Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of experiential training on group counsellor development Hoover, Stuart Michael 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_september_hoover_stuart.pdf [ 2.05MB ]
Metadata
JSON: 24-1.0166291.json
JSON-LD: 24-1.0166291-ld.json
RDF/XML (Pretty): 24-1.0166291-rdf.xml
RDF/JSON: 24-1.0166291-rdf.json
Turtle: 24-1.0166291-turtle.txt
N-Triples: 24-1.0166291-rdf-ntriples.txt
Original Record: 24-1.0166291-source.json
Full Text
24-1.0166291-fulltext.txt
Citation
24-1.0166291.ris

Full Text

THE IMPACT OF AN EXPERIENTIAL TRAINING PROGRAM ON GROUP COUNSELLOR DEVELOPMENT by  Stuart Michael Hoover M.A., The University of British Columbia, 2011.  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Counselling Psychology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2015  © Stuart Michael Hoover, 2015ii  Abstract To lead effective psychotherapy groups, counsellors must perform a complex array of skills with both confidence and competence.  Best practices in training group counsellors advocate an integration of didactic and experiential learning (Barlow, 2012; CACREP, 2009).  Experiential training is considered an invaluable training method for group counselling and has been shown to develop skill usage and understanding of group process in trainees (Stockton, Morran & Krieger, 2004).  However, no study to date has experimentally examined the impact of experiential training on counsellor trainee development of confidence and competence.  Therefore, this dissertation introduces an experiential training program seated in social cognitive theory and provides qualitative and quantitative evaluations of the impacts of the program on counsellor trainee confidence and competence variables.    iii  Preface This dissertation is an original intellectual product of the author, S. Hoover.  All clinical and research fieldwork was supervised by Dr. Marv Westwood.  The fieldwork reported in Chapter 3 was collected by S. Hoover and J. Gerlitz and analysed by S. Hoover, and covered by The University of British Columbia’s Behavioural Research Ethics Board Certificate number H09-03384, entitled “An evaluation of experiential training on counsellor development”.  The data collection and analysis reported in Chapter 4 were conducted by S. Hoover with the assistance of a paid research assistant, M. Cayley, and covered by The University of British Columbia’s Behavioural Research Ethics Board Certificate number H13-01991, entitled “Experiam: A Framework for Developing Group Facilitation Competency”.  iv  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iii Table of Contents ......................................................................................................................... iv List of Tables ............................................................................................................................... vii List of Figures ............................................................................................................................. viii Acknowledgements ...................................................................................................................... ix Chapter 1: The Impact of Experiential Training on Group Counsellor Development ...........1 1.1 Why is this question important? ..................................................................................... 2 1.2 What is the counter argument to my research question? ................................................ 4 1.3 Constructs and variables ................................................................................................. 5 1.3.1 Group counselling ....................................................................................................... 5 1.3.2 Counselling groups. .................................................................................................... 5 1.3.3 Group leadership. ........................................................................................................ 6 1.3.4 Experiential group training. ........................................................................................ 6 1.3.5 Confidence variable 1: Group Counsellor Self-Efficacy (GCSE). ............................. 8 1.3.6 Confidence variable 2: Anxiety. ................................................................................. 9 1.3.7 Competence variable: Performance. ......................................................................... 10 1.4 Structure of the dissertation .......................................................................................... 12 Chapter 2: Literature Review .....................................................................................................14 2.1 The efficacy of group counselling ................................................................................ 14 2.1.1 What does group counselling seek to do? ................................................................. 14 2.1.2 How did group counselling evolve? ......................................................................... 15 2.1.3 The curative factors of group counselling. ............................................................... 17 2.2 Counsellor characteristics found to most effectively create efficacious groups ........... 21 2.2.1 Are group leaders necessary? .................................................................................... 22 2.2.2 Leader performance variables. .................................................................................. 23 2.2.3 Leader confidence variables ..................................................................................... 28 2.2.4 Influence of SCT on group leadership. ..................................................................... 31 2.2.5 How are group counsellors created? ......................................................................... 35 2.2.6 Training the group counsellor: A historical perspective. .......................................... 35 2.3 Experiential learning and group counsellor training ..................................................... 37 v  2.3.1 The impact of experiential group training on counsellor development. ................... 42 2.3.2 Self-efficacy from experiential training. ................................................................... 42 2.3.3 Experiential training where students actually lead. .................................................. 47 2.4 Gap in the literature ...................................................................................................... 51 Chapter 3: Experiam: A Peer-Based Experiential Training Program for Group Leader Development .................................................................................................................................52 3.1 Rationale for Experiam: Peer-based simulated practice ............................................... 52 3.1.1 Foundational thinking and mechanisms of change ................................................... 53 3.1.2 Application of foundational thinking: The Integrated Model of Mastery (IMM) .... 54 3.2 A manual for Experiam ................................................................................................. 58 3.2.1 Overview of Experiam and glossary. ........................................................................ 59 3.2.2 Design overview. ...................................................................................................... 61 3.2.3 Session procedure ..................................................................................................... 62 3.2.4 Roles ......................................................................................................................... 63 3.2.5 Scenarios ................................................................................................................... 64 3.2.6 Lab leaders ................................................................................................................ 65 3.2.7 Integration with the course ........................................................................................ 68 3.3 A preliminary study of Experiam ................................................................................. 69 3.3.1 Results ....................................................................................................................... 70 3.3.2 Involvement of the expert. ........................................................................................ 71 3.3.3 Working as peers....................................................................................................... 71 3.3.4 Leading scenarios...................................................................................................... 72 3.3.5 Post-leading learning. ............................................................................................... 73 3.3.6 Role simulation and observing others. ...................................................................... 73 3.3.7 Group design. ............................................................................................................ 74 3.3.8 Lab leader role. ......................................................................................................... 74 3.4 Discussion ..................................................................................................................... 74 3.4.1 Practical implications of the study ............................................................................ 76 3.4.2 Limitations and future research ................................................................................ 79 Chapter 4: Evaluating Experiential Training on Group Counsellor Development ..............80 4.1 Introduction ................................................................................................................... 80 4.2 Method .......................................................................................................................... 84 4.2.1 Participants ................................................................................................................ 84 4.2.2 Materials ................................................................................................................... 86 4.2.3 Measures ................................................................................................................... 86 vi  4.2.4 Procedure .................................................................................................................. 90 4.2.5 Course procedure. ..................................................................................................... 