WI “WU! W NHIHHMWHIWII WW I 1 ‘40—; IN IN mmw IGAN STATE UNIVE ITY LIBRARIES iiiiii ii millwill 3 1292\‘1 (W683 8850 This is to certify that the thesis entitled GROUP MHMQERS' PERCEPTIONS OF LEADER ATTRIBUTES AND BEHAVIOR AND OUTCOME IN INTERPERSONAL GROUPS presented by Jonathan Neil Weller has been accepted towards fulfillment of the requirements for __M..A,__degree in _Bs¥chology MN. m Major professor Date 7/29/9 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE lN RETURN BOX to remove this checkout from your record. TO AVOID FINE return on or before date due. DATE DUE MTE WE DATE DUE JAN 11. 3990 1/” WM“ GROUP MEMBERS’ PERCEPTIONS OF LEADER ATTRIBUTES AND BEHAVIOR AND OUTCOME IN INTERPERSONAL GROUPS By Jonathan Neil Weller A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1998 ABSTRACT GROUP MEMBERS’ PERCEPTIONS OF LEADER ATTRIBUTES AND BEHAVIOR AND OUTCOME IN INTERPERSONAL GROUPS By Jonathan Neil Weller This study investigated variables that might explain client change in group psychotherapy by looking specifically at leader effects in numerous experiential groups. One hundred ninety-seven (57 male; 140 female) undergraduates who enrolled in an upper-level psychology course (Small Interpersonal Groups for Experiential Learning [SIGEL]), participated in thirty individually led experiential groups. Group members rated themselves and same-group peers on two Acceptance scales (Acceptance vs. Rejection of Self and Others [ARS/ARO]) after 23 and 46 hours of group interaction. Group affiliativeness was measured using the Group Climate Questionnaire at the end of each session. Members also rated their leaders’ behaviors and personal attributes using an adjective checklist. Large positive correlations were found between afliliativeness scores and member perceptions of leaders as Caring, Charismatic, and Skillful (_r= .73, .74, and .63, respectively). Much smaller correlations were found between member changes in acceptance and congruence with the leader ratings given. ACKNOWLEDGMENTS I would like to thank and recognize all those who have given of their time, energy, and knowledge in supporting my completion of this project. I would especially like to express my deep gratitude to Dr. John Hurley for the use of his database from which this study was conducted and for the support, encouragement, and inspiration he provided me. I would like to express further appreciation to Dr. Norman Abeles for his role, as acting chairman, for his guidance, and constructive critique of my work. And would also like to recognize Dr. Bertam Karon for his willingness to be a committee member and for the advice and knowledge he imparted to me during this project. Finally, I would like to acknowledge my wife, Shauna and daughter, Rainier from whom I have frequently received inspiration, love, patience and boosts of determination during the ups and downs of this endeavor. iii TABLE OF CONTENTS LIST OF TABLES ........................................................................ v INTRODUCTION ........................................................................ 1 LITERATURE REVIEW Group Leadership ................................................................ 2 The Influence of Leaders ............................................... 2 Group Member Perceptions of Peer and Leader Behavior ......... 4 Previous Research Findings on Leader Factors ...................... 5 Interpersonal Competence, Congruence, and Measurement ................ 7 RATIONALE AND HYPOTHESES Hypotheses ........................................................................ 1 0 METHOD Participants ........................................................................ l 1 Measures ........................................................................... 12 Procedure ........................................................................... 15 Data Analysis ...................................................................... 17 RESULTS Leader Adjective Measure ....................................................... 18 Hypotheses ........................................................................ 19 DISCUSSION Hypothesis 1 ...................................................................... 22 Hypothesis 2 ...................................................................... 24 Hypothesis 3 ...................................................................... 26 Sources of Error .................................................................. 28 RECOMMENDATIONS Future Research .................................................................. 29 APPENDIX A ............................................................................ 40 APPENDIX B ............................................................................ 42 APPENDIX C ............................................................................ 44 LIST OF REFERENCES ................................................................ 47 LIST OF TABLES Bead Table 1 - Sample-Specific Reliablities of Leader Adjective Measure Scales (N=74). 32 Table 2 - Group Demographic Data and Descriptive Statistics for Averaged Group Members’ GCQ13 and Leader Adjective Measure Ratings ................................. 33 Table 3 - Partial Correlations Between Leader Adjective Measure Scales and Aggregated GCQ13 Scores (Controlling for Number of Sessions) ........................................ 34 Table 4 - Demographic Data for Individual Group Members (N=74) and Descriptive Statistics for Self- and Peer-Based Acceptance Ratings ..................................... 35 Table 5 - Descriptive Data and Intercorrelations of Individual Changes in Acceptance for 74 SIGEL Group Members (Controlling for Number of Sessions) ........................... . 36 Table 6 - Partial Correlations of Individual Changes in Acceptance and Leader Adjective Ratings Given for 74 SIGEL Group Members (Controlling for Number of Sessions). 37 Table 7 - Descriptive Statistics for Measures of Congruence and Changes in Congruencies Between Self and Average Peer-Based Acceptance Ratings 38 Table 8 - Intercorrelations Between Ratings Given on the Leader Adjective Measure and Congruence Between Self and Peer—Based Acceptance Ratings for 74 SIGEL Group Members .......................................................................................... 39 INTRODUCTION With the ever increasing focus on treatment efficiency and managed care in today’s market, demand for group psychotherapy services has and will continue to increase (Burlingame & Barlow, 1996). However, according to a survey reported in American Psychologist (Sayette & Mayne, 1990), graduate level training programs in clinical psychology rarely offer their students significant amounts of training in this useful modality of clinical intervention. So, where are students interested in groupwork to turn for relevant experiences and clinical skill development, when few graduate programs offer much in the way of specific training in group psychotherapy? In addressing group therapist training, Yalom (1995) includes experiential group participation in what he suggests are the minimal essential components of a program to train group therapists. He suggests that it is wise for trainees to have an experience in which they deal with basic group and interactional dynamics before moving on to therapy groups with highly specialized patient groups. At Michigan State University, both undergraduate and graduate students in psychology have had the opportunity to participate in and lead Small Interpersonal Groups for Experiential Learning (SIGEL) for over 20 years, as part of an upper-level, elective psychology course offered by Professor John Hurley. The benefits of this course have been multi-faceted: first, it offers participants an opportunity for personal growth and the development of