Observer Data: Leader Behavior and Member Benefit (LBQOBS). . . . . . lOO Observer Data: Leader Behavior and Member Benefit (IACOBS). . . . . . . . . . . lOl Halo Effect . . . . . . . . . . . . . 101 Independent Measures . . . . . . . . . . 102 Interpersonal Checklist . . . . . . . . . l02 Course Grade . . . . . . . lO3 Observer Ratings of Self— -Exploration . . . . . lO4 Multiple Regression Analyses . . . . 104 Proposed Rank Order of Predictor Variables . . . l05 Typal Analysis of Predictors . . . . . . 106 Description of an Effective Group Leader . . . . lO7 Redundancy Among Predictor Variables . . . . lO9 Leader Variables Unrelated to Benefit According to Regression Analyses . . . . . . . . . . 110 Effects of Self- Disclosure . . . . . . . . lll Groups Nested in Leaders . . . . . ll4 Comparison to Ohio State Leadership Studies . . . ll4 Comparison to Yalom, Lieberman, and Miles . . . . ll5 Comparison to Carl Rogers . . . . . . . . . ll7 Comparison to Hurley Studies . . . . . ll8 -Interpersonal Communication Skills Taught in the IPL . . . . . . . ll8 Suggestions Concerning IPL Structure . . . . . 122 Future Research . . . . . . . . . . . . 124 Conclusion . . . . . . . . . . . . . . 125 vi Page APPENDIX . . . . . . . . . . . . . . . . . . l27 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . l63 vii Table l. LIST OF TABLES A multitrait-multimethod matrix to examine convergent and discriminant validity for LBQPAR, LBQOBS, IACOBS . . . . . . . . . . . . . All product-moment correlations of rated leader effec- tiveness and total member benefits Partial summary table for stepwise multiple regression for significant leader effectiveness measures with participant benefit measures . . Partial summary table for stepwise multiple regression for Learning and Satisfaction with LBQPAR . Partial summary table for stepwise multiple regression for Type l to Type 4 with LBQOBS . . . . . viii Page 7T 80 89 91 91 Figure 0-1-wa Typal Typal Typal Typal Typal analysis analysis analysis analysis analysis LIST OF FIGURES of GMBQPAR items of GMBQOBS items (of LBQPAR's component scales of LBQOBS's component scales of IACOBS's component scales ix Page 76 77 85 86 87 Appendix A. Participant and Observer Questionnaires . B Scoring Groups for Leader Behavior Questionnaire . C. Analyses of Variance D Correlation Matrices E. Regression Analyses F. An Observer's Report LIST OF APPENDICES Page l28 I37 I39 l43 ISO 156 INTRODUCTION Much has been written about the importance of empathy, genu- ineness, respect, self-disclosure, confrontation, concreteness, and immediacy in counselor/client relationships. Counselors and human beings in general are more fully functioning, more self-actualizing, and are more successful in therapy if they have high levels of these variables (Carkhuff, l969; Carkhuff and Truax, l964; Jourard, T964; Maslow, T954, l962; Rogers, 1957, l96l; Truax, l966). Very little research has been undertaken to ascertain what variables are impor- tant for successful group leadership in small groups teaching interpersonal communication skills (Campbell and Dunnette, I968; Egan, 1970). What leader behaviors are facilitative of group member satisfaction and personal growth and what leader behaviors cause group member dissatisfaction, or even "casualties" (Yalom and Lieberman, 197l)? The purpose of this research is to begin to clarify the answer to this question. The hypotheses to be explored here are based on the assumption that a successfully functioning group leader in groups teaching positive interpersonal communication skills needs many of the same interpersonal skills as a successful psychotherapist (Egan, 1970). The particular leadership variables examined here include Respect, Empathy,Genuineness,Self-Disclosure, Concreteness, Immediacy, Confrontation, Risk Taking, Non-verbal Behavior, Group Climate, Group Structure, and Trust. CHAPTER I LITERATURE REVIEW Encounter Group Theory Growth of Group Phenomena Carl Rogers (1968) has called the encounter group experience perhaps the "most significant social invention of this century." Popular articles on encounter groups are appearing in numerous books and magazines. Business leaders have used small group laboratories for years to increase managerial skill, human relations acuity, and productivity in their organizations (Benne, 1964; Blake and Mouton, 1964; Buchanan, 1965; Campbell and Dunnette, 1968; House, 1967). The use of small groups to teach positive interpersonal communica- tion skills in education (Fox and Lippitt, 1964) and in the wider community (Arnold and Stiles, 1972; Klein, 1965) is on the rise. Two journals, Group_and Organization Studies and Small Group, Behavior, are specifically devoted to the field. Rogers (1967) feels that the reason behind the proliferation of such groups is that ordinary people find it alleviates loneliness and permits them to grow and change. In an increasingly anonymous society it permits and teaches the growth of real relationships. Traditionally, edu- cation is supposed to develop a normal man's potentialities, but there is evidence to suggest that formal education has failed to achieve this function (Jacob, 1957; Miles, 1964; Rogers, 1961). School often discourages or represses creativity (Guilford, 1962; Holland, 1961). Friedenberg (1961) states that education is an) instrument of conformity rather than liberation. Indeed, fuller interpersonal living is rarely one of the results of the 8 to 16 years of formal education. Group Process Commensurate with the rising popularity of these groups is the development of a wide range of names, emphases, and procedures. This partial listing of different types of groups is taken from Carl Rogers (1970). The T-group emphasizes human relation skills. The encounter group emphasizes personal growth and the improvement of interpersonal communications through an experiential process. Sensitivity groups may resemble either of the above groups. Task- oriented groups are used in industry and focus on the actual task of the group in its interpersonal context. Sensory awareness, body awareness, and body movement groups focus on physical awareness and expression through movement and dance. Organizational development groups are used by industry to help create more skillful leaders. Team building groups are used in industry to build close-knit and effective work teams. Gestalt groups emphasize a Gestalt therapeutic approach where the leader typically focuses on just one individual at a time. The Synanon grogp is used to treat drug addicts and is characterized by violent attacks on participant's defenses. Interpersonal Process Laboratories (IPL) is the name given the interpersonal communication skills groups in the Education 200 program at Michigan State University and is the object of the present research. Each organization or institution that decides to use the intensive small group experience to develop positive interpersonal communication skills creates its own name and format. For convenience the term encounter group will be used to describe the research in this litera- ture review unless the research cited or the organization referred to uses another name. A description of the group process specific to the IPL is given on page 56 and in more detail starting on page 118. What is the structure of an encounter group and what are its goals? The brief summary to follow is based on the theory and research of Bradford, Gibb, and Benne (1964); Buchanan (1965); Bolman (1974); Campbell and Dunnette (1968); Egan (1970); Maslow (1968); and Schein and Bennis (1956). Generally, encounter groups are traditionless, unstructured, primary groups of about ten people. They meet periodically for a set period of time and then disband. Often the participants are strangers. The leader does not lead in the conventional sense. He does not define the process of the group. Rather, it is the here-and-now activities of the group itself that determine the course of action the group will take. The group leader acts more as a resource person and as a member of the group than as an authoritarian figure imposing goals and behaviors on group members. He is a facilitator rather than a leader. In fact, the leader is usually referred to as the "trainer" or "facili- tator" rather than as the group leader. In practice, leaders vary from being highly structured and highly authoritarian to those who are completely non-directive. Certain elements seem necessary for a successful group. The most basic element is feedback. The participants are required to give and receive meaningful feedback concerning their behavior and feelings as it affects themselves and others. It is the real, here- and-now perceptions and feelings that group members have of them- selves and other group members that are the core of the group process. Often the group's beginning is tense, unproductive, super- ficial, and perhaps even silent. Inevitably, group members become dissatisfied with this, and with the conventional social roles and facades that they generally present in ordinary social intercourse. They begin to reveal their real inner selves with all their imper- fections, as well as their true strengths. A climate of tension or anxiety generally pervades the group in the early stages before facades and defenses are lowered and group members begin interacting on a more genuine and intimate level. Despite the anxiety, members are often willing to open up to the group about their innermost fears, joys, and hopes. To accomplish this it is necessary for the group to establish a climate of "psychological safety." This means that no matter what a person reveals about himself in the group, he must feel that the group will be supportive and non-evaluative of him. He must feel that it is safe to expose his feelings, drop his defenses, or try out new behaviors in the group. The leader can help facilitate a group member into his real feelings but he cannot order him to be real. This is a decision that must be made by each member of the group. The leader acts as a model for the behaviors deemed neces- sary for maximum learning to occur. He gives and accepts feedback non-defensively. He is accepting of recognized imperfections but confronting of inconsistencies. He expresses his own feelings openly and honestly and encourages others to do so also. The leader can and often does introduce exercises to facilitate participants down into their real feelings. Exercises may be verbal or non-verbal. Leaders may rely heavily on exercises or they may use none at all. Group Goals The goals of the encounter group experience are not all agreed upon, but a few are basic to all encounter groups. 1. Increased self-insight or self-awareness. This means increased understanding of one's own feelings and behaviors and their impact on others. 2. Increased interpersonal sensitivity. This means a greater knowledge of the feelings and behavior of others and of their impact on you. 3. Understanding the conditions that inhibit or facilitate group functioning. What behaviors and attitudes make for a pro- ductive and satisfying group experience and what behaviors are destructive of a positive group experience? 4. Development of diagnostic skill for individual, group, and organizational behavior. This means learning the skills to be an effective facilitator yourself. In a critical review of the literature, Parloff (1970) sug- gests the goals of T-groups are (a) enhanced organizational effi- ciency, (b) enhanced interpersonal skills, and (c) an enhanced sense of well-being. Some writers would substitute the terms personal growth and/ or self-actualization for the above list. These terms refer to developing your own unique potentialities to their greatest extent while still relating successfully with other people. In practice, the skills mentioned in the above list are a major part of personal growth or self-actualization. Appropriate Group Behavior In a pilot study using an operant conditioning analysis to observe and classify the actual behavioral requirements of a T-group, Marshall and Coleman (1974) observed and categorized crucial behav- iors that were consistently positively or negatively reinforced. Positively reinforced verbal behavior included current feelings, risk taking, self-disclosure, "useful" interpretations, questions concern- ing other's current feelings, requesting group attention to one's own current feelings, humorous comments to relieve tension, empathy and sympathy responses, and compliance to leader suggestion. Physical touching, with the exception of genital contact, was positively reinforcing non-verbal behavior. Negatively reinforced verbal behavior included extended silences; "intellectual" statements about attitudes and feelings; strong direct expression of aggressive feelings; flippant, casual, "non-feeling" remarks; defensive behavior; discussion of the past irrelevant to the here-and-now; and lack of emotional involvement in the group. Negatively reinforced non-verbal behavior includes insin- cere or rejecting affect, inconsistencies in verbal and non-verbal egg: ll‘((.{{ I; I'- ’l l 11' III]. II expression, physical tension or rigidity, and maintaining physical distance from the group. Positive reinforcers were expression of positive feelings to another, expressions of empathy, agreement to content of a statement asking someone to share current feelings, positive comments on another's individual style, and comments on an individual's importance to the group. Non-verbal positive reinforcers included structured physical group interactions, spontaneous touching and embracing, and casual physical contact. Negative reinforcers included expression of negative feel- ings to another, doubting another's sincerity, disagreeing with the content of another's expression, and ignoring another. Negative non-verbal reinforcers were decrease or withdrawal of physical contact. In an interesting article examining encounter groups from a psychoanalytic viewpoint, Freundlich (1972) suggests that “the intensity of emotional support and confrontation (in encounter groups) is more profound than that which occurs during traditional psychotherapy." He emphasizes the extended time period, non-verbal behaviors, and member expectations as reasons for this greater ‘ intensity. Encounter Group Research Attaining Group Goals: Positive Results How successful are encounter groups in attaining these goals? According to current research, this depends on who you ask. The majority of participants are satisfied that it is an enjoyable and worthwhile learning experience (Campbell and Dunnette, 1968; Bunker, 1965; Gibb, 1970; House, 1967; Miles, 1965; Stock, 1964: Yalom, Lieberman,and Miles, 1973). In a comprehensive and well- controlled experiment, Yalom, Lieberman, and Miles (1973) found that 65 percent of all participants felt that some positive change had taken place and that these changes would endure. Six months later the degree of enthusiasm concerning positive benefits dropped some- what. Initially, following the group the ratio of satisfied to dissatisfied members was 4.75:1. Six months later, the ratio was reduced to 2.33:1. Participants reported they were more open, hon- est, intimate, self-aware, aware of others, spontaneous, confident, and talkative. Leaders were even more optimistic than the participants con- cerning the positive effect of the encounter group. They saw improvement in 90 percent of the participants and high improvement in 30 percent. However, correlations between leader ratings, self- evaluations, and coparticipant ratings were near zero. In a six-month follow-up study, questionnaires were mailed to participants' close friends, associates, and relatives asking them to describe any changes in his/her behavior since the encounter group. Change differences, either positive or negative, for both encounter group subjects and control subjects were insignificant. Finally, a Composite Change Index was created reflecting changes in participants from all perspectives. Roughly one- third showed positive changes following the encounter group, lO one-third was unchanged, and one-third showed negative change. The picture was roughly the same six months later. In summarizing his extensive review of the literature, Gibb (1970) concluded that "intensive group training experiences have therapeutic effects." Changes occur in sensitivity, ability to man- age feelings, directionality of motivation, attitudes toward the self, attitudes towards others, and interdependence. Sensitivity implies awareness of one's own feelings and the feelings and per- ceptions of others. It also involves openness, authenticity, and spontaneity. Managing feelings refers to the ownership of one's feelings and congruence between feelings and behavior. Direction- ality of motivation refers to self-actualization or self- determination. Attitudes towards self include self-esteem, self- confidence, and congruence of perceived and ideal self. Attitudes towards others includes a decrease in authoritarianism, reduced emphasis on structure and control, greater acceptance of others, and emphasis on participative management. Interdependence refers to interpersonal competence, teamwork in problem solving, and being a good group member. In comprehensive studies, Bunker (1965) and Miles (1965) had encounter group subjects rate themselves on a perceived change mea- sure and had six to eight associates of each subject respond to the same perceived change measure eight months after training. Two control groups were used. One "matched" group was nominated by the subject himself. The “random" control group was created by ‘(Eiil‘g (I [1‘ (I! l I l 1" Il‘ It All I | 11 randomly choosing names from a national listing. Percentages of self-reported changes were 82 percent for the experimental group, 33 percent for the matched controls, and 21 percent for the random controls. Observers reported behavior changes for 30 percent of the experimentals, and 10 percent and 12 percent for each of the con- trols, respectively. The perceived change measure was an open-ended questionnaire asking the respondent to report any changes in the subject's behavior over the past year in comparison to the previous year. An informal content analysis (Miles, 1965) concluded that changes included increased sensitivity to others, heightened equalitarian attitudes, greater communication and leadership skills, and patterns of increased consideration and relaxation in decision-making. Argyris (1965) and Boyd and Ellis (1962) compared the effec- tiveness of T-group training and the lecture/discussion method in producing positive change in the interpersonal sensitivity of the subjects and both studies found T-groups to be more effective. Boyd and Ellis (1962) found observers reported significantly more behavior changes and significantly more positive behavioral changes in T-group trained managers than in control groups of lecture/ discussion trained managers and non-trained managers. T-group par- ticipants were also observed to have made more undesirable changes as well. Finally, in their book devoted to T-group research Cooper and Mangham (1971) conclude with some reservations that ". . . there 12 is moderately strong evidence of lasting and effective change in individuals participating in T-group training." Attaining Group Goals: Negative Results Unfortunately, not all the studies of encounter group out- comes give positive results. Yalom, Lieberman, and Miles (1973) studied 16 encounter groups and found vast differences in effective- ness. One of the worst groups had three dropouts, two psychological casualties, and one positive learner. One of the more successful groups had ten persons with a positive learning experience and no one with a negative experience except, perhaps, one dropout. As noted earlier, the magnitude of positive results depends on the person reporting. Invariably, self-reports of positive changes are greater than observer or peer evaluations. Leader evaluations are most positive of all. Yalom et a1. (1973) reported that one-third of all subjects experienced negative changes. In fact, 9.4 percent of all experimental subjects were classified as casualties (Lieber- man and Miles, 1971). A casualty was defined as someone suffering from significant negative outcomes as a result of encounter group participation. The best predictor of casualties were multiple coparticipant selection and, secondly, self-selection. Leader judgments about casualties were unreliable. In an analysis of leader styles it was found that aggressive, charismatic leaders were most likely to have severe casualties. The mode of injury was generally verbal attack or rejection by the leader or the group or both. Failure to attain unrealistic goals 13 was also cited as a reason. Some examples included a girl who was verbally and physically assaulted by the leader and the group in an effort to have her "get in touch” with her anger. Attack by the group also came from cool, distant, laissez-faire leaders. Here the group is usually non-cohesive and non-supportive and there is a "pressure to produce." Leaders described as "energizers” had a 17% casualty rate, and ”providers" only a 3% casualty rate. In terms of casualties, the worst leader said the members were "infantile--not grown up enough to make an adult contract. I'll have a good time for myself." One "provider" had no casualties. He said, "I knew there were some fragile people. I pulled my punches." Group Trainingpin Industry: Positive Results Many researchers have examined the utility of encounter group training in industry (Campbell and Dunnette, 1968; Buchanan, 1965; House, 1967). Two previously described studies (Bunker, 1965; Miles, 1965) which reported positive results used co-workers to describe behavior changes in elementary school principals and mana- gers. Valiquet (1968) compared former T-group participants to untrained controls inside the same company and found risk taking and "function flexibility"--the ability to be an effective group member and to accept change--were shown to be higher for former T-group members. Marrow, Bowers, and Seashore (1967) compared Harwood Manu- facturing Company and Weldon Manufacturing Company and found Harwood 14 to be superior in man-hour productivity, turnover rate, waste and morale. Consultants were hired to change conditions in Weldon. Over three years, productivity and morale improved significantly at Weldon. The changes were attributed to (a) individual counseling and training of low performing workers, (b) termination of employees with chronically low production rates and high absences, (c) train- ing of supervisors and staff in interpersonal relations using T-groups, and (d) introduction of group problem-solving meetings between supervisors and workers. Seashore and Bowers (1963) evaluated a change program in an industrial firm which included T-group training for managers, organizational feedback meetings, and a control group composed of departments which did not participate. They found that productivity did not change for either group, but employee attitudes became more positive in the experimental groups. Gropp Training in Industry: Negative Results Many studies analyzing the effectiveness of encounter groups in industry have produced problematic results (Campbell and Dunnette, 1968; House, 1967). For example, Deep, Bass, and Vaughn (1967) studied the effects of encounter group training on simulated mana- gerial behavior. They found that work groups that were intact encounter groups did worse on the Carnegie Institute of Technology Management Game than either control groups which had no encounter group training or splinter groups composed of persons who had group training but who had never grouped together. Subjectively, Bass 15 attributed the poor performance of the intact encounter groups to the neglect of the control/management function. No one bothered to check if the others were carrying out their assignments. Underwood (1965) asked observers to rate behavior changes according to their effects on job performance. He found that observers noted 250 percent as many changes for the experimental group as for the control group. However, the ratio of changes judged to increase effectiveness to those judged to decrease effec- tiveness was 4:1 for the control group. For the experimental group the ratio was only 2:1. The suggestion is that while encounter groups lead to more observable changes in member's job behavior, they also produced a higher percentage of unfavorable changes in respect to ratings of job performance. These outcomes reflect one of the major questions and prob- lems concerning the effectiveness of encounter groups in industry or elsewhere. This is the problem of transfer of training. Given that group member opinion is generally favorable towards groups and that group members feel they have increased their own self-awareness and their interpersonal sensitivity, why is it that so few studies show such favorable outcomes as experienced by business associates or friends? Problems Attenuating Encounter Group Effectiveness in Industgy Bolman (1970) originally set out to study the efficacy of laboratory versus lecture in training executives in interpersonal skills. The experimental subjects, who initially participated in a 16 one-week encounter group, reported satisfaction with the group and increased interpersonal sensitivity. A control group, which studied interpersonal sensitivity via the lecture/discussion format for one week, exhibited no change on the pre-post-test. However, the experimental group felt that their encounter group experiences were invalidated and discredited in the remaining five weeks of the program, which was similar for both experimental and control groups. Encounter group members felt themselves slipping back to their old behaviors as a result of this invalidation. Campbell and Dunnette (1968) and Schein and Bennis (1965) suggest that the problem of demonstrating transfer of training effects from encounter group to job are magnified because many of the supportive elements of the encounter group process are removed back on the job where values, norms, leadership, and organizational structures are different. It is possible that Lowin's (1968) warning to researchers of participative decision-making (PDM) to establish a sound PDM program where top level management is solidly behind its implementation should also be extended to experimenters attempting to study the efficacy of sensitivity training in industry. Friedlander (1967) had real—life, completely intact, on-the- job work teams participate in T-group training and compared the results to similar control groups in the same company who received no training. It was felt that by providing laboratory training for actual work groups, many of the problems created by returning from the laboratory environment to the actual job environment could 17 be more easily resolved. Analysis Showed the training group increased significantly on team effectiveness in problem-solving, mutual influence among group members, and members' sense of personal involvement and participation in group meetings. Perhaps the greatest problem encounter group proponents face in attempting to convince their more "tough-minded" colleagues is the methodological shortcomings apparent in most encounter group research (Campbell and Dunnette, 1968). Helmreich (1972) suggested three basic methodological problems facing encounter group research: (a) inadequate or absent control groups, (b) unwarranted causal inferences from correlational data, and (c) reliance on subjective self-reports or possibly biased ratings by peers, subordinates, or supervisors. Campbell and Dunnette (1968) came to much the same con- clusions. They refused to draw conclusions concerning encounter group effectiveness because of missing or inadequate control groups. They also felt that studies using observers in the back-home set- tings were contaminated because observers were often nominated by the subject, often had interactions with each other, and perhaps discussed the subject; and the observers apparently knew the subject they were describing had gone through the encounter group process. Finally, they suggest that assessment of what goes on in encounter group training "must first cope with the problem in measuring this elusive phenomenon called interpersonal sensitivity." 