91 4.2.6 Experiam procedure. ................................................................................................. 91 4.2.7 Data analysis. ............................................................................................................ 94 4.3 Results ........................................................................................................................... 94 4.3.1 Confidence variables: Self-efficacy and anxiety. ..................................................... 95 4.3.2 Competence variable: Performance. ......................................................................... 98 4.3.3 Ancillary analysis.................................................................................................... 101 4.4 Discussion ................................................................................................................... 107 4.4.1 Limitations and implications for research and pedagogy ....................................... 111 Chapter 5: General Discussion .................................................................................................116 5.1 Summary of findings................................................................................................... 116 5.2 Novel contributions ..................................................................................................... 117 5.3 Implications for research............................................................................................. 123 5.4 Implications for practice and pedagogy ...................................................................... 128 5.5 Limitations and delimitations ..................................................................................... 134 5.5.1 Limitations to Experiam ......................................................................................... 136 5.6 Conclusion .................................................................................................................. 138 Bibliography ...............................................................................................................................139 Appendices ..................................................................................................................................157 Appendix A : Detailed methodology for primary study (chapter 4) ........................................... 157 Appendix B : Coding scheme for GPIRS (adapted from Burlingame 2013) ............................. 171 Appendix C : Revised GPIRS Scale (Burlingame, 2013) ........................................................... 172 Appendix D : Design for primary study (chapter 4) ................................................................... 174  vii  List of Tables Table 3-1: Critical incident and wish list categories (N = 188) .................................................... 71 Table 4-1: Participant group experience ....................................................................................... 85 Table 4-2: Means and standard deviations for dependent variables: Group Leader Self-Efficacy (GLSI) and State Anxiety (STATE) ..................................................................... 95 Table 4-3: Utterances and rated items in the GPIRS (N=28) ....................................................... 98 Table 4-4: Descriptive statistics for performance data (GPIRS) (N=28) ..................................... 99 Table 4-5: Frequency of skills across two subscales and three domains of performance (GPIRS)............................................................................................................................. 100 Table 4-6: Group Leader Self-Efficacy (GLSI) for follow-up period ........................................ 102  viii  List of Figures Figure 3-1: Integrated model of mastery (IMM) .......................................................................... 56 Figure 3-2: Integrated model of master (IMM) ............................................................................ 58 Figure 4-1:  Mean Group leader self-efficacy (GLSI) for treatment and control groups over three measurement periods during the experimental period. ......................................... 96 Figure 4-2: Frequency and clarity of selected GPIRS items between first and sixth sessions (N = 28). ...................................................................................................................... 105 Figure 4-3: Skill clarity (GPIRS) between session 1 and session 6, coded in five 2-minute time intervals. .............................................................................................................. 106   ix  Acknowledgements As I wrote the discussion chapter to this dissertation, an Experiam session was taking place down the hall, in the infamous group room “278B”.  My vision of Experiam was to enable peers to learn group work as a group.  As I sit here writing its inaugural academic work, I am inspired that students are struggling together to increase group leading confidence and competence in themselves and each other.  Experiam is emerging and evolving in that room, embracing what Stockton (2010) called the “art and science” of group counselling. I begin by acknowledging the funding I received in pursuit of this academic endeavour. In particular, the Social Sciences and Humanities Research Council, Dofasco Inc., and UBC aboriginal graduate student awards. Next, I would like to sincerely thank the folks at BYU for their support of my study.  Specifically to Gary Burlingame, Katherine Bruner, and Chris Chapman, who all responded generously to my requests for help and treated me as one of their own students. My dissertation committee was the best I could have asked for. Drs. Marv Westwood, Bruno Zumbo, and Dan Cox provided a balance of challenge and support that encouraged me to take the dissertation farther than I would have on my own. Indeed, their collective ability to see beyond my immediate problems was most appreciated. They taught me what a dissertation is and what it is not, and for this I thank them, each one. To the peers who helped form this program, I hope you got as much for yourselves as you gave to me. In particular, I acknowledge my co-developer and first group co-leader, Julia Gerlitz, who believed in me and in what Experiam could become.  Finally, I thank my family for their support. All the hours we did not spend together in lieu of this document – these were not wasted. With their patience, I have sought the “church in the East” that Rilke spoke of, hopefully sparing them the burden of unfulfilled expectations.  1  Chapter 1: The Impact of Experiential Training on Group Counsellor Development The Greek word for anarchy is ‘without a ruler’. (Billow, 2010, p. 66). Good group leaders are not born. They are trained. (Barlow, 2004, p. 113). Psychotherapy groups within the field of counselling psychology employ a strength-based focus to foster awareness, education, and skill-training in clients (Bedi, Haverkamp, Beatch, Cave, Domene, Harris, & Mikhail, 2011).  In preparing counsellor trainees to lead effectively, clinicians, researchers and professional associations agree that group counsellor training requires the integration of didactic and experiential group learning methods (CACREP, 2009; Feiner, 1998).  Unfortunately, many counsellor training programs focus on the development of the individual counsellor, which can result in group leaders applying skills and interventions that do not harness group dynamics (Ohrt, Ener, Porter & Young, 2014; Yalom & Leszcz, 2005).  In addition, while experiential training is considered an invaluable training method, it carries some ethical concerns that may explain its low prevalence in training programs (Kottler, 2004; Shumaker, Ortiz & Brenninkmeyer, 2011; Stockton & Toth, 1996).  Finally, even when experiential group training is offered in counselling programs, only about half of students actually lead the group (St. Pierre, 2014).  Therefore, experiential learning, while thought to be a critical group counsellor training method, suffers from issues of underemployment, misapplication, ethical and resource barriers, and even definition. Research on experiential training began with anecdotal supervision case studies, which evolved to quantitative outcome studies of skill acquisition from training interventions and qualitative examinations of trainee experiences that helped them attain group leading proficiency (Stockton et al., 2004).  These authors emphasized the importance of training leaders in efficacious behaviours, yet acknowledge a dearth of research in this area.   2  Examining predictors of such behaviours may better demonstrate the importance of experiential training to counsellor development.  Therefore, this study evaluated the impact of an experiential group training program, called Experiam, on counsellor trainee development of confidence and competence in group leading.  Confidence as a construct was examined via self-efficacy and anxiety variables, and competence as a performance variable.   1.1 Why is this question important? Groups with good therapeutic outcomes require high cohesion, which is a member-member, member-group, and member-leader construct that group leaders directly and indirectly facilitate (Burlingame, Fuhriman, & Johnson, 2001).  This curative factor is fostered by interactions not directly involving the leaders, yet mediated by the leader.  Creating optimum conditions for good group outcomes requires that leaders harness group dynamics with high confidence and competence; a blend of the art and science of group counselling (Stockton, 2010).   Groups involve multiple levels of communication and counsellor trainees may experience these interactions as complex, rapid, unpredictable and uncontrollable (Barlow, 2004; Page, Pietrzak, & Lewis, 2001).  Perceiving an event as unpredictable or uncontrollable can generate anxiety and fear which is an indicator of, and contributor to, low self-efficacy and performance (Bandura, 1982; Bandura, Reese, & Adams, 1982). Developing confidence and competence is explained through the social cognitive theory (SCT) concept of counsellor self-efficacy (CSE), which describes one’s beliefs that they can orchestrate the multiple skills involved in effective counselling (Larson & Daniels, 1998).  While there is evidence that CSE is associated with performance in individual counselling, the link between self-efficacy and performance in group counselling contexts is still a  3  theoretical one (Larson, Suzuki, Gillespie, Potenza, Bechtel & Toulouse, 1992; Hoyt, Murphy, Halverson & Watson, 2003; Page et al., 2001).   Increasing group counsellor self-efficacy (GCSE), therefore, seems to be an effective way to prepare trainees to facilitate effective groups.  