greater interpersonal competence; second, it has served as an interesting and prolific source of data for the development of and further validation of interpersonal theory (Hurley, 1976, 1978, 1989, 1997a); third, it offers the group leaders the kind of experience mentioned previously as prerequisite to further training in groupwork and opportunity to develop the Skills, attributes, and self-awareness that are implicated in the successful outcome of both experiential and psychotherapy groups. The purpose of this study is to come to a clearer and more specific understanding of these attributes and what effects these have on group outcome. It is hypothesized that many of the same leader attributes and skills that have been suggested as key to the relief of client symptomology in therapy groups will also correlate with member gains in congruence and interpersonal competence in experiential groups. LITERATURE REVIEW Group Leadership The Influence of Leaders Early outcome studies looking at therapist factors often focused on the technique or theoretical orientation of the therapist. In summarizing the results of such research, Yalom (1995) states that leaders’ ideological school meant relatively little in describing their behavior and that outcome could not be predicted by the leaders’ membership in a particular school of ideology. However, the effectiveness of a group was, in large part, a function of its leaders’ behavior. Lieberman , Yalom, and Miles (1973) suggest that leaders have both direct and indirect roles in shaping the outcome of groups. First, the leader attempts to directly facilitate change in members by his/her personal interactions with them. Evidence suggests that group therapists influence the communicational patterns in their groups by modeling specific behaviors such as support and/or self-disclosure (Yalom ,1995). The role of social learning (or the development of basic social skills) cannot be understated, as it is cleme one of the therapeutic factors operating in all forms of therapy groups; though the nature of the skills taught and the process by which they are learned may be different between types of groups. Second, the therapist or group leader indirectly influences change in the group by helping to construct an environment that will be an effective agent of change. A cohesive environment has been cited in the conceptual literature as the “curative factor” in group psychotherapy. While this may overstate the importance of cohesion, closely associated with the cohesiveness of a group is the positivity of the therapist-client relationship, which has also been found to be a substantial predictor of client change in a number of group psychotherapy studies (Neimeyer, Hatter, & Alexander, 1991). The leader serves as a focal person in terms of members’ attraction to the group and the subsequent cohesion developed, this implies that individuals’ attitudes toward the group should be positively and significantly associated with their attitudes toward the leader (Lundgren, 1979) While making efforts to illuminate the subtlety and importance of indirect leader factors, Lieberman, et a1 (1973) did not deny that the leader’s person and behavior loomed large in the course of the group. They stated that the group members often endow the leader and his behaviors with surplus meaning and power. This gives the leader considerable power to help or harm. Lundgren’s (1979) study found that group members’ ratings of the leader were more positive than of themselves or of the group as a whole on a variety of semantic differential rating scales. In general, the findings indicate that members tend to hold highly positive, if not idealized, attitudes toward the trainer during both early and later periods of group history. It seems likely that this pattern of widespread and strong attraction to the trainer provides major leverage in terms of member involvement in the group setting, acceptance of influence from the trainer, and member change. On the negative side, Kaplan, Obert ,and Van Buskirk (1980) discussed the potential for casualties in encounter groups and suggested that the characteristics of the group leader may be one of the background factors that can provide the context for injurious episodes to occur. Group therapists and leaders have within their groups the powerful potential to bring about corrective emotional experiences in ways unavailable to the individual therapist, but in order for this to take place members must perceive the group as a safe and supportive environment and members must be sufficiently engaged and honest with each other when giving feedback in order to allow effective reality testing Yalom (1995). Group Member Perceptions of Peer and Leader Behavior Small groups emphasizing interpersonal skill development offer an encouraging opportunity to study interpersonal behavior and the perceptions that group members develop within them. The issue of who is most suitable to rate an ordinary group member’s behavior is a topic that has received substantial coverage. Thorndike and Hagen (1977) identified the ideal rater as “a person who had a great deal of opportunity to observe the person being rated in those situations in which he would be likely to show the qualities on which ratings are based.” Ratings of group participant’s conduct by aggregated small group peers prove particularly useful, as Lewin and Zwany (1976) found peers’ ratings “superior to all other measures available.” Further research supports these findings as they apply to the leadership of encounter, experiential and therapy groups. Lieberman et. a1 (1973) found that often times very competent leaders are unaware of significant factors that are largely responsible for their success. In their study of perceptions within treatment groups of incest survivors, Neimeyer, et al. (1991) found that clients’ perceptions of other group members were only predictive of outcome when assessed early in therapy and that by the end of therapy the clients’ construing of the therapist became much more important. Yalom (1995) acknowledges that member responses to the group leader will be diverse. As a result of transference, the same therapist will be experienced by different members as warm, cold, rejecting, accepting, competent, or bumbling. In recognizing that this phenomenon will likely also take place between group members and experiential group leaders, the current study decided to focus on leader factors from the perspective of the group members. Previous Research Findings on Leader Factors Reviews on therapist contributions to outcome have concluded that those variables that are specific to or developed within a given treatment relationship are more potent in determining outcome than are global variables developed independently of the therapy (Lafferty, Beutler, & Crago, 1989). A number of such aspects of leader behavior that have been found to impact group outcome are cognitive input, stimulation, transference, leader responsibility, and clinical skills. In a study comparing differences between more and less effective therapists, Lafferty, et a1. (1989) found that therapists’ measured level of empathy was strongly related to their effectiveness. Lending additional support to the importance of leader empathy, Lieberman et al. (1973) stated that leaders “must attempt to view the [group] experience[s] from the experiential position of each of the members.” These researchers and others (MacKenzie et. al, 1987) have also discussed group leader charisma as an influential factor in members outcomes. However, mixed results have been found in this area. Some research has shown charisma and a leader’s overreliance on it to be harmful, while other studies suggest that when paired with empathy and skillfulness it can have beneficial effects. Support and