18 ,Social Psychological Leadership Research The industrial/social psychological literature suggests the great complexity inherent in the study of leadership. Although not directly related to leadershipirlgroups teaching interpersonal com- munication skills, this literature gives some insight into the nature of effective leadership and successful groups. Definition of a Group Kurt Lewin (1939) defined a group in terms of the inter- dependence of its members. Cattel (1951) defined the group as an instrument for satisfaction of individual needs. Proshansky and Seidenberg (1965) define a group as two or more individuals who share a common set of norms, beliefs, values, and goals which must be such as to give satisfaction to the needs of the individual mem- bers of the group. Definition of Leadership Leadership is even harder to define than groups. Present. thought refutes the idea that leadership ability in one situation also means leadership ability in another situation. Rather, leader- ship is seen as a set of group functions which must occur in any group if it is to effectively satisfy the needs of its members (Gibb, 1969). Generally, the leader is perceived as the individual in a,given office. The problem here is the difference between leadership and headship or formal and informal leadership (Gibb, 1969). In organizations such as industry, education, and the military, group structure and processes, including leadership 19 hierarchies, are formally defined. In less structured traditionless groups there may be no formally defined leaders. Even in formally organized bureaucracies, the informal organization produced by mem- bers' interpersonal relationships exercises.great social control over its members (Gibb, 1969; Seashore and Bowers, 1963). Scott (1952) and Stogdill and Koehler (1952) present evidence that sug- gests that the morale and effectiveness of the organization depend on the extent of the discrepancies between formal and informal organization. This is a dilemma that confronts any researcher who examines group leadership. Even when there is a formal leader, actual leadership activities can be undertaken by any of the group members. Hemphill (1952) defined leadership in terms of leadership acts which "initiate a structure in the interaction of others as part of the process of solving a mutual problem." Leaders are iden- tified by the relative frequency in which they engage in these acts. Collins and Guetzkow (1964) define leadership as a "scattered activity." One member is influential at one time in one set of circumstances and another is influential at another time in a dif- ferent set of circumstances. Most groups have many leaders. Only rarely, and then only in highly structured groups, can we identify "the leader" (Gibb, 1969). Leadership, Friendship, Popularity There is good evidence that group members can reliably iden- tify, through sociometric techniques, those members who have 20 greatest influence on them. These perceptions are closely correlated with leaders identified by external observers and other criteria (Gibb, 1950; Wherry and Fryer, 1959). However, sociocentrality is not necessarily leadership. One must be careful about the nature of the sociotelic question asked. Is one measuring leadership or simply friendship or popularity? Hollander and Webb (1955) found friendship choices to be more closely related to followership than leadership. Gibb (1950) had group members rate themselves on three sociometric questions: (a) With whom would you like to spend your leisure time? (b) Whom would you like to work with? (c) Identify the person whose removal from the group would bring about the largest group change. Sociometric choice on the third criteria had by far the highest correlation with observer ratings of leader- ship. Leadershjp Traits Early leadership research concentrated on the search for leadership traits. The clear implication here is that leadership is specific in an individual and constant in all situations. The recent concept of distributed leadership is quite contrary to this earlier position. Perhaps this is why the earlier research on leadership traits was fairly unproductive. Reviews by Bass (1960), Mann (1959), and Stogdill (1948) show positive, but small, correla- tions between certain personality traits and leadership. Mann (1959) summarizes the most positive findings: "The positive relationship of intelligence, adjustment, and extroversion to 21 leadership are highly significant. In addition, dominance, mascu- linity, and interpersonal sensitivity are found to be positively related, while conservatism is negatively related" (p. 252). Ohio State Leadership Studies Many studies have attempted to examine what it is that leaders actually do. The most comprehensive of these are the Ohio State Leadership Studies conducted by Hemphill and his colleagues (1950). Nine a priori dimensions of leader behavior were postula- ted, and questionnaire scales designed to measure these dimensions were factor analyzed by Halpin and Winer (1952, pp. 21-31), sug- gesting four dimensions that were shown to be crucial by later research. These four dimensions and their percentage of total variance are: 1. Consideration: 49.6%. The extent to which the leader is considerate of his followers. This factor is characterized by warmth of interpersonal relationships, mutual trusting, readiness to explain actions, and willingness to listen to subordinates and allow them to participate in decision-making. 2. Initiating structure: .33.6%. Scale items with high positive loadings include "maintains definite standards of per- formance, assigns crew members to particular tasks, and asks that the crew follow standard operating procedure." 3. Production emphasis: 9.8%. The leader motivates the group members by emphasizing the job to be done. 22 4. Sensitivity: 7.0%. This factor stresses the leader being a socially acceptable individual in interactions with other members. Subsequent studies have confirmed that Consideration and Initiating Structure are the two major dimensions of leader behavior (Fleishman, 1962; Fleishman, Harris, and Burt, 1955; Halpin, 1955; Halpin, 1957; Oaklander and Fleishman, 1964; Rush, 1957). Since a leader low in Consideration was almost never an effective leader, whether he was high or low on structure (Fleishman and Harris, 1962), Gibb (1969) has suggested that perhaps Consideration is thg_most crucial leadership variable. Leader Evaluation Leadership at a given level in an organization is evaluated in different terms by those above and those below that level. Halpin and Winer (1952) found that administrative superiors tend to associ- ate high Initiation of Structure with "good" leadership and high Consideration with "poor" leadership. The American Soldier (Stouffer et a1., 1949) presents strong evidence that privates and officers hold different attitudes toward authority. Although they did not use the language of the Ohio State Leadership Studies, Stouffer and his associates found that enlisted men preferred their immediate superiors to be high in Consideration, whereas higher ranking officers wanted these same men to be higher in Initiating Structure. 23 Participative Decision-Making Participative decision-making (PDM) is a theory of manage- ment which emphasizes supportive and considerate leadership on the part of management and participation in decision-making on the part of subordinates (Likert, 1961; Vroom, 1969). Some research shows a positive relationship between PDM and productivity (Bavelas, in French, 1950; Coch and French, 1948; Strauss, in Whyte, 1955) and other researchers have found no relationship between PDM and pro- ductivity (French, Israel, and As, 1960; Sales and Rosen, 1965). However, almost all research has shown a definite positive relation- ship between PDM and employee satisfaction with work (Morse and Reimer, 1956; Seashore and Bowers, 1963). In a typical study, Seashore and Bowers (1963) evaluated a change program designed to implement Likert's (1961) theory of effective management. Likert emphasizes PDM, consideration, and effective feedback. The organi- zational change program included sensitivity training for all mana- gers. Results indicated that productivity did not change for either experimental or control groups, but employee attitudes became more positive in the experimental group where all managers went through sensitivity training. PDM seems a real-life actualization of the distributed leadership orientation emphasized in current social psychology. The methods and goals of PDM are also congruent with those of encounter groups. Often, today, in implementing PDM programs in industry, some form of sensitivity training is used (Buchanan, 1965; Campbell and Dunnette, 1968; House, 1967). 24 Caron (1964) found that in comparison to a group of controls, experimental subjects placed more emphasis on Consideration and less emphasis on Initiating Structure following an encounter group experience. Caron used the Leader Opinion Questionnaire (Fleishman, Harris, and Burt, 1955) which asks the subject to describe the "ideal" leader. Research directly relating encounter groups in the indus- trial context to goals of personal satisfaction in the job setting and increased productivity has shown that consideration and morale increase following the encounter group experience and that the relationship between encounter groups and productivity is problem- atic (Campbell and Dunnette, 1968; Vroom, 1969). Here it might be helpful to delineate the differences between "tactical" and "organic" PDM (Lowin, 1968; March and Simon, 1958). Tactical PDM refers to the temporary espousing of PDM philosophy by management in an effort to induce the perception of PDM on the part of employees, while in actuality management is cool or even hostile to many of the ideas of PDM. They are interested in the possible potential results of PDM, but afraid of the changes necessary in attitude and behavior for PDM to become a reality. This is the point of conflict between what McGregor (1960) calls Theory X and Theory Y organizational paradigms. Theory X is the traditional hierarchy with formal superior/subordinate positions. Managerial attitudes associated with this view are of employee irresponsibility, untrustworthiness, incompetence, and need to be closely supervised. Theory Y emphasizes horizontal rather than 25 vertical organization. Theory Y managerial attitudes toward employees are that they are responsible, trustworthy, competent, and do not need to be closely supervised (McGregor, 1960). The gap between Theory X and Theory Y is so great that in his review of PDM research Lowin (1968) speaks of a "systematic antagonism" to PDM research, and suggests that the experimenter must first prepare a supportive environment in which PDM acts are not aborted. Precon- ceived expectations on the part of management concerning the value of employee suggestions, or else fear of a loss of control, under- mine the effectiveness of PDM. Many researchers have emphasized the importance of agtual_organic PDM rather than attempting to create a superficial PDM simply to influence the perceptions of the employees (Benne, 1961; Bennis, Benne, and Chin, 1961; Marrow and French, 1946; Schein and Bennis, 1965). The implication seems to be that the greatest increase in satisfaction and productivity can be expected where there is a strong, "organic" commitment to PDM on the part of management. Inconsistencies in applying PDM will destroy the effectiveness of ROM. However, in its ideal case, PDM is a theory of leadership in logical harmony with theory concerning high level encounter group leadership. The structure of the encounter group encourages distributed leadership, leader-as-member, consideration, and the development of interpersonal skills. Leadership and Decision-Making In Leadership and Decision-Making, Vroom and Yetton (1973) study the effect of participation in decision-making by subordinates. Mill-[(571 A III. 1'1 l1 (III 1 (’1 I 26 They suggest that no one leadership style is applicable to all situations. Individual leaders show great variance in the leader- ship styles they use across different problems. Based on their own research and the literature pertinent to PDM and leadership styles they created a normative model of decision-making in the form of a "decision tree." First, they categorize decisions as either group problems or individual problems. In either category the leader's decisions can be autocratic, consultative, group, or dele- gated. In autocratic decisions the leader solves the problem him- self based on his own information or that obtained from subordinates. In consultative decisions the leader shares the problem with the relevant subordinates, perhaps as a group, and makes the decision alone. For group decisions the leader shares the problem with the relevant subordinates as a group. The group attempts to reach a consensus and the decision may or may not reflect the ideas of the leader. Delegated decisions usually concern only one subordinate. The leader delegates his decision power to the subordinate. Vroom and Yetton have prepared eight "yes" or "no" questions for the leader to ask himself concerning the appropriate decision- making process. 1. If the decision were accepted would it make a dif- ference which course of action was adopted? Some problems have various good solutions and some have only one solution. 2. Do I have sufficient information to make a high quality decision? 3.- Do the subordinates have more information that is valuable in making a correct decision? 00 they have a preference? 27 4. Is the problem structured? 5. Is the acceptance of the decision by subordinates critical to effective implementation? 6. What is the prior probability that the leader's autocratic decision will receive acceptance by subordinates? 7. Are subordinates motivated to base solutions on organizational goals and objectives? 8. Is conflict among subordinates likely among preferred solutions? Depending on how the leader answers these eight questions, he can follow the decision tree to the feasible set of problem solu- tion styles for that problem. Some problems have only one solution and others have several, ranging from autocratic to group-oriented. The number of man-hours to solution and the potential group compo- sition are factors that can help in choosing among alternatives in a feasible set. Encounter Group Leadership Research Early Studies Research directly dealing with encounter group leadership adds a little more insight into the nature of effective encounter group leadership. An early survey of encounter group research (Stock, in Bradford, Gibb, and Benne, 1964) suggested that only one study involving the role of the trainer had been undertaken to that date (Deutsch, Pepitone, and Zander, 1948). This study first sug- gested that the trainer served as a model for the kinds of affect expressed in the group. It also found that trainers expressed more positive than negative feelings. Stock (1964) concludes that "the 28 trainer's role is one of the relatively unexplored areas." This conclusion was later echoed by Campbell and Dunnette (1968) and Egan (1970). These conclusions are even harder to understand given the primary importance of the group leader to the group processes. Campbell and Dunnette (1968) describe the trainer as "a dominant technological element that bears on the group's effectiveness for giving and receiving feedback." The trainer "exhibits for considera- tion the very processes deemed necessary for maximum learning to occur." Culbert (1966) feels the trainer's behavior is the single most powerful determiner of the group's behavior. The trainer not only provides the group's goals, but also the processes necessary for achieving them. Based on empirical research into the normative structure of encounter groups, Luke (1972) concludes that members perceive the trainer to exert more influence than members in deter- mining group norms. The primary norms are supportive of interper- sonal encounter. Leader As Group Member Most authors also emphasize that the leader is also a mem- ber of the group (Bradford, Gibb, and Benne, 1964; Egan, 1970). Some suggest that the leader overtly rejects his leadership role (Campbell and Dunnette, 1968). Egan (1970) speaks of the diffusion of leadership, and although he emphasizes the importance of initial structuring on the part of the leader, he suggests that the leader becomes, over time, less and less a leader, and more and more a group member. 29 Leaderless Groups Many groups are "leaderless." Gibb (1964) has been experi- menting with them for a number of years. Harrow et a1. (1967) found unled groups to be more warm and supportive than led groups. Salzburg (1967) found unled groups to be more spontaneous, yet they produced fewer problem relevant responses. Berzon and Solomon (1966) found unled groups had more confrontation than led groups. There are also "instrumented" training laboratories (Blake and Mouton, 1962; Solomon, Berzon, and Davis, 1970) that do not use a trainer, but provide self-administered evaluation forms to provide the participants with feedback. Solomon et a1. (1970) even struc- tured a nine-day group on tape with exercises, ground rules, and feedback instruments. However, when he is present, the leader is a powerful member of the group. T-Group Leadership Cooper and Mangham (1971) divided the empirical studies of the T-group trainer into one of four categories: 1. The relationship of trainer personality and trainer style. After reviewing several studies, they are only able to con- clude that personality does influence style. 2. Participant perceptual change and the trainer. These studies (Lohman, Zenger, and Weschler, 1959; Vansina, 1961) examine member self-perceptions and member perceptions of the trainer. Mem- ber "ideal" versus "actual" self-perceptions are also included. These studies show that member self-perception and member perception of 30 the trainer are significantly closer at the end of the group experi- ence. The trainer is initially perceived as an ideal. Later on he is still perceived as more adequate than members view themselves, but idolization is diminished. Descriptions of ideal versus actual self are closer at the end of the group experience. 3. Trainer impact of group development. Psasthas and Herdert (1966) conclude the trainer is the most important source of group norms. They had participants write down the most significant trainer intervention. These included analyzing group process, per- sonal feelings, feedback, and acceptance/concern about membership. In the beginning, analyzing group process was highest; later, acceptance/concern was most important. 4. Trainer influence and participant change. Bolman (1971) suggests that trainer congruence and empathy are related to partici- pant learning. Culbert (1966) suggests trainer self-disclosure is important, especially in the beginning of the group. Peters (1966) suggests that learning occurs to the extent the member identifies with the trainer. Smith and Pollack (1968) believe that internaliza- tion leads to more lasting change than compliance or attraction. O'Day (1976) suggests that leader willingness to encourage and accept the direct expression of member hostility toward himself is a necessary precondition for members to become independent of the leader and to develop supportive, trusting relationships with each other. Finally, Cooper and Mangham (1971) question whether the trainer is the sole source of influence. They feel future research 31 should examine the effects of group composition and intragroup dynamics in relationship to trainer behavior and participant change. Egan (1970) describes the functions of the leader/member: 1. Initial structuring. He suggests that the leader should be warm and accepting from the beginning, rather than aloof and ambiguous. 2. Putting his knowledge and experience at the service of the group. The leader is a social engineer who encourages positive group processes and forestalls processes that do not contribute to the goals of the group. The leader encourages interactions that increase the cohesiveness of the group and confronts behaviors that -limit interaction. 3. Dealing with the major problems of the group as natural. One problem is that of handling the trainer. Bennis (1964) divides participants into three categories: dependents, counterdependents, and independents. Dependents look to the trainer for cues, counter- dependents solve their dependency needs by opposing the leader, and independents are not threatened by the prospect of intimacy. Another problem is the prevention of "tacit understandings," either covert or overt, on the part of group members not to discuss a particular issue. Any kind of hidden agenda in the group muddies communica- tions. Finally, Egan discusses "lowest common denominatorism." Whitman (1964) claims the group can move along only'as rapidly as its slowest member. Even one unmotivated, indifferent member can retard the goals of the group. The leader must see to it that a deviant member does not absorb all the group's energies. 32 Bolman (1970) expresses six educational functions of a leader in an experiential group: 1. Providing feedback to members about the conse- quences of their behavior. Questioning to stimulate participants to think or behave in new ways. Modeling so participants can learn by imitation or identification. Supporting participants to provide a climate of psychological safety to encourage lowering of defenses and risk taking. Providing a conceptual framework to help under- stand behavior. Structuring the environment or behavior through the introduction of exercises. In a unique and informative operant analysis of encounter groups, Marshall and Coleman (1974) suggest the group leader is the most important source of positive and negative reinforcement in the group. He is constantly shaping the desired behaviors of the group members and is used as a behavioral model to imitate by the members. In this pilot study the authors observed two encounter groups in an attempt to classify the behavioral requirements of the T-group, so its goals, operation, and method could be more clearly delineated; behavioral requirements that are consistently positively or nega- tively reinforced. The authors classified the types of reinforcers, both positive and negative, used by members and leader, behavior that was differentially reinforced, modeling behavior by the leader and members, and examples of behavior modification processes such as 33 shaping and extinction. The authors emphasize there was no fre- quency count of behaviors or reliability checks of observations. They felt the leader was of primary importance in shaping appropriate behavior. He was watched and modeled. He modeled risk taking, self-disclosure, and reinforced positive and negativeverbal and non-verbal behavior. In the beginning, he reinforced weak forms of encounter behavior and in later sessions he reinforced only stronger encounter behavior. The most recent and most comprehensive study of encounter groups (Yalom, Lieberman, and Miles, 1973) also studied some of the effects of leadership in encounter groups. Based on member and observer ratings of leaders, they divided leaders into six categories. 1. Energizers (Gestalt, Synanon, and Psychodrama). These leaders are charismatic, authoritarian, struc- tured, confronting, and highly self-disclosing. 2. Providers (NTL, Personal growth, Transactional analysis, Marathon eclectic). These leaders are typified by high amounts of caring, loving, and accepting. ' 3. Social engineers (NTL, Rogerian, Psychodrama). These leaders concentrated on directing the group as a whole, rather than on the problems of individual mem- bers. They were low on self-disclosure and confron- tation. 4. Laissez-faire (Psychodrama, Transactional analysis). These leaders were perceived as technicians. They were generally cool and distant. 5. Cool aggressive stimulators. These leaders were less aggressive and less authoritarian than ener- gizers, but were similar to energizers and laissez- faire. 34 6. High structure (Sensory awareness). This single leader was perceived as highly authoritarian and used many exercises. 7. Tape. The instrumented, leaderless groups were perceived as supporting warm, and deemphasizing interpersonal confrontation. In terms of satisfaction and learning, the participants rated energizers the highest and then providers. However, over a six-month period after the group experience, ratings for energizers tended to drop whereas ratings for providers remained stable. It is significant to note that the casualty rate for energizers was 17 percent and only 3 percent for providers. This compares to an average casualty rate of 9.4 percent. All the leaders were considered "experts" in their field and were well-paid to participate in the study. The data showed no dif- ference in leadership quality as a function of the ideology of the leaders. For example, one Gestalt and one transactional group were rated near the top, whereas another Gestalt and another TA group were rated near the bottom in terms of member outcome. This par- ticular finding emphasizes the need for research that deals with effective leadership characteristics without regard to ideological orientation. The group name or the leader's label is irrelevant to successful teaching of interpersonal communication skills in small groups. Factor analysis of the 48 scales which observers and members used to rate leaders resulted in four basic leadership dimensions: Emotional Stimulation, Caring, Meaning Attribution, and Executive Function. Leader behavior is Emotional Stimulation to the extent 35 that it elicits a response from a participant. Emotional Stimula- tion might be in the form of feedback or self-disclosure. Caring involves friendship, love, protection, support, praise, and encour- agement. Meaning Attribution involves giving meaning to the experi- ences participants undergo. These may be ideas, concepts, or values about what goes on in an encounter group. Executive Function involves managing the group as a social system. The most effective leaders were high in Caring, moderate in Stimulation and Executive behavior. Meaning Attribution showed a strong correlation with positive outcomes, correlating r_= .67 for productive groups. Hurley (1976) and Hurley and Force (1973) have suggested that Acceptance/Rejection of Self (SAR) and Acceptance/Rejection of Others (ARO) are the two prepotent dimensions for assessing the effects of trainers on their encounter groups as well as for understanding all interpersonal behaviors. Hurley and Force's (1973) results stated that participants' within-lab effectiveness ratings of their trainers correlated highly (r_= .98) with six-month postlab multi- observer, multi-method change scores. Effective interpersonal behavior was defined here as high scores on both self- and other- acceptance. Participants' ratings of trainer effectiveness were also highly correlated with ratings of trainer self—disclosure and feedback seeking. Self-reports of laboratory gain were most favor- able, but generally in agreement with observer ratings made by either an intimate or colleague nominated by the participant. Using the SAR and OAR paradigm to assess participant gain, Pinches (1972) found that self-reports of participant gain were highly related to 36 participant ratings of amount of trainer attention paid to group members, perceived trainer ego-strength, and activity. This suggests that directive, involved trainer behaviors were more constructive than passive leadership. In a reanalysis of Bolman's (1973) data, Hurley (1975) presents new statistical analyses on various question- naire items that suggest trainer ARO and SAR is a stronger predictor of participated rated leader effectiveness than Bolman's (1973) empathy/congruence. Interpersonal Sensitivity The importance of interpersonal sensitivity to effective leadership has been demonstrated by many researchers in many environ- ments. Campbell and Dunnette (1968) suggest that the problem of measuriwginterpersonalsensitivity is the first problem that needs exploration if we are to understand what goes on in an encounter group. They also bemoan the complexity inherent in measuring this elusive phenomenon. They feel these problems have been stumbling blocks for all encounter group research to date. The most comprehensive study of interpersonal sensitivity, its definition, measurement, and prediction is found in the litera- ture on high- and low-functioning therapists. Rogers and his associates (1967) outlined certain qualities that were empirically demonstrated to be necessary for a successful therapist. These include warmth, genuineness, accurate empathy, and a willingness to be known. Rogers used the term "unconditional positive regard." 37 Unconditional positive regard includes not only warmth, but also a deep caring for the therapeutic client as a person. Several authors (Jourard, 1964, 1968; Mowrer, 1964; Rogers, 1961) have written on the importance of full client self-disclosure for successful therapy. Truax and Carkhuff (1965) have reported significant correlations between patient and therapist self- disclosure. Further, level of patient self-disclosure appears to be a predictor of final case outcome. Jourard (1969) has argued that disclosure by experimenters will result in greater honesty in subjects and prevent experimenters from acting like spies and manipulators. The experimenter who asks for self-disclosure from subjects and reveals nothing of himself may have created a situation of inequity. Jourard and Friedman (1970) suggest that disclosing experimenters elicit more self- disclosure from subjects and are rated more trustworthy and more positive in general than non-disclosing experimenters. Two excel- lent literature reviews of the empirically investigated parameters of self-disclosure are provided by Cosby (1973) and Goodstein and Reinecker (1974). Others (Berenson, Carkhuff, and Moravec, 1968; Berenson and Mitchell, 1968) have added confrontation to the repertoire of high- functioning therapists. Truax and Carkhuff (1965) showed that dif- ferent degrees of concreteness offered by high-functioning thera- pists is related to patient progress. 