However, counsellor training programs often focus on preparing the individual counsellor at the cost of other areas including group counselling (Robison Jones & Berglund, 1996; Yalom & Leszcz, 2005).  The assumption that high self-efficacy from individual counselling experiences will necessarily improve GCSE has not been demonstrated in empirical research; instead theorists and clinicians agree that such skills do not clinically map from one domain to the other (Barlow, 2004; Yalom & Leszcz, 2005).   Several models of experiential group training have been developed and evaluated, although the links between such programs and group leader self-efficacy have limited empirical evidence (Page et al., 2001).  A problem in defining experiential training was revealed in a recent survey of group training programs.  St. Pierre (2014) found that while most counsellor trainees observed an experiential group, nearly fifty percent did not actually lead a group.  Self-efficacy is argued to be most impacted by a combination of direct enactment of behaviours and vicarious observation of others, with the former being the more impactful of the two.  Therefore from a SCT perspective, observation-only programs would be less effective in raising GCSE than those placing the trainee in the driver’s seat (Bandura, 1984).  Clearly, not all experiential training programs are the same, and evaluations of such programs may not be generalizable or comparable. In summary, increasing counsellor trainee self-efficacy in harnessing group dynamics is important and assumed to be best accomplished when didactic training is coupled with experiential training programs that involve direct experience in leading a group.  Establishing  4  the link between leading an experiential group and counsellor self-efficacy remains a gap in the literature that requires addressing. 1.2 What is the counter argument to my research question? Wicker (1998) suggested that research problems are improved through exploration of their counter arguments.  A counter argument to this research problem is that participating in experiential training does not improve, or may even decrease, counsellor confidence and competence.  This argument is supported by evidence that such training could cause ethical boundary violations (such as dual relationships) that may raise anxiety of students or may lead to negative relationships with their peers for having participated in the group (Feiner, 1998).   In addition, counsellor trainee self-efficacy and performance may be lowered if they perceive they are failing.  Larson and Daniels (1998) stated that the ideal level of self-efficacy is somewhat higher than actual performance, thus challenging counsellor trainees to take risks that are outside, but not too far outside, their abilities.  They stated that this ideal level is related to the optimal level of anxiety for task performance.  Therefore, experiential training that imposes risk on trainees could exceed tolerable levels of control and anxiety in some trainees and reduce GCSE.  Finally, Larson and Daniels (1998) stated that self-efficacy is impacted by supervisory feedback, meaning that critical feedback from an instructor could also reduce GCSE in some trainees.  No study was found that examined the impact of programs that impose too little rish on trainees, meaning that the effects on GCSE of lower-than-optimal levels of anxiety remain theoretical.  5  1.3 Constructs and variables Three domains of literature that underpin the research problem, and three constructs by which confidence and competence was evaluated are now defined. 1.3.1 Group counselling  The term “group counselling” has been used interchangeably with group psychotherapy and can be applied to both the product (the counselling group) and the means of delivering the product (group leadership), and so is employed in this dissertation only when referring to the broader understanding of both product and delivery. Example include the “group counselling course”, the history of the field, and outcome research on the overall impact of the therapeutic experience on participants. 1.3.2 Counselling groups. A counselling group is the product that counsellor trainees in Experiam were taught to competently deliver to the public.  Barlow (2012) stated that a counselling group is a category of therapeutic groups that is broadly synonymous with group psychotherapy and group treatments, and differentiated from leaderless or large format therapy groups.  She defined it as the application of evidence-based group dynamics to therapy and prevention settings.   There are many kinds of groups within the helping professions, and four specializations are defined for group counsellors, namely psychoeducation, psychotherapy, counselling, and task groups (ASGW; 2000).  Gladding (2012) stated that while these group specializations each have their own differential qualities, some groups offer a mix of these specializations at different stages of treatment.   6  In this study, a counselling group was defined as group leadership provided by a trained counsellor primarily working with non-clinical populations, with the goal of behavioural change and development for issues such as problems in living or interpersonal relationships (Bedi et al., 2011; Gladding, 2012).  The participants in this study were not trained group counsellors, but were in the process of training such skills. It should be noted that the first course in group counselling did not prepare trainees as specialists in one category of group work, but provided a first exposure to group theory and skills that may apply to many contexts.   1.3.3 Group leadership. Group leadership is the way in which counsellors deliver counselling groups.  In reviewing the literature on groups, many overlapping and often ill-defined terms are used to describe what leaders do in groups.  Examples include therapy, psychotherapy, counselling, leading, psychology, and management.  Barlow (2012) stated that group leadership skills and interventions are those founded in group dynamics and supported by professional organizations.  For the purposes of this dissertation, Shechtman and Toren’s (2009) definition of group leadership was adapted as the combination of intrapersonal traits and observable group leadership skills necessary to efficaciously facilitate therapeutic change in counselling groups.  The former are defined as confidence variables: self-efficacy and anxiety, while the latter are defined as a competence variable: performance.  Participants in this study were learning to develop these variables by way of experiential group training, described next. 1.3.4 Experiential group training. Published educators in group leadership differentiate between didactic and non-didactic learning methods, with the latter including experiential training, observation of  7  experts, supervision, skills training, and personal psychotherapy in groups (Corey, 2012; Feiner, 1998, Yalom & Leszcz, 2005).  However, these components overlap, leading to confusion about what is experiential training or how to measure it (Kline, Falbaum, Pope, Hargraves & Hundley, 1997; Osborn, Daninhirsch & Page, 2003).  Confusion is exacerbated when there is no leading component to the training, such as programs that limit experiential training to only observing a group (St. Pierre, 2014).  To add further confusion, didactic training is not independent of experiential learning in group classes.  Most graduate programs in North America have only a single group counselling course, and so experiential activities must necessarily be built into the classroom (Furr & Barret, 2000).   Shumaker et al. (2011) stated that experiential groups explore group processes and formation from a first-hand perspective.  Corey (2012) added that experiential training is a combination of leading small groups, followed by supervision feedback by the instructor and discussion by the entire group.  Therefore, being a participant-observer is not enough to qualify as experiential leading.  The Experiam program reflects Kiweewa, Gilbride, Luke and Seward’s (2013) definition of experiential group training as one that employs group leading as a primary, but not the sole, learning method to explore trainees’ personal issues relevant to group work while at the same time learning group processes and skills.  Experiential group leading is therefore an pedagogy that addresses both components of effective group leadership described above: skills-training (competence) and self-efficacy (confidence).  In other words, while the primary focus of such training is to learn about group process and skills, trainees may also engage in some personal development (Merta, Wolfgang & McNeil, 1993).   8  Three variables used to measure underlying constructs of confidence and competences are now defined. 1.3.5 Confidence variable 1: Group Counsellor Self-Efficacy (GCSE).  Self-efficacy, as it is used in this dissertation and in the psychological research literature, is a term pioneered and developed by Albert Bandura and applied in SCT.  In his seminal 1982 article, he defined self-efficacy as self-referent judgments of one’s generative capability to organize cognitive, social, and behavioural skills into integrated courses of action.  However, he stated that capability must be paired with execution, which involves self-judgments of action as well as knowledge of one’s ability to perform the requisite skills.  Therefore, self-efficacy is judgment of one’s ability to orchestrate multiple skills in service of an action, as well as expectations that such action will result in a particular outcome.  He illustrated this definition by contrasting the constituent skills of driving a vehicle, such as shifting gears, steering, braking, and changing lanes, to the belief in one’s ability to navigate a congested street or twisting mountain road (Bandura, 1984).  The integration of discrete skills expands beyond behavioural actions to include personal factors (cognitive, motivational, and affective processes) and environmental influences, which are referred to as the triadic reciprocal causation for human agency (Bandura, 1989; 2001).  Cognitive processes include planning and prediction that integrate memory with higher order executive function.  Motivational processes reflect the exertion of effort to initiate, maintain, and persist in effort towards a challenge.  Affective processes describe arousal and emotion states that feed into cognitive and motivational processes.    Counsellor self-efficacy (CSE) is defined as the beliefs or judgments counsellors have about their ability to effectively counsel a client in the near future (Larson & Daniels, 1998).   