acceptance from the group leaders has also been found in many cases to lead to positive member gains. Therapists, who were found to be less effective, perceived their patients as more involved and themselves as more supportive and directive. They also placed greater value on their own prosperity and stimulation as valued end states, and placed less value on being intellectual. The authors of this research thought that this may have effected the less effective therapists’ ability to refine and develop ideas and skills that could impact patient growth (Lafferty, et al.,1989). Lieberman et. al (1973) concluded that when leaders demanded more expression of emotion through “evocative, challenging, and 9, stimulating behavior” they were more inclined to induce “casualties. Kaplan, et a1. (1980) cite coercion, aggressive stimulation, unrealistic expectations, and passivity, as ingredients of poor group outcome. Both affilliativeness and cohesion within groups have also been found to be positively correlated with group outcome. Highly cohesive groups have greater levels of self-disclosure. Lieberman et. a1 (1973) concluded that much of the personal gains group members achieved were through self-disclosure and expression of positive feelings, m they were accompanied by cognitive insights. These insights and the positive outcomes related to them were found to be facilitated by group leaders who provided conceptual organization for the meaning of individuals’ behavior within the group. By having a conceptual background from which to work, these leaders are able to operate with a degree of comfort and effectiveness not otherwise realized. They are also modeling for the group members the important Skill of organizing their personal experiences. However, leader activity was found to correlate with outcome in a curvilinear fashion, too much or too little activity on the part of the group leader results in poor outcome. The more structured exercises the leader uses the more competent he/she is deemed by the members, but the results at follow-up were less positive. Interpersonal competence. congruence, and measurement In studying the perceptions of self and others within groups, Hurley (1978) suggests Acceptance vs. Rejection of Others (ARO) and Self-Acceptance vs. Rejection (ARS) are the principle dimensions that undergird a wide variety of specific interpersonal behaviors, especially in a psychotherapeutic group context. These two interpersonal dimensions have been considered akin to other popular circumplex model dimensions (Benjamin, 1995; Leary, 1957), such as affiliation-disaffiliation, love-hate or dominance- submission. Affilliativeness has been found to be closely associated with the effectiveness of psychotherapeutic groups, encounter groups, and even psychiatric wards (Hurley, 1989). Hurley further states that empirical evidence has shown the ARC and ARS dimensions to be generally independent lending them to orthogonal representation. When juxtaposed the ARC and ARS scales produce what Berne described as the four basic “life positions” (I’m OK, you’re OK; I’m OK, you’re not OK; I’m not OK, you’re OK; and I’m not OK, you’re not OK). The fourth life position (I’m not OK, you’re not OK) has been identified with psychopathology and the first position (I’m OK, you’re OK) with mental health. Interpersonal competence can be measured along the diagonal line drawn from the least healthy position (I’m not OK, you’re not OK) to the most healthy position (I’m OK, you’re OK) on a graphic, two-dimensional representation of ARO and ARS. It has been found that the more consensually valid a person’s self-images are the more mentally healthy and psychologically integrated they tend to be. One of the most beneficial elements of group therapy and arguably interpersonal groups are there ability to modify distorted views of one’s self through the process of consensual validation—that is, through comparing one’s interpersonal evaluations with those of others. One achieves mental health to the extent that one becomes aware of one’s interpersonal relationships. “Psychiatric cure is the expanding of the self to such final effect that the patient as known to himself is much the same person as the patient behaving to others,” (Yalom ,1995). The Small Interpersonal Groups for Interpersonal Learning (SIGEL) of this study offer an opportunity for the participants to gain valuable feedback of this kind from “knowledgeable peers” not only in the form of group process, but also in the form of quantifiable interpersonal measures like Hurley’s ARO and ARS scales. These scales lend themselves well as outcome measures. Over the last two decades, they have been used successfiilly to link a variety of in-group behaviors to outcome in a number of studies (Hurley, 1976, 1978, 1990, 1997a; Hurley, Feintuch, & Mandell, 1991). RATIONALE AND HYPOTHESES Yalom (1995) states that the most common research strategy for determining therapeutic factors is to correlate in—therapy variables with outcome. Having discovered which variables are significantly related to successful outcome, one can establish a reasonable base from which to delineate the therapeutic factors. He states that both the measurement of outcome and the selection and measurement of in-therapy variables are inherently difficult and problematic. Yalom further encourages researchers to regard the results of past research as guidelines and that much work is still needed to test and deepen the results. This is the intent of the current study. Though the setting for this study was interpersonal groups aimed at developing greater levels of acceptance for self and others, many of the group dynamics and potent leadership factors are thought to be at least similar if not the same, as those existing in psychotherapy groups. If this is in fact true and can be empirically demonstrated, then experiential groups would be excellent training settings for students interested in groupwork, especially in programs where group psychotherapy experiences may not be available. Building on a previous study by MacKenzie, Dies, Coche Rutan, and Stone (1987), the current study will seek to obtain further data regarding leader characteristics linked to positive group outcomes that will be of value to training programs and individuals implementing either psychotherapy or experiential, interpersonal groups. In Hurley’s (1988) critique of the MacKenzie, et al. (1987) study, he pointed out the lack of theoretically well-grounded outcome measures and the difficulty this poses in drawing valid generalizations from this work. Thus, in designing the current study this criticism was addressed by choosing well established measures (group climate questionnaire and acceptance vs. rejection of self and others [ARO and ARS]) from the interpersonal and group theory literature. The purpose of this study was to investigate theorized and minimally researched variables that might explain client change, by looking specifically at leader effects in a large number of experiential groups. Of particular interest was the relationship between group members’ perceptions of their leaders and the individual member and overall group gains achieved. Hypotheses H1: The level of affiliativeness within Small Interpersonal Groups for Experiential Learning, as measured by the aggregate score of the final item on the 13- item version of the Group Climate Questionnaire (GCQ13), will be positively correlated with group leaders being perceived by the group as Caring and Charismatic and negatively correlated with perceptions of them being Inhibiting and Condescending, as measured by the Leader Adjective Measure. H2: Gains or losses in individual ARO and ARS levels, as measured by percentage of possible gain or loss on both self- and peer-based ARO and ARS ratings, will be positively