38 Carkhuff (1969) summarizes the core of conditions shared by all helping processes and, indeed, all human relations. He has also created valid indices for assessing an individual's capacities for high or low level functioning as a helper (Carkhuff, 1969; Carkhuff and Berenson, 1967; Carkhuff, Kratochvil and Friel, 1968). The conditions deemed necessary for effective helper/helpee relationships include high levels of empathy, respect, concreteness of expression, genuineness, confrontation, self-disclosure, and immediacy of rela- tionship. Empathy refers to the helper's ability to understand sensitively and accurately the helpee's inner experiences. The helper should understand both the affect and content of the helpee's message, even to the point of knowing what has been left out, yet implied by the helpee's message. Empathy is the key ingredient to effective helping (Carkhuff, 1969; Rogers, 1967). Respect is simi- lar to what Rogers (1967) calls "unconditional positive regard," although Carkhuff emphasizes the positive rather than the uncondi- tional. Warmth is considered a vehicle for communicating respect. Respect refers specifically to the helper's view of the helpee's worth as a person and his ability to resolve his problems. Helper emphasis on concreteness of expression facilitates the helpee in discussing all personally relevant feelings and experiences in specific and concrete terms rather than discussing his problems in an abstract or intellectual manner. Facilitative genuineness refers to the helper's being completely himself in his relationship to the helpee. The helper's feelings and behaviors are completely congruent, both positive and negative; yet his relationship with 39 the helpee is non-exploitative and non-destructive. Facilitative self-disclosure involves the helper volunteering intimate and detailed information concerning his own personality. However, self- disclosure must be appropriate. It must be in keeping with the needs of the helpee. Confrontation involves the helper's experiences of discrepancies and inconsistencies in the helpee's behavior, which the helper perceives and communicates to the helpee. Confrontation may take place in the area of the helpee's weakness or his strengths. High level functioning helpers tend to confront the helper's strengths more often than his weaknesses (Berenson and Mitchell, 1969; Carkhuff and Berenson, 1967). Immediacy of relationship refers to the intense, spontaneous quality of a here-and-now, per- sonal and intimate relationship. The immediate relationship between the two is the focus of attention. The helper acts upon what he sees going on between himself and the helpee. Self-exploration (SE) refers to the quality of participant self-disclosure. SE is another scale created by Carkhuff (1969, Vol. II) to measure helpee self- exploration. Level of self-exploration should be positively cor- related to helper effectiveness. Facilitative and Action- Oriented Conditions Carkhuff (1969) emphasizes that effective helping processes involve the discrimination and communication of facilitative and action-oriented conditions. Facilitative conditions elicit helpee Self-Exploration and include Empathy, Respect, and Concreteness. Action-oriented conditions are Genuineness, Self-Disclosure, 4o Confrontation, and Immediacy. Once the helpee has explored his inner self he experiences a need to come out of himself and interact with the world in a constructive fashion. Sensitive discrimination allows the helper to (a) discern the helpee's level of functioning and dysfunctioning, and (b) make prescriptions and prognoses con- cerning alternative modes of behavior. Effective communication by the helper allows him to express his discriminations to the helpee. One of the goals of treatment is to increase the helpee's discrimina- tion of his internal and external environment and to increase his ability to communicate with himself and the world. Based on this theoretical and empirical perspective, two sets of seven scales each were created to measure a potential helper's ability to discriminate and communicate each of the seven core conditions. One scale was created to measure helper Self- Exploration. Each scale has five levels, with one representing the lowest level of functioning and five representing the highest level of functioning. A rating of three indicates that the helper is capable of communicating the minimally acceptable facilitative con- ditions for effective helping. The purpose of these two indices was to assess the present level of functioning of a trainee in any of the helping professions. These selection indices were based on the assumption that the best indicator of a future level of func- tioning as a helper is an indicator of his present level of func- tioning as a helper. 41 Index of Accurate Discrimination To assess discrimination the subject is requested to listen to tapes (audio, video, or even typewritten manuscripts) that are excerpts from actual counseling sessions. After listening to a taped helpee expression, the subject hears four different helper responses and is asked to rate them from most to least facilitative. In the standardized discrimination index there are 16 excerpts. All helper responses have been previously rated by experts on a scale of one to five according to whether the response was indicative of low or high levels of the core conditions necessary to effective help- ing. llmasubject's present level of functioning can be ascertained by comparing his answers to those of professionals (Carkhuff, 1969; Carkhuff and Berenson, 1967; Carkhuff, Kratochvil, and Friel, 1968). Index of Accurate Communication (IAC) Communication is assessed by having the subject listen to the same 16 excerpts and give his own response to the helpee expres- sion. Communication is assessed first, before discrimination, to avoid biased responses from viewing four possible responses. The subject's responses are then rated on level of communication of the core conditions by raters who are practiced and expert in the art of discrimination. Again, the rating is made on a l to 5 scale with 3 a rating that indicates the subject is communicating the core condi- tions at a minimally acceptable level. The SE scale is similarly scored except the target person is the helper. An alternative method to measure communication is to cast prospective helpers in the helping role in a role-playing exercise using "standard helpees," trained for 42 their role, and trained raters to rate the helper on the core condi- tions. This method is preferred, but both should give valid results. Validity Research studies undertaken to assess the validity of these two scales have proved the scale for assessment of accurate communica- tion to be "the most valid standard index for selecting persons equipped to function in the helping role" (Carkhuff, 1969). The Index of Accu- rate Communication, in both forms, has been used successfully to assess training in rehabilitative counselor education (Anthony and Carkhuff, 1969), guidance counseling (Martin and Carkhuff, 1968), clinical psychology training (Carkhuff, Kratochvil, and Friel, 1968), nurses (Kratochvil, 1969) and lay volunteers (Pierce, Carkhuff, and Berenson, 1967). These results are consistent with Carkhuff's model that suggests the best students for graduate training in the helping professions, as judged by experts in the different fields tested, are those who obtain high scores on the Index of Accurate Communication (IAC) prior to entering graduate training. Results from the present study show the IAC ratings of IPL leaders taken during an actual group setting provided adequate convergent and divergent validity. This was ascertained using Campbell and Fiske's (1939) multitrait- multimethod matrix to test IACOBS and LBQOBS. This is discussed fur- ther in the results. Studies relating discrimination assessment to success in the helping role have been less fruitful. With low level functioning helpers, discrimination is as likely to be high as it is to be low, while communication is consistently low. With high levelfunctioning 43 helpers, both discrimination and communication are high. The ability to discriminate accurately is necessary for interpersonal sensitivity but not adequate alone to enable effective communication. Accurate discriminations are also essential to the rating procedures used in ascertaining high or low level functioning on the indices of discrimination and communication. The research of Carkhuff and his associates into the core conditions of interpersonal sensitivity is the theoretical and empirical base from which IPL group leader behavior will be explored in this present study. Carkhuff (1969) states that the core condi- tions are the bases for all human relations, as well as those for fulfilling the helping role. The present research proceeds on the assumption that IPL group leaders are helpers and that high levels of these conditions must be communicated by the IPL group leader if -he is to be successful in his role. Other variables not explored by Carkhuff and his associates, yet acknowledged as important by other researchers, are also explored. These other variables include trust, risk taking, non-verbal behavior, group climate, and group Structure. Hypotheses concerning interpersonal sensitivity will be examined in detail in the next chapter, and hypotheses concerning group structure and other variables not included in the Carkhuffian model will be examined. CHAPTER II FORMULATION OF HYPOTHESES The basic hypothesis of this research is that group members who rate themselves as highly satisfied with their IPL group experi- ence will also give their group leaders high ratings on the 12 predictor variables. Observer ratings of member benefit will also correlate highly with observer ratings of leader effectiveness on the 12 predictor variables. The predictor variables are Empathy, Respect, Genuineness, Self-Disclosure, Concreteness, Immediacy, Confrontation, Risk Taking, Non-verbal Behavior, Group Structure, Group Climate, and Trust. The criterion is a measure of participant benefit that was completed by participants, observers, and close ‘friends of the participants. The 12 predictor variables are dis- cussed in depth in the next two sections. Group Leader Interpersonal Sensitivity and Group Member Benefit Empathy and Respect The IPL group leader is the behavioral and attitudinal model for the group members. It is his responsibility to exhibit those behaviors deemed necessary for maximum participant growth and satisfaction. It is also necessary that the group members be able to discriminate these behaviors in order to include them in their (ww1behavioral repertoire. It is hypothesized here that groupleaders 44 45 must exhibit high levels of the core conditions necessary for any effective helping relationship. These include high levels of empathy, respect, self-disclosure, genuineness, confrontation, concreteness, and immediacy. The most important variables here should be empathy and respect. Carkhuff (1969) and Rogers (1967) conclude from compre- hensive theoretical and empirical studies that empathy is the key ingredient in any helping relationship. It is imperative that the IPL group leader be perceived as highly empathic to each individual member of the IPL group. Empathy refers to the helper's ability to understand sensitively and accurately the helpee's inner experience and to communicate this understanding to the helpee. Rogers (1967) suggests that it is necessary for the therapist to have an "uncondi- tional positive regard" for his client. Carkhuff (1969) feels that high positive respect from the helper to the helpee is critical to an effective helping relationship. Respect is here defined as warmth or caring. It is essential that the IPL group leader be perceived by the individual group members as warm and caring about them personally. These two hypotheses are consistent with much other encounter group theory and research (Egan, 1970). These hypotheses are also consistent with the theory and research in the field of PDM which emphasizes supportive and con- siderate leadership on the part of management. The Ohio State Leadership studies (Hemphill, 1950) also conclude that Consideration, which includes all aspects of interpersonal sensitivity, is one of the two most important fundamentals for effective leadership. The other ingredient for effective leadership is Initiating Structure. 46 Gibb (1969) suggested that Consideration is thg_most crucial leader- ship variable since leaders low in Consideration are never effective whether they are high or low on Initiating Structure. Stouffer et a1. (1949) suggest that subordinate soldiers preferred immediate superiors who were high on Consideration. Self-Disclosure Self-disclosure involves the helper volunteering appropriate intimate information concerning his own personality. Numerous researchers have confirmed that self-disclosure on the part of the helper and helpee is important in group therapy, individual therapy, psychological research, and encounter groups (Carkhuff, 1969; Culbert, 1968; Egan, 1970; Hurley and Force, 1973; Jourard, 1964, 1969; Jourard and Friedman, 1970; Mowrer, 1964; Rogers, 1961, 1967; Truax and Cark- huff, 1965). Rogers (1961, 1967) emphasized the importance of full client self-disclosure to successful therapy. Later research (Cark- huff, 1969; Culbert, 1968; Hurley and Force, 1973; Jourard, 1969) emphasized the importance of therapist or leader self-disclosure to positive client or helpee outcomes. Carkhuff (1969) has cautioned that helper self-disclosure must be appropriate to the needs of the helpee. Culbert (1968) feels that trainer self-disclosure is more appropriate in the beginning rather than in the latter part of the group life. In keeping with Jourard's (1964) curvilinear theory of self-disclosure, it is hypothesized here that both very high and very low levels of self-disclosure on the part ofthe IPL group leader should , be unsatisfactory to the participants. Too much or too little dis- closure leaves little room for an empathic relationship with other 47 group members. IPL group leaders should be perceived by group members as self-disclosing, but not overly so. The leader should be discriminating in his self-disclosure and keep it appro- priate to the needs of the group members. He should also actively encourage self-disclosure on the part of the group members. The actual self-exploration of group members is scored on the SE scale. SE should be positively correlated with leader self-disclosure. Genuineness Encounter group leader genuineness is related to all other variables examined here (Carkhuff, 1969). It is hypothesized that the group leader must be perceived as highly genuine by the group members. The leader must be completely himself in all that he does. His empathy, respect, and self-disclosure, all that he says or does in the group must be perceived as real by the group members. He cannot use a facade either professional or personal. If he embraces a weeping group member it must be because he genuinely cares for this person, not because he knows he should from reading encounter group literature. Although admittedly difficult to measure, genuineness is often operationally defined in terms of con- gruence or consistency (Bolman, 1973; Egan, 1970; Rogers, 1961). Bolman (1973) suggests that trainer congruence/empathy is the key factor in successful encounter group leadership. The leader must be perceived by the members as congruent or consistent in his simul- taneous verbal and non-verbal expressions, as well as consistent across time in his beliefs and actions. The leader must be himself, always owning up to both his strengths and his weaknesses (Egan, 48 1970; Rogers, 1967; Stimson, 1970). Being genuine is the opposite of "faking it." Confrontation Confrontation refers to the leader confronting the group members with discrepancies and inconsistencies in their verbal and non-verbal expressions, or inconsistencies in the members' behavior across time. If the leader perceives discrepancies in a member's feelings and actions toward another (for example, at one time he is warm and accepting of another on the verbal level, yet sits with his back partially toward the other person and refuses to look him in the eyes), the leader will communicate the discrepancy to the member and suggest the member work on awareness, understanding, and communi- cating his different feelings toward the other group member. If a member claims he is deeply and positively involved in the group Texperience, yet sits outside the group, the leader confronts the discrepancies in this behavior. Berenson and his associates (1968) have shown confrontation to be an effective therapist tool. Cark- huff and Berenson (1967), Berenson and Mitchell (1968), and Carkhuff (1969) suggest that empathy, respect, and genuineness are not incon- sistent with confrontation. The leader can confront specific nega- tive behaviors without rejecting the whole person. Berenson and Mitchell (1968) suggest that better helpers confront helpee's strengths more than his weaknesses. It is hypothesized that the~leader should be perceived as confronting, yet not constantly confronting. 49 Confrontation in this context seems logically similar to Ini- tiating Structure and production emphasis as described by Hemphill (1950) in relation to business leadership. The task of the encounter group is personal growth. By confronting member inconsistencies, the encounter group leader maintains standards of performance, assigns particular tasks to members, motivates group members, and emphasizes the job to be done. Concreteness and Immediacy Concreteness of expression and immediacy of relationship are also part of leader-initiated structures. Helper concreteness enables the helpee to discuss all personal feelings and experiences in concrete rather than abstract or intellectual terms. Immediacy refers to the intense, spontaneous quality of a here-and-now, personal and intimate relationship. It is hypothesized that the leader should also be per- . ceived by his group members to be high on concreteness of expression, not only for himself, but for all group members. Concreteness is also related to immediacy of expression and it is hypothesized the leader must be perceived as both concrete and immediate in his self-disclosure and feedback to other members. It is imperative for a successful IPL group to express and examine their current (here-and-now) feelings about themselves, each other, and the group process, rather than to engage in superficial conversation or intellectual abstractions about people and feeling in general. Egan (1970) explains the difference between "history" and "story." Story is the mode of involvement; history, of uninvolvement. Self-disclosure that is mere history is pseudo-self-disclosure. It is 50 often a long, analytical account complete with facts and interpreta- tions, but leaves both speaker and listener untouched. History does not unite listener and speaker; rather, the words become an obstacle between them. Story is authentic self-disclosure that invites involve- ment with others. The speaker drops his defenses and stands somewhat naked in his own eyes and in the yes of others. Story is emotional. Risk Taking Although not included in the Carkhuffian paradigm (1969), risk taking, trust, and non-verbal behavior are considered here as important dimensions for establishing effective interpersonal rela— tionships. Egan (1970) states that risk is an essential feature of encounter groups. Most people hesitate to disclose themselves in a group. The possibility of rejection following self-disclosure is ever present. Engagement in life and the activity of self-disclosure -.has always involved risks. Therefore, it is essential that the leader act as a model for the other members by taking risks of self- disclosure and intimacy with the other members. It is hypothesized here that the effective IPL group leader is perceived as a high risk taker by the group members. IEEEE Initial risk demands a climate of trust. Gibb (in Bradford, Gibb, and Benne, 1964) and Schein and Bennis (1965) refer to the need for a climate of "psychological safety" in the encounter group. Each individual must feel that it is safe to expose his feelings, drop his defenses, and try out new behaviors. He can only do this 51 if he trusts the group and the leader to be supportive, non- evaluative, and non-rejecting of his "real" person. As the model for the other group members, it is especially important that the leader be perceived as trustworthy. He must take active steps to implement a climate of psychological safety. There is also a body of social psychological research relating trustworthiness to suc- cessful persuasion attempts (Hovland and Wiess, 1951). It is here hypothesized that the IPL group leader must be perceived by his group members as highly trustworthy before a successful experi- ence can be completed. Non-Verbal Behavior Non-verbal behavior is often an important dimension of encounter group behavior (Egan, 1970; Rogers, 1970). Many research- ers have demonstrated the importance of non-verbal communication to the entire communication process (Birdwhistle, 1952; Hall, 1964, 1966; LaBarre, 1964). Hasse and Tepper (1972) showed that non- verbal components such as eye contact, trunk lean, body orientation, and distance accounted for twice as much of the variance in empathy ratings as did verbal components. Bader (1972) found that praises, touches, and special techniques were the categories of behavior that separated effective group therapists from non-effective therapists. Non-verbal behavior is present in the encounter group both adver- tently and inadvertently. The most controversial and perhaps most important is non-verbal communication with bodily contact (Egan, 1970). This behavior includes touching, holding, kissing, hugging, 52 pushing, swinging, playing, wrestling, etc. Bodily contact may occur spontaneously or it may be the result of some non-verbal exercise suggested by the leader. For example, after a particularly deep self-disclosure a participant begins to cry and another puts his/her arms around him/her to give support. The leader may intro- duce an exercise such as "backtalk" (Gunther, 1968) in which parti- cipants communicate by touching backs, to stimulate contact and communication in non-conventional, non-role-related ways. Non- verbal exercises can be very helpful in increasing intimate group communication if the exercise fits into the here-and-now activities of the group. Exercises can also be used mechanically by the leader to substitute for the give-and-take of verbal interaction. This may merely reflect the anxiety of the leader over dragging group processes (Egan, 1970). Spontaneous non-verbal communication via bodily contact is often an indicator of a highly cohesive, highly productive encounter group experience (Egan, 1970; Rogers, 1970). It is here hypothesized that spontaneous non-verbal behavior on the part of the leader, especially touching, is necessary for effective IPL group processes. It is also hypothesized that the use of some leader-structured non-verbal exercises increase group member satisfaction and personal growth. Group Structure and Group Climate in Relationship to Partipipant Benefit The classic study of the effects of various leadership structures on groups was undertaken by White and Lippit (1943). Adult leaders were trained to lead children's clubs in either an 53 authoritarian, democratic, or laissez-faire manner. In the authori- tarian group the leader determined all policy and dictated activity steps one at a time, as well as work group membership. All demo- cratic policies were discussed and decided by the group with the active participation of the leader giving suggestions and encourage- ment. Members chose their own work groups. The leader tried to be a regular group member. In the laissez-faire condition the leader participated minimally. There was complete freedom for the group. Results showed the laissez-faire group accomplished very little. Members were very bored and often broke into horseplay. The autocratic group was more productive. Indeed, when the leader was present the autocratic group was the most productive of the three groups. However, if the leader was absent there was more hostility expressed between group members than in the other groups, and work stopped. There was less hostility expressed in the democratic group than in any of the others whether the leader was present or not. Democratic participants were more satisfied with their experi- ence and with their leader than were any of the other groups. Later research on group member reaction to different group structures or climates has been summarized elsewhere in the section on participa- tive decision-making. PDM, which is similar to democratic decision- making, invariably produces greater member satisfaction. Cohesion is another important dimension relating to group structure or group climate. Cartwright and Zander (1960) define cohesion as "the resultant of all forces acting on all members to remain in the group." Often cohesion and interpersonal attraction 54 are used synonymously (Collins and Raven, 1969). Theoretical and empirical data indicate that there is a positive relationship between cohesion, communication, and conformity (Hollander, 1971; Homans, 1950). Highly cohesive groups are highly satisfied groups. Thibaut and Kelley (1959)theorize that an individual remains in a group because the outcomes he receives from that group are greater than the outcomes he would receive elsewhere. The relationship of cohesion to productivity is equivocal. Bjerstedt (1969) has shown increased productivity in highly cohesive work groups. Ray (1952), Whyte (1955), and Marquis, Guetzkow, and Heyns (1951) have demonstrated no relationship, or a negative relationship, between cohesion and productivity. Seashore (1954) concluded that work group norms could favor either high or low group productivity depending on whether the group perceived company management as sup- portive or non-supportive. Runyan (1974) found that high cohesive groups were significantly more risky in making group decisions than low cohesive groups. Group_Structure and Group Climate It is hypothesized here that IPL groups characterized by democratic, non-authoritarian group structure and a highly cohesive group climate will produce more positive outcomes in terms of learning and satisfaction, as perceived by the group members. The importance of democratic leadership and high group cohesion to satisfaction has been demonstrated elsewhere, and should generalize to this situation. High learning is here analogous to high 55 productivity. Collins and Raven (1969) state that the causal relationship between interpersonal attraction (cohesion) and commu- nication are among the best established in social psychology. Carkhuff and Berenson (1967) and Carkhuff (1969) have demonstrated that open communication is the most important factor in producing positive outcomes in the helping relationship. The importance of open communication, cohesion, and democratic leadership to effec- tive encounter group outcomes has been theoretically stated by many authors (Bradford, Gibb, and Benne, 1964; Egan, 1970; Rogers, 1970). These hypotheses are also consistent with Caron's (1964) finding that following an encounter group experience employees placed more value on consideration and less on initiating structure in comparison to a pre—test on the Leader Opinion Questionnaire (Fleishman, Harris, and Burt, 1955). CHAPTER III METHOD, INSTRUMENTS, SUBJECT POPULATION, AND OPERATIONAL DEFINITIONS Sample POpulation The sample population for the present study consisted of undergraduate education majors at Michigan State University. Edu- cation 200, the first required course in MSU's teacher education program, is the initial exposure of future teachers to the idea that interpersonal communication skills are necessary for effective teach— ing. Education 200 defines experiences that deal with self-growth as the personal demands of teaching and the experiences of helping others grow as the task demands of teaching. Task demands are taught through reading materials, individualized carrel activities, lectures, and tutorial settings; and the personal demands are taught through Interpersonal Process Laboratories (IPL). These small groups of about 15 students and one leader, typically an education graduate student, meet twice weekly and explore the demands placed on teachers with respect to the interpersonal skills they need to communicate effectively. IPL groups are partially lecture, par- tially leader-initiated experiential exercises dealing with commu- nication skills, and partially unstructured group process using the various communication skills for self-exploration and giving and receiving responsible interpersonal feedback. Eight different 56 57 group leaders, teaching 15 different IPL groups volunteered their groups as subjects for this study. It was decided to use only group leaders who taught two different IPL groups since this would increase the N_of the sample population and minimize the effect of an unusual group on the scores for any particular leader. At the start of one of the groups, the members decided not to participate in the research, resulting in 15 groups instead of 16. There were 36 IPL groups altogether, but only group leaders who volunteered to be studied, and who led two groups, were actually included. Some leaders refused to be part of the study. They stated that their group members did not want an observer present in the group. They believed this would be disrup- tive. They did not want to take group time to fill out the question- naires. Seven of the group leaders were male and one was female. Four of the five observers were male (see page 69). Participants of both sexes were represented in approximately equal numbers. Instruments The fundamental hypothesis of this research is that group mem- ber benefit from the small group experience designed to teach inter- personal communication skills is highly correlated to effective leader interpersonal behavior. Participant's benefit and leader effective- ness are measured by questionnaires (GMBQPAR and LBQPAR) specifically designed for this study. The first is designed to assess each parti- cipant's perception of his own learning and satisfaction. The second questions each group member about his own relationship to the group leader. 