9  The authors stated that counsellors are expected to be efficacious with their clients, and that CSE beliefs are the primary causal determinant of effective counselling action, making CSE a valuable construct for evaluating counsellor training.  Integrating Bandura’s definition, they stated that CSE affect counselling actions by affecting the choices of response to clients, the level of effort expended, the persistence in the face of failure, and risk-taking.  Group CSE, or GCSE, is thought to operate in a similar manner to individual therapy, but with some differences unique to group work, such as micro-skills particular to group work (e.g., responding constructively to an attack), process tasks, (e.g., screening members, setting norms, providing structure), and diversity issues (Page et al., 2001).  Therefore, GCSE is defined as the beliefs or judgements group counsellors have about their ability to effectively lead a group in the near future. 1.3.6 Confidence variable 2: Anxiety. Groups are believed to confront members with anxiety, particularly in the initial stage (Corey, 2012, Yalom & Leszcz, 2005).  Anxiety has theoretical and clinical consensus as an essential condition for change in groups, but if too high, can erode member perceptions of membership and safety, thus reducing motivation to participate in therapeutic tasks (Bion, 1961; Schutz, 1958; Yalom & Leszcz, 2005).  However, anxiety is also present in group leaders, and can cause ineffective behaviours such as freezing or resistance, particularly as beginner leaders are learning group dynamics (Billow, 2001; Corey, 2012, Yalom, 1966). In SCT, anxiety or fear is viewed as perceived inefficacy in coping with a potentially adverse event (Bandura, 1982; Bandura et al., 1982).  Once a sense of control over the event can be established or predicted, fear is argued to diminish.  How control is established is argued to arise from behavioural and cognitive sources.  Behavioural control has a history of empirical  10  evidence demonstrating that successful action reduces uncertainty of feared events.  However, for anticipated events, cognitive control must be employed to manage fear.  Anticipatory thought in humans involves the ability to predict the outcome of potential future events, which Bandura argued has functional value if the thought does not exceed realistic bounds.  If it does, rumination can lead to decreased self-efficacy.  Such processes explain how one’s thoughts can defeat oneself before the event takes place.  Increased self-efficacy is argued to reduce anxiety through increased cognitive control over a feared event.  Larson and Daniels (1998) cited evidence that decreased anxiety was a predictor of increased CSE.  Therefore, anxiety is defined here as the combination of physiological, emotional, and cognitive responses to one’s perceived inefficacious ability to cope with a potentially aversive future event/incident.   1.3.7 Competence variable: Performance.  The construct of confidence in counsellor trainees can be thought of as largely unobservable, internal processes, while competence is the observable delivery of group leadership. The last variable is performance, which underlies the construct of competent group leadership. In a review of individual counselling studies, Larson and Daniels (1998) found eight measures of performance, including demonstration of basic to advanced micro skills, and evaluation of these skills from the perspective of the trainee and the supervisor.  Counsellors exert direct and indirect influence on group dynamics and therapeutic outcome, thus the definition of counsellor performance has multiple perspectives. Group dynamics are a global perspective of performance and most commonly conceptualized by Yalom’s 11 therapeutic factors, of which the most empirically-supported factor is cohesion (Barlow,  11  2012). An alternative dynamics that leaders may be judged to have fostered is the degree of interpersonal conflict in the group (Kivlighan & Tarrant, 2001; Romano & Sullivan, 2000).   A more specific perspective of leader performance involves the delivery of discrete skills such as linking members, making interactions more present-focused, and encouraging emotional self-disclosure between members (Toth & Stockton, 1996). Once learned, specific behaviours must be delivered appropriately, based on several factors including the stage of the group, the goals of the group, counsellor and member personal characteristics, and multicultural issues, such as convergent or divergent identities within the group.  The literature on leader actions or behaviours uses terms such as skills, micro skills, behaviours, and interventions to describe group leadership.  However, the term intervention can be confused with a collection of skills delivered over a period of sessions, such as a manualized treatment intervention for anxiety.  Similar confusion can arise from the term behaviour, which could include non-verbal gestures, such as a frown, and which may be too granular to map to therapeutic outcomes.  Therefore, the term skill will be used in this dissertation to describe leader actions that are discrete, measurable, and that have empirical evidence as markers of group leadership performance. Integrating the global perspective, the Experiam program focused primarily on those micro skills that establish and maintain cohesion in groups. Burlingame, Fuhriman and Johnson (2002) classified three domains of leader interventions that have an empirical basis in the development of cohesion: group structure, verbal interactions, and establishing and maintaining the emotional climate.  These domains have been recently mapped to a behavioural rating scale called the Group Psychotherapy Intervention Rating Scale, which enables group skills to be rated according to frequency and clarity of delivery (GPIRS;  12  Chapman, Baker, Porter, Thayer & Burlingame, 2010).  Therefore, the actions captured by this scale most closely represent effective group leadership performance.    A final distinction must be made regarding performance in this study.  The Experiam program was designed for trainees to practice newly-acquired group micro skills that develop and maintain group cohesion. Therefore, this evaluation will measure performance as the appropriate choice and timing of counselling behaviours, called skills, that have empirically-supported impact on therapeutic outcomes in group leadership.  More specifically, performance will be a function of the frequency of skills particular to establishing and maintaining group cohesion, as well as the clarity with which acquired skills are demonstrated. 1.4 Structure of the dissertation This dissertation was written in manuscript style.  Chapter two reviews three domains of literature relevant to the program and its evaluation.  Chapter three is a stand-alone manuscript that provides (a) an introduction to foundational thinking and theory underpinning Experiam, (b) a manual for delivering the lab, and (c) a preliminary qualitative focus group study.  Chapter four is also a stand-alone manuscript detailing the primary study of the dissertation.  An in-depth methodology section of the study in chapter 4 is provided in appendix A.  Chapter five is a general discussion of the Experiam program with a focus on the primary study. The purpose of this dissertation is to introduce and evaluate the Experiam program using SCT as an underlying framework.  The evaluation comprises a preliminary qualitative focus group study and a primary randomized control study.  Note that this study examined the impacts of training using experiential methods, and not training in experiential methods.   13  For example, experiential group training means that training in group leading that applies an experiential method, and not training in how to teach counsellors how to use experiential groups to teach other concepts.      14  Chapter 2: Literature Review Boote and Beile (2005) stated that doctoral literature reviews must satisfy three requirements: To present results of similar studies; to relate the present study to ongoing dialogue in the literature; and to provide a framework for comparing this study with other studies.  This literature review will achieve these requirements through the integration of three interrelated sections.  The first introduces the broad concept of group counselling and what makes it therapeutically efficacious.  The second reviews the counsellor characteristics found to most effectively create these curative group elements.  The third section reviews how such counsellor characteristics are most effectively taught to trainees.  The final section integrates these domains to describe a gap in the literature that provides a rationale for the primary study. 2.1  The efficacy of group counselling  This section defines the purpose and goals of group counselling, and then describes the historical evolution of group counselling.  Finally, factors of group counselling known to be curative are reviewed. 2.1.1 What does group counselling seek to do?  To answer this question, one must first understand how the purposes and goals of counselling psychology apply to group psychotherapy.  The purpose of group counselling can be defined as enhancing and promoting positive growth, well-being, and mental health of individuals, families, groups, and the community, focused primarily on subclinical populations (Bedi et al., 2011).  This is does by addressing distress in life events and transitions, decision making, career/education, relationships, and mental and physical health concerns through a therapeutic change process that comprises remediation, prevention,  15  psycho-education and advocacy.  Group counselling addresses these issues using a strength-based, culturally-sensitive focus towards client awareness building, education, and skill training that applies group dynamics as a theoretical foundation for change (Barlow, 2012). The core values of counselling psychology can be applied to group in the following ways.  First, group members are viewed as agents of their own change and group cohesion as a central mechanism of change.  Second, case conceptualization of the group takes social context and culture into account.  Finally, treatment is client-centred and respects unique cultural variables with each person and that exist within the group.  Training group counsellors should be congruent to these values (Bedi et al., 2011; Grigg, 2006).  In contrast to this definition, the clinical assessment function of counselling psychology does not apply to counselling groups at the master’s level. 2.1.2 How did group counselling evolve? Group counselling emerged from the evolution of group therapy from the 1900s to today.  