correlated with the group members’ perception ratings of their group leaders on the Caring scale (specifically the accepting, open, genuine, and empathic subscales) of the Leader Adjective Measure and negatively with perception ratings of their leaders on the Controlling scale. H3: Gains or losses in individual, interpersonal Congruence, as measured by changes in the distance between self-based and mean peer-based ARO/ARS ratings plotted on a two-dimensional graph, will be positively correlated with the individuals’ perception ratings of their group leaders on the Skillful, Charismatic (specifically the brilliant, inspiring, and resourceful items of this scale), and Caring scales of the Leader Adjective Measure and negatively correlated with the Inhibiting scale of that same measure. METHOD Participants The participants were 197 ( 57 male; 140 female) undergraduates who were enrolled in an elective, upper-level psychology course (Small Interpersonal Groups for Experiential Learning [SIGEL]) instructed by Professor John Hurley at Michigan State University. This study employed data from all solo-led SIGEL groups since Fall term 1988 to Spring term 1997. These 30 groups consisted of three to eight members, usually juniors or seniors (average age approximately 22 years) plus the one leader. ll The groups to be used in this study were exclusively solo-led, so as to avoid interactive effects of a co-facilitator. Leaders (9 Males; 21 females) ranged in age fi'om 19 to 35 with a mean age of approximately 24 years. For the majority of groups, the leaders were students who had taken the course in the past and exhibited an above average level of interpersonal competence. A majority of these undergraduate student leaders spent a term in preparation and training for leadership. This consisted of direct observation of SIGEL groups, readings and discussions of the small group literature, and participation in an advanced SIGEL group. As this group of leaders were being asked to lead their respective groups alone, a majority of them had also previously co-led at least one group in the past. Occasionally, clinical psychology graduate students led groups. Measures Leader Adjective Measure: A modified version of MacKenzie, et al.’s (1987) Leader Adjective Measure (see Appendix A) was given to and completed by group members within the last three sessions of the term in order to collect descriptions of leader attributes and behaviors. This measure consists of 25 adjectives (loading on five factors) which had been derived from clinical and research reports concerning critical leadership variables. Adjectives on the checklist included such words as “skillful”, “controlling”, and “genuine.” In addition to MacKenzie’s original 24 adjectives, “condescending” was added to the list of words presented to group members. The participants were asked to indicate which adjectives best described their leader by rating how descriptive each item was of him or her on a scale of 0 to 6 (0 being not at all and 6 being extremely). 12 To the knowledge of the researcher, MacKenzie, et al. (1987), had only used this measure once previously. AS such, a confirmatory factor analysis was employed using the SIGEL group data in an attempt to further validate the measure and to see if the same factors apply to this data. MacKenzie, et al. found five factors (Caring, Charismatic, Skillful, Inhibiting, and Controlling) that collectively accounted for 82 percent of the variance in the data they reported. Thirteen-item version of Group Climate Questionnaire(GCQ13): Participants were routinely asked to fill out the 12-item version of the Group Climate Questionnaire (MacKenzie, 1983) extended by one item, “Everything considered, I gained something of value from today’s session.” (see Appendix B), at the close of each regular 90-minute group meeting and at the beginning of the next 90-minute group meeting following the marathons. Each item on the GCQ13 was followed by a response scale anchored at “Not at all—0”, “A little bit—1”, “Somewhat—2”, “Moderately—3”, “Quite a bit—4”, “A great deal—5”, and “Extremely—6”. In a 1997 report, Hurley stated that the 12 basic GCQ items could be factored into two scales reflecting affiliative and disaffiliative tendencies in the group, which could then be collapsed to create a single Affiliativeness composite. This composite scale was found to correlate .80 with the “everything considered...” item. It was further noted that this item correlated at .41 and .44 with the late group (Hr. 46) Self-Acceptance and Other- Acceptance scales, respectively. 9’ Given its direct focus on personal benefits, the “everything considered. . .. item seems a clear outcome index. Its firm link to members’ and leaders’ views of group Affiliativeness mark this composite as a viable alternate measure of outcome. These findings of Hurley’s (1989) study suggest that this single item provides a rough but useful outcome index if conscientiously administered after each group session in a confidential manner. Thus, for simplicity’s sake, only the aggregate of the 13th item on the GCQ13 was used as an outcome measure in this study. Acceptance vs. Rejection of Self & Others (ARS & ARO): Participants and leaders will rate the within-group conduct of each same-group member, including themselves, on eight bipolar subscales presented in a semantic differential format. Self Acceptance (ARS) was assessed by Hides feelings --- Shows feelings, Guarded ---Expressive, Active-«Passive, and Submissive-«Dominant. Acceptance of others (ARO) was assessed by Warrn---Cold, Helps others-«Harms others, Harsh--- Gentle, and Accepts others-«Rejects others (see Appendix C). Ten equally spaced marks separated each pair of anchors, permitting a possible composite score for each of the scales ranging from 0-36 (Hurley, 1997b). All participants were required to complete rating booklets containing these scales twice during the life of each group, following the “marathon” sessions at approximately 23 hours and 46 hours of meeting. The initial scale presented in these booklets, Liked-«Disliked, did not contribute to ARS or ARO, but provided the participants the opportunity to express strong feelings that might otherwise interfere with their ability to rate their peers accurately without a strong emotional bias . The items of the ARO and ARS scales were alternated in the booklet and their more- and less-favorable anchors were staggered to dampen response sets. Instructions within the booklet suggested that the person mark the 14 space between each pair of the bipolar anchors that “best represents your personal impression of each member’s actual behavior within the group sessions up to now.” The instructions further stated that it “will be most useful if you use the full range of possible ratings.” The participants were advised beforehand that all ratings would be fully shared among the same-group members about one week after their completion. (Hurley & Rosenberg, 1991; Hurley, 1997b) Hurley (1989) has been able to support the construct validity of both the ARO and ARS scales. The ARO scale correlates positively and significantly (r=.55, p<.OOl) with both the LOV (love minus hate) factor of LaForge and Suczek’s (1955) Interpersonal Check List (ICL) and the Affiliation factor of Lorr and McNair’s (1965) Interpersonal Behavior Inventory (IBI-4) (r=.74, p<.001) but insignificantly with the orthogonal DOM (dominance minus submission) factor of the ICL and IBI’s Control factor. These two factors correlate positively and significantly with the ARS scale (r=.63 ,p<.001 with Dominance) which correlated very weakly with the ICL’s LOV(r=.18) and the IBI’s Affiliation factors (F.00). Concerning internal consistency, May(1989)reported Cronbach alpha coefficients of .86 and .81, respectively, for ARS and ARO