58 Several independent assessments of leader level of functioning were completed to temper possible bias effects frdm participant reports of leader behavior. Observers rated participant benefit in the group (GMBQOBS). Near the completion of the group, friends nominated by the participants assessed participant benefit in the group (GMBQFR). An established personality test, the Interpersonal Checklist (ICL), was completed by each group member at the beginning and end of the group to check on possible participant change. The written version of the Index of Accurate Communication (IAC) was completed by all group lead- ers before the group started (PREIAC) to ascertain initial level of leader functioning. The LBQOBS was completed by an observer following each observation to measure leader effectiveness. The observer also measured leader effectiveness after each observation according to the IAC (IACOBS). For the IACOBS observers rated the actual in-group per- formance of each leader. Participant self-exploration (SE) was rated by the observer after each observation for each participant. SE was expected to relate positively to both leader effectiveness and partici- pant benefit. The instruments created specifically for this study include (a) Leader Behavior Questionnaire (Participant and Observer) (see Appendix A.3), and (b) Group Member Benefit Questionnaire (Partici- pant, Observer, Friend) (See Appendices A.4, A.l, and A.2). Leader Behavior Questionnaire The LBQ is a 60-item, 7-point, Likert-type questionnaire designed to assess perceptions of leader interpersonal sensitivity. Items are designed to assess Empathy, Respect, Genuineness, 59 Self-Disclosure, Immediacy, Concreteness, Confrontation, Risk Taking, Trust, Non-verbal Behavior, Group Structure, and Group Climate. The LBQ has five items to measure each of the variables considered. The author created all the items on the LBQPAR and LBQOBS. The scoring groups were created according to the author's judgment (Appendix B). The definition for seven of the variables were taken from Carkhuff (1969). Definitions for the Risk Taking, Trust, Non-verbal Behavior, Group Climate and Group Structure were taken from the appropriate research cited in this study. All questions have equal weight, so to obtain an Empathy score, for example, the score on all questions mea- suring Empathy would be added together (see Appendix A.3). Some items on the LBQ were phrased negatively. For scoring purposes, the data from these items were reversed. Group Structure was scored so that high scores indicated a perception of low Group Structure. Group Climate was scored so that high Group Climate indicated a perception of a cohesive Group Climate. LBQOBS is identical to LBQPAR. Group Member Benefit Questionnaire The GMBQPAR is a l4-item, 7-point, Likert-type questionnaire designed to assess participant benefit. Items are designed to assess participant satisfaction and learning. GMBQOBS is nearly identical to the GMBQPAR: GMBQOBS has only 12 questions. Questions numbered (63), (65), and (69) on the GMBQPAR do not appear on the GMBQOBS. Question (6) on the GMBQOBS does not appear on the GMBQPAR. Item (75) on the GMBQPAR asked each participant to mark a "1" if he passed the course and "2" if he received no credit from the course. Item (75) is not part of the GMBQPAR but is used as the item that measures the variable 60 Grade which will be discussed later. The GMBQFR is identical to GMBQPAR (Appendix A.2). Near the end of the group, each participant was wasked to give the GMBQFR to someone "who knows you well and would know of your IPL group experience.“ Index of Accurate Communication Five non-participant observers with prior experience leading and observing groups teaching interpersonal communication skills were trained as raters on the IAC. Two observers gave each leader a sin- gle overall rating of effectiveness on the written PREIAC rather than a score for each of the seven variables. Each group was visited five times throughout the ten-week term by at least three different observers. The IACOBS, the LBQOBS, and the GMBQOBS were each com- pleted following each observation. For the IACOBS the observers rated the leaders on each of the seven variables. Observers also rated the participants on self-exploration (SE). Interpersonal Checklist The Interpersonal Checklist (ICL) was used as still another independent data source. The ICL is a 134-item, true or false, self-descriptive personality inventory. The ICL was completed by all participants at the beginning and at the end of the group. Raw ICL scores were transferred to a circumplex of 16 different behavioral categories according to a theory of interpersonal behavior described by Leary (1957). The two most similar categories were combined together to form octants labeled, in clockwise order around a circle, (a) Managerial-Autocratic, (b) Responsible-Hypernormal, (c) Cooperative-Overconventional, (d) Docile-Dependent, 1" (’llflII I (It 1" (47") . illlllll‘. '- .II} 61 (e) Self-Effacing-Masochistic, (f) Rebellious-Distrustful, (g) Aggressive-Sadistic, and (h) Competitive-Narcissistic behavior. These octant categories can be reduced to two orthogonal bipolar dimensions: (a) Dominance-Submission (DOM) and (b) Love- Hostility (LOV). These two dimensions lie on the vertical and horizontal axes of the circle, respectively. The degree of relationship between any two variables is a decreasing function of their separation on the perimeter of the circle. The closer the variables are on the perimeter of the circle, the more highly they should be correlated. A varying degree or intensity of any one of the 16 variables can be represented by the distance from which it is scored along the radius from the center of the circle. Traits represented near the center of the circle are considered to be unusually intense. A concise scoring can be made by summarizing all scores in terms of the two major axes: Dominance- Submission (DOM) and Love-Hate (LOV). Four different forms of the ICL have been devised and tested by LaForge and his associates (LaForge and Suczek, 1955). The reli- ability and validity of the ICL has also been established in factor analytic studies (Lange, 1970; LaForge, 1963; Foa, 1961). Lange (1970) reports that his study "supports the assumption that two bipolar dimensions (Dominance-Submission, DOM, and Love-Hate, LOV) underlie the ICL and the original formulation of LaForge and Suczek (1955) concerning the interpersonal variables taken to be their measure are correct." Lange (1970) also suggests that "the contents and weights assumed by LaForge and Suczek do not seem essential" to 62 the computation of LOV and DOM and may be inappropriate in view of the factor loadings found. In the present study, all items on the ICL are unweighted. Computation of the Love and Dominance scores were made using the following formulas given by LaForge and Suczek (1955) using the 16 original categories: LOV=M-E+.9(M+L-D-F)+.7(O+K-C-G) + .4 (P + J - B - H) DOM = A - I + .9 (B + P - H - J) + .7 (C + O - G - K) +.4(D+N-F L) Many other independent researchers have come to very similar conclusions concerning the two major bipolar dimensions used to describe interpersonal behavior. Hurley and his associates (1972, 1973) label the two dimensions Acceptance/Rejection of Self (DOM) and Acceptance/ Rejection of Others (LOV). Hurley and Force (1973) use the ICL as well as their own instruments to assess gains in interpersonal compe- tence following an encounter group. Interpersonal competence was de- fined as Self-Acceptance x Other Acceptance. Participant gains were correlated highly with within-group effectiveness ratings of their encounter group trainers. Other researchers have labeled these dimen- sions (a) Solidarity and Status (Brown, 1965) and (D) I'm (Not) OK-- Your're (Not) OK (Berne, 1965). Dominance and Love are also logically similar to Initiating Structure and Consideration. These last two di- mensions are those designated by the Ohio State Leadership studies (Hemphill , 1950) as the two major factors of effective industrial leader- ship. 63 For the purposes of the present study, positive interper- sonal behavior is hypothesized to fall in the high LOV/DOM quadrant of the Interpersonal Behavior Circle as measured by the ICL. Parti- cipants who claim personal gain and satisfaction from the IPL group experience should have high scores in the LOV/DOM quadrant of the ICL and also should rate their leaders as high on the LBQ. A check on participant gains can also be made in a pre/post comparison. Operational Definitions The purpose of this research is to study leadership effective- ness in small groups designed to teach interpersonal communication skills. The primary hypothesis is that participants who express strong feelings of personal benefit from their IPL group will perceive their leader as a significantly more positive model of interpersonal communication skills than participants who perceived less personal benefit from their IPL group. The important communication skills the leader must model to a high degree are Empathy, Respect, Self-Disclo- sure, Genuineness, Confrontation, Immediacy, Concreteness, Risk Tak- ing, Trust, and Non-verbal Behavior. The leader must also help create a democratic Group Structure, and a supportive, cohesive Group Cli- mate. These are the twelve predictor variables. The criterion vari- able, particpant benefit, is defined as satisfaction with the group experience and learning the interpersonal communication skills. Instrument Key LBQ is the Leader Behavior Questionnaire (Participant, Obser- ver). It is designed to measure the 12 predictor variables: Empathy, 64 Respect, Genuineness, Self-Disclosure, Concreteness, Immediacy, Con- frontation, Risk Taking, Trust, Non-verbal Behavior, Group Climate and Group Structure. GMBQ is the Group Member Benefit Questionnaire (Participant, Observer, Friend). It is designed to measure group member learning and satisfaction. ICL is the Interpersonal Checklist (pre- and post-group, par- ticipant rates self). It is a standardized personality inventory used to ascertain participant benefit (LaForge, 1955, 1974). IAC is the Index of Accurate Communication (pre-group and in- group observer ratings of the leader). The written version was admin- istered pre-group to group leaders (Carkhuff, 1969). First, observers gave each leader a single, overall rating of effective leadership. Second, in-group observer ratings were made on actual leader behavior in the group. Ratings were based on the seven scales designed to mea- sure the first seven predictor variables. The seven variables measured by the IACOBS are Empathy, Respect, Genuineness, Self-Disclosure, Concreteness, Immediacy, and Confrontation. SE is Self-Exploration (Observer), a standardized scale of participant self-exploration (Carkhuff, 1969). Hypotheses The criterion variable of participant benefit will be first operationally defined as high or low scores on the GMBQPAR, and sec- ond, on the GMBQOBS. The hypotheses are identical. Using correla- tions obtained from independent measures of leadership effectiveness and member benefits, i.e., observer measures of leader effectiveness 65 (LBQOBS) and participant measures of benefits, the same high correla- tions should be found between benefits and the twelve predictor vari- ables of leader effectiveness. The most crucial predictor variables are hypothesized to be Empathy, Respect, Self-Disclosure, Risk Taking, and establishing a supportive, cohesive Group Climate. The findings on Group Structure are those least likely to support the porposed hypotheses. Authori- tarian or democratic Group Structure could both lead to high partici- pant benefit if the leader his high on Empathy and Respect. A leader- ship style perceived as laissez-faire will not lead to hufliparticipant satisfaction. The first seven variables are measured by the LBQPAR, LBQOBS, and IACOBS. The last five variables are measured only by the LBQPAR and LBQOBS. Ratings of leaders' level of Empathy, Respect, Genuineness, Self-Disclosure, Immediacy, Concreteness, and Confrontation: 1. from participant's assessments (LBQPAR) will correlate positively with all measures of participant's benefit (GMBQPAR, OBS, FR). 2. from observer's assessments (LBQOBS) will correlate positively with all measures of participant's benefit (GMBQPAR, OBS, FR). 3. in group IAC assessments (IACOBS) will correlate posi- tively with all measures of participant's benefit (GMBQPAR, OBS, FR). 4. from all measures of leader effectiveness (LBQPAR, OBS, IACOBS) will correlate positively with participant's ICL gains. Ratings of leader's level of Risk Taking, Trust, Non-verbal Behav1or, and capacity to create a cohesive Group Climate, and Group Structure: 1. from participant's assessments (LBQPAR) will correlate p051tively with all measures of participant's benefit (GMBQPAR, OBS. FR). 66 from observer's assessments (LBQOBS) will correlate positively with all measures of participant's benefit (GMBQPAR, OBS, FR). from all measures of leader's effectiveness (LBQPAR, 085) will correlate positively with participant's ICL gains. Ratings of participant' 5 level of Self- -Exploration from observer' 5 assessments (SE): 1. will correlate positively with all measures of parti- cipant's benefit (GMBQPAR, OBS, FR). 2. will correlate positively with observer's ratings of leader Self-Disclosure (LBQOBS, IACOBS) and partici- pant's ratings of leader's Self-Disclosure (LBQPAR). 3. will correlate positively with all measurements of leader effectiveness (LBQPAR, OBS, IACOBS). 4. will correlate positively with participants' ICL gains. Group leader PREIAC scores will be positively correlated with: 1. all measures of participant's benefit (GMBQPAR, OBS, FR . 2. all measures of leader's effectiveness (LBQPAR, OBS, IACOBS). 3. participant's gains on the ICL. A participant's grade in the course (pass/no credit) will correlate positively with: 1. all measures of participant's benefit (GMBQPAR, 085, FR). 2. all measures of leader's effectiveness (LBQPAR, OBS, IACOBS). 3. participant's gains on the ICL. All measures of leader's effectiveness will be positively correlated: l. as measured by the (LBQOBS, LBQPAR, IACOBS, PREIAC). 67 All measures of participant's benefit and learning will be positively correlated: l. as measured by (GMBQPAR, 085, FR, ICL gains). Finally, several stepwise multiple regression analyses on all leader effectiveness data (LBQPAR, LBQOBS, IACOBS), using the three different criterion measures (GMBQPAR, OBS, FR), were used to ascer- tain the variables accounting for the most variance on the criterion measures. The variables with the most predictive power can be ascer- tained and the effect of any high intercorrelation among the predic- tor variables can be accounted for. CHAPTER IV RESULTS Introduction Pearson product-moment correlations were used to ascertainthe relationships between the predictor and criterion variables. One-tailed significance tests were used. Multiple regression analyses were com- pleted to determine the best subset of predictor variables. Analyses of variance were used to examine the homogeneity of observers' scores on the same leaders and the same group members from a film during the last training session. The reliability and validity of the predictor and criterion measures were examined using McQuitty's (1957) typal analysis, Campbell and Fiske's (1959) multitrait-multimethod matrix, and Cronbach's (1951) alpha. Analysis of variance was used to ascer- tain variance due to main effects of groups nested in leaders. Deci- mal points are omitted in all correlation tables. Predictor Variables The 12 predictor variables of leader effectivenss are Empathy, Respect, Genuineness, Self-Disclosure, Concreteness, Immediacy, Con- frontation, Risk Taking, Non-verbal Behavior, Group Climate, Group Structure, and Trust. The predictor variables were measured by the participants and the observers, respectively. The IACOBS measuresthe first seven predictor variables: The PREIAC is a global measure of leader functioning taken before the group. The IAC measures were 68 69 rated by the observers. Participant Self-Exploration (SE) was also measured by the observers. Criterion Variables Participant benefit during the group is the criterion vari- able. The GMBQPAR, GMBQOBS, and GMBQFR are nearly identical measures of participant benefit completed by the participants, the observers, and a friend of each participant from outside the group. The Inter- personal Checklist (ICL) was completed pre- and post-grouptn/the par- ticipants to check for gains or losses on the LOV and DOM factors (LaForge and Suczek, 1955). Reliability and Validity Interrater Reliability The five observers were initially selected because of their previous experience as group leaders and observers, and their willing- ness to make the time commitment to be trained and then to observe the ten-week-long groups. Three observers were undergraduates and two were graduates. The observers used role-playing techniques to famil- iarize themselves with the IAC, the LBQOBS, and the GMBQOBS. In the final training session the observers viewed a film of a T-group in process involving two leaders and nine group members. Their ratings of the nine participants and the two leaders in this film served as the basis for the interrater reliability analyses. Two analyses of variance for within subjects were conducted to examine interrater reliability for the observers;-one for the LBQOBS and one for the GMBQOBS (Appendices C.1 and C.2). Each obser- ver rated each of the nine participants on the total GMBQOBS and the 70 two leaders on the 12 variables of the LBQOBS. The purpose of the ANOVAs is to examine the variance due to the ratings given different group members and leaders in comparison to the variance due to the five different observers' ratings of the same group member or leader. For the GMBQOBS, a one-factor within-subjects design, a significant main effect for total scores was obtained for group members nested in observers (E_= 3.71, p_< .Ol, gf_= 8,32). There is significantly more variance in the ratings given the nine different participants in comparison to the five ratings made for each individual parti- cipant. For the LBQOBS, a two-factor within-subjects design, a sig- nificant main effect for total scores was obtained for leaders nested in observers (f_= 2.38, p_< .03, gf_= 11,48). One observer did not provide complete data on the LBQOBS. The ratings made by the four observers showed significantly more variance for the two leaders than the ratings made by the four observers on the twelve variables for each leader individually. Predictor Variables Convergent and discriminant validity was examined for the three main leader effectiveness measures, the LBQPAR, LBQOBS, and the IACOBS, using Campbell and Fiske's (1959) multitrait-multimethod matrix (Table l). The validity diagonal values (monotrait-multimethod) for the seven predictor variables measured by the LBQOBS and the IACOBS reveal correlations between 3_= .53 and r_= .76. These substantial correlations offer evidence of convergent validity for the seven predictor measures common to the LBQOBS and the IACOBS. '71 .OOOOOO an nouumccou .OcomO.O xO.O..O>O OO OO OO Nm .. OO NO Om N. ON OO ,NN. OO OO OO ON OO Om m. m. O. O. OO ,mq,,.O OO O. OO- OO NO. LOOOO OO ON O. .O NO .. ON .O .N .O OO.,mO,,.O OO Om OO NO N. N. N. O.- m. OOV,OO,,OO OO- ON- NO- ..4 OOOE. O ON ON O. OO OO OO O. O. N. OO OO.,ON.,NN mO NN OO Nm .O NO N. .O O.- m.;.ON, O. N.- m.- O. OOOO O OO ON OO NO ON ON O. OO OO O. mN.,mN,,Om NN OO Om Om Om N. NO OO mO OO..NW. ON O. ON OO O OO Om OO OO ON .O- O. .O OO O. mm,,OO.,OO OO O. NN N. O. Om O. .O .O Ow,.mN-.Oo- O. OOO O NO mO NN OO OO NO- O. NO ON Nm OO..~O..NO Om O. OO ON mm O. ON OO NO Om.,Ow,,mN OOOO O OO .O .O .O O. NO OO NO NO OO OO...mO O. O. NO N. OO OO- OO OO O. O.- OO,,MN. OEO . O. OO OO .O OO OO OO OO OO ON ON ..N,,OO OO O. OO ON NN OO NO ON ON NO OOOL. OO- Om N. N. OO O. ON Om OO OO Om,.OO,,mN NN ON NN NN OO OO OO NO O. OOLOO .O- ..- Om NN OO O. mm N. .O OO Ow,,O.-.r.- OO- O.- OO- O. OO- OO- N.- O. a..O OO mO OO N. N. OO Om Om NO .. .O,,OO, NN O. O. .O Om mO OO OO O): OO NO O. OO .. N. Om Om N. OO OO,,-,.ON OO .O ON OO .. NO .O O OO NO Om Om ON .O Om OO NO Om NO,,OW,,ON NO OO- OO ON OO NOOOO O OO OO .O OO NO O. OO NO OO OO OO,.OO,.OO .O NO NO OO OOEE. O NN mm .O OO m.- N. O.- NO- Om OO- OOV,OO,,OO NO- ON- O. OOOO O .O mm OO .O NO .O OO OO Nm OO On,,ON-.OO .N NO OO O OO OO Om OO NO NO N. m. O. OO .w,,OO,.ON .O OOO O OO OO N. ON .O .N OO NO- NO mN mOv.O.,.O. OOOO O. O. NN NN Om OO- OO- OO ON OO- OO,.Ow, OOO .. OO OO OO OO OO OO OO .O ON NO OOOL. O. O. O. OO NO O. NN OO OO OO OOLOO OO OO ON OO NO .O OO OO Om e..O Om Om NO NO Om ON NO .O O): Om mO NO OO OO OO ON .O O OO OO .O ON .O Om LNOOO O .O OO OO OO NO OOEE. O OO OO NO NO OOOO O OO NO NN OO O OO NO _ OOO O . mm amom mmmmmmmmmmmmmwmmmmMmmmmwm Own a OOO Own a OOO u OOOOO. .. l .11... I! . TILI. 1.1.1“ c mmou<. .mmOOmO .m Ocec.s.LuO.uucOacmmcm>cou OOOEOxm ou x_.ume vocOwe_u.OE-OOONO_O.OE < p o.aa~ 72 Discriminant validity can be demonstrated by lower correlations be- tween different traits measured by a different method (multitrait- multimethod) in comparison to the validity diagonal values showing cor- relations between the same traits using different methods (monotrait- multimethod). The validity diagonal value for Genuineness in the LBQOBS and IACOBS matrix is [-= .54 and the correlation between Genu- ineness on the IACOBS and Respect on the LBQOBS is L: .65. This is the greatest deviation from adequate discriminant validity revealed by this method. Another technique to examine discriminant validity is to compare the correlations of two different measures of the same trait (monotrait-multimethod) with correlations between different traits using the same method (multitrait-monomethod). Only three of the seven predictor variables, Self-Disclosure, Concreteness, and Confrontation, offer good discriminant validity for the LBQOBS and the IACOBS using this method. The other variables are only margin- ally valid. A final indicator of discriminant validity is if the pattern of trait intercorrelations for the monomethod block is simi- lar to the pattern of trait intercorrelations in the heteromethod block. Inspection of the correlations for LBQOBS and IACOBS reveal a general consistency in trait interrelationships. In summary, for the LBQOBS and the IACOBS convergent validity seems satisfactory and divergent validity is less satisfactory due to the number of high intercorrelations between the seven predictor variables. Using the multitrait-multimethod matrix to examine the 12 LBQ predictor variables measured across data providers (participants and observers), very little convergent or discriminant validity can 73 be demonstrated. An examination ofthe validity diagonal values for the 12 predictor variables measured by the LBQPAR and the LBQOBS reveals correlations ranging from §_= .00 to [_ .75. Only Self-Disclosure and Risk Taking offer good convergent and discriminant validity between observers and participants. Non-verbal Behavior and Empathy offer marginal convergent validity in the validity diagonal. Com- paring the seven predictor variables measured by the IACOBS and the LBQPAR, only Respect, Immediacy, and Self-Disclosure show evi- dence of promising convergent and discriminant validity. Immediacy shows marginal convergent validity on the validity diagonal. A possible explanation for the differences between observer and participant ratings of leader effectiveness is that participant ratings of the leader were made at each of five different observation periods spaced throughout the life of the group. Each observer mea— sure was based on leader behavior at one meeting whereas participants rated the leader on the basis of his behavior across the entire term. In summary, observers and participants appear to have different inter- pretations of the definitions of the 12 predictor variables. Two independent instruments scored by the observers provide reasonably valid results according to the Campbell and Fiske multitrait-multi— method matrix for determining convergent and discriminant validity. Nearly identical instruments completed by participants and observers provide little evidence of validity. Multiple regression will be used to differentiate the best subset of predictors from the highly intercorrelated sets. 74 Overall leader effectiveness measures were created by adding together scores for all the items on a particular leader effective- ness measure. Group Structure was scored so that high scores indi- cated a perception of low group structure. Correlations among overall leader effectiveness measures substantiate the high validity between LBQOBS and IACOBS and the lower validity between measures taken from participants versus observers. The correlation between overall LBQOBS and overall IACOBS is §_= .85, p_< .001. The cor- relation between overall LBQPAR and overall LBQOBS is [_= .48, p_< .05. The correlation between overall LBQPAR and overall IACOBS is r_= .25 (ns). The PREIAC is a single overall observer-based measure of leader functioning taken before the start of the group. The corre- lation between PREIAC and overall IACOBS is [_=.64 , pl< .05, and with overall LBQOBS is I: = .45, p_< .05. However, the correlation between PREIAC and the overall LBQPAR is only r_= .18 (us). The PREIAC proved a poor predictor of participant benefit and was not used in subsequent analyses. No correlations between PREIAC and any measure of participant benefit approached significance. Criterion Measures The GMBQPAR and GMBQOBS are nearly identical measures of member benefit. The items of both measures were subjected to a McQuitty (1957) elementary linkage analysis for isolating typal structures. Elementary linkage analysis, or typal analysis, is a method of clustering which yields results similar to rotated factor 75 analytic solutions. A type consists ofefll items that intercorrelate more highly with at least one other item of that type than they do with any items of a different type. Results for the GMBQPAR are given in Figure 1 (see Appendices 0.1 and 0.2). Most of the items grouped together in two primary clusters. Type 1 was labeled Satisfaction and Type 2 was labeled Learning. Stepwise multiple regression analyses using first Type 1 and then Type 2 as the criterion measures revealed the same results as a regression analysis using the GMBQPAR in its entirety. Although two distinct clusters can be differentiated, their predictive power is no different than that of the GMBQPAR as a whole. In addition, Cronbach's (1951) alpha reliability coefficient was r_= .86 for the entire scale. Therefore, it seems reasonable to conclude that the GMBQPAR is a satisfactorily valid and reliable measure of participant benefit. Results for the typal analysis for the GMBQOBS is given in Figure 2 (see Appendices 0.3 and 0.4). lflnn‘different clusters could be identified. The content of each cluster overlapped so much that meaningful labeling was difficult. Stepwise regression analysesuming each of the four clusters as criterion measures for the 12 observer- rated leader effectiveness variables yielded results quite similar to a regression analysis using the whole GMBQOBS to predict leader effec- tiveness. For Types 1, 2, and 4, Group Climate was entered first and Group Structure second. Type 3 yielded one unexpected result in the regression concerning Non-verbal Behavior, which was entered first followed by Group Structure and Group Climate. Cronbach's alpha reli- ability coefficient for GMBQOBS was r_= .85. Thus it appears that (63) (68) (69) (71) (72) (61) (62) (64) (65) (66) (67) (70) (73) (74) 76 TYPE 1: Feeling of Satisfaction __._L (68)— 49 _. (69) 80 (72) V—_‘ ‘ 42 \ 58 71 \ ( ) (63) I feel better about myself as a result of my participation in the IPL group experience. Group interaction in my IPL was often superficial or boring. I can honestly recommend the IPL course to my friends as a valuable learning experience. Group interaction in my IPL was often negative or destructive. I have basically positive feelings about my IPL group experience. TYPE 2: Learnings and Understanding (66) My IPL experience has enhanced my awareness and understanding of my own feelings and behavior. I have a greater understanding of the processes that facilitate group functioning. The IPL group process was often confusing to me. As a result of my IPL experience it is now easier for me to relate to people outside my group. I have a greater understanding of the processes that inhibit group functioning. I have a greater understanding of the communication skills necessary to be a good group leader. My IPL experience has enhanced my awareness and understanding of other people's feelings and behaviors. TYPE 3: Active and Constructive .____L (73) 74 (74) "______. I feel I was an active participant in the IPL group process. I feel I was a constructive participant in the IPL group process. Figure l.--Typal analysis of GMBQPAR items. 77 TYPE 1: Value g (7) 72 (10){ (HF 58 \ / (8) (9) (7) The IPL group process often seemed confusing to him/her. (8) His/her interactions had a negative or inhibiting effect on the IPL group process. (9) He/she demonstrates a good understanding of the communication skills necessary to be a good group leader. (10) Group interaction in the IPL ofen seemed superficial or boring to him/her. TYPE 2: Involvement __L (6) 64 (11) ,fl O;mm .OO.m>-coz 11 11 11 um «tic. .v we: mm No: mgzuuzgpm mo Oo mm m.