The earliest pioneers in group therapy were European, including Sigmund Freud’s Wednesday night meetings, Alfred Adler’s group treatments for social struggles, and Jacob Moreno’s Theatre of Spontaneity (Barlow, Fuhriman & Burlingame, 2004; Gladding, 2012).  However, the first written account of what was to become group therapy was from American physician John Pratt.  In 1905, he ran a “thought control class” for tuberculosis patients and discovered that the participants were in good spirits and that 75% of them exhibited recovery from the terminal illness (Barlow, Burlingame, & Fuhriman, 2000, p. 115).  Psychiatrists Trigant Burrow and Lewis Wendel also published seminal papers in 1928 and 1938, respectively, which introduced here-and-now interactions that counteracted isolation and specified dynamics that were curative in inpatient groups (Kline, 2003).  With the advent of  16  World War II, UK psychoanalysts Wilfred Bion and Siegfried Foulkes treated veterans and conceptualized the “group-as-a-whole”, where the group itself is analysed instead of the individuals comprising it (Hinshelwood, 2007, p. 344).  Following WWII, Kurt Lewin, considered the father of group research, co-established major centres for group research and therapy, including the Tavistock institute in the U.K. and the National Training Labs (NTL) in the U.S. (Burnes, 2004).   For the first half of the 20th century, group therapy was dominated by two theories: psychoanalysis and psychodrama, and was primarily applied to inpatients in psychiatric hospitals (Barlow et al., 2000).  However, the 1950s heralded many new theories and therapies for groups, as well as an overlap between theory and therapy, due to the group itself being considered a curative factor.  In 1946, Kurt Lewin pioneered a new kind of group, called the human relations training group, or T-group.  Yalom and Leszcz (2005) dedicated an entire chapter of their text to the history and significance of these groups on the identity of modern group counselling.  They stated that Lewin’s first T-group heralded a wave of clinical and research interest in experiential learning and human relations training.  Clinicians in both analytic and humanistic schools were influenced by the buzz of human relations laboratory training, leading to experimentation in interpersonal interaction and personal change in the newly-formed NTL.  Such groups emphasized here-and-now interactions, feedback, active member engagement, challenging belief systems, and cognitive aids, such as Luft and Ingham’s (1955) Johari Window.   Labelled encounter groups by Carl Rogers in the 1960s, they marked the beginning of group work for more normal populations, claiming that anyone could participate and benefit from groups.  The authors stated that the goals of encounter groups evolved from providing  17  education around a specific issue, to achieving personal growth or complete human potential across broader, existential Gestalts.  These groups were not initially considered therapy groups, and so practitioners were not required to have clinical training; however, they were considered flamboyant and reflected the “let it all hang out” attitude prevalent in 1960s American culture (Yalom & Leszcz, 2005, p. 530).  This, along with excessive and unethical use of interventions, led to acrimony with the psychological community, who considered the movement to be reckless and potentially harmful.  However, encounter groups infused group psychotherapy with a new energy and made group work applicable to more functional populations and for many everyday problems in living.  Although such extreme groups continue until this day, they are considered at the fringes of the mainstream therapeutic community.  In 1991, the Association for Specialists in Group Work (ASGW) produced a set of standards for group workers and classified specializations within group work that endures to the present time (ASGW; 2000).   2.1.3 The curative factors of group counselling. The first paragraph of Irvin Yalom’s magnum opus on group psychotherapy stated that group therapy is “at least” as effective as individual therapy to help clients, with the implication that it can be more effective (Yalom & Leszcz, 2005, p. 1).  The authors outlined eleven therapeutic factors (TFs) that define therapeutic change in groups.  These factors are widely researched with cohesion and interpersonal learning having broad clinical and empirical acceptance (Kivlighan, Coleman & Anderson, 2000). Research on the curative factors of groups is divided between common factors versus specific interventions perspectives (Greene, 2003).  The author stated that former examines the dynamic relationships between constructs and is more suitable to process research, while  18  the latter uses descriptive measures to reveal the particular therapeutic factors that were valuable to group members, and is used in many outcome studies.  However, outcome studies have dominated process research in group counselling over the past two decades (Corey, 2012).  Reasons include methodological difficulties in capturing group process relationships, and that the current era of managed care has caused policy makers and researchers to explore the question “who would do well in which kind of group led by which kind of leader” (Barlow et al., 2004, p. 18). In a review of group counselling research, Kivlighan et al. (2000) stated that meta-analyses of outcome research demonstrated that group counselling is indeed efficacious, although these studies ranged in leader qualities, population, and interventions measured.  The authors also examined process research of the eleven TFs with emphasis on cohesion and interpersonal learning, as these were frequently-measured factors of change, although it was also shown that the type of group moderated the importance of the TF to participants.   Fuhriman, Drescher, Hanson, Henrie, and Rybicki (1986) conducted a factor analysis of Yalom’s curative factors and discovered that four emerged as key factors across group formats and compositions.  These are cohesion, interpersonal learning, catharsis, and insight.  They stated that these factors are necessary but not sufficient conditions for change in all groups.  The authors further stated that the other therapeutic factors are conceptually nested within these factors.  For example universality may be a specific aspect of cohesion, but the realization that one is not alone is a form of insight.  Finally, they identified that interpersonal learning was correlated to all factors, suggesting that it is not a distinct factor.  They concluded that cohesion is a highly curative factor, and that catharsis may be a necessary condition for insight to occur.  Yalom and Leszcz (2005) explained that the primary task for  19  the group therapist is the creation of a cohesion climate, which acts as a therapeutic factor in itself, as well as a necessary condition for other therapeutic factors to operate.  Multiple meta-analyses in group cohesion have demonstrated it as the evidence-based element of group therapy that is associated with good therapeutic outcome (Norcross & Wampold, 2011).  The omnipresence and omnipotence of cohesion (relative to other TFs) in research, clinical manuals, and training programs provides a strong rationale to train group counsellors in skills and interventions that foster it. Cohesion is the group analogue of the therapeutic relationship in individual counselling, defined broadly in terms of a force that causes a sense of “esprit de corps”, or “we-ness” (Yalom & Leszcz, 2005, p. 55).  It is also defined specifically in terms of attractiveness or tripartite relationships comprising member-group, member-member, and member-leader bonds (Burlingame et al., 2002).  In a review of group cohesion, Cota, Evans, Dion, Kilik and Longman (1995) stated that cohesion is a complex construct that eludes strict definition or measurement.  They defined it unidimensionally as a field of forces, and multidimensionally as having primary elements common to all groups, such as consensual rules among members about values and behaviours and resistance to disruption, and secondary elements particular to certain types of groups, such as risk-taking, roles members take, and elements specific to the group’s composition and goals.   Although the construct of cohesion is not as precisely defined as many researchers would like, there is ample evidence that the presence of cohesion is associated with good outcome in group therapy.  Studies of the lifespan of groups have shown a linear relationship between cohesion and outcome, while others have demonstrated higher outcome when cohesion is higher in the initial stages of the group, but lower during the middle, working  20  stage as members tolerate the intense work needed for therapeutic benefit (Riva, Wachtel, & Lasky, 2004).  In a review of group cohesion research, Burlingame et al. (2002) stated that its impact on process and outcome in group therapy has been validated in “literally hundreds of articles” (p. 72).  Outcomes included lower levels of attrition, decreased symptomology, and improved interpersonal relationships within and beyond the group.  Process variables included higher levels of member self-disclosure, higher tolerance for feedback and conflict, and higher member-member listening, empathy and support.  In a specific study of the mediating influence of cohesion, Kivlighan and Tarrant (2001) found that curative group climates are characterized by high engagement of participants, particularly in the early stage of group development, and by low avoidance or conflict between members.  They also stated that engagement increases and avoidance/conflict decreases as the group progresses.  Such process variables are shown to correlate with high levels of cohesion, but also describe cohesion, suggesting that the many dimensions of cohesion become stronger in the presence of each other in a virtuous cycle. This body of evidence would suggest that groups are universally beneficial.  However, Barlow (2004) cautioned to not blindly believe that groups always work for all people and leaders in all contexts.  Indeed, groups can fail.  Lieberman (1990) provided an empirical study of this caveat by examining how groups fail to provide measurable benefit to their members.  He stated that group outcome literature is dominated by studies demonstrating what is efficacious, leaving a gap in our understanding of what does not work in groups.  