ratings based on group peers’ ratings of individuals. Procedure All data for the present study were extracted from Professor John Hurley’s collection of SIGEL group data for consecutive academic terms from fall of 1988 through the spring of 1997. The course was not offered during summer terms. At each terrn’s initial meeting of the full class, students were assigned to small groups according to the following criteria: a) students were not to be well acquainted with any other member in their group, especially the leader; b) the student’s ability to meet at the scheduled times; and c) balancing the male/female ratio in each group. The SIGEL class met for a one hour lecture and two 90-minute small group sessions weekly as well as two extended l2-hour small group “marathon” sessions usually held after at least 12 and 34 hours of the groups meeting together. SIGEL’S goal is to enhance participants’ awareness of discrepancies between self-images and how they are perceived by knowledgeable peers within groups that blend a high level of mutual acceptance and support with constructive interpersonal confrontations while maintaining a here-and-now focus (Hurley, 1989). The textbook, Integpersonal Living: A S_kills/Contract Approach to Humflelatioqs Training in Groups, by Egan (1976), helps set group norms in relation to issues such as the here-and-now focus, self-disclosure, empathic listening, and constructive confrontation. At the end of each 90-minute session, participants were required to complete the Group Climate Questionnaire (GCQ) (MacKenzie, 1983) with the previously mentioned thirteenth item, “all things considered. . .”, added to it. Following both marathon sessions, participants were asked to rate all other group members and themselves on the ARO and ARS scales. These scales were given to the participants in mini-booklet form to complete on a take-home basis. Printed instructions asked the participants to mark on the scales their “personal impressions of each member’s actual behavior within the group sessions up to now.” The instructions also stated that, “ These ratings will be more useful if you use the full range of possible ratings of each scale.” Prior to their receiving the rating 16 scales participants were informed that the ratings will be shared with all group members at a later time. Three to four sessions prior to the end of the academic term, participants were given the leader adjective checklist and asked to indicate to what degree (on a scale of zero to Six, with six being “extremely” and zero being “not at all”) each of the 25 adjectives described their leader. Data gralvsis Aggregate group ratings on the GCQ13 had been previously collected for all of the participating SIGEL groups. These were correlated with group member mean ratings of the leader on each of the items and the five factors of the Leader Adjective Measure. This produced correlations that reflected the relationship between the averaged group members’ perceptions of the leader's behavior and the level of affiliativeness of the groups climate. Helping to facilitate an affiliative group climate could be considered one of the primary tasks or goals of SIGEL group leaders and evidence suggests that group members’ interpersonal gains, as measured by ARO and ARS, are closely connected to such a climate existing (Hurley, l997a). For every participant in this study, scores were calculated reflecting 23-hour to 46-hour, self- and average peer-rated ARO and ARS shifts (expressed as the percentage of possible gains or losses attained), these attempted to measure gains in Acceptance and Interpersonal Competence. Having employed the ARO and ARS ratings more than once in the groups, we will be able to track the change or lack of change in group members’ perceptions of themselves and others over time. Past observations of such shifts have shown that more constructive groups quite consistently register movement over time toward higher Interpersonal Competence and diminishing discrepancies between ratings received from other members and self-ratings (Hurley, 1978). A graphic, two- dimensional representation of ARO and ARS scores, allowed values to be calculated that determined the amount of change in congruity for each participant by subtracting the two- dimensional distance of ARO/ARS self and ARO/ARS same-group peers’ ratings at 46 hours from the distance at 23 hours. These scores were then converted so as to reflect the percentage of total possible gains or losses achieved. Each of these values were correlated with the individual group member’s ratings of his/her leader on the Leader Adjective Measure to determine the relationship between the individual’s perceptions of his/her group leaders’ and their behavior and the individual’s personal gains in acceptance, congruence, and interpersonal competence. Measuring these specific changes and their relationship to group leader attributes is desirable due to the connection these outcome variables have with well founded psychological theories (Hurley, l997b; Rogers, 1972; and Lundgren, 1979). RESULTS Leader Adjective Mea_sure Since the Leader Adjective Measure was not well established, a confirmatory factor analysis was employed to determine whether the scales, as stated in the MacKenzie, et a1. (1987) article, would fit the current study’s data. The results of the confirmatory factor analysis supported the use of his factor model (Chi-square=527 .5, df=252, p<.001). Having confirmed the existence of these factors, they were then used as 18 scales for the current study and yielded internal consistencies ranging from .89 to .76 (see Table 1 for a list of all scale reliabilities), which seemed reasonable. Hypotheses The results of hypotheses 1-3 are summarized in Tables 3, 6, and 8. For both hypotheses one and two, partial correlations controlling for the number of group sessions were used in order to avoid confounding the results due to more or less hours of interaction with specific group leaders and the use of aggregated data across differing lengths of time. H1: The level of affiliativeness within Small Interpersonal Groups for Experiential Learning, as measured by the aggregate score of the final item on the l3-item version of the Group Climate Questionnaire (GCQ13), will be positively correlated with group leaders being perceived by the group as Caring and Charismatic and negatively correlated with perceptions of them being Inhibiting and Condescending, as measured by the Leader Adjective Measure. The results fully confirmed this hypothesis. Four out of the five Leader Adjective Measure subscales were correlated with the aggregate score from the thirteenth item of the GCQ at statistically significant levels. Both the Caring and the Charisma scales were strongly correlated in a positive direction as was predicted (r=.73 and r=.74, p<.01, respectively). These two scales were found to be highly intercorrelated (r=.8 1 , p<.01) and negatively correlated with both the Inhibiting and Condescending scales. The Inhibiting and Condescending scales were also found to be negatively correlated with the groups’ level of affiliativeness as measured by the GCQ13 (r=-.49, p<.01 and r=-.34, p<.05, respectively). Again, these two scales were also significantly intercorrelated (r=.62, p<.01). Additionally, Skillfulness was also positively correlated with the GCQ13. Interestingly, the Skillful scale was more highly correlated with Charisma than Caring (r=.87 vs. r=.67, both p<.01). Only the Controlling scale did not meet a statistically significant level of correlation with the thirteenth GCQ item, although it was highly intercorrelated with the Inhibiting scale. Overall, these results suggest that group members’ perceptions of a group facilitator are related to the level of affiliativeness achieved in the group. H2: Gains or losses in individual ARO and ARS levels, as measured by percentage of possible gain or loss on both self- and peer-based ARO and ARS ratings, will be positively correlated with the group members’ perception ratings of their group leaders on the Caring scale (specifically the accepting, open, genuine, and empathic subscales) of the Leader Adjective Measure and negatively with perception ratings of their leaders on the Controlling scale. The results partially supported this hypothesis. Gains or losses for averaged peer ratings of ARO and ARS were positively correlated at a statistically significant level with the Caring scale (r=.21 and r=.20, p<.05, respectively). However, shifts in self-ratings of ARO and ARS were not significantly correlated with the Caring scale and the magnitude of the correlations was much lower (r=.06 and r=. 