- so «.m m. on «Om.m co.umu:oc.cou moi ¢.1 ..1 .o m. .N fimom «cw «mime AQMwquE. mo O. m. mo- «mm mo «mm «O.o «OO.O Ommcmpmgucou - - - .N- OON NO O.N «OONO OOONO OOOO..O OOOOO mm O. .N mo OOOm mm mo mm «OO.m OspOOEm mo «k¢m .m m. ¥OONM mv «#mm *ONQ ¥k*.¢ mmwcmcwzcmw 11 11 11 v. «kkcm mm 0. «*nm ¥«¥Nm ungh ¥«¢© *xmom *me .m «mmme m. «*mm Ompm ¥¥¥mm puwammm .O.mz. .NNma. .O.mm. .O.ma. .NNmm. .O.ma. .NOma. .O.mm. .OO.LOO .OOO.OOO O.NOOOO 2m. mmo m.mmao om; .ONm>.mOOo om; .OHOOO.O.ONOO .o.>O;mm .mumm. mo Om:.pmm Ou.wmcma .mnsms .Ouop ucm Ommcm>.pum..m .mumm. umpO. .o Oco.uO.m..ou ucmsosupuauocq ..< N m.nmh 81 Observer Data: Member Benefits and Leader Behavior (IACOBSi Observer data on group member benefit and observer data from the seven predictor variables measured by the IACOBS revealed leader Respect ([.= .50, p_< .001) and Genuineness ([_= .34, p_< .002) to have the two highest correlations with participant benefit. Independent Measures An analysis of independent measures, i.e., using observer data to predict participant or friend data, etc., was completed. For these analyses it was necessary to aggregate all data to the group level (N_= 15). Participant-rated leader effectiveness predicted observer- rated benefit quite well. Correlations for all 12 variables between §_= .31 and r_= .82. Group Climate (§_= .82), leader Genuineness ([-= .62), Respect ([_= .61), Concreteness (§_= .61), and Trust (g_= .57) correlated with observer-rated benefit at the p_< .01 level of confidence. Two more, Self-Disclosure and Immedi- acy, reached the .05 level. Friends' ratings of participants' benefits and participants' ratings of leader effectiveness show significant correlations (Ef:.05) with leader Respect (§f=.23), Genuineness (§;=.29), Self-Disclosure (r (I; .17), Concreteness ([_= .23), Immediacy, ([_= .30), Risk Taking .20), Non-verbal Behavior (:_= .18), and Group Climate (§f=.21). Participant-rated benefits were correlated with observer rat- ings of leader effectiveness on the 12 predictor variables. The LBQOBS revealed leader Confrontation (§f=.51, pg<.05) and Self- 82 Disclosure (§f=.44, p51.05) to be the most highly correlated with par- ticipant self-ratings of benefit. This is the strongest evidence in this study relating the importance of Confrontation to participant benefit. There were no significant correlations between friend ratings of participant benefits and the observer ratings of leader effective- ness on the LBQOBS. Using the IACOBS, Respect was the only significant correlation with friends' benefit ratings at [_= .64 (p_< .01). The correlation between GMBQPAR with the overall measure for LBQPAR was 3= .84 (p_ < .001); with LBQOBS, 3= .47 (p < .05). The cor- relation between GMBQOBS and LBQOBS was Ef=.55 (p_< .05) and for LBQPAR, rf=.50 (p_< .05). The correlation between GMBQOBS and overall IACOBS was not significant. Interpersonal Checklist The ICL was administered at the beginning and end of the groups. It was hypothesized that participant gains or losses in mental health as measured by the LOV and DOM factors of the ICL might be re- flected in their satisfaction with the group or their rating ofleader effectiveness. Correlations range from §f=-.O7 to £==.l7. Only Respect (§f=.l4), Immediacy (§f=.15), and participant-rated benefit (§f=.15) showed significant correlations (p_< .05) with gains on the LOV factor. Group Climate ([f=.17) had significant correlations (Bf:.05) with DOM gains. All correlations between ICL gains and all measures of leader effectiveness were quite low so no further analyses 83 are reported. The difference between the means on the pre- and post- test for LOV is .63 and for DOM, 1.39. Operating on the hypothesis that smaller standard deviations on the post-scores of the original 16 scales of the ICL in compari- son to the standard deviations on the pre-scores might be indicative of improved mental health, the ICLdata were re-analyzed using stan- dard deviations instead of means as raw data. No significant correlations were obtained. Self-EXploration Participant Self-Exploration (SE) was hypothesized to be related to participant benefit and leader effectiveness. However, SE as rated by the observers correlated only §_= .20 (p_< .05) with LBQOBS, and §_= .16 (p_< .05) with LBQPAR. Furthermore, SE corre- lated only §_= .14 (p_< .05) with participant-rated benefit. The correlation between SE and GMBQOBS was r_ = .16 (p < .05). There was no sigificant correlation between SE and GMBQFR. SE correlated with IACOBS measures of Self-Disclosure ([_= .35, p_< .001), Imme- diacy (§_= .34, p_< .OOl), Confrontation (r_= .31, p_< .004), and Empathy ([_= .20, p_< .05). Course Grade It was also hypothesized that a participant's self-rating of benefit and of leader effectiveness might be a function of his/ her grade in the course which was graded on a pass/no-credit basis. The correlation between grade and benefit was 1= .20 (p < .01). Dividing the population into pass/no-credit groups showed virtually 84 no difference in the correlations between participant-rated benefit and the participant-rated leader variables when examining only the pass group, only the no-credit group, or all participants together. Typal Analysis: LBQPAR, LBQOBS, IACOBS A typal analysis was performed on each of the three correlation matrices and revealed three similar types of clustering (see Figures 3, 4, and 5; also Appendices 0.5-0.7). The heavy linkages in the figures depict the actual typal bonds. These three figures also show the actual g values for the typal bonds. All additional correlations of .E = .40 or higher are depicted by finer linkages. As shown in Figure 3, the two primary types for the LBQPAR centered first around an g= .74 correlation between Respect and Trust, and second, around an rf=.64 correlation between Risk Taking and Self- Disclosure. Concreteness, Empathy, Genuineness, Non-verbal Behavior, and Group Climate cluster around Respect. Immediacy and Confrontation cluster near Trust, and Group Structure clusters near Self-Disclosure in the second type. Figure 4 shows the typal analysis for LBQOBS. Again, the Type 1 cluster had Respect and Trust as its primary components with Empathy, Genuineness, Group Structure, Concreteness, and Group Climate also part of Type 1. Type 2 showed a strong relationship between Con- frontation and Immediacy with Risk Taking, Self-Disclosure, and Non- verbal Behavior also part of Type 2. Figure 5 shows two clusters for the IACOBS. Type 1 consisted of Concreteness and Confrontation with Empathy and Immediacy. Type 2 consisted of Genuineness, Self-Disclosure, and Respect. 85 .Oucon .OO.. ma umuu.gmu oO.O ONO m>onm No ow. O .o mco..O.wN.ou .Oco.p.uOO ..onw Lo ow. .1..- .._.o mcowuwmeLou chowpwbvm _._.<% O .ANNu z. Om.OUO pcmcoqeoo O.mmoom. mo O.Om.O:O .OOm.-.O me:m.u 9.33.3181 953.38 ..mm mm 9.33.33... oe \ S. \\N\ 8535.928131. O [Immllv .OO.. omammmullmml //////// fm Omocmc.=:mw OO /OO/ N_O O/O pagan . Omcmumcucou he . .e we .' ov me 87 .Om>.m O0.0 ONO m>OOO .0 cc. O .o Oco.um.m..ou ucOu...cm.m« O..N. "mm. Om.OOO .OOOOOEOO O.mmou<. .o O.OO.O:O .OOO --.m m.:m.u l Numammmlmwl 322.23 om 233.35-. .mm OOO.OoEE. ///// ow ow .\ NO/./- .111 El 5.32228 mo 328.228 Al. 8 2.2.... // om /‘|lhd\)\\1 N mam. . mam. 88 Multiple Regression Regression analyses were completed on all nine possible combinations of predictor measures and criterion measures. High intercorrelations among the 12 predictor variables tended to obscure each one's true relationship to the criterion measure. Stepwise regression‘was used to ascertain the best subset of predictors from the possible 12. It extracts the variable or variables that explains the most common variance that several variables hold in common with participant benefit. The only variables considered here are those significant at the p_< .05 level of confidence (Table 3; also see Appendix E.3). Using participant data on leader effectiveness and member benefit (LBQPAR, GMBQPAR), the best four predictor variables-were Respect, Group Climate, Empathy, and Group Structure, respectively. Cumulative 32 = .50. 32 for Respect alone was .39. Using observer data on leader effectiveness and member benefit (LBQOBS, GMBQOBS), the best three predictor variables were Group Structure, Group Climate, and Immediacy. Cumulative R2 = .58. For Group Structure alone, 3? = .48. Using observer data on the IACOBS to predict leader effec- tiveness and observer data on member benefit (GMBQOBS), the best leader variable was Respect. 32 = .25. Using independent measures, the best predictor variables were Immediacy and Genuineness, respectively, for participant-rated leader effectiveness and observer-rated member benefit. Cumulative _R_2 = .13. The best predictor variable was Genuineness for 89 Table 3 Partial summary table for stepwise multiple regression for significant leader effectiveness measures with participant benefit measures Measures Step Variable f_to Enter Significance R? GMBQPAR 1 Respect 124.56 .0001 .39 and LBQPAR 2 Group Climate 24.96 .0001 .47 3 Empathy 9.71 .002 .49 4 Group Structure 4.98 .03 .50 GMBQOBS 1 Group Structure 65.41 .0001 .48 and LBQOBS 2 Group Climate 9.92 .002 .55 3 Immediacy 5.44 .03 .58 GMBQOBS 1 Respect 23.10 .0001 .50 and IACOBS GMBQFR l Immediacy 9.10 .003 .09 with LBQPAR 2 Genuineness 4.17 .05 .13 GMBQOBS l Genuineness 6.93 .01 .26 with LBQPAR GMBQFR 1 Respect 8.94 .01 .41 with IACOBS 9O participant-rated leader effectiveness and friend-rated benefit; 3? = .07. The best predictor variable'was Respect for observer-rated leader effectiveness (IACOBS) and friend-rated member benefit (3? = .41). LBQOBS with GMBQPAR, LBQOBS with GMBQFR, and IACOBS with GMBQPAR did not reveal any significant results. GMBQPAR with Learningiand Satisfaction Dividing the GMBQPAR into two different criterion measures based on the typal analysis produced results identical to LBQPAR with GMBQPAR. Cumulative 3? = .41 for Learning and 3? = .43 for Satisfaction (Table 4; see also Figure 1 and Appendix E.2). GMBQOBS with Types 1 to 4 Dividing GMQOBS into four clusters based on the typal analy- sis produced similar outcomes for Types 1, 2, and 4. The only dif- ference is that Group Climate was entered first and Group Structure was entered second. For GMBQOBS Group Structure was first and Group Climate was second. Each variable explained approximately the same amount of variance as they do in the regression equation for GMBQOBS as a whole. Type 3 introduced Non-verbal Behavior as the best pre- dictor variable with Group Structure and Group Climate, respectively. 2 R_ = .23 for Non-verbal Behavior (Table 5; see also Figure 2 and Appendix E.l). Analysis of Variance for Groups Nested in Leaders A unique experimental design in this study in which each leader wasiricharge of two groups allowed for the examination of 91 Table 4 Partial summary table for stepwise multiple regression for learning and satisfaction with LBQPAR (Figure 2)* Measures Step Variable f_to Enter Significance Learning 1 Respect 92.01 .000 . with 2 Group Climate 14.00 .000 .37 LBQPAR 3 Empathy 8.21 .005 .40 Satisfaction 1 Respect 93.29 .000 .33 with 2 Group Climate 21.30 .000 .40 LBQPAR 3 Empathy 6.32 .013 .42 4 Group StrUcture 3.89 .050 .43 *Only significant variables are included (p_< .05). Table 5 Partial summary table for stepwise multiple regression for Type 1 to Type 4 with LBQOBS (Figure l)* Measures Step Variable E_to Enter Significance R? - Type 1 1 Group Climate 16.41 .000 .19 with 2 Group Structure 6.68 .012 .26 LBQOBS Type 2 1 Group Climate 15.59 .000 .18 with 2 Group Structure 9.56 .003 .28 LBQOBS Type 3 1 Non-verbal Behavior 21.80 .000 .24 with 2 Group Structure 14.27 .000 .37 LBQOBS 3 Group Climate 13.37 .000 .47 Type 4 1 Group Climate 50.70 .000 .42 with 2 Group Structure 6.50 .013 .47 LBQOBS *Only significant variables are included (p_< .05). 92 the main effect for groups nested in leaders. A priori, it seemed logical to assume that different leaders would produce more variance than different groups led by the same leader. Main effect for groups was highly significant (E_= 6.30, p_< .001, gf_= 7,84). Eta square (n2), a measure of the variance which each source variable accounted for, is .10 for leaders, .31 for groups, and .59 for par- ticipants. One leader who led only one group was eliminated from this analysis. Participant scores on Respect were randomly selected from each group to give equal cell Nfs (N_= 7) (see Appendix C.3). Respect was chosen as it appeared to be the best single predictor of leader effectiveness. CHAPTER V DISCUSSION Introduction The basic premise underlying the present research was that perceived group leader effectiveness is highly correlated with group member benefit in the small group experience designed to teach inter- personal communication skills. This study also posited that 12 skills or variables are necessary for effective leadership hithese groups. Although it has never been completely empirically verified that the member's positive perception of the leader is directly related to the learning of interpersonal communication skills, past research indi- cates that how members perceive the leader is correlated with the degree to which members become involved in the group, express feel- ings, engage in meaningful analysis, achieve independence from the leader, and establish cooperative relationships (Bolman, 1971; Cooper and Mangham, 1971; Culbert, 1961; Hurley, 1972; Lohman et a1., 1959; O'Day, 1976). Two studies, Bent, Putman, Kosler, and Nowicki (1976), and Strupp, Fox, and Lessler (1969), provide strong data that show that client's attitude towards his/her therapist was closely related to success in therapy. Clients who were very satisfied with their therapy and felt it had a positive, noticeable, generalizable effect on their behavior rated their therapist as significantly 93 94 warmer, more likable, and more involved than clients who were dis- satisfied with therapy. Reliability_and Validity of Predictor Measures The IAC has been shown to be a reliable and valid predictor of high functioning helpers in numerous research studies (Carkhuff, 1969). The standard procedure is to role-play, and cast the helper in the helping role with a "standard helpee" who is trained to role- play the helpee's role. Trained raters then rate the helper on the seven core conditions alleged to be necessary to effective helping: empathy, respect, genuineness, self-disclosure, concreteness, imme- diacy, and confrontation (Carkhuff, 1969). This procedure has been used effectively in rehabilitation counselor education (Anthony and Carkhuff, 1969), guidance counseling (Martin and Carkhuff, 1968), clinical psychology training (Carkhuff, Kratochvil, and Friel, 1968), and with nurses (Kratochvil, 1968), among others. In the present study, trained observers rated the group leaders on the seven vari- ables in a real-life helping situation (IACOBS). Carkhuff (1969) has also created a written index of communi- cation in which the helper responds in writing to 16 standard helpee expressions. Although role-playing is the preferred technique for assessing helpers, written responses are also valid indicators of helper level of functioning when rated by trained raters (Antonnuzzo and Kratochvil, 1968; Greenberg, 1968). The standard written form of the IAC was administered to all group leaders prior to the group experience (PREIAC) and each leader given a single global rating of 95 effective functioning. The PREIAC correlated positively with both LBQOBS (§_= .45, p_< .05) and IACOBS (§.= .64, p_< .05). The PREIAC showediuisignificant correlations with any measures of parti- cipant benefit. Possibly this is the result of the modification of using only a single rating of leader effectiveness instead of rating each of the seven variables separately as is usual, and which was done for the IACOBS. However, it is possible that the written ver- sion completed prior to the group experience simply would not be related to participant benefit even if ratings on all seven scales were made. Using the multitrait-multimethod matrix to examine the validity of the predictor measures from observer reveals sufficient convergent validity but marginal discriminant validity for the LBQOBS and IACOBS (see Table 1). Examination of measures completed by independent sources (LBQPAR with LBQOBS) clearly supported the validity of only Self- Disclosure and Risk Taking. Empathy and Non-verbal Behavior showed some convergent validity. For LBQPAR with IACOBS, Self-Disclosure revealed adequate validity and Respect and Immediacy showed some con- vergent validity. External evidence for the validity of other mea- sures taken from the LBQPAR and LBQOBS was weak. However, the con- vergent validity of the LBQOBS was substantially upheld by the IACOBS which measures 7 of the 12 predictor variables. It is also possible that the lack of substantial agreement between the LBQPAR and the LBQOBS was not due to the invalidity of the instruments, but to discrepancies 96 in perceptions and definitions held by the observers in comparison to those held by the participants. Observers and participants strikingly disagreed about which variables were highly correlated with benefit (see Table 2). Group Structure correlated .70 with benefit as rated by the observers. Group Structure correlated -.02 with benefit as rated by the parti- cipants. Participant-rated leader Self-Disclosure, Immediacy, Con- frontation, and Risk Takingwere signigicantly correlated with participant-rated benefit but wereranperceived as related to par- ticipant benefit by the observers. IACOBS measures on Self- Disclosure, Immediacy, and Confrontation were not significantly correlated with benefit, either. Self-Disclosure, Immediacy, and Confrontation measured by LBQOBS and IACOBS revealed nearly identical correlations. Although observers'and participants' ratings of Self- Disclosure and Risk Taking showed respectable convergent and divergent validity, these two source groups disagreed onlwnvSelf-Disclosure and Risk Taking related Unbenefit. They also disagreed on benefits' relationships to Group Structure, Immediacy, and Confrontation. It is possible that all three leader effectiveness measures are more valid than they appear, but that observers and participants held different definitions of many variables. The observers were chosen because of considerable prior experience in leading and observing similar groups. The participants were virtually all experiencing aninterpersonal communication skills group for the first time. This lack of experience on the part of the participants 97 in comparison to the observers may partly account for the differ- ences obtained on the measures taken from the two sources. Another explanation for the large differences manifest in observer and participant ratings of leader effectiveness is the fact that participant ratings were made once at the end of the term and observer ratings were made after each of five observations spread across the term. Therefore, each observer evaluation was based on leader behavior at only one group meeting, whereas participant evalu- ations were based on leader behavior throughout the entire term. This is true for participant benefit measures, also. Therefore, it is unlikely that participant and observer measures would be identi- cal. Observers were rotated to different groups as randomly as schedule conflicts permitted. Each group was observed by at least three different observers so that they would not become emotionally involved with any particular group. Given these procedures to main- tain the objectivity of the observers, the lhnited agreement between observers-and participants was understandably low. The high degree of convergent validity between the two observer measures of leader effectiveness and the modest indicator of interrater reliability offered by the analysis of variance on observer ratings of the LBQOBS are fair evidence of reliability. The analysis of variance showed observers agreed more when they rated the same leader than when they rated different leaders on the T-group training film (see Appendix C.2). 98 Reliabilityiand Validity of Criterion-Measures The nearly identical GMBQ questionnaries for participants and observers were designed to measure one variable: participant “benefit. Items pertained to participant satisfaction and learning in the group. In order to establish validity it was necessary to show that all items on the questionnaire do indeed measure the same thing. Also, both questionnaires should have agreed with each other. The GMBQ items for participants were subjected to a typal analysis and all but two items formed typal structures that could reasonably be labeled as Learning or as Satisfaction (see Figure 1). Using first Learning and then Satisfaction in a stepwise regression analysis with the 12 predictor variables of the LBQPAR revealed identical results for Learning and Satisfaction and the GMBQPAR as a whole (Appendix E.2). It was concluded that GMBQPAR was a valid instrument to measure perceived participant benefit. Data from this study show that ratings of leader effectiveness made by participants correlates equally well with participant ratings of their own Learn- ing and their own Satisfaction in small groups designed to teach interpersonal communication skills. The GMBQ items for observers were similarly subjected to a typal analysis and four item types were identified (Figure 2). Types 1, 2, and 4 reveal similar results to the total GMBQOBS score, except that Group Climate was entered first and Group Structure second. For the whole GMBQOBS Group Structure was entered first 99 and Group Climate second. For Type 3, however, Non-verbal Behavior explained more variance than did any other variable. This was the only time that Non-verbal Behavior proved a significant predictor of participant benefit. The correlation between GMBQPAR and GMBQOBS is §_= .27, p_< .001. This indicates very modest agreement between observers' and participants' views of member benefit. These sources disagreed on the definitions of some leader effectiveness variables, especially Group Structure. This modest correlation further underscored the differing viewpoints of the participants and observers. The reason- ably high correlations that friends' benefit ratings showed with both observers' and participants' benefit ratings provided some external support for the validity of the benefit questionnaires. The amount of agreement "my' be reasonable given that observers rated their benefit questionnaires after a single observation during the term, participants rated their benefit questionnaire once at the end of term, and friends rated their benefit questionnaires at the end of term without ever observing a group meeting. A modest indicator ofinterrater reliability on the GMBQOBS is suggested by an analysis of variance that showed observers agree more when they rate the same group member than when they rate different group members (Appendix C.1). Iizwas concluded that GMBQOBS shows pro- mise of validity as an instrument to measure participant benefit. GMBQFR was not subjected to a typal analysis due to its low correlations with the overall leader effectiveness measures made from 100 either participant or observer data. It correlated more highly with both GMBQOBS and GMBQPAR than they do with each other (£f=.47,‘[='.55, p_< .001, respectively). These moderate correlations offered some external evidence for the validity of the participant benefit measures. Hypotheses Participant Data: Leader Behavior and Member Benefit (LBQPARjfii The major hypothesis of this research was that the 12 predic- tor variables would correlate highly with participant benefit. Cor- relations were all highly statistically significant with p_< .001 for all variables except Group Structure. The correlations ranged from §_= .28 for Self-Disclosure to [_= .63 for Respect. It was hypothe- sized that Group Structure and Self-Disclosure were the variables least likely to correlate significantly with benefit. Observer Data: Leader Behavior and Member Benefit (LBQOBS) The observer ratings of leader effectiveness and participant benefit produce significant correlations, p_< .05, for 8 of the 12 variables. Leader Self-Disclosure, Risk Taking, Confrontation, and Immediacy failed to correlate significantly with participant benefit according to the observers. Group Structure had the highest corre- lation with benefit for the observers (g_= .70). High scores on Group Structure reflected a perception of ipw_levels of leader- initiated structure. 101 Both observers and participants agreed that at least 7 of the 12 predictor variables were positively correlated to participant benefit (p_< .05). These variables were Respect, Trust, Empathy, Genuineness, Group Climate, Concreteness, and Non-verbal Behavior (Table 2). Observer Data: Leader Behavior and Member Benefit (IACOBSi The Index of Accurate Communication measured 7 of the 12 predictor variables and was rated by the observers from group obser- vations. These variables were Empathy, Respect, Self-disclosure, Genuineness, Concreteness, Immediacy, and Confrontation. Correla- tions of these observer-rated variables with observer-rated benefit revealed significant linkages of the latter with Respect (r_= .50, p < .001) and Genuineness (§_= .34, p_< .002). Halo Effect Since data taken solely from one frame of reference, whether participant or observer, might suffer from a halo effect (i.e., a positive attitude toward one group or one leader could lead to gen- eralized positive scores on all variables), correlations from inde- pendent measures of leadership and benefit were made. An examina- tion of the problem of halo effect is given in an article entitled "Correlational Bias in Observer Ratings" by Borman and Kinney (1976). They suggest that the correlational bias, or logical error, or halo effect, is present in all research studies that use observers to rate behaviors or traits. They also suggest that 102 their research shows that there is a very high correlation between accuracy and halo effect. To completely eliminate the halo effect, correlational analy- ses (and multiple regression analyses) were made across independent measures of leader effectiveness and participant benefit. Data on benefit obtained from either participants, observers, or friends were correlated with leader effectiveness data obtained from another source, either participants or observers. Also, many items were phrased negatively to offset a generalized yea-saying response. Independent Measures Friends' rating of participants on the GMBQ showed only one significant correlation for any predictor variable on the IACOBS. The variable was Respect (5 =.64, p_< .Ol). Correlations between GMBQPAR and LBQOBS reveal substantial correlations for Self- Disclosure and Confrontation (r_= .44, [_= .51, p < .05, for both). Correlations between GMBQOBS and LBQPAR revealed Respect, Genuineness. Self-Disclosure, Concreteness, Immediacy, Group Climate, and Trust had significant correlations with benefit (see Table 2). Thus, the data from independent measures generally confirm the relationships between member benefits and all leader effectiveness measures except Group Structure and Empathy (see Table 2). Interpersonal Checklist A self-description personality inventory, the Interpersonal Checklist, was administered to all participants at the IPL group's beginning and end. Scores on the ICL were summarized to yield 103 measures of two independent factors: Love-Hate (LOV) and Dominance- Submission (DOM). It was hypothesized that participant's ICL gains would correlate positively with participant benefit in the group, and also with positive leader effectiveness ratings. However, cor- relations between gains on the LOV and DOM factors with both leader effectiveness and participant benefit hovered near zero. That there was little general change between the two test administrations was evidenced by the very small differences between the pre- and post- group means of ICL scores. Perhaps the pre- and post-test scores were so similar because the IPL group experience was not powerful enough to register any significant changes in personality. It is also possible that the ICL did not validly assess behavioral or attitudinal changes in the present situation. Answering true or false on such self-description items as "businesslike," "faithful follower," "always pleasant and agreeable," and "can be strict if necessary" might be unaffected by even a very positive or negative group experience. Perhaps if the leaders or observers rated each member pre- and post-group on the ICL, differences in ICL scores would positively correlate with leader effectiveness and member benefit. Course Grade The correlation of §_= .20, p_< .001 between course grade and participant-rated benefit suggests that those who received "no- credit" grades for not mastering the IPL skills were among those who tended to be less satisfied with the IPL group and the group leader. 104 Observer Ratings of Self-Exploration Only small correlations were obtained between SE and any measures of leader effectiveness or participant benefit. SE did show moderate correlations with leader Self-Disclosure on the LBQPAR (r_= .43, p_< .05, N_= l5) and IACOBS (r_= .35, p_< .001, N.