He examined 36 homogenous peer support groups of new mothers versus controls, but discovered no significance on tests of mental health and marital relationships; instead he discovered a trend that participation in the groups increased parental distress! To  21  examine this counter-intuitive phenomenon (and failed experiment), he conducted post-hoc surveys of group cohesion, social support, discussion topics, and curative factors in the groups.  He found four central processes that may explain why these groups failed.  First, the group lacked emotional importance to members and agreed-upon norms were not observed (i.e. the ideal conditions of cohesion).  Second, the group was not a seen as unique or salient to members.  Third, the group did not help restructure the way people solve their problems.  Fourth, the group did not offer a variety of experiences to contain the individual differences among members.  He concluded that for peer groups, these criteria must be present for them not to fail.  These criteria also map to the elements of cohesion identified by the above studies, providing further support to this factor. What makes groups curative seems to centrally involve a cohesive climate, evidenced by the presence of certain underlying components and the absence of others.  Leader characteristics that foster a cohesive climate are now examined. 2.2 Counsellor characteristics found to most effectively create efficacious groups Since cohesion has been shown to be a primary therapeutic factor of groups, the processes by which group counsellors create cohesive groups are now reviewed.  The first section justifies the importance of the group leader.  The following two sections delineate effective leadership using two interrelated domains of social cognitive theory that define performance and self-efficacy.  First, leader performance will be reviewed using outcome expectancies, which Bandura (1978a) defined as self-estimates that a given behaviour will lead to certain outcomes.  Second, efficacy expectations are convictions that one can successfully execute the behaviour itself, and are reviewed as leader confidence variables best described by what is termed self-efficacy.  22  2.2.1 Are group leaders necessary? Are group leaders necessary for effective groups? The literature says yes (Coco, Gullo, Prestano & Burlingame, 2015; Riva, et al., 2004).  A way to evaluate the importance of the leader is to contrast the efficacy of leaderless groups, such as self-help groups (e.g., alcoholics anonymous) and mutual support groups (e.g., parenting support groups).  Gladding (2012) stated that such groups either rotate leadership between members, or a leader emerges as the group develops.  He stated that the efficacy of such leaders is developed through trial and error, which has potential to be harmful.  Yalom and Leszcz (2005) provided anecdotal clinical evidence that such groups may lose the therapeutic depth and degrade into a “cocktail party” level of conversation (p. 313).  Within leaderless peer supervision groups, Counselman & Weber (2004) cited limitations including drifting from the task of the group, competition, advice-giving, over-supportiveness, and fear of being criticized.  These findings highlight specific elements that trained group leaders provide. Stockton (2010) commented that effective group leadership is a blend of art and science.  The latter arises from the application of research findings of what is efficacious in groups, while the former is less easily defined or internally valid, but no less vital to efficacious leading.  He stated that the art of group leadership goes beyond knowing what intervention to perform in a specific situation to reflect mastery in the delivery of the intervention.  He concluded that research must maintain clinical relevance by being practical, realistic, and adaptable.  I believe he is referring to intrapersonal elements of the leader, some which cannot easily change, such as trait personality characteristics, and others which can, such as self-efficacy.  Gladding (2012) stated the difference succinctly: “Personal qualities of effective leaders have traditionally been explained through examining their personality traits  23  or learned skills” (p. 60).  The following sections will be divided similarly into efficacious leader behaviours and leader confidence variables. 2.2.2 Leader performance variables. What is a group leader? Barlow (2004) summarized research of expert group leaders and found that they were perceived as natural helpers, with the distinction that they knew when and how to strategically heighten member struggles to allow the group to address them.  The impact of cohesion on group outcome inspired literature examining how it is developed through group leader interventions and structured exercises (Kivlighan et al, 2000).  Regarding the former, the authors stated that high cohesion is associated with the early setting of group norms by the therapist, a term called ‘conditionality’, while cohesion in the working stage required more warmth and caring from leaders.  Regarding the latter, the authors stated that highly structured groups at an early stage have higher cohesion, particularly for risk-averse members, but that such structure must adjust as the group develops so as not to stifle member to member interactions.  This finding was demonstrated experimentally by Lee and Bednar (1977), who found high leader-facilitated structure and high risk-taking (especially in early sessions) were determinants of participants engaging in the highest levels of meaningful interpersonal communication in later sessions.  They concluded that leader-created group structure regulates participants’ exposure to interpersonal risks, which enables them to take more risks and thereby gain therapeutic benefit. Brabender (2010) reviewed the development path to expertise in group psychotherapy and found three behaviours that distinguish the expert group leader.  First, experts formulate more complex conceptualization of the group process.  This means that they view members  24  multi-dimensionally and can conceptualize multiple interventions to help them achieve their goals without a priori bias to one particular intervention.  Second, experts foster more open groups that lead to higher levels of self-disclosure, which in turn fosters therapeutic benefit.  One way they do this is though well-timed and delicate use of personal self-disclosure.  Finally, experts have a superior ability to keep the group focused on the here-and-now and on process over content issues.  While these findings were not directly related to therapeutic outcome, their use by those considered to be experts in the field provide clues to what leader variables are likely to be the most efficacious.   The seminal study on group leader behaviours was conducted by Lieberman, Yalom, and Miles in 1973.  They employed a quasi-experimental study to examine 210 participants in 18 encounter-style groups that showcased ten diverse group theories, including psychodynamic, Gestalt, humanistic, and psychodrama.  Leader behaviour was coded in real-time by trained raters who observed the groups.  Outcome was evaluated using a battery of psychological instruments measured at pre, post, and at six month follow-up.  They discovered that participants improved over non-participating controls, and that certain leader styles emerged as more efficacious across groups.  In particular, they found a linear relationship between good outcome and leaders displaying caring and providing a cognitive framework for understanding feelings or experiences.  This means that the more a leader could display these behaviours, the higher the outcome over time.  Interestingly, they also discovered curvilinear relationships between outcome and two leader behaviours.  The first was when leaders activated the emotional climate of the group.  This meant that effective leaders must provide some emotional stimulation to the group, but not too much.  This finding is analogous to Ogden, Pain, and Fisher’s (2006) metaphor for information  25  processing called the window of tolerance, where individuals can have either too much arousal, called hyper-arousal, or too little, termed hypo-arousal, and that we process information optimally in the window between these extremes.  The second curvilinear relationship was when leaders used executive control over the groups, such as setting goals, managing pace, and enforcing norms.  This meant that leaders should provide some structure and control, but also be adaptable and judicious in their use of control as too much lowers the therapeutic outcomes of the group.  These results have been replicated in other studies, and the findings have become constructs used in measures of other group processes.  For example, Kivlighan et al. (2000) reviewed outcome studies and discovered a balance of leader-imposed structure was associated with the highest levels of cohesion.   Shechtman and Toren (2009) conducted a study, inspired by the Lieberman et al.’s (1973) findings, to examine leader behaviours on process and outcome in group counselling.  The authors hypothesized that leader interventions would have a direct impact on outcome, and also an indirect impact, mediated by process variables.  Over two hundred counselling students participated in 14 groups led by experienced leaders and data was collected by self-report measure and by transcript analysis of the sessions.  They defined effective leader interventions from Lieberman et al.’s (1973) study, and measured these interventions using member-completed questionnaires as well as leader transcript analysis.  Four process variables were measured, including the degree of self-disclosure, resistance to group process, member attachment to the leader and the group, and impressions of therapy.  The first two variables were assessed by self-report measure and transcript analysis, while the last two were self-report only.  Three outcome variables were measured, including difficulties in interpersonal relationships, risk-taking in intimate relationships, and self-esteem.  They were  26  self-reported before and after the group treatment.  There was no control group for comparison.  Using a three-level hierarchical model, the authors found some support for their hypotheses.  Regarding the direct effects of leader interventions on outcomes, the authors found that executive management of the group decreased risk-taking, but with inconsistent results between the questionnaires and the transcript analyses.  They posited that group members may perceive their group behaviours to be different from how they are actually observed.  Regarding the mediating relationship of the process variables, emotional stimulation by leaders increased member self-disclosure, which in turn was positively associated with self-esteem.  