13, p>.05, respectively). Neither shifts in the averaged peer ratings or the self ratings of ARO and ARS were correlated with the open, genuine, and empathic items from the Caring scale individually. Group members’ 20 ratings given to their leaders on the accepting item from the Caring scale were correlated with shifts in the average peer ratings they received for ARS (r=.20, p<.05). Shifts in self and peer based ratings of ARO and ARS were not correlated significantly with the Controlling scale as predicted. In fact, results also showed that the non-statistically significant correlations were in a positive rather than negative direction. The largest statistically significant correlation between ARO/ARS ratings and Leader Adjective Measure scales was found between the ratings given on the Charisma scale and gains in ARO based on averaged peer ratings (r=.37, p<.01). The only other statistically significant correlation was found between gains in self-based ARS ratings and ratings given to leaders on the Inhibiting scale (r=.21, p<.05). H3: Gains or losses in individual, interpersonal Congruence, as measured by changes in the distance between self-based and mean peer-based ARO/ARS ratings plotted on a two-dimensional graph, will be positively correlated with the individuals’ perception ratings of their group leaders on the Skillful, Charismatic (specifically the brilliant, inspiring, and resourceful items of this scale), and Caring scales of the Leader Adjective Measure and negatively correlated with the Inhibiting scale of that same measure. The results did not support this hypothesis. The Skillful, Charismatic and Caring scales were not significantly correlated with shifts in the congruence between self and peer based ratings of ARO and ARS. Further, the brilliant, inspiring and resourceful items were not correlated with changes in member congruence either. In fact, all of these 21 correlations, though not statistically Significant, were found to be negative rather than positive (see Table 8). The Inhibiting scale was significantly correlated with shifts in the members’ levels of congruence between self-based and peer-based ratings of ARO and ARS (r=.27, p<.05). However, this correlation was in the opposite direction from what had been hypothesized. Similarly, the Controlling scale was also positively correlated with congruence shifts (r=.25, p<.05). This is not surprising as the Controlling and Inhibiting scales are highly intercorrelated (r=.68, p<.01). DISCUSSION Hypothesis 1 The confirmation of Hypothesis 1 is indicative of the great influence group leaders have on the development of an affiliative and productive climate within experiential groups. This finding strongly supports previous research (Lieberman, et al., 1973; Dies, 1983; and MacKenzie, et al., 1987) which has emphasized that the perception by members of positive and supportive behaviors from a leader correlates with favorable outcome. In the case of Hypothesis 1, outcome was measured in terms of how strongly group members endorsed having “gained something of value” from their group sessions, which is also related to the affiliativeness of the group climate. That sense of edification was shown to be strongly, positively correlated with group members’ perceptions of the leader as both caring and charismatic and negatively correlated with perceiving the leaders’ behaviors as inhibiting and condescending. 22 The caring scale of the Leader Adjective Measure seems to tap well leader behaviors falling within the domain of the classic Rogerian concepts of acceptance, genuineness, warmth and empathy. Much research has been done concerning these concepts and many studies point to such behaviors as non-specific, but potent therapist variables that contribute to positive outcome. Thus, the relationship found between the Caring scale and the measure of group affiliativeness employed in this study seems well founded. The charismatic scale is comprised of items that seem to relate to the leaders’ directiveness and their ability to motivate and emotionally stimulate the group. The high correlation between the charismatic scale and group affiliativeness corroborates similar findings of MacKenzie et al. (1987) who found that the members of the most successful of fifty-three American Group Psychotherapy Association (AGPA) Institute experientially-based training groups rated their group leaders using more charismatic scale adjectives than the members of less successful groups. However, some authors (Lieberman, et al., 1973; Dies 1983) have cautioned against too much “ therapeutic directiveness” or emotional stimulation citing higher casualty rates in groups where leaders were rated highly on such dimensions. As the results of the present study reflected a strong intercorrelation between the skillful scale and the charismatic scale, it may be that our sample actually viewed the charismatic scale items as reflecting their respective leaders abilities to facilitate interaction and gently guide the direction of the group into helpful directions. 23 The inhibiting scale and condescending item were two of the three negative dimensions tapped by the Leader Adjective Measure. Since, the trust, safety and supportiveness of a group can be greatly compromised by a leader who disrespects group members or who behaves defensively, it was not surprising that the level of affiliativeness in groups was found to negatively correlate with members ratings of their leader on the inhibiting and condescending scales. As development of a supportive, affiliative, open and safe interpersonal space for group members could be conceptualized as one of the primary goals of group psychotherapists and experiential group leaders alike, the confirmation of this study’s first hypothesis lends further support to the notion that experiential groups can be an adequate and useful analogue in which to train students and help them develop a behavioral repertoire applicable to firture therapy group facilitation. Hypothesis; Hypothesis 2 was partially supported by the results, although changes in self- based ratings of acceptance of self and others failed to significantly correlate with any of the hypothesized leader adjective scales. In fact, only group members’ ratings of their leaders as inhibiting (F21) were significantly correlated with any shifts in self-based ratings (in this case self-acceptance). This is likely due to the fact that self-ratings had a higher degree of variability and instability, as compared to averaged peer ratings, since self ratings are based on an N of one. Additionally, it could be that peoples’ perceptions of themselves are more easily changeable than are peer perceptions of them. 24 Shifts in averaged peer ratings of acceptance of self and others were found to be significantly correlated with members’ perceptions of their group leaders as caring, as was hypothesized. However, correlations with more specific items from the caring scale (accepting, Open, genuine, and empathic) only yielded significance between shifts in peer based self acceptance ratings and members’ ratings of their leader as accepting. Hurley (1997) found that shifts in acceptance of self were moderately correlated with an affiliative group climate. In as much as Hypothesis 1 established a link between leader caring and affiliativeness, it seems reasonable that leader caring would also be, as was found, correlated with shifts in members’ levels of acceptance of self and others. The current study’s findings are also similar to Hurley and Rosenberg’s (1990) findings that members’ pooled ratings of their group leaders “warmth” were most strongly correlated with residual gains in acceptance of self and others (.61 and .71, respectively). Leader caring seems intuitively related to leader warmth, though it may be a larger, more inclusive construct. This similarity strengthens this study’s finding that residual gains in acceptance were correlated with leader caring. The more specific correlation between members’ ratings of their leader as accepting and the same members’ individual gains in acceptance seems best explained by behavioral modeling on the part of the leader and social learning on the part of the group members. Though it was hypothesized that group member ratings of their leader as controlling would be negatively correlated with residual member gains in acceptance of self and others, the results failed to confirm this. In fact, while correlations were not statistically significant, they were positive and thus, in the opposite direction of the 25 hypothesis. This finding could be the result of inadequate sample size or some other source of error. However, since the correlations failed to reach statistical significance, they cannot be considered valid without further research or analysis. Hypothesis 3 The results of Hypothesis 3 are somewhat puzzling, as none of the Leader Adjective ratings given by group members to their group leaders were found to correlate in the anticipated directions with the members’ gains in congruence between self and peer based acceptance ratings. In fact, all of the scales correlated in the opposite direction from what was anticipated, though only the correlations for the Inhibiting and Controlling scales reached statistical significance. There are a number of potential explanations for this. First, little prior research has been conducted using the measures employed by this study to look at the interpersonal congruence between self and peer based acceptance ratings. Thus, the hypothesis was not founded on a large body of research and was primarily exploratory in nature. Though correlational directions were predicted, the research reviewed did not provide a sufficiently firm foundation on which to base these predictions were largely based on the investigator’s expectations. Second, the reliability of self ratings are deficiant, since they are based on only one data point at 23 hours and 46 hours (especially when compared to averaged peer- ratings which were typically based on the average of four or more individual peers’ ratings). Subsequently, the self-based portion of the measure of congruence employed in this study were prone to be more volatile than the peer-based portion. Since these two points (self rated ARS/ARO and peer rated ARS/ARO) were being plotted on a two 26 dimensional graph and the distances measured between them in order to determine congruence, the effects of one of these points moving too easily or drastically could have impacted the statistical analyses employed with this data. Finally, an individual’s perception of their experiential group leader may not have any influence over whether they accept or adopt the way other group members view their interpersonal competence (or level of acceptance for self and others). While links may exist between group members’ perceptions of their leaders and the level of group affiliativeness (Hypotheses 1) or individual group member gains in acceptance (Hypothesis 2), these links may be unique to those constructs. Though one of the goals of SIGEL is to provide a forum wherein group members can modify distorted views of themselves through the process of consensual validation, individual group members’ ability to actually modify such distortions may be much less a function of how their group leaders are perceived and more a function of the member’s personality traits, activity level, or the quality of feedback received (to name a few alternatives). Shifts in the level of congruence between group members’ self-appraisals and appraisals of them by their peers is an intriguing way to approach studying outcome in interpersonal groups (and potentially psychotherapy groups), but the evidence presented by this study’s results is not supportive of linking leader factors with such shifts. Only the perception of leaders as Inhibiting and Controlling were positively correlated with gains in congruence at a statistically significant level. It is difficult to explain how or why such a connection would exist. Controlling and inhibiting are descriptors, which would seemingly be used to depict behavior typically viewed as counterproductive to group process and interpersonal development, and yet, these descriptors were found to relate to an increased level of congruence in group members. Caution and skepticism with regard to these specific findings is justifiable and even advisable, as both the Controlling and Inhibiting scales had very skewed and restricted distributions. Thus, these findings may 27 simply be the artifacts resulting from a non-normal distribution of the data on these two scales. If taken at face value, the results of Hypothesis 3 could be conceptualized to represent a phenomenon, wherein less congruent group members might be more likely to idealize their leader (leading to a halo effect). As the members became more congruent, they may have been more realistic in their perception of their leader and rated them accordingly. This would lead the positively oriented scales (Caring, Charismatic, and Skillful) to be negatively correlated and the negatively oriented scales (Inhibiting and Controlling) to be positively correlated with member gains in congruence. It is possible that this did happen to some degree, although it’s likely that group members respond quite differentially to the Small Interpersonal Groups for Experiential Learning (SIGEL) experience, as is similarly encountered with patients in therapy outcome assessment. Thus, further analysis is necessary before clear-cut conclusions can be derived from these results. Sources of Error There are a few sources of potential error that pose a threat to the validity of this study’s results. First, there are a number of issues involving the sample. The sample consisted primarily of college upper-classman in psychology, which leads to some difficulty with regard to the generalizability of the findings. This sample also had approximately two female subjects to every one male, though gender effects were virtually undetectable when partialed out. Of additional concern was that only one third of the potential sample of all members from solo-led groups was able to be used. This 28 was due to the acceptance ratings and Leader Adjective ratings being collected separately, with the Leader Adjective ratings being collected anonymously. Only about one third of the participants’ data was then able to be paired by the SIGEL instructor after original collection. It is unclear whether some systematic error may have entered into the data as a result. However, this sample’s means and standard deviations were similar to other studies that employed the same measures with larger samples of SIGEL participants. Another area of concern is measurement. The variability of self ratings (on