= 72). This tends to confirm that leader Self-Disclosure is related to participant Self-Exploration or Self-Disclosure. This finding has been reported for therapists and clients and the suggestion has been made by several authors that therapist and client Self-Disclosure is related to client benefit (Carkhuff, 1969; Culbert, 1968; Hurley and Force, 1973; and Jourard, 1969). The present findings do not confirm that leader Self-Disclosure or participant Self-Exploration are related to participant benefit. Multiple Regression Analyses High intercorrelations among the 12 predictor variables tended to obscure the true relationship between each one and the criterion variable. The participant-rated measures show that both Respect and Trust correlate substantially with benefits (r_= .63, r.= .52, p_< .001, respectively). Yet Respect and Trust as measured by the LBQPAR correlate r_= .74, p_< .001. It is possible that the variance which Trust shares with Benefit is actually the same variance it shares with Respect, and that high leader Respect is a better predictor of benefit than high leader Trust. Trust and 105 Respect might actually be nearly identical measures with Respect accounting for the most common variance with benefit. Stepwise mul- tiple regression analyses were completed for all nine possible com- binations of predictor and criterion variables to ascertain which subset was actually the best predictor of benefit. The only variables considered as adequate predictors were those that had an f_signifi- cant at the p_< .05 level of confidence in any of the nine regres- sions (see Table 3 and Appendix E.3). Proposed Rank Order of Predictor Variables Respect was the salient variable in three of the regression equations. Using GMBQPAR as the criterion variable and LBQPAR for the predictor variables, Respect accounts for 39 percent of the vari- ance. Using IACOBS and GMBQOBS, Respect accountedikn~25 percent of the variance. Using GMBQFR and IACOBS, Respect accounted for 41 per- cent of the variance. Respect was the best predictor of benefits according to participants, observers, and in one across-source (observers and friends) comparison. The most powerful predictor of participant benefit in this study appeared to be Respect. Group Structure, Group Climate, Genu- ineness, Immediacy, and Empathy added significantly to predicting benefit. Self-Disclosure, Concreteness, Risk Taking, and Non-verbal Behavior appeared irrelevant to predicting participant benefit. Trust was highly correlated with Respect. Respect was important whether 106 participants, observers, or friends evaluate participant benefit. Group Structure was important when observers evaluated partici- pant benefit. Group Climate was somewhat important to both observers and participants in predicting benefit. Genuineness and Immediacy played a small but significant role in explaining parti— cipant benefit using independent measures of leader effectiveness and participant benefit. Empathy was significantly related to benefit but explained only 3 percent of the variance between GMBQPAR and LBQPAR. Typal Analysis of Predictors Typal analysis is a method for clustering data into distinc- tive structures. The purpose of clustering is to reduce a variety of tests and ratings to a small number of representative variables. Type typal analysis added some insight into the meaning of the pre- dictor variables (see Figures 3, 4, and 5). For LBQPAR, Type 1 was centered around an r_= .79 correlation between Respect and Trust. Empathy, Group Climate, Group Structure, and Genuineness, the other variables that significantly predicted benefit using the LBQPAR and GMBQPAR in regression analysis, along with Concreteness, complete Type 1. The centroid of Type 1 appeared to be Respect and Trust. Trust seemed to play a large role in the definition of Respect. Mutually perceived Genuineness, Respect, Empathy, and Trust between the leader and the members, and among the members them- selves (Group Climate), appear to lead to high self-reports of benefit for participants. 107 Type 1 for LBQOBS centered around an r_= .74 correlation between Respect and Trust. Again, these two variables were highly related. Group Structure was a part of Type 2. Group Structure was the best predictor of participant benefit according to observers, with Group Climate, and Immediacy adding somewhat to the predictive power of the equation. The typal analysis for IACOBS was less clear. Type 2 centered aound an r_= .50 correlation between Self-Disclosure and Genuineness with Respect hanging on to Type 2. The regression equation using IACOBS showed Respect to be the best predictor of participant bene- fit. Observer-rated Respect explained 25 percent of the variance on GMBQOBS and 41 percent of the variance on GMBQFR. Description of an Effective Group Leader The picture that emerges of an effective group leader for small groups working on interpersonal communication skills from these data is of a person who is perceived as a warm human being and who cares about the feelings and experiences of each of the group members. It seems important that each group member feel that the leader cares about him/her personally. The leader helps him/her express himself/herself positively to the group. Leader Respect for the participants is the variable which best predicts participant benefit. The five items that were used to score Respect are (a) My TA (Teaching Assistant) communicated to me a deep caring about my feelings, experiences, and potential; (b) I often felt my TA was judging my attitudes and behaviors too harshly; (c) I feel my TA has mostly warm feelings about me; (d) My 108 TA helped provide me with opportunities to express myself in a positive manner to the group; and (e) I sometimes felt my TA was ignoring me. Scores on (b) and (c) were reversed and all five scores added together for a Respect score. The items are identical on the LBQPAR and LBQOBS and similar for the IACOBS. The descrip- tions of each predictor variable that follows are also taken directly from the questions of the LBQ questionnaires and the rating scales of the IAC (see Appendices A.3 and B). Low level of Group Structure is the strongest predictor of participant benefit for the observers. A leader who is rated low on Group Structure allows much of the group's activity to be spontaneously created by the needs and interactions of the group mem- bers. A leader who is perceived as high on Group Structure structures most of the group's activities in the form of lectures, discussion topics, or group exercises. High or low Group Structure was irrele- vant to participant ratings of their own benefit, however. High scores on Group Climate indicated that all participants were perceived as being involved in the group actively and construc- tively at least part of the time. The Group Climate must be supportive. A group characterized by hostility, passive observers, or objections to group norms or activities is not conducive to participant benefit. Participants must develop a genuine liking for one another. A Genuine leader was perceived as sincere, spontaneous, non- defensive, congruent in his behaviors, attitudes, and feelings, and he is not seen as playing the role of group leader; rather, he is a warm, fully functioning human being. 109 Immediacy of relationship refers to the leader establishing personal, one-to-one relationships with each participant. Immediacy involves sharing of authentic positive or negative feelings between the leader and each group member. Empathy was defined as the ability of the leader to under- stand and reflect back to the participant the affect and content of the participant's message. Empathy is the ability of the leader to clarify, and add significantly to, the participants' understanding of their own feelings and behaviors. Redundancy Among Predictor Variables _ The high intercorrelations between many of the predictor variables suggest considerable redundancy among the 12 predictors. Trust and Respect seem highly related. It is possible that all of the variables shown to be significantly related to participant bene- fit in the regression analyses are actually all measuring leader Respect. Muehlberg, Pierce, and Drasgow (1969) factor-analyzed the seven variables rated for the Index ofAccurate Communication (IAC). Their results affirm the high redundancy among these predictor variables that was also revealed in this study. They found that empathy, respect, genuineness, concreteness, and self-disclosure intercorrelated r_= .78 to r_= .91. They concluded that a single major factor accounted for practically all of the observed correla- tions among the variables. Therapists rated high on one dimension were rated high on all dimensions. The common factor was "being a good guy, i.e., likable, friendly, helpful." In the present study 110 perhaps Respect is the factor underlying the predictor variables related to participant benefit. Leader Variables Unrelated to Benefit According to Regression Analyses A Trustworthy leader is supportive, honest about his own feelings, and helps create a climate of psychological safety in which participants can feel free to express their own feelings. Trust is highly correlated with participant benefit by data from both observers and participants. It also forms the center of Type 1 with respect for both participants and observers. It appears neces- sary for the leader to be Trustworthy, but respect predicts partici- pant benefit better. Leader Concreteness is the ability to keep discussions to specific feelings and experiences of personally relevant material, rather than on abstract or general conceptions of what constitutes interpersonal communication skills. Risk Taking was defined as the leader's willingness or ability to take initial risk in self-disclosure; establishing imme- diate, intimate relationships with group members; giving and receiv- ing positive and negative feedback; and using spontaneous non-verbal communication with group members. Risk Taking involves the danger of hurt through rejection for a variety of possible reasons. How- ever, if the group is to delve meaningfully into some of the problem areas of interpersonal communication it is necessary for the group leader and the members to take these risks. These data suggest that this kind of risk taking on the part of the group leaders was infre- quent or resulted in an unsatisfactory experience. 111 Confrontation involves examining the discrepancies between the leader's perceptions of the group members and their own percep- tions of themselves. Although the term Confrontation may evoke images of negative feedback, to be effective, Confrontation must be sensitive and perceptive. It may involve confronting the parti- cipant with a more positive perception of himself, on the part of ,the leader, than the participant actually holds for himself. These data suggest that Confrontation was very infrequent or resulted in an unsatisfying experience. Non-verbal Behavior refers to such things as simply under- standing that various non-verbal cues such as body position, ges- tures, and tone of voice were indicative of a person's true emotional state, to the leader's willingness to establish eye contact during interaction. It also refers to the leader suggesting non-verbal exercises. Informal feedback sessions, as well as the actual data, from both observers and group leaders, suggested to the author that these were important factors in the IPL groups. However, other more intimate forms of Non-verbal Behavior such as group leader touching or hugging another member, or touching or embracing among group members themselves, were very infrequent, if not absent altogether. Perhaps this explains the smaller amount of variance explained by this variable. Effects of Self-Disclosure According to the stepwise regression analyses, Self-Disclosure was not related to participant benefit. A great deal of research has led to differing conclusions about the importance of Self-Disclosure. 112 A short review of this lterature may help in understanding Self- Disclosure. Self-Disclosure is the amount and quality of information the leader reveals about himself. Does he reveal personal information about himself at all, and if so, is it enough to reveal him as a unique human being? Does he reveal information that would be embar- rassing if revealed to an outsider? Is he willing to examine his most difficult areas of experience? Judging from the present data these group leaders were not greatly self-disclosing, or Self- Disclosure is not of great importance to participant benefit, at least in comparison to other variables. Simonson (1976) suggests that therapists who were only mod- erately self-disclosing, i.e., demographic data, elicited more self- disclosure from their clients than did those who were either highly self-disclosing or who did In} self-disclosing. A cold therapist received no self-disclosures regardless of his own level of self- disclosure. Gittes and Blackman (1976) found that high self- disclosers were more apt to be accurate and less likely to gild (lie or distort) than were low self-disclosers. Superficial information was less likely to be gilded than intimate information, and friends received more accurate information than acquaintances. Chelune (1975) suggests highly affective self-disclosure is more self- revealing than emotionless self—disclosure. Flexibility in self- disclosure, i.e., when to be revealing and when to be closed, is another factor in understanding self-disclosure. Weigel, Dinges, Dryer, and Straumford (1972) provide data that show that therapist mental health is negatively correlated (r_= -.83) with therapist 113 self-disclosure as perceived by clients and co-therapists early in group psychotherapy. Member self-disclosure and mental health are positively correlated in both member and therapist ratings. Neigel et al. (1972) suggest that therapist self-disclosure in group psycho- therapy violates client role expectations for therapists. The LBQ Self-Disclosure items were concerned with intimate Self-Disclosure from the leader in a group teaching positive inter- personal communication skills. It was hypothesized that intimate Self-Disclosure would be important in such a group. It was also suggested that leader Self-Disclosure would have to be appropriate to the needs of the group members. Leader Self-Disclosure correlated positively with participant Self-Exploration, but neither was related to participant benefit. The findings of the present study agree with those of Weigel et al. (1972) and Simonson (1976) that high therapist or leader self-disclosure is not related to client or mem- ber benefit. The present findings also suggest that participant Self-Exploration is not related to participant benefit. This is contrary to the findings of Carkhuff (1969), Culbert (1968), Hurley and Force (1973), and Jourard (1969) who suggest that therapist and client self-disclosure is positively related to client benefit. The present findings do confirm that leader Self-Disclosure is posi- tively related to participant Self-Exploration. In summary, the positive and/or negative effects of self-disclosure are still not fully understood. The empirical data from many studies are conflict- ing. The present study suggests that leader and participant benefit are not related to Self-Disclosure. 114 Groups Nested in Leaders In this study each group leader led two different groups. Analysis of variance for groups nested in leaders reveals that rat- ings made for the same leader on Respect show more variance than ratings of Respect across different leaders. Other studies have been criticized for evaluating leader effectiveness on the basis of only one group. This design was incorporated into the present study to account for such criticisms. The highly significant E_obtained for leaders nested in groups suggests this criticism is valid. The same leader can be evaluated as effective in one group and much less effective in another. This effect even appears to overshadow evalu- ations of different leaders. Further research is definitely indica- ted to verify or discredit these results. If the same leader, behaving in a similar manner,is evaluatedso differently across groups, new parameters must be researched to accurately evaluate group processes in training programs to teach interpersonal skills. Comparison to Ohio State Leadership Studies The Ohio State Leadership Studies (Hemphill, 1950) found effective leadership in industry to be related to Consideration and Initiating Structure. The results of the present study show Con- sideration in the form of Respect, a supportive Group Climate, Imme- diacy of relationship, Genuineness, Empathy, and perhaps Trust to be vitally important to successful leadership in small groups teaching interpersonal communication skills. The importance of Initiating Structure is less certain. Trained observers, who were themselves group leaders, showed a marked preference for low structure on the 115 part of the leader. Participants found level of Group Structure was not related to participant benefit. Participants were equally satisfied by either high or low structure. These findings seem to agree with Caron (1973) who found that after participating in sensitivity groups employees felt Consideration was more important than Initiating Structure when rating their supervisors. Immediacy, Concreteness, and Confrontation are logically analogous to Initi— ating Structure. Only Immediacy was significantly related to par- ticipant benefit in the multiple regression analyses. Comparison to Yalom, Lieberman, and Miles Yalom, Lieberman, and Miles (1973) found Caring and Meaning Attribution to be the best predictors of participants' benefit in encounter groups. The present leader Respect findings parallel their results for Caring. Yalom et al. found effective leaders were also moderate on Emotional Stimulation and Executive Function. None of the successful predictor variables in this study are directly analogous to Meaning Attribution. Concreteness appears related to Meaning Attribution but did not link to participant benefit in the present study. Risk Taking, Self-Disclosure, Confrontation, and Non- verbal Behavior seem analogous to Emotional Stimulation, but these variables were unrelated to participant benefit in this study. Yalom et al. suggested moderate Emotional Stimulation to be related to effective encounter group leadership. Executive Function seems to be related only to Group Structure in this study. Observers preferred low structure and participants found either high or low 116 structure equally satisfying. Yalom et al. found the effective leader should be moderate in Executive Function. The IPL groups were not intended to be encounter groups. Those in charge of the Education 200 program criticized encounter groups because of their relative lack of structure in training par- ticipants in the interpersonal communication skills, and the lack of any attempt at a cognitive understanding of them. The strong emphasis placed on Respect, supportive Group Climate, Genuineness, and Immediacy might ultimately prove the IPL groups are more beneficial than the groups characterized by high Emotional Stimulation from charismatic leaders. Yalom et al. (1973) found charismatic leaders high on Emotional Stimulation produced the most group casualties (17 percent) whereas group leaders high on Caring had the fewest casualties (3 percent). ' Casualties were not directly measured in this research, unless we view students who received "no-credit" grades as casualties. Yalom et al. found that group leaders were least likely to be per- ceptive of potential casualties. The best predictor of casualties proved to be co-participant selection. The author was unaware of any formal or informal report of casualties except perhaps during one group session in which the observer and group leader both reported that Dr. Lopis, director of Education 200, had to personally inter- vene. The observer also reported that Dr. Lopis was exceptionally skillful in resolving the problem. Education 200 is a required course for all potential teachers at Michigan State University and enrolled about 1,000 students per 117 year over the past five years. No severe casualties have been repor- ted. This seems a remarkable record in light of the reported 10 per- cent casualty rate in the encounter groups studied by Yalom, Lieber- man, and Miles (1973). However, the casualty rate reported by Yalom et al. is considerably higher than the rate reported by other inves- tigators. Rogers (1970) and Egan (1970) summarized the existing literature and reported casualty rates of 1 percent or below. Comparison to Carl Rogers The results of the present study tend to support some findings of Rogers and Truax (1967). They suggested that an effective thera- pist needed to be rated as high on positive regard (Respect), empathy, genuineness, and a willingness to be known (Self-Disclosure). The present results do not support the importance of Self-Disclosure, but leader Respect and Genuineness appear to be authentic predictors of participant benefit. The present study found Trust to be highly cor- related with Respect and seems important to an understanding of Respect. Respect is related to high positive regard but not "uncon- ditional" positive regard. Empathy proved only peripherally related to benefit. This result is puzzling given the prominence of Empathy in the literature. Hurley (1975) suggests that trainer Acceptance/ Rejection of Others is a stronger predictor of participant-rated bene- fit than Empathy. Empathy was an important skill taught as Active Listening in the IPL. This was done in a step-by-step fashion and might have resulted in a mechanical view of Empathy for the partici— pants rather than as a genuine ingredient in a positive relationship with another. 118 Comparison to Hurley Studies Hurley (1976) and Hurley and Force (1973) have suggested that Acceptance/Rejection of Self (SAR) and Acceptance/Rejection of Others (ARO) are the two pre-potent dimensions for assessing leader effects on their encounter groups. Participant ratings of leader acceptance of encounter group members were highly correlated with participant gains from self-reports, observers, and intimates. The findings of the present study that leader Respect for participants, and leader Trustworthiness are highly related to participant bnefit as rated by participants, observers, and friends corroborates the Hurley findings concerning the importance of leader acceptance of encounter group members (ARO). The present study did not assess leader acceptance and rejection of self (SAR). Interpersonal Communication Skills Taught in the IPL The author of the present research has suggested 12 variables important to leader effectiveness in interpersonal communication groups. The data have shown some of these variables to be highly correlated with participant benefit and some are not. A dis- cussion of the actual skills taught in the IPL groups might be helpful in understanding the present results. The following discussion is taken from a paper by Dr. John Lopis, director of the program, and is entitled I'Group Process and and Interpersonal Communication Strategies for Teachers." Seven specific interpersonal skills are seen as necessary for effective group processing. Interaction analysis is defined as the ability 119 to assess the speaker's message as cognitive or content-oriented or else affective, in which case the feelings of the speaker are of more concern than the actual content of the message. The listener should be able to respond to the content and the affect of the speaker's message. Related to this skill is active listening which involves maintaining a focus on the sender and paraphrasing and show- ing perceptions tentatively. The listener actually facilitates the exploring of the cognitive and affective messages of the speaker. Exploratory questioning is the third IPL skill. This questioning is to further explore and clarify the speaker's message. Questions should be non-cued, i.e., questions which do not impose the listen- er's values on the speaker. The observation skill acts to clarify various non-verbal behaviors. The objective is for the listener to recognize and interpret diverse modes of non-verbal communicating, i.e., hands, face, postures, gestures, etc. Observation skill is considered an important tool for active listening. Self-disclosure is the fifth IPL skill. It means an awareness of one's ideas, opinions, and feelings, and a communication of that awareness. The giving and receiving of positive and negative feedback is the sixth IPL skill. This is considered an important skill and is develOped at some length in this paper concerning the difference between responsible and irresponsible feedback. The intent of responsible feedback is constructive rather than hurtful; it should deal with a specific behavior and not be an indictment of the whole person; it should deal with the consequences of the specific behavior in question, and, finally, the feedback should be timely. Value 120 awareness is the seventh IPL skill. Value awareness or clarifica- tion involves checking for consistency or inconsistency in the value systems we profess in words or express in behavior. Finally, IPL participants are expected to learn the process model for personal and interpersonal growth. The four steps in the model are assessment, goal setting, strategies, and evaluation. First, an assessment is made of what skills need working on and then realistic goals are set for improving that skill. Strategies are then identified to meet the goal. Finally, the individual evaluates his progress toward his goals. In a further analysis of the process model in a group situ- ation (in this case, teacher and students), Dr. Lopis suggests the teacher take into account the trust or risk level of the group. He suggests that the teacher model all of the interpersonal skills and also that the level of group comfort or rapport is significant to the perceived risk in working on affective goals. He suggests that the first couple of group sessions be devoted to establishing group rapport and then there should be opportunity for the group to receive feedback from the leader. The skills should be presented and worked on in order of risk, with those involving the least risk first. Feedback is considered a high risk skill. Later on in the life of the group, the leader should reduce the amount of structure in the group so that the students can learn to be self-directive, and responsible for their own growth. Initially, the leader may use simulated encounter situations, but eventually the leader will capitalize on the real-life encounters that occur during the group or class, and resolve them using the appropriate skills. Dr. Lopis 121 is concerned that the IPL groups not be confused with sensitivity or encounter groups, toward which he feels negatively. He cites the teaching of the particular interpersonal communication skills in the IPL sessions as quite different from the completely unstructured group process often found in encounter or sensitivity groups. Par- ticipants were also required to pass three written exams taken from appropriate textbook and carrel listening programs. Although the terminology and the emphasis is sometimes dif- ferent, the IPL groups, based on their own written theoretical orien- tation, suggest the importance of Empathy (active listening), Self- Disclosure, Non-verbal Behavior, Group Climate, Group Structure, Risk Taking, Confrontation, and Concreteness. There is no direct emphasis placed on Respect, Trust, Genuineness, and Immediacy as is the case with this research. Paradoxically, the variable found to be most directly related to participant benefit in the IPL groups is the communication of Respect, which is not included as one of the impor- tant interpersonal communication skills to be taught in the IPL groups. Self-Disclosure, Non-verbal Behavior, Risk Taking, Confrontation, and Concreteness are taught but are unrelated to any measures of partici- pant benefit in this study. Immediacy, Genuineness, and Trust which are related to parti- cipant benefit are also not directly taught. Only Empathy, Group Cli- mate, and Group Structure are directly taught in the IPL groups and are all shown to be related to participant benefit as rated by the participants, observers, or friends. A first-hand account of how ob- servers perceived the group leaders is given in Appendix F. 122 Suggestions Concerning IPL Structure Based on informal feedback sessions between the author and the observers and the author and the group leaders (the author did not personally observe any of the groups), the author will offer certain suggestions about the IPL group format as a basis for explaining the lower correlations between the benefit measures with Self-Disclosure, Non-verbal Behavior, Risk Taking, and Confrontation. However, the strong possibility that these variables are less rele- vant to participant benefit cannot be lightly dismissed. A possible explanation for the low overall predictive power of Self-Disclosure, Non-verbal Behavior, Risk Taking, and Confronta- tion is that the IPL groups were not as "deep" or "intimate" as other interpersonal communication groups. Education 200 is a required course for a large number of students who probably expected a more traditional course. Many probably would not have chosen to join an IPL group voluntarily. The strong emphasis placed on more traditional, didactic teaching allowed for less time for personal and interpersonal exploration. The emphasis on passing written tests and behaviorally displaying the correct interpersonal communication skills to the satisfaction of the group leader could lead to feelings of competition between students and a fear of failure. Risk taking in the form of self-disclosure of deep, personally meaningful experi- ences, ideas, and feelings, or confrontation of personal differences, were behaviors that were not observed as frequently or intensely by the observers. 