In addition, caring leader behaviour was associated with several process variables, which in turn were associated with improved interpersonal relationships.  The authors concluded that member self-esteem and interpersonal relating are impacted by a group climate high in self-disclosure, low resistance, and group bonding, and that this climate is influenced by leaders displaying caring and emotionally stimulating behaviours.  Disappointingly, even though the title of the article contains the word “effect” the authors did not discuss effect sizes to determine the magnitude or clinical significance of these associations.  The authors stated that a limitation to the study was the use of graduate students who did not present with many interpersonal problems to begin with.  This result highlights that group counselling efficacy requires that the treatment be matched to the appropriate population.   In contrast to researching what therapists do that works, Bugenthal (1988) wrote an article dedicated to failure in individual psychotherapy.  Using case illustrations of individual, existential-humanistic therapy sessions, he differentiated between outright and  27  partial failure, where the former usually results in client harm or termination from therapy and the latter are actions that slow the client’s pace of therapeutic change.  He identified several discrete therapist behaviours, such as putting one’s own needs before the client’s, not investing in a resistant client, allowing therapy to go on too long, making poorly timed interpretations, and topping the client’s level of insight, thus robbing them of the personal discovery.  He stated that the common denominator to therapeutic failure is the hesitation of the therapist to invest as fully, and be as present, as the client needs.  He said such failure goes beyond technique or theory to reflect the level of sincerity with which the therapist seeks the humanity within his/her client.  Stockton, et al. (2004) applied this last point to group contexts by positing the belief that leaders must be trained in behaviours salient to group work, such as being able to face attack from group members without becoming defensive.  However, facing and attack and not becoming defensive are behaviours that combine intrapersonal state variables with observable skills.   Finally, Smokowski, Rose, and Becallao (2001) conducted a qualitative study of what creates a “group casualty”, defined as a member suffering psychological distress directly attributable to a group experience and enduring longer than six months after the group ended (p. 224).  The authors interviewed thirty-three people who met this definition and found three categories of factors associated with damaging group experiences.  These comprise factors concerning the group leader, the group climate, and the casualties themselves.  Leader behaviours that contributed to group casualties included being too passive with structure, norms, and task orientation, which could contribute to a damaging group climate.  In addition, group members were vulnerable to leaders who were overly active, such as encouraging confrontation, stimulating intense emotions, and pressuring members to disclose  28  and receive feedback.  Groups were polarized into an in-group and out-group by such leader behaviours.  This provides further support that effective leadership requires skills that balance between member engagement and member regulation.  The authors advocated that leaders actively structure group norms, tasks, and member roles, and refine their ability to identify and integrate members displaying quietness or temerity. These studies demonstrate the indirect impacts of leader behaviour on group counselling outcomes.  In other words, effective group leading requires attention to the creation of an environment that heals, and not to consider oneself the direct object of healing for the client, as happens in individual counselling.  This phenomenon highlights how the training of group counsellors requires a systemic worldview of the change process, compared to individual counsellor training methods that directly link counsellor behaviour to client outcome.   2.2.3 Leader confidence variables Effective group counselling requires more than knowledge of theory and enactment of skills.  Experts in psychotherapy emphasize being self-aware and fully present to clients, which can be translated as the intrapersonal leader variables (Bugenthal, 1988; Harpaz, 1994; Kivlighan & Tarrant, 2001).  These authors are describing qualities of the leader that go beyond observable skills to reflect Stockton’s (2010) concept of the “art” of group counselling.  These qualities are reviewed in specific and global terms, and then seated in SCT. Specific intrapersonal leader qualities can be described using a trait approach that vest ideal qualities in personality characteristics (Gladding, 2012).  These include poise, judgment, empathy, ego strength, low anxiety, desire to help, tolerance for frustration and  29  others.  In contrast, unproductive group leaders exhibit anxiety, fear, dread, distance, resistance, and poor personal boundaries (Billow, 2001; Markus & Abernathy, 2001; Shechtman & Toren, 2009; Yalom, 1966).  Harpaz (1994) examined failures in psychodynamic group psychotherapy as arising primarily from unresolved, irrational fears within the therapist.  The author stated the fear of abandonment and the fear of engulfment are major contributors to failure in group psychotherapy, defined as premature termination not attributable to external or situational factors.  The fears occur at opposite ends of the attachment spectrum, with abandonment fear causing a desire to increase attachment to others, and engulfment fear causing a desire to break attachments.  He summarized how such fears can cause failure at three stages of group treatment.  In particular, he stated that therapists who have not worked through such fears will fail at the initial phase of therapy, because their countertransferential reactions will be highest, relative to an undeveloped therapeutic bond.  The author stated that such inquiry is not prevalent in the literature due to the discomfort such topics may bring up in therapists, leading instead to over-attention on client characteristics as the cause of therapy failure.  He stated that therapists must determine, through rigorous self-scrutiny, personal psychotherapy, and supervision, the degree to which their fears get activated in a group.  If not understood, he argued that such fears can lead to poor choice and timing of therapeutic interventions and increased risk of therapeutic failure and harm to group members.   Viewing effective group leaders by specific traits is a popular approach, but resists empirical validation and are not empirically supported as efficacious for leadership (Gladding, 2012).  Instead, this author argued that the interaction of multiple personal qualities correlates with effective leadership, suggesting personality profiles that may predict  30  effective group leadership.  This presents a conundrum for research, as personal qualities in the leader seem important, particularly when absent, but are hard to measure in isolation or in combination with other qualities. Kivlighan and Tarrant (2001) provided empirical support for a global leader variable by examining how leaders’ intentions impact group outcomes.  They examined how group climate mediated the leader to outcome relationship, and that this climate was created by leader intentions including focus on therapeutic work, creating a safe atmosphere, maintaining structure, and encouraging interpersonal communication.  They examined the intention of leaders, rather than explicit behaviours, citing studies supporting leader intentions as more closely related to outcomes than specific behaviours.  This means that regardless of what the leader says or does, the degree to which he/she means it has the weight of therapeutic benefit for clients.  The authors employed an exploratory factor analysis of 43 youth groups to differentiate group leader intentions, and discovered four factors.  The first factor was therapeutic work focus, which were leader intentions to create insight and change, to focus on awareness of feelings, and to challenge participants.  The second was group structure, comprising intentions to focus on participant behaviours and on clarifying the session structure.  The third was termed interpersonal to reflect intentions to overcome relational obstacles such as resistances and therapist needs.  The fourth was called safe atmosphere, categorized by intentions to set limits while also supporting participants.  The authors measured group climate using two empirically-supported constructs, the degree of members reported they were actively engagement, and the degree of conflict/distance between members.  Outcome was assessed using a self-report measure of the benefits of therapy and their perceived relationship to the therapist.    31  The authors proposed three hypotheses.  First, that the group leader intentions would be associated with group climate, second that this climate would, in turn, be associated with client outcome, and finally that the relationship between intentions and outcome would be fully mediated by the climate.  They examined 43 group leaders running 2-hour groups for at-risk youths, aged 13 to 15 years.  The unit of analysis was the group, and participant measures were aggregated for each session.  They performed a hierarchical level model of the climate variables and determined that the first hypothesis was partially met, in that the final model showed significant correlation between the two exogenous variables, therapeutic work, and a safe atmosphere, and a climate of active engagement, and the two exogenous variables, group structure, and interpersonal, and the climate of low conflict/distance.  They found that the second hypothesis was likewise partially met, in that the active engagement climate was associated with therapeutic benefit and the low conflict/distance climate was associated with a good therapist relationship.  Finally, they found that the third hypothesis was mostly supported, in that the group climates fully mediated the relationship between the two exogenous variables associated with it and the respective outcome outlined above.  The results of this study show that intentions play a role in therapeutic outcome.  While it makes a case that intrapersonal leader variables have impact, this study would have had stronger findings had there also been a comparison between leader intentions and behaviours. 2.2.4 Influence of SCT on group leadership. Intrapersonal leader qualities may be best understood through the lens of social cognitive theory.  