ARS and ARO), as was highlighted previously, may have contributed to the puzzling results of Hypothesis 3. Of even greater concern was the Leader Adjective Measure’s susceptability to a couple of response sets. Generally, most group members appeared to have positive perceptions of their leaders. It is hard to tell if a “halo effect” may have been at work, while participants rated their leaders or if the ratings reflect unbiased impressions of the leaders’ behaviors and attributes. Related to this is the issue of certain scales being truncated due to ceiling and floor effects. This is particularly relevant in the case of the Leader Adjective scales and it may have led to deflated correlations and/or spurious relationships in some cases. RECOMMENDATIONS Futge Reseacrch The current study seems to have raised as many questions as it has answered. There is clearly much room for firrther research and analysis in this area. Hypothesis 1 seems to have supported the notion that experiential groups can be a practice ground for graduate student trainees to work on developing their groupwork skills prior to taking on 29 an actual therapy group. An interesting future study could look at whether training and practice in leading experiential groups prior to leading a psychotherapy group can lead to quantifiable improvement in a clinical trainee’s technique and ability to facilitate the development of a therapeutic group climate. Another area where follow-up research could be helpful would be in the further development and improvement of measures that tap group leader (or therapist) behaviors in ways that avoid the “halo effect” and restriction of data. This study reflects in many ways the need to further refine and develop the Leader Adjective Measure. The scaling of this measure allows for a great deal of overlap and intercorrelation between scales, which can lead to confusing results. Future measures could not only attempt to tap perceived leader behaviors, but potentially leader motivations, personality traits and other relevant factors that may impact the development of group members. Finally, of great interest would be’studies that follow-up with experiential group participants to see if gains in interpersonal competence, as perceived by group peers, are maintained and translated into “real world” benefits and aptitudes. Few studies to date have done such follow-up, and yet, it would seem important to evaluate outcome in the long-term, as well as, the short—term. 30 TABLES 31 Table l. Sample-Smcific Reliablities of Leader Adjective Measure Scales (N=74) Scale Ajplg CARING .89 CHARISMA .82 SKILLFUL .76 INHIBITING .77 CONTROLLING .78 32 Table 2. Group Demographic Data and Descriptive Statistics for Averaged Group Members’ GCOl 3:.1nd Leader Adjective Measure Ratings Mean. .82 TOTAL SESSIONS 23.33 4.20 NUMBER OF MEMBERS 6.57 1.14 % OF MALE MEMBERS 29.03 18.31 GCQ13 (aggregated score) 104.28 20.45 Leader Adjective Measure Scales CARING 34.18 4.10 CHARISMA 23.31 4.78 SKILLFUL 15.24 1.71 CONTROLLING 3.39 2.26 INHIBITING 3.37 1.82 CONDESCENDING .62 .61 * Group Climate Questionnaire (l 3th-item only) 33 Table 3. Partial Correlations Between Leader Adjective Measure Scales and Aggregated GCQI3 Scores (Controlling for Number of Sessions) CARING CHARISMA SKILL INHIBIT CONTR CONDE CARING _ CHARISMA .81** SKILLFUL .67** .87M INHIBITING -.62** -.43* -.25 __ CONTROL -.25 .01 .10 .70“ CONDESCEND -.49** -.21 -.07 .62** .46M _ GCQ13 .73** .74** .63** -.49** -.25 -.34* *p<.05, one-tailed test. **p<.Ol , one-tailed test. 34 Table 4. Demographic Data for Individual Group Members (N=74) and Descrpitive Statistics for Self- and Peer-Based Acceptance Ratings M_ea_n $2 Total Sessions 26.01 3.72 Number of Peers 5.68 1.24 Self-Ratings ARO, Hr. 23 27.09 5.35 ARS, Hr. 23 23.34 6.74 ARO, Hr. 46 27.54 5.05 ARS, Hr. 46 25.69 6.56 Avemge Peer firings ARO, Hr. 23 27.19 4.42 ARS, Hr. 23 22.74 6.29 ARO, Hr. 46 27.89 3.75 ARS, Hr. 46 24.42 4.66 35 Table 5. Descriptive Data and Intercorrelations of Individual Changes in Acceptance for 74 SIGEL Group Members (Controlling for Number of Sessions) SelZ-Ratings ARO GAIN 1. % of Possible Gain ARS GAIN 2. % of Possible Gain A veraged Peer-Ratings ARO GAIN 3. % of Possible Gain ARS GAIN 4. % of Possible Gain Mean .45 12 2.35 21 .71 13 1.68 14 S2 4.03 .29 5.17 .31 3.48 .22 4.69 .24 |'-‘ .41" .09 .15 IN lb.) .18 .28“ .62" *p<.05, one-tailed test. * *p<.01, one-tailed test. 36 Table 6. Partial Correlations of Individual Chpnges in Acceptance and Leader Adjective Ratings Given for 74 SIGEL Group Members (Controlling for Number of Sessions) Leader Adjectives l 2 3 £1. 5 .6. Z 8 2 .19 Change Measures Self-Ratings % ARO GAIN .06 .12 .04 -.04 .07 .12 .07 .06 .11 -.17 %ARS GAIN .13 .16 .00 .13 .12 .02 .18 .21* .12 -.06 Average Peer Ratings % ARO GAIN .21* .02 .09 .18 .17 .37** .14 -.18 .16 .04 % ARS GAIN .20* .18 .11 .15 .20* .16 .00 -.18 .09 -.13 *p<.05, one-tailed test. **p<.01, one-tailed test. 1 = Caring; 2 = Empathic; 3 = Genuine; 4 = Open; 5 = Accepting; 6 = Charisma; 7 = Controlling; 8 = Inhibiting; 9 = Skillful; 10 = Condescending 37 Table 7. Descriptive Statistics for Measures of Congruence and Changes in Congruencies Between Self and Average Peer-Based Acceptance Ratings .M_wn S_Q Congruence of Averaged Peer and Self ARO/ARS (23 Hrs.) 6.33 3.59 Congruence of Averaged Peer and Self ARO/ARS (46 Hrs.) 6.44 4.24 Change in Congruence Between 23 and 46 Hrs. -.10 4.65 Percent of Possible Change in Congruence Achieved .16 .34 38 Table 8. Intercorrelations Between Ratings Given on the Leader Adjective Measure and Congruence Between Self and Peer-Based Acceptance Ratings for 74 SIGEL Group Members l .2. 2 fl 1. 23 Hr. Congruence _ 2. 46 Hr. Congruence .30** _ 3. Raw Shift in Congruence .49** -.67** _ 4. % of Possible Gains Achieved .33** -.72** .92" _ Leader Adjective Measure Scales/Subscales CARING -.24* .00 -.18 -.12 CHARISMA -.29** -.05 -.17 -.14 (BRILLIANT) -.24* -.01 -.l8 -.l4 (INSPIRING) -.35** -.15 -.13 -.04 (RESOURCEFUL) -.06 -. l 3 .07 .04 CONTROLLING .13 -.20* .29" .25"' INHIBITING .26* -.12 .32** .27* SKILLFUL -.17 .01 -.14 -.12 CONDESCENDING .1 1 .15 -.05 -.07 * p<.05 (one-tailed test), an: p<-01 (one-tailed test) 39 APPENDIX A 40 APPENDIX A GROUP CLIMATE QUESTIONNAIRE (l3-item Version) Instrpctions: Read each statement carefully and try to think of the group as a whole. Using the Rating Scale as a guide, circle the response to each statement that best describes this group during today’s session. Please avoid omissions and encircle only one response per statement. Name: Date: The members liked and cared about each other. The members tried to understand why they do the things they do, tried to reason it out. The members avoided looking at important issues going on between themselves. The members felt that what was happening was important and there was a sense of participation. The members depended upon the group leader(s) for direction. There was friction and anger between the members. The members were distant and withdrawn from each other. The members challenged and confronted each other in their efforts to sort things out. The members appeared to do things the way they thought would be acceptable to the group. The members distrusted and rejected each other The members revealed sensitive personal information of feelings. The members appeared tense and anxious. Everything considered, I gained something of value from today’s session. 41 H02] -l> >1 mf‘u-I—i—w- [r'r'> OO O _ l-a—w le>=gmzofl N N [=urnoog] mat—CO] > -I>m:oo >] l—I—w A A45 lr>mu LII LII [ >1 mr—a—l—r‘ OOOOO OOOOOOOOOO [rt-'3; OOOOOOOOOO y—hp—lp—‘p—bh—I HH—s—ay—a—a—a—ap—ap— l—r>:r:gmzo_m] NNNNN NNNNNNNNNN NNNNNNNNNN |wmuofl m-I—COJ > wwwww wwwwwwwwww wwwwwwwwww -I>m:oo >] I—r—w h-b-h-b-b ##A-h-hA-h-h-h-h Ahhhthh-fih mmmmmmmmmm IP>WU MMMMM MMMMUIMMLIIMLI’I [