123 The participants significantly (p_< .OOl) linked their bene- fitin IPL groups with Risk Taking, Self-Disclosure, and Confrontation. However, in a rank order of magnitude of correlation with benefit, they placed these variables 9, 10, and 11, respectively. The observers explained to the author that often it was hard to judge participant Self-Exploration (SE) in the IPL groups as much group time was used for a didactic presentation on communication skills by the leader or preparing for a written test. The mean score for SE, which is a part of the IAC, was 2.75 with a standard deviaion of 1.00. The mean score for the seven predictor variables covered by the IAC were between 3.44 and 5.50 with a standard deviation between 1.41 and 1.79. The observers rated the level of participant SE lower than any of the seven predictor variables of leader effective- ness. SE did correlate significantly with three predictor variables from the IAC (Self-Disclosure [r_= .35, p_< .OOl], Immediacy [r_= .34, p_< .001], and Confrontation [r_= .31, p_< .OO4]). These three variables were not perceived by the observers as significantly related to participant behavior. Self-Disclosure, Confrontation, Non-verbal Behavior, Risk Taking, and participant Self-Exploration are either infrequent activities in IPL, or activities associated with participant dissatis- faction. In order to explore interpersonal communication problems at a deep individual level these activities need some work. Perhaps the intimacy required by these variables is inappropriate in a teaching context or perhaps they could be worked on more effectively in an advanced IPL group. 124 Future Research Some cautions must be exercised in interpreting these results. Correlation is not causation. High correlation indicates only that as scores on one variable go up, scores on the other vari- able also increase. Only a strict experimental design could estab- lish for certain that a leader who is seen as respectful actually gau§g§_participants to have a satisfactory experience, or to exhibit a high degree of interpersonal competence. The best experimental design would call for the leader to exhibit behavior designed to be respectful in one group and to behave in an opposite manner in another group. Each predictor variable could be systematically examined in this manner. However, given the real-life function of the IPL groups, this would be unethical, if not impossible. Perhaps such an experimental design could be simulated in the laboratory. The interpretation offered here concerning the relative unimportance of certain variables, i.e., Self-Disclosure, Risk Tak- ing, Confrontation, and Non-verbal Behavior is only an hypothesis that requires further testing. It would be beneficial to compare the benefit of participants in the IPL groups to interpersonal communication groups unconstrained by such traditional structures as grades, tests, lectures, and emphasis on mastering communication skills step-by-step. Less didactically structured groups might be freer to establish relationships characterized by more intimate Self-Disclosure, Self-Exploration, Confrontation, Risk Taking, and Non-verbal Behavior. 125 An objective measure of participant benefit would also be valuable. Observer, participant, and friend ratings of participant benefit are insightful but lead to discrepancies and uncertainties in assessing true member benefit. An objective, behavioral measure of benefit would be more conclusive. The inability to measure any change on the ICL is disappointing. Significant correlations between gains on the LOV/DOM factors and participant benefit and leader effectiveness variables would be reasonably objective evi- dence relating the leader variables to benefit. Perhaps if the observers or group leaders had rated the group members pre- and post-test on the ICL, greater differences in the LOV/DOM means would have been obtained along with significant correlations with leader effectiveness and member benefit. The present findings affirm the importance of the Hurley studies relating leader acceptance of members to encounter group member benefit (ARO). Future research should examine further the importance of leader acceptance or rejection of himself to group member benefit (SAR). Hurley's data suggest that this dimension is equally important. Conclusion It appears essential that the leader of a group teaching interpersonal communication skills be perceived as being Respectful, Genuine, Immediate, and Trustworthy in his relationship to the participants. Leader Empathy is valuable but seems a little less important. The Group Climate must be perceived as supportive and 126 cohesive. The relationship between these leader variables and par- ticipant benefit is agreed upon by both observers and participants. Observers believe the group leader should be low on the structure they initiate for the group. They prefer the group activities to be created directly from the needs of the participants. The partici- pants perceive Group Structure to be unrelated to benefit from the group. Data from friends of the participants reveal leader Respect is the best predictor of member benefit. Leader Concreteness, Self- Disclosure, Confrontation, Risk Taking, and Non-verbal Behavior appear unrelated to participant benefit. APPENDICES 127 APPENDIX A PARTICIPANT AND OBSERVER QUESTIONNAIRES 128 APPENDIX A PARTICIPANT AND OBSERVER QUESTIONNAIRES A.l. Group Member Benefit Questionnaire (Observer) The purpose of this questionnaire is to ascertain your feel- ings about each participant's learning in and satisfaction with his/ her IPL group. Write the number which best describes how you feel next to each statement. If you use the full range of the seven-point scale as much as possible, your answers will be more informative. Includeymnu~rating of the participant's Self-Exploration and the leader's ratings on the seven scales of the IAC. ‘ Q) Q 3 5 (2) ® Strongly isagree Slightly No Slightly gree Strongly Disagree Disagree Opinion Agree Agree Student Number TA Code Number Section #____ (l) He/she has basically positive feelings about his/her IPL group. (2) (3) (6) (7) (8) He/she exhibits a good understanding of his/her own feelings and behavior. Group interaction in the IPL had a negative or destructive effect on him/her. He/she exhibits a good sensitivity to other people's feelings and behavior. His/her interactions were generally facilitative of positive IPL group functioning. In my estimation, he/she took full advantage of the IPL group experience. The IPL group process often seemed confusing to him/her. His/her interactions had a negative or inhibiting effect on the IPL group process. 129 130 (9) He/she demonstrates a good understanding of the communication skills necessary to be a good group leader. (10) Group interaction in the IPL often seemed superficial or boring to him/her. (ll) He/she was an active member in the IPL group process. (12) He/she was a constructive member in the IPL group process. A.2.. Group Member Benefit Qgestionnaire (Friend) This questionnaire is part of the data that will be used in a research project on participant satisfaction in the Ed. 200 IPL' group program. The purpose of this questionnaire is to ascertain your feelings about your friend's attitudes and behaviors since his/ her participation in the IPL groups. Circle the number which best describes your feelings. If you use the full range of the seven- point scale as much as possible, your answers will be more informa- tive. Please do not discuss your answers with your friend until you have returned the questionnaire. 1 2 3 4 5 6 7 Strongly Disagree Slightly No Slightly Agree Strongly Disagree Disagree Opinion Agree Agree Student # of participant TA code # (l) He/she has expressed basically positive feelings about his/her IPL group experience. (2) He/she has gained a greater understanding of his/her own feel- ings and behaviors. (3) He/she has recommended that his/her friends take the IPL course or a similar course. (4) Group interaction in the IPL often seemed negative or destruc- tive to him/her. (5) He/she has gained a greater understanding of other people's feelings and behavior. (6) As a result of his/her participation in the IPL group it is easier for him/her to relate to people outside of his/her group. (7) He/she has a greater understanding of the processes that facili- tate group functioning. 131 (8) He/she feels better about himself/herself as a result of his/her IPL group experience. (9) The IPL group process often seemed confusing to him/her. (10) He/she has gained a greater understanding of the processes that inhibit group functioning. (ll) He/she has a greater understanding of the communication skills necessary to be a good group leader. (12) Group interaction in the IPL often seemed superficial or boring to him/her. (l3) He/she was an active participant in the IPL group process. (14) He/she was a constructive participant in the IPL group process. A.3. _Leader Behavior Questionnaire (Participant and Observer) This questionnaire is designed to assess effective leader- ship in small groups designed to teach positive interpersonal commu- nication skills. Mark the number on your IBM answer sheet which best describes how you feel about your IPL group or you TA. If you use the full range of the seven-point scale as much as possible your answers will be more informative. 0N your IBM answer sheet fill in the box marked student num- ber. In the box marked Day_fill in your TA's code number. Fill in your responses to the 60 items of the Leader Behavior Questionnaire in boxes l-60. Fill in your responses to the Group Member Benefit Questionnaire (Participant) in boxes 61-74. In box 75 put the grade you expect to receive in Ed. 200. Fill in l for pass or 2 for no- credit. - l 2 3 4 5 6 7 Strongly Disagree Disagree No Slightly Agree Strongly Disagree Slightly Opinion Agree Agree Please write in your student # and TA code # on this sheet also. Student Number TA Code Number (1) My TA helped me add significantly to my understanding of my own feelings and behaviors. (2) My TA communicated to me a deep caring about my feelings, expe- riences, and potential. (3) (4) (5) (6) (7) (8) (9) (10) (ll) (12) (13) (14) (15) 132 I felt my TA was sincere and spontaneous in his interactions with me regardless of whether his response to me was positive or negative. My TA rarely volunteered personal information about himself. My TA generally kept group discussion on a strictly abstract and intellectual level. My TA helped me understand my own feelings and experiences by relating to me his own feelings and experiences in similar circumstances. At times my TA confronted me directly with discrepancies between his perception of me and my own perception of me. My TA was willing to disclose and work on his own personal prob- lems in the group. My TA sometimes suggested non-verbal exercises for group members to do. I would describe the general group climate as being supportive. Most group meeting time was structured by the TA. My TA helped create a climate of "psychological safety" in which I could feel comfortable exposing my feelings and trying new behaviors. I believe my TA to be supportive and non-rejecting of my real self. I would describe my TA as an active participant rather than as an observer of the group. Group meetings were often characterized by hostility, passive observers, or objections to group norms or activities. (16) My TA was uncomfortable maintaining steady eye contact in groups. (17) My TA was one of the first group members to express strong (18) positive feelings towards another group member. I do not feel that my TA confronted me in a sensitive or per- ceptive manner. (19) My TA did not really establish a personal, one-to-one relation- ship with most group members. (20) (21) (22) (23) (24) (32) (33) (34) (35) (36) (37) 133 My TA involved me in detailed discussion of my own personal feelings, experiences, and behaviors regardless of their emo- tional content. My TA disclosed enough information about his own experiences and feelings to reveal himself as a unique individual. I sometimes felt my TA was being defensive in his interactions with me. I often felt my TA was judging my attitudes and behaviors too harshly. My TA communicated back to me a minimally acceptable under- standing of the affect and content of my communications in the group. My TA often did not understand my frame of reference. I feel my TA has mostly warm feelings about me. My TA often responded according to his "prescribed role" rather than expressing what he personally felt. My TA sometimes volunteered intimate personal information that might be embarrassing to him if revealed to an outsider. My TA was quite adept at guiding general discussions to spe- cific feelings and experiences of personally relevant material. My TA was one of the first group members to express strong nega- tive feelings towards another group member. My TA never confronted me with discrepancies in my attitudes or behavior. I felt comfortable sharing with my TA my own personal feelings about him. My IPL group has actually formed together in a group circle or embrace to express mutual caring. I generally felt secure and free enough to be myself at group meetings. My TA often used structured exercises to teach effective inter- personal communication skills. My TA has been completely open and honest with me. I do not trust my TA enough to share my real feelings with him. 134 (38) The subject matter of group meetings was often spontaneously created by the interactions and activities of group members. (39) Most group members genuinely like one another. (40) My TA was one of the first group members to touch or embrace another group member. (41) At times my TA confronted me with discrepancies between my "real" self and my "ideal" self. (42) My TA shared with me his own personal feelings about me. (43) Our TA had us discuss real feelings, but these discussions were mostly abstract and intellectual. (44) Personal feelings and experiences revealed by our TA were often irrelevant or destructive in relationship to the needs of the group members. (45) I felt my TA was usually congruent in what he said and what he appeared to be feeling. (46) My TA helped provide me with opportunities to express myself in a positive manner to the group. (47) My TA did not seem interested in me. (48) My TA helped me learn to effectively discriminate and communi- cate back the affective tone and content of another's message. (49) My TA sometimes expressed his feelings towards a group member by appropriate, spontaneous, physical contact. (50) I sometimes felt my TA was ignoring me. (51) I generally felt my TA was deeply and freely himself during group meetings. (52) My TA seemed open to examining the most difficult areas of his own experience. (53) Reflections and interpretations offered by our TA concerning individual attitudes or behaviors were concrete, specific, and understandable. (54) We (the TA and myself) did not attempt to analyze or understand our own immediate relationship. (55) At times my TA confronted me with his perception of me that was more positive than my own perception of myself. 135 (56) My TA was generally not willing to disclose himself at a deep emotional level to the group. (57) I find that postures, bodily positions, gestures, and tone of voice are excellent indicators of what a person is really thinking or feeling. (58) Whether working together or playing together, all group members were active and constructive participants at least part of the time. (59) My TA used mostly didactic techniques (lecture/film/discussion/ reading) to teach interpersonal communication skills. (60) I do not trust my IPL group enough to share my real feelings with them. A.4. Group Member Benefit QuestionnaireA(Participant) The purpose of this questionnaire is to find out how you feel about your IPL group experience. Mark the number of your IBM answer sheet that best describes how you feel. If you use the full range of the seven-point scale as much as possible your answers will be more informative. 1 2 3 4 5 6 7 Disagree Disagree Disagree No Slightly Agree Strongly Strongly Slightly Opinion Agree Agree (61) My IPL experience has enhanced my awareness and understanding of my own feelings and behavior. (62) I have a greater understanding of the processes that facilitate group functioning. (63) I feel better about myself as a result of my participation in the IPL group experience. (64) The IPL group process was Often confusing to me. (65) As a result of my IPL experience it is now easier for me to relate to people outside my group. (66) I have a greater understanding of the processes that inhibit group functioning. (67) I have a greater understanding of the communication skills necessary to be a good group leader. 136 (68) Group interaction in my IPL was often superficial or boring. (69) I can honestly recommend the IPL course to my friends as a valuable learning experience. (70) My IPL experience has enhanced my awareness and understanding of other people's feelings and behaviors. (71) Group interaction in my IPL was often negative or destructive. (72) I have basically positive feelings about my IPL group experi- ence. (73) I feel I was an active participant in the IPL group process. (74) I feel I was a constructive participant in the IPL group process. (75) What is your expected grade for Ed. 200? Mark 1 for pass. Mark 2 for no credit. APPENDIX B SCORING GROUPS FOR LEADER BEHAVIOR QUESTIONNAIRE 137 APPENDIX B SCORING GROUPS FOR LEADER BEHAVIOR QUESTIONNAIRE Empathy: 1, 24, 25, 47, 48 Respect: 2, 23, 26, 46, 50 Genuineness: 3, 22, 27, 45, 51 Self-Disclosure: 4, 21, 28, 44, 52 Concreteness: 5, 20, 29, 43, 53 Immediacy: 6, 19, 32, 42, 54 Confrontation: 7, 18, 31, 41, 55 Risk Taking: 8, 17, 30, 56, 4O Non-verbal Behavior: 9, 16, 33, 49, 57 Group Climate: 10, 15, 34, 39, 58 Group Structure: ll, 14, 35, 38, 59 Trust: 12, 13, 36, 37, 60 138 APPENDIX C ANALYSES OF VARIANCE 139 APPENDIX C ANALYSIS OF VARIANCE C.l.--ANOVA to examine interrater reliability with observers nested in group members (GMBQOBS). Total Score for GMBQOBS Group Members Observers l 2 3 4 5 6 7. 8 9 l 58 64 13 56 51 47 57 6O 56 2 57 49 33 51 53 53 56 6O 57 3 62 55 34 39 41 43 55 60 55 4 45 48 43 54 53 46 62 54 57 5 59 50 36 53 49 40 56 53 62 ANOVA Variable Hypothesis Mean Square [- p_< Group members 192.07 3.71 .01 Observers/ 51.79 group members df = 8,32 140 V1 All" 141 C.2.--ANOVA to examine interrater reliability with observers nested in group leaders (LBQOBS). Twelve Variables of LBQOBS Variable Observers l 2 3 4 5 6 7 8 9 10 ll 12 Leader 1 Observer 1 20 23 20 15 21 17 21 16 23 25 23 23 2 24 23 24 17 24 20 22 16 21 21 21 24 3 25 23 21 17 21 21 24 16 24 23 22 22 4 21 20 l6 17 18 ll 15 l4 19 23 25 21 Leader 2 Observer l 25 24 27 21 23 22 23 21 26 25 25 25 2 24 23 25 19 24 20 22 18 21 21 23 24 3 25 26 26 22 26 22 25 15 25 23 23 23 4 27 27 23 27 26 26 19 26 23 27 25 21 ANOVA Variable Hypothesis Mean Square f_ p_< Leader/LBQOBS 5.27 2.38 .03 Leader/LBQOBS/ 2 2] Observers , ' df = 11,48 142 C.3.--ANOVA for groups nested in leaders. Respect Scores on LBQPAR Leaders I III VII Groups 1 2 3 4 5 6 7 8 9 10 ll 12 l3 14 l 25 12 34 21 35 31 30 32 34 31 27 35 21 34 a: :3 E; 2 26 27 26 23 24 26 28 35 3O 30 23 29 28 28 can .§.'? 3 29 23 33 23 31 20 26 27 34 32 26 17 29 30 U44 3E §_ 4 30 29 28 26 28 30 19 28 31 16 25 23 22 31 Mm “:2 5 3022302031272924342519242923 6 25 29 30 291 31 31 24 29 26 27 28 33 31 32 7 24 27 29 35 28 22 29 34 32 22 32 28 11 28 ANOVA Variable Hypothesis Mean Square .5 p_< Groups/Leaders 50.3 6.3 .001 Participants/Groups/ 8 0 Leaders ' df = 7,84 APPENDIX D CORRELATION MATRICES 143 144 ._ mgzmwm mmme mm S R mm 2 R 8 mm 8 8 S 8 E 8 Jfiwmezu ex on e_ on mm “P am e_ mm mm mm mm mm Aeev e“ we m_ mm mm mp up eP m_ mm em mm m, Ammv on we we mm om we mm mm mm N_ mm mm mm Amev m_ me we Fm em on m_ em mm ON mm mm m, A_NV om mm mm Fm Fe em oe em em op mm mm we Aoev em am ow em _e me mm ON me om mm me we Amev N_ m_ we mm em me om em on em mm em _e Amev mm N_ mm m_ oe mm om ee mm _N am me em “Nev e_ e_ mm em em om em ee _e m_ mm me om Aeev mm m_ mm mN em me mm mm Fe mo mm me am Amev mm mm N_ ON e_ om 0N _N mp mo NO Nm op Remy mm KN mm mm em mm mm ow mm mm No we em Amev mm mm mm mm mm me em me me we mm Ne mm “Nev mm m. an m_ we me _e em on mm op mm mm A_ev Aeev Amev ANNV “Fey Aoev Amev Amev Remy Aeev Amev Aeev Amev Amev Apev e.Amm_ u.av mz OO ON OO OO OO OO NO OO OO OO ON OO OO OO ON OO OO OO OO PO ON NO OO OOOOO OO NO OO NO OO OO _O OO OO OO NO OOeeN OO O_ NO NO NO OO NO OO OO NO NO OOOO OO NN NO OO OO NO OO OO OO NO NN OO NO OO OO ON OO ON OO OO OO OO NO OOO ON OO OO NO OO _O OO NO NO OO OO OOOO NO OO OO NO ON OO NO NO NN NO OO OeO OOONO oOeOO ONOO O>z NO OOOOO OoeeO OOOO OO OOO OOOO OOO «..Ama— u 171; ”Emma..— .._.o mm—nmwgw> LOHUwUmLQ NF Low xwprE cowpwmeLOUe—mucmil.m.o 148 .O mgzmmu mam.O ON OO OO _O OO OO OO OO OO ON ON OOOOO O, OO- OO- N_- OO NN OO OF OO N_ .O OONO OO OO- OO N_ NN OO O_ ON OO OO OO OOOOO OO NO- OO OO OO OO NF N_ OO OO NO O>z PO __- NN OO OO NO ON OO N_ NO OO OO OO OO N_ OO OO OO NO OO OO ON NO OOOOO OO NN OO OO NO OO OO OO NO OO NO OOeeN OO OO O_ N, O_ NO OO NN OO HO OO OOOO OO O_ ON N_ OO OO OO NN PO OO OO OO OO OO OO OO _N OO NO OO NO OO OO OOO ON N. OO OO N. ON OO _O OO OO OO OOOO ON _O OO NO OO _O NO OO OO OO OO OEO OOOOO EOOO OoOeOO O>z OO OOOOO OOeeO OOOO OO OOO OOOO OeO «..ANN u 113 mmoomA $0 mmpnmem> LopuwumLQ N_. Low. XPLHME comeFwL500LchHll.m.o 149 D.7.—-Intercorrelation matrix for seven predictor variables of IACOBS (N_= 72).* Emp Resp Gen 50 Conc Immed Confr Emp 42 3O 4O 56 32 46 Resp 42 45 28 19 Ol 11 Gen 30 45 so ' 29 1o 37 SD 40 28 50 20 26 40 Cone 56 19 29 20 50 66 Immed 32 01 10 26 50 59 Confr 76 ll 37 40 66 59 *See Figure 5. APPENDIX E REGRESSION ANALYSES 150 151 E.l.--Stepwise multiple regression summary table for Type 1 to Type 4 with LBQOBS.* (See Figure 2.) Step Var. E_ Sig. Mult. R R2 R2 Change Simple R Type 1 With LBQOBS 1 Clim 16.409 .000 .436 .190 .190 .436 2 Struc 6.676 .012 .511 .261 .071 .233 3 NVB 2.881 .094 .540 .291 .030 .018 4 50 1.639 .205 .555 .308 .017 .015 5 RISK .512 .477 .560 .314 .005 -.086 6 COHfF .765 .385 .467 .322 .008 -.005 7 Trust 1.150 .288 .578 .336 .012 .324 8 Resp 1.087 .301 .587 .345 .011 .272 9 Gen .227 .636 .589 .347 .002 .234 10 Emp .130 .720 .590 .349 .002 .228 11 Immed .059 .809 .591 .349 .001 .102 12 COHC .042 .838 .591 .350 .000 .222 lype 2 with LBQOBS 1 Clim 15.595 .000 .427 .182 .182 .427 2 Immed 9.565 .003 .531 .282 .100 -.143 3 Gen 3.796 .056 .565 .320 .038 .268 4 Resp 3.525 .065 .595 .354 .034 .180 5 Trust 1.417 .238 .606 .367 .014 .253 6 Emp 2.278 .136 .623 .389 .021 .012 7 NVB .661 .419 .628 .395 .006 .109 8 R1Sk 1.590 .212 .640 .410 .015 -.205 9 Struc 1.361 .248 .650 .423 .013 -.143 10 50 1.169 .284 .658 .433 .011 .031 11 Confr .525 .471 .662 .483 .005 -.068 12 Cone .261 .611 .664 .441 .002 -.069 Type 3 Wlth LBQOBS 1 NVB 21.796 .000 .487 .237 .237 .487 2 Struc 14.274 .000 .607 .368 .131 .358 3 Clim 13.377 .000 .687 .472 .104 .460 4 Resp 1.302 .258 .694 .482 .010 .309 5 Risk .903 .345 .699 .489 .007 .246 6 Gen .350 .556 .701 .492 .003 .201 7 Emp .263 .610 .703 .494 .002 .428 8 Confr .038 .847 .703 .495 .000 .349 9 Immed .014 .905 .703 .495 .000 .400 Type 4 with LBQOBS ” 1 Clim 50.702 .000 .648 .420 .420 .648 2 Struc 6.498 .013 .686 .470 .050 .172 3 Confr 3.222 .077 .703 .494 .024 .014 4 $0 .672 .415 .706 .499 .005 .083 5 Immed .817 .369 .711 .505 .006 .241 6 Trust .402 .528 .713 .508 .003 .442 7 Emp .330 .568 .715 .511 .003 .316 8 Gen .220 .641 .716 .512 .002 .287 9 Resp .062 .804 .716 .513 .000 .424 10 RISK .021 .886 .716 .513 .000 -.054 {E for some variables below default and not entered in analysis. 152 E.2.-—Stepwise multiple regression summary table for learning and satisfaction with LBQPAR;* (See Figure l.) Step Var. 5 Sig. Mu1t. R R2 R2 Change Simple R Learning with LBQPAR l Resp 92.013 .000 .570 .325 .325 .570 2 Clim 13.999 .000 .609 .371 .046 .495 3 Emp 8.210 .005 .631 .398 .026 .474 4 Struc 3.669 .057 .640 .409 .012 -.227 5 Cone 4.352 .038 .650 .423 .013 .442 6 NVB 3.347 .069 .658 .433 .010 .422 7 Immed .704 .402 .659 .435 .002 .475 8 Confr .877 .350 .662 .438 .003 .277 9 Gen .403 .526 .662 .439 .001 .358 10 SD .564 .454 .664 .441 .002 .292 11 Risk .888 .347 .656 .443 .003 .241 12 Trust .150 .699 .666 .444 .000 .503 Satisfaction with LBQPAR l Resp 93.294 .000 .523 .328 .328 .573 2 Clim 21.305 .000 .629 .396 .068 .534 3 Emp 6.329 .013 .665 .415 .020 .459 4 Struc 3.892 .050 .654 .427 .012 -.023 5 Risk 1.671 .198 .658 .432 .005 .297 6 SD .793 .374 .659 .435 .002 .237 7 Gen .649 .422 .651 .437 .002 .414 8 Trust .573 .450 .662 .439 .002 .468 9 Conc .267 .606 .663 .439 .001 .337 10 Immed .090 .765 .663 .440 .000 .406 ll NVB .068 .794 .663 .440 .000 .347 12 Confr .048 .827 .663 .440 .000 .252 *f_for some variables below default and not entered in analysis. 153 E.3--Stepwise multiple regression tables for all participant benefit and leader effectiveness measures. 2 Step Var. §_ Sig. Mult. R R R2 Change Simple R GMBQPAR with LBQPAR 1 Resp 124.558 .000 .628 .395 .395 .628 2 Clim 24.957 .000 .682 .455 .070 .559 3 Emp 9.707 .002 .701 .491 .025 .510 4 *Struc 4.982 .027 .710 .504 .013 -.023 5 Risk 1.577 .197 .713 .509 .004 .312 5 NVB .449 .504 .714 .510 .001 .404 7 Trust .097 .755 .714 .510 .000 .523 8 Immed .105 .745 .714 .510 .000 .482 9 Conc .051 .805 .715 .511 .000 .407 10 Confr .050 .807 .715 .511 .000 .305 11 SD .051 .821 .715 .511 .000 .284 12 Gen .033 .855 .715 .511 .000 .407 GMBQOBS with LBQOBS 1 Struc 55.412 .000 .595 .483 .483 .595 2 Clim 9.915 .002 .740 .548 .055 .200 3 Immed 5.443 .023 .753 .582 .034 .151 4 Trust 2.025 .159 .771 .594 .012 .540 5 Emp 1.870 .175 .778 .505 .011 .337 5 Resp .792 .377 .781 .510 .005 .445 7 Gen .383 .538 .782 .512 .002 .371 8 Confr .384 .538 .784 .514 .002 .072 9 SD .254 .515 .785 .515 .002 .137 10 Risk .214 .545 .785 .517 .001 - 052 11 Conc .033 .857 .786 .518 .000 .221 GMBQOBS with IACOBS 1 Resp 23.101 .000 .498 .248 .248 .498 2 Immed 1.882 .175 .518 .268 .020 -.137 3 Conc 1.921 .170 .537 .288 .020 .143 4 Gen 1.199 .277 .548 .301 .013 .335 5 Emp .547 .424 .555 .308 .007 .159 5 SD .079 .779 .555 .308 .0 .153 7 Confr .034 .854 .557 .309 .000 .038 E.3.--Continued. 154 Step Var. f_ Sig. Mult. R R2 R2 Change Simple R GMBQOBS with LBQPAR 1 Gen 6.927 .010 .261 .068 .068 .261 2 Struc 1.635 .204 .290 .084 .016 .111 3 SD 1.360 .246 .312 .097 .013 .037 4 Clim .503 .480 .319 .102 .005 .154 5 Emp .254 .615 .323 .105 .003 .101 6 NVB .410 .524 .330 .109 .004 .029 7 Risk .155 .694 .332 .110 .002 -.011 8 Trust .090 .765 .333 .111 .001 .137 9 Resp .209 .649 .336 .113 .002 .171 10 Conc .217 .642 .340 .115 .002 .085 GMBQFR with LBQPAR 1 Immed 9.102 .003 .296 .087 .087 .296 2 Gen 4.175 .044 .355 .126 .039 .291 3 Trust 1.111 .295 .370 .137 .010 .160 4 NVB .657 .420 .378 .143 .006 .179 5 Confr .363 .549 .382 .146 .003 .129 6 Clim .506 .479 .388 .151 .005 .211 7 Conc .364 .548 .393 .154 .003 .230 8 Struc .331 .567 .397 .157 .003 -.O41 9 Risk .336 .564 .401 .169 .003 .198 10 SD .379 .540 .405 .164 .004 .159 11 Emp .164 .686 .407 .166 .002 .088 12 Resp .038 .847 .408 .166 .000 .226 GMBQPAR with LBQOBS 1 Confr 4.529 .053 .508 .258 .258 .508 2 Struc 1.719 .214 .593 .351 .093 .408 3 Resp 3.546 .086 .714 .509 .158 .173 4 Immed 3.650 .085 .800 .641 .131 .214 5 Risk 2.160 .176 .843 .710 .070 .415 6 Gen 2.033 .192 .877 .769 .059 .425 7 SD 1.145 .320 .895 .801 .033 .438 8 NVB 1.190 .317 .913 .834 .033 .423 9 Trust 1.032 .356 .929 .863 .028 .327 10 Clim 1.332 .313 .947 .897 .034 .066 ll Conc 1.697 .284 .967 .934 .037 .062 12 Emp .201 .698 .970 .940 .006 .350 E.3.--Continued. 155 2 Step Var. f_ Sig. Mult. R R R Change Simple R GMBQFR with OBQOBS l Struc 2.074 .173 .371 .138 .138 .371 2 Confr .802 .388 .438 .192 .054 -.146 3 Resp .931 .355 .505 .255 .063 .306 4 NVB 1.208 .297 .579 .335 .080 .336 5 Gen 1.771 .216 .667 .444 .109 .162 6 Conc 2.891 .127 .769 .592 .148 -.O45 7 Clim 3.104 .121 .847 .717 .125 -.207 8 SD .539 .491 .861 .741 .023 .129 9 Trust 1.411 .288 .893 .798 .057 .136 10 Risk 6.885 .059 .962 .926 .128 -.011 ll Immed .308 .617 .966 .933 .007 .010 12 Emp .132 .751 .968 .937 .004 .088 GMBQPAR with IACOBS l Resp 2.921 .111 .428 .183 .183 .428 2 Confr .719 .413 .479 .230 .046 .334 3 Emp 3.020 .110 .629 .396 .166 .207‘ 4 Immed 4.892 .051 .770 .594 .199 -.108 5 SD .793 .396 .792 .627 .033 .314 GMBQFR with IACOBS l Resp 8.936 .010 .638 .407 .407 .638 2 50 2.023 .180 .702 .493 .085 .023 3 Immed .484 .501 .717 .514 .021 -.O47 4 Confr 1.507 .248 .760 .578 .064 .026 5 Emp 1.242 .294 .793 .629 .051 .287 6 Conc .826 .390 .815 .664 .035 .055 7 Gen .078 .789 .817 .667 .004 .058 *f_for some variables below default so not entered into analysis. APPENDIX F AN OBSERVER'S REPORT 156 APPENDIX F AN OBSERVER'S REPORT The Leader's Impact on Interpersonal Groups By Mike Farris For the past ten weeks, a study has been conducted to research several aspects of group dynamics and how groups are affected by the behavior of the group leader or facilitator. This was done by observing approximately 20 groups offered as an introduc- tory class for education majors. The observations primarily focused upon the facilitator, and were based on the following criteria: empathy, self-disclosure, respect, genuineness, concreteness, imme- diacy, and confrontation. This seven-part scale is drawn from Robert C. Carkhuff's "Index of Accurate Communication," found in his book Helping and Human Relations. It is Carkhuff's contention that all of these qualities must be present at a minimally acceptable level in the facilitator in order for him to be considered competent or effective. His research has borne this out; the groups that were rated most helpful had group leaders that scored highest on these seven indices. At the outset of the study, the observers were familiarized with the "Index of Accurate Communication," and we strove to reach 157 158 a point of standardization within Carkhuff's framework. After our ratings appeared to be consistent with one another, we each observed approximately six to ten groups, visiting each one at least twice. The most striking observation I have found has been the extent to which the character or function of the group is dependent upon the behavior of the leader. In a sense this is Carkhuff's con- clusion, although not exactly. I didn't find a group leader to be ineffective or incompetent if he appeared to lack strength in one or two of the aforementioned categories. I did, however, notice that the character of the group was affected in what seemed to be direct relation to the balance of skills within the leader. Throughout the term, I rated Group Leader A highest in gen- uineness, respect, and self-disclosure and lowest in confrontation and immediacy. Concreteness and empathy fell between the others. I observed two of his groups a total of four times, and noticed several striking similarities between them, all of which seemed to be in relation to Leader A's behavior. Both groups seemed open and willing to share each other's ideas and feelings, but always respectful of the limits each individual set for himself. Decisions about what activities would take place, what topics to discuss, and how the group Spends its time were, for the most part, made by consensus. There was a patience that seemed to exist and a willingness to spend as much time as was needed for everyone to feel comfortable. Leader A's influence came through gentle prodding and suggesting, to which the group was very receptive. The groups remained at a low-intensity level, and confrontations usually took the form of "giving helpful 159 and constructive feedback to one another." Discussing feelings was accepted and encouraged in both groups, although the feelings were not often pursued at any great length, and these discussions most often led to cognitive conclusions. It seemed to be a group com- mitted to each other's growth and well-being, yet always careful to keep the pace comfortable. Leader B scored lowest in respect, self-disclosure, and genuineness, highest on confrontation and immediacy, with empathy and concreteness falling in the middle ground. I observed one of his groups twice. For the first group meeting, Leader B arrived one and one-half hours late for the three-hour group meeting. This was after I had noticed that members came in at various intervals for the first half hour of the group. When questioned about his late arrival, Leader B stated that he had been in the hallway the entire time, and had planned his late entrance in order to find out whether or not the group would "do anything without him." It seemed that several of the group members were angry with him, although this was apparent only non-verbally; anger was never discussed in the group. The group members consistently raised their hands and waited for Leader B to call on them before they Spoke. Leader B responded to confrontation by redirecting the focus onto the confronter and, by the second observation, there seemed to be a marked decrease in willingness to confront. Members no longer pursued confrontation--they seemed to accept their own blame as confronters. Group decisions were primarily arrived at by dictum; this was accepted by the group. 160 At one point in the second meeting, a group member was talk- ing about a painful experience he had had and began crying while he was talking. Leader B immediately went over to him and put his arm around the group member, while the rest of the group observed. He continued talking for approximately one hour, and the group did not take part in this at all, except to ask a few clarifying ques- tions. The group member repeatedly apologized for breaking down and for taking up group time. He received no support from the group. Leader B's high confrontiveness and immediacy coupled with his low scores on genuineness and self-disclosure seemed to keep him in the focus as a judge or authority a great part of the time. This combination of skills seemed to stifle assertiveness and initiative within the group; members remained comfortable in the knowledge that support, confrontation, and structure would come from the leader. Leader C's strongest areas seemed to be concreteness and genuineness. His lowest scores were confrontation, empathy, and respect. Self-disclosure and immediacy fell in the mid-range. I observed one of this groups three times. This group was unique in that it was the only group that was hostile to the study and to having observers present. There was one group member (Member C) who seemed to dominate throughout, and he seemed unusually hostile and controlling. In this group, a balance of skills may have been more essential than in the others; at least the leader's confrontation skills needed to be more in line with his higher skills in order for the group to have benefited from the group experience. There seemed to be a support system set up among this dominant group member, 161 another male group member (Member 0), and the leader. Member C was often sarcastic and mocking of other group members, to which Member 0 and Leader C usually responded with laughter. At one point, Mem- ber C attempted to foil the present study by incorrectly filling out one of the forms required. When I asked him about this, he became hostile and refused to take part in the study any longer. The rest of the group followed suit, one by one, until the entire group had decided to break its contract and withdraw from the study. The dis- cussion of this issue lasted approximately an hour, with Leader C laughing frequently throughout, and treating the subject lightly. Not once did Leader C confront Member C or the group on following his precedent. Because of this lack of confrontation, it iS doubt- ful that the group will be aware of Member C's control or their compliance to it. As a matter of fact, Leader C had several ways of positively sanctioning Member C's behavior. For example, the group participated in an exercise in which members filled out and discussed forms which asked questions such as "Who do I feel closest to?" and "Who makes me happy?" Leader C's responses to both these questions were “Member C." This seemed to impart the idea that Member C's behavior is acceptable and that the group's reaction to him is unacceptable. In the past ten weeks, many of these different leader behaviors and group reactions have become apparent to me. Leader A, whose group I felt quite comfortable observing, seemed to have the combination of skills that is most conducive to growth and develop- ment of the group. The atmosphere was safe and comfortable yet 162 exploratory and insightful. Leader B's skill combination seemed to provide the least amount of opportunities for growth in the group. For the most part, initiative and growth in group members seemed to detract from Leader B's prominence, which was accepted and valued by the group. Leader C seems to be a special case and points to the need for flexibility in a facilitator. The need to adapt to the needs of the particular group or situation was most strikingly shown in this case; a much higher level of confrontation may have been use- ful for Leader C in this group. These are but a few examples of how different leader behaviors influence the character and development of a group. These particular incidents are the most striking I have seen in the past ten weeks. There are many otherS--perhaps less dramatic, but equally profound. BIBLIOGRAPHY 163 BIBLIOGRAPHY Anthony, N., and Carkhuff, R. R. The effects of rehabilitation counselor training upon discrimination, communication, and helping attitudes. Rehabilitation Counseling Bulletin, 13(4), 1970, 333-342. / Antonnuzzo, R., and Kratochvil, O. Level of functioning and written responses to written and verbal stimulus expressions. Unpub. research, State University of New York at Buffalo, 1968. Argyris, C.‘ Explorations in interpersonal competence--I. Journal of Applied Behavioral Science, 1965, 1, 58-84. Arnold, William R., and Stiles, B. A summary of increasing use of "group methods" in correctional institutions. International Journal of Group Psychotherapy, 1972, 22, 77-92. Bach, G. R. The marathon group: Intensive practice of intimate interaction. Psychological Reports, 1966, 18, 995-1002. Bader, Ellen. Towards a taxonomy of effective group leader behavior. Unpub. M.A. thesis, Michigan State Univ., 1973. Bass, B. M. Leadership, Psychology, and Organizational Behavior. New York: Harper, 1960. Benne, K. 0. History of the T-group in a laboratory setting. In L. 0. Bradford, J. R. Gibb, and K. D. Benne (eds.), T-Group Theory and Laboratory Method. New York: Wiley, 1964. Benne, K. D. Deliberate changing as the facilitation of growth. In N. G. Bennis, K. D. Benne, and R. Chin (eds.), The Planning of Change. New York: Holt, 1961, pp. 230-34. Berenson, B. G., and Mitchell, K. M. Therapeutic conditions after therapist initiated confrontation. Journal of Clinical Psychology, 1968, 24, 363-64. Berenson, B. G., Mitchell, K. S., and Moravec, J. A. Level of thera- pist functioning, patient depth of exploration, and type of confrontation. gpyrnal of Counseling Psychology, 1968, 15, 136-39. 164 “IIIP'IIIV‘JJI'V .1 1 165 Berne, E. Principles of Group Treatment. New York: Oxford Univer- sity Press, 1966. Berzon, B., and Solomon, L. N. Research frontiers: The self directed therapeutic group: three studies. Journal of Coun- seling Psychology, 1966, 13, 491-97. Birdwhistle, R. L. Introduction to Kinesics. Louisville: Univer- sity of Kentucky Press, 1952. Bjerstedt, A. Preparation, process, and product in small group interaction. Human Relations, 1961, 14, 183-89. Blake, R. R., and Mouton, J. S. The instrumented training laboratory. In I. R. Weschler and E. H. Schein (eds.), Issues in Train- ipg, Washington, D.C.: National Training Laboratories, 1962, pp. 61-76. Bolman, Lee. Laboratory vs. lecture in training executives. Journal of Applied Behavioral Science, 1970, §(3), 323-25. Bolman, Lee. Leader effectiveness in group dynamics. In C. L. Cooper (ed.), Group Dynamics Training in Management Education. In press. Borman, J., and Kinney, D. Correlational bias in observer ratings. Journal of Personality and Social Psycholong 1976, 3(2), 263-73. Boyd, J. D., and Ellis, J. 0. Findings of research into senior management seminars. Toronto: The Hydroelectric Power Commission of Ontario, 1962. Bradford, L. P., Gibb, J. R., and Chin, R. (eds.). T-Group Theory and Laboratory Method. New York: Wiley, 1964. Brown, R. Social Psychology. New York: The Free Press, 1965. Buchanan, P. C. Evaluating the effectiveness of laboratory training in industry. In Explorations in Human Relations Training and Research, #1. Washington, D.C.: National Training Laboratories, National Education Association, 1965. Bunker, D. R. Individual applications of laboratory training. Journal of Applied Behavioral Science, 1965, 11, 131-48. Campbell, J. P., and Dunnette, M. 0. Effectiveness of group experi- ences in managerial training and development. Psychological Bulletin, 1968, 19, 73-104. 166 Campbell, 0. T., and Fiske, D. W. Convergent and discriminant validation by the multitrait-multimethod matrix. Psycho- logjeal Bulletin, 1959, §§(2), 81-105. Carkhuff, R. R. Helping and Human Relationships: A Primer for Lay and Professional Helpers. Vol. 1. Selecting and Training. Vol. 2. Practice and Research. New York: Holt, Rinehart and Winston, Inc., 1969. Carkhuff, R. R., and Berenson, B. G. Beyond Counseling and Therapy. New York: Holt, Rinehart, and Winston, Inc., 1967. Carkhuff, R. R., Kratochvil, D., and Friel, T. The effects of pro- fessional training: The communication and discrimination of facilitative conditions. Journal of Counseling Psychology, 1968, 15, 67-74. Carkhuff, R. R., and Truax, C. B. Concreteness: A neglected vari- able in research in psychology. Journal of Clinical Psy- chology, 1964, 11, 67-71. Caron, T. J. Human relations training and attitude change: A vector analysis. Personnel Psychology, 1964, 11, 403-24. Cartwright, D., and Zander, A. Group Dynamics, 3rd ed. Evanston, 111.: Row, Peterson, 1968. Cattell, R. B. New concepts for measuring leadership in terms of group syntality. Human Relations, 1951, 4, 161-84. Chelune, G. J. Self-disclosure: An elaboration of its basic dimen- sion. Psychological Reports, 1975, pp, 79-85. Coch, L., and French, J. R. P., Jr. Overcoming resistance to change. Human Relations, 1948, 1, 512-32. Collins, B. E., and Raven, R. H. Group structure: attractions, coalitions, communication, and power. In G. Lindsay, and E. Aronson (eds.), Handbook of Social Psychology, Vol. 4, 2nd ed. Reading, Mass.: Addison Wesley, 1969, 102-205. Cosby, Paul C. Self disclosure: A critical review. Psychologjcal Bulletin, 1973, 19(2), 73-91. Cooper, C. L., and Mangham, I. L. T-Groups. A Survey of Research. New York: Wiley-Interscience, 1971. Cronbach, L. J. Coefficient alpha and the internal structure of tests. Ppychometrika, 1951, 1p, 297-334. 167 Culbert, 5. Trainer self disclosure and member growth in two T- groups. Journal of Applied Behavioral Science, 1967, 4, 47-74. Deep, S. 0., Bass, B. M., and Vaughn, S. A. Some effects of busi- ness gaming of previous quasi T-group affiliations. Journal of Applied Psychology. 1967, 51(5), 426-431. Egan, Gerard. Encounter: Group Processes for Interpersonal Growth. California: Wadsworth Publishing Co., 1970. Fox, R. S., and Lippit, R. The innovation of classroom mental health practices. In M. B. Miles (ed.), Innovation in Edu- cation. New York: Columbia University Teacher's College, 1964. Pp. 271-97. Fleishman, E. A. Leadership climate, human relations training, and supervisory behavior. Personnel Psyehologx. 1956, p, 205-22. Fleishman, E. A., Harris, E. F., and Burtt, H. E. Leadership and Supervision in Industry. Columbus: Ohio State University Press, 1955. Foa, V. G. Convergences in the analysis of the structure of inter- personal behavior. Psychological Review, 1961, pp, 341-353. French, J. R. P., Jr. Field experiments changing group productivity.“ In J. G. Miller, ed., Experiments in Social Process: A Symposium on Social Psychology. New York: McGraw Hill, 1950. Pp. 79-96. French, J. R. P., Israel, J., and As, D. An experiment in participa- tion in a Norwegian factory. Human Relations, 1960, 1;, 3-9. Freidenberg, E. Z. Coming of Age in America. New York: Random House, 1963. Freundlich, D. A psychoanalytic hypothesis of change mechanisms in an encounter group. International Journal of Group Psycho- therapy, 1972, 22, 42-53. Friedlander, F. The impact of organizational training laboratories upon the effectiveness and interaction of ongoing work groups. Personnel Psychology. 1967, 29, 289-307. Gibb, C. A. The sociometry of leadership in temporary groups. Sociometpy, 1950, 1;, 226-43. 168 Gibb, C. Leadership. In G. Lindsay and E. Aronson (eds.), Hand- book of Social Psychology, Vol. 4, 2nd ed. Reading, Mass.: Addison Wesley, 1969, pp. 205-83. Gibb, J. R. The effects of human relations training. In A. E. Bergen and S. L. Garfield (eds.), Handbook of Psychotherapy and Behavior Change. New York: Wiley, 1970, pp. 2114-76. Gittes, G., and Black, H. IS self-disclosure self-revealing? Journal of Counseling Psychology, 1976, 21(4), 327-32. Goodstein, L. D., and Reinecker, V. M. Factors affecting self- disclosure: A review of the literature. In B. H. Maher (ed.), Progress in Experimental Personality Research, Vol. 7. New York: Academic Press, 1974. Greenbaum, S. Discrimination of written responses by low level com- municators. Unpub. research, State University of New York at Buffalo, 1968. Greenberg, B. The differential ratings of counselor responses to client expressions in a written, audio, and real-life form. Unpub. research, State University of New York at Buffalo, 1968. Guilford, J. P. Factors that aid and hinder creativity. Teachers College Record, 1962, 63, 380-82. Gunther, B. Sense Relaxation: Below Your Mind. New York: Collier, 1968. Haase, R. F., and Tepper, D. T., Jr. Nonverbal components of empathic communication. Journal of Counseling Psychology, 1972, 12(5), 417-24. Hall, E. T. Silent assumptions in social communication. Disorders of Communication, 1964, 42, 41-55. Hall, E. T. The Hidden Dimension. Garden City, N.Y.: Doubleday, 1966. Halpin, A. W. The leader behavior and effectiveness of airplane com- manders. In R. M. Stogdill and A. E. Coons (eds.), Leader Behavior: Its Description and Measurement. Columbus: Ohio State University Press, 1955, pp. 52-64. Halpin, A. W. The leader behavior and leadership ideology of edu- cational administrators and airplane commanders. Harvard Educational Review, 1955, gp, 18-32. 169 Halpin, A. W., and Winer, B. J. The Leadership Behavior of the Airplane Commander. Columbus: Ohio State University Research Foundation, 1952. Harrow, M., Astrachan, B. M., Becker, R. E., Milli, J. C., and Schwartz, A. Influence of the psychotherapist on the emo- tional climate in group therapy. Human Relations, 1967, 20, 49-64. Helmreich, R., Bakeman, R., and Scherwitz, L. Study of small groups. Annual Review of Psychology, 1973, 24, 337-54. Hemphill, J. K. Leader Behavior Description. Columbus: Ohio State University Personnel Research Board, 1950. Holland, J. L. Creative and academic performance among talented adolescents. Journal of Educational Psychology, 1961, pg, 136-47. Hollander, E. H. Principles and Methods of Social Psychology; 2nd ed. New York: Oxford University Press, 1971. Hollander, E. P., and Webb, W. B. Leadership, followership, and friendship. Journal of Abnormal Social Psychology, 1955, 59, 163—67. 3 Homans, G. C. The Human Group. New York: Harcourt, Brace, and World, 1950. House, R. J. T-group education and leadership effectiveness. A review of the empirical literature and a critical evaluation. Personnel Psychology. 1967, 29, 1-32. Hurley, J. R. Some effects of trainers on their T-groups recon- sidered. Journal of Applied Behavioral Science, 1975, 11(2). 191-96. Hurley, J. R., and Force, E. J. Interpersonal competence in 1- groups: A multi-observer, multi-method assessment. Inter- national Journal of Group Psychotherapy, 1973, g§(2), 166-76. Hurley, J. R., and Force. E. J. Two Prepotent Interpersonal Dimen- sions and the Effects of Trainers on T-groups. Small Group Behavior, 1976, 1, 77-98. Jacob, P. E. Changing Values in College. New York: Harper, 1957. Jourard, S. M. The Transparent Self. Princeton: Van Nostrand, 1964. 170 Jourard, S. M. The effects of experimenter self disclosure on subject's behavior. In C. Spielberger (ed.), Current Topics in Community and Clinical Psychology. New York: Academic Press, 1969. Jourard, S. M., and Friedman, R. Experimenter-subject distance and subject self disclosure. Journal of Personality and Social Psychology, 1970, 1p, 278-82. Klein, 0. C. Sensitivity training and community development. In E. H. Schein and W. G. Bennis (eds.), Personal and Organi- zational Change Through Group Methods. New York: Wiley, 1969, pp. 184-200. Kratochvil, 0. Changes in values: Interpersonal functioning of counselors in training. Counselor Education and Supervision, 1969, 8, 104-7. La Barre, W. Paralinguistics, Kinesics, and Cultural Anthropology. In T. H. Sebeck, A. S. Hayes, and M. C. Bateson (eds.), Approaches to Semiotics. The Hague: Mouton, 1964, pp. 191- 200. La Forge, R., and Suczek, R. F. The interpersonal dimension of per- sonality. III. An interpersonal checklist. Journal of Personality, 1955, 24, 94-112. Lange, P. E. Validation of the orthagonal dimensions underlying the ICL and the octant constellation assumed to be the measure. Journal of Projective Techniques and Personality Assessment, 1970. a, 519-27. Leary, Timothy. Interpersonal Dipgnosis of Personality. New York: Ronald, 1953. Lewin, K. Field theory and experiment in social psychology: Con- cepts and methods. American Journal of Sociology, 1937, 44, 868-96. Likert, R. New Patterns of Management. New York: McGraw Hill, 1961. Lippit, R., and White, R. K. The "social climate" of children's groups. In R. J. Barker, J. S. Kienen, and H. F. Wright (eds.), Child Behavior and Development. New York: McGraw Hill, 1943. pp. 485-538. Lohman, K. Zenger, J. H., and Weschler, I. R. Some perceptual changes during sensitivity training. Journal of Educational Research, 1959, pg, 28-31. 171 Lowin, A. Participative deicsion making: A model, literature cri- tique, and prescription for research. Organizational Behavior and Human Performance, 1968, 3, 68-106. Luke, R. A., Jr. The internal normative structure of sensitivity training groups. Journal of Apined Behavioral Science, 1972. 3(4), 421-37. McGregor, D. The Human Side of Enterprise. New York: McGraw Hill, 1960. McQuitty, L. L. Elementary linkage analysis for isolating orthogonal types and typal relevancies. Education and Psychological Measurement, 1957, 11, 207-21. Mann, R. D. A review of the relationship between personality and performance in small groups. Psychological Bulletin, 1959, 56, 241-70. March, J. G., and Simon, H. A. Orgenizations. New York: Wiley, 1968. Marquis, D. G., Guetzkow, H., and Heyns, R. W. A social psycho- logical study of the decision making conference. In H. Guetzkow (ed.), Groups, Leadership and Men: Research in Human Relations. Pittsburg: Carnegie Press, 1951, pp. 55- 67. Marrow, A. J., Bowers, D. G., and Seashore, S. E. Management by Participation. New York: Harper and Row, 1967. Marrow, A. J., and French, J. R. P., Jr. A case of employee par- ticipation in a nonunion sh0p. Journal of Social Issues, 1946. 2. 29-34. Marshall, K. F., and Coleman, A. D. Operant analysis of encounter groups. A pilot study. International Journal of GrouprPsychotherapy, 1974, 35(1), 42-54. Martin, J. C., and Carkhuff, R. R. The effects of training on personality and behavior. Journal of Clinical Psycholggy, 1968, 24, 107-110. Maslow, A. H. Motivation and Personality. New York: Harper, 1954. Maslow, A. H. Toward a Psychology of Being, New York: Van Nostrand, 1962. Miles, M. B. Changes during and following laboratory training: A clinical experimental study. Journal of Applied Behavioral Science, 1965, 1, 215-42. 172 Miles, N. B. Human relations training: Processes and outcomes. Journal of Counseling Psychology. 1960, 1, 301-6. Morse, N. C., and Reimer, E. The experimental change of a major organizational variable. Journal of Abnormal Social Psy- chology, 1956, 33, 120-29. Muehlberg, N., Pierce, R., and Drasgow, J. A factor analysis of therapeutically facilitative conditions. Journal of Clinical Psychology, 1969, 33, 93-95. Oaklander, H., and Fleishman, A. E. Patterns of leadership in hospital settings. Administrative Science Quarter1y, 1964, 8, 521-32. Parloff, M. B. Group therapy and the small group field. Inter- national Journal of Group Psychotherapy, 1970, 33, 267-304. Peters, 0. R. Identification and personal change in laboratory training groups. Unpublished Ph.D. thesis, Alfred P. Sloan School of Management, M.I.T., 1966. Pierce, R., Carkhuff, R. R., and Berenson, B. G. The different effects of high and low functioning counselors upon coun- selors in training. Journal of Clinical Psychology, 1967, 23, 212-15. Pinches, S. K. Linkages between'Fgroup trainer personality vari- ables and T-group effectiveness. Unpublished M.A. thesis, Michigan State University, 1972. Pollack, H. B. Homogeneous and heterogenous T-groups. Journal of Consulting and Clinical Psychology, 1971, 31, 60-66. Proshansky, H., and Seidenberg, B. Basic Studies in Social Psy- chology. New York: Holt, Rinehart, and Winston, 1965. Psasthas, G., and Herdert, R. Trainer interventions and normative patterns in the T group. Journal of Applied Behavioral Science, 1966, 3, 149-70. Rogers, C. R. The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting Psychology, 1957, 21, 95-103. Rogers, C. R. On Becoming a Person. Boston: Houghton Mifflin, 1961. Rogers, C. R., and Truax, C. B. The therapeutic conditions antecedent to change: A theoretical view. In C. R. Rogers (ed.), Ipe_ Therapeutic Relationship and ItS Impact. Madison: The University of Wisconsin Press, 1967, pp. 97-108. illllllll‘lfll 7.! 173 Rogers, C. R. Interpersonal relationships: USA 2000. Journal of Applied Behavioral Science, 1968, 3(3), 265-80. Romanoff, B., Lewittes, D., and Simmons, W. Accuracy of "Significant others'" judgments of behavior change. Journal of Coun- seling Psycholpgy, 1976, 33(5), 409-13. Roy, 0. Quota restriction and gold bricking in a machine shop. American Journal of Sociology, 1952, 31, 427-42. Runyan, D. The group risky shift effect as a function of emotional bonds, actual consequences, and extent of responsibility. Journal of Personality and Social Psychology, 1974, 33(5), 670-76. Rush, C. H., Jr. Leader behavior and group characteristics. In R. M. Stodgill and A. E. Coons (eds.), Leader Behavior: Its Description and Measurement. Columbus: Ohio State Univer- sity Press, 1955, pp. 52-64. Sales, S. M., and Rosen, M. A. A laboratory investigation of the effectiveness of two industrial supervisory patterns. Unpub- lished manuscript, Cornell University, 1965. Salzburg, H. C. Verbal behavior in group therapy with and without a therapist. Journal of Counseling Psychology, 1967, 11, 24-27. Schein, B. H., and Bennis, W. D. (eds.). Personal and Organizational Change Through Group Methods: The Laboratory Approach. New York: Wiley, 1965. Scott, E. L. Perceptions of Organizations and Leadership Behavior. Columbus: Ohio State University Research Center, 1952. Seashore, S. Group Cohesiveness in the Industrial Work Group. Ann Arbor: University of Michigan, Institute for Social Research, 1954. Seashore, S. E., and Bowers, D. G. Changing the Structure and Func- tion of an Orgenization. Ann Arbor: Survey Research Center, University of Michigan, Monograph #33, 1963. Simonson, Norman R. Impact of therapist disclosure on patient dis- closure. Journal of Counseling Psychology, 1976, 33(1), 3-6. Solomon, L. N., Berzon, B., and Davis, D. P. A personal growth pro- gram for self directed groups. Journal of Applied Behavioral Science, 1970, 3, 427-52. 174 Stimson, Roger. Six conceptual distinctions useful in group encounter. Unpublished manuscript, Michigan State Univer- sity, 1970. Stock, 0. A. A survey of research on T-groups. In L. P. Bradford, J. R. Gibb, and K. D. Benne (eds.), T-Group Theory and Laboratory Method. New York: Wiley, pp. 395-441. Stogdill, R. M. Personal factors associated with leadership. Journal of Psychology, 1948, 25, 35-71. Stogdill, R. M., and Koehler, K. Measures of Leadership Structure and Organization Change. Colombus: Ohio State University Research Foundation, 1952. Stouffer, S. A., Suchman, E. A. DeVinney, L. C., Star, S. A., and Williams, R. J., Jr. The American Soldier, Vol. 1, Princeton: Princeton University Press, 1949. Thibaut, J. W., and Kelley,li.1L The Social Psychology of Groups. New York: Wiley, 1959. Truax, C. B. Therapist empathy, warmth, genuineness, and patient 'personality change in group psychotherapy: A comparison between interaction unit measures, time samples, and patient perception measures. Journal of Clinical Psychology. 1966, 22, 225-29. Truax, C. B., and Carkhuff, R. R. Toward Effective Counseling and Psychotherapy: Training and Practice. Chicago: Aldine Publishing Co., 1967. Tyler, B. B. A study of factors contributing to employee morale. Unpublished M.A. thesis, Ohio State University, 1950. Underwood, W. J. Evaluation of laboratory method training. Train- ing Director's Journal, 1965, 13(5), 34-40. Valiquet, M. I. Individual change in a management development pro- gram. Journal of Applied Behavioral Science, 1968, 3, 313-25. Vansina, L. Research concerning the influence of the T-group method on the formation of participant's social values and openness. In Evaluation of Supervisory and Management Traininngethods. Paris: Organization for Economic Cooperation and Develop- ment. Vroom, V. H. Some Personality Determinants of the Effects of Par- ticipation. Englewood Cliffs, N.J.: Prentice Hall, 1960. 175 Vroom, V. H. Industrial social psychology. In G. Lindsay and E. Aronson (eds.), Handbook of Social Psychology. Vol. 5, 1969, 196-268. Vroom, V. H., and Yetton, P. Leadership and Decision-Making. Pittsburg: University of Pittsburg Press, 1973. Weigel, R., Dinges, N., Dryer, R., and Straumford, A. A. Perceived self-disclosure, mental health, and who is liked in group treatment. Journal of Counseling Psychology, 1972, 13(1), 47-52. Wherry, R. J., and Fryer, D. H. Buddy rating: Popularity contest or leadership criteria? Personnel Psychology, 1949, 3, 147-59. Whitman, R. M. Psychodynamic principles underlying T-group processes. In L. P. Bradford, J. R. Gibb, and K. D. Benne (eds.), T-Group Theoryrand Laboratory Method. New York: Wiley, 1964, pp. 310-35. Whyte, W. F. Money and Motivation: An Analysis of Motives in Industry. New York: Harper, 1955. Yalom, I. D., and Lieberman, M. A. A study of encounter group casu- alties. Archives of General Psychiatry, 1971, 33, 16-30. Yalom, I. D., Lieberman, M. A., and Miles, M. Encounter Groups: First Facts. New York: Basic Books, 1973. I i..:.'n I F4—2 F4 .l.n..'.a. ’. .1. "11111111111111.1111“