In particular, self-efficacy explains individual differences in leaders more specifically than by leader intentions, and allows for people to change beyond their traits.  Self-efficacy is defined as the judgments one makes of how well he/she can execute courses  32  of action to deal with a situation, both in decisions to initiate action and in decisions to sustain action in the face of adversity (Bandura, 1982; 1989).  Beutler, Machado and Neufeldt (1994) stated that counsellor self-efficacy, along with counsellor conceptual level and ego development, are stable and reliable personal characteristics that can be used to train counsellors. Bandura (1989) argued that self-generated activities such as these beliefs are at the heart of causal processes in humans.  Causal processes are described as being within the control of people, which he said was deterministic.  He qualified determinism, stating that while humans are not free of external constraint or coercion, they have control over self-influence, through regulative and reflective thought, one’s skill-set, and the ability to select courses in a predictive way.  Unlike traits, self-efficacy is a changeable characteristic in people, which is described as a process of triadic reciprocal determinism (Bandura, 1978b).  He stated this process expands from Kurt Lewin’s equation behaviour = f(person, environment) to allow each component to reciprocally influence the others.  Personal factors include cognitive, affective, motivational, and biological events (Bandura, 2001).  Therefore, the environment provides people with feedback on their self-efficacious behaviours, and that the feedback in turn influences emotional and somatic reactions, thoughts, motivation, and forethought, which alters self-efficacious beliefs of how one may act in the future (Bandura, 1989).  He stated that self-efficacy can be influenced by four experiential sources, namely mastery, modelling, social persuasion, and affective arousal.  He defined mastery as the successful enactment of behaviours toward an outcome, and modelling as the vicarious experiencing of others’ enactive attainment.  He further stated that mastery experiences impact self-efficacy more than modelling, but that a combination of the two was most  33  impactful.  He also summarized factors that can reduce self-efficacy in people, including the presence of a highly efficacious person, attending to what is strange versus familiar in a new task, or being cast in a subordinate role.  Finally, he hypothesized that this may occur because it introduces an illusory self-precept, and cited a study demonstrating how a formidable looking competitor can lower the self-efficacy of athletes. Larson and Daniels (1998) reviewed counsellor training using SCT and defined the construct of counsellor self-efficacy (CSE) as a counsellor’s beliefs about his/her capabilities to effectively counsel a client in the near future.  They stated that counsellors are expected to be efficacious with their clients by orchestrating multiple sub-skills to respond to dynamic circumstances.  Counsellors set goals for themselves and their clients through the reciprocal processes of forethought and feedback.  Barnes (2004) described three assumptions of CSE.  First, that CSE is the primary mechanism through which effective counselling can occur.  Second, strong CSE beliefs result in trainee perseverance in difficult tasks.  Third, counsellor trainees with high CSE are better able to receive and integrate evaluative feedback. Larson and Daniels (1998) also found that CSE had minimal relationship to most counsellor demographic variables including personality, aptitude, achievement, social desirability, age, sex, theoretical orientation, and time spent as a client.  However, CSE was found to have a moderate to strong correlation with other self-reflective variables, such as perception of fraudulence, self-concept, and self-consciousness.  Finally, the degree of experience had a strong positive relationship to CSE, yet was not found to hold for the counsellors’ level of training, beyond the initial stages of training.  This gives support to the changeable nature of self-efficacy that a training program can embrace.    34  The authors reviewed interventions impacting self-efficacy, anxiety, and performance in counsellors and found that most studies examined mastery and modelling on CSE, with a few studies examining anxiety arousal.  They concluded that performance was most improved by mastery and modelling experiences, but that neither was superior to the other.  Finally, they stated that performance feedback had a direct relationship to anxiety levels in counsellors.  The absence of social persuasion studies may be due to the difficulty in measuring the construct.  Group cohesion may resemble social persuasion, suggesting that group work can provide a theoretical fit for this CSE building block. Page et al., (2001) stated that CSE has a natural extension to group environments, as emotionally-charged group environments require the confident orchestration of multiple skills; they defined this as group CSE, or GCSE.  In a recent article on group leader development, Ohrt, Robinson, & Hagedorn (2013) were unable to find studies of group leader self-efficacy, despite the creation of the GCSE scale.  The authors also related GCSE to performance and anxiety in counsellors and hypothesized these self-referent thoughts as contributors to the degree of anxiety practicum trainees’ experience.  Regarding anxiety, Bandura (1978a) stated that people fear and avoid situations that they believe will exceed their coping skills, calling this a low self-efficacy belief.  He also stated that those who cease coping efforts in the face of such fears will retain low self-efficacy regarding the situation over time.  This mirrors Harpaz’s (1994) statement about the persistence of leader fears of their groups and illustrates the importance of self-efficacy in counsellor trainees to ensure that they are not overwhelmed by their groups.  Regarding this last point, Bandura (1978a; 1982) stated that performance-based treatments go beyond new behavioural  35  accomplishments to diminish the cognitive component of fear arousal, thus raising self-efficacious beliefs that one can cope with events in the future. A fundamental assumption of SCT is that future performance is mediated by perceptions of past performance (Larson & Daniels, 1998).  This assumption provides a theoretical link between the concept of self-efficacy and performance.  The reciprocally determining nature of GCSE to performance suggests that successful early-group experiences can seed efficacious beliefs of future successful group leading.  Such experiences also act as an antidote to the fear arousal that trainees may experience in group environments.  Like a motor that runs after the rip cord is pulled, counsellor trainees require successful initial group training experiences to foster the self-efficacy to persevere in more threatening subsequent ones. 2.2.5 How are group counsellors created?  This section describes the field of group counsellor training, then situates experiential training within the field, and concludes by examining domains of such training.   2.2.6 Training the group counsellor: A historical perspective. Merta and Sisson (1991) stated that customary methods for teaching group skills include academic, observation, experiential, and supervision.  The first is the didactic instruction of theory and knowledge of group process and skills.  The second involves students observing an ongoing group, either by video recording or in vivo.  In the third method, students participate in various roles as members in a therapy or personal growth group.  The fourth involves clinical supervision of students’ group leading skills.  Barlow (2004) provided a sequenced three-year plan for training doctoral students’ beginner to advanced group skills.  She stated that group training presents an “essential tension” because  36  it requires students “unlearn” some individual psychotherapy skills in order to be effective in groups (p. 113).  She stated that group training should be layered in a sequence, beginning with the student participating in a therapeutic group as member, as this provides a felt-sense of the group that cannot be conveyed didactically.  Next, students should participate in a theory course to bring understanding to the group experience.  She said that such knowledge must equally balance between domains of research (theory) and practice (therapy).  She invoked Melanie Klein’s metaphor of a scissor approach, where one blade is intellectual knowledge and the other is personal experience.  Her multi-year plan combines the four training methods at beginner, intermediate and advanced stages of trainee development, leading to a matrix of activities and objectives from which her plan emerges.  Unfortunately, this comprehensive training plan may exceed the training abilities of many programs that do not specialize in group training (Stockton & Toth, 1996).   From a developmental career perspective, Brabender (2010) provided a five-stage framework to assess a counsellor’s level of group expertise.  The first stage involves a person’s decision to begin training.  Stage two is the trainee stage, where most experiential training programs occur. Stages three to five cover the practicing novice through to the achievement and maintenance of expertise.  Regarding stage two, the author stated that didactic and experiential components are critical to the trainee, but that such training is not offered in all programs.  She reviewed U.S. programs and concluded that 30% of clinical PhD programs and 60% of PsyD programs offered only the most basic didactic instruction, leaving many students to pursue basic instruction at conferences or extra-curricular workshops.  She concluded that experiential training is an important follow-up to knowledge  37  acquisition, but that trainees experience high anxiety in fledgling leadership roles, particularly in the absence of supportive supervision.   Whether some or all of these methods are used, a group training program requires a theoretical foundation (Kormanski, 1991).  The author stated that this has been a rare occurrence in training programs.  This leads to many unanswered questions from students and results in group training that is incomplete.  He described several theories, including individual growth, personality, and group dynamics.  He noted that Kolb’s (1984) experiential theory is the most commonly adopted theory in group training programs, due to its relationship to purpose and its learning-by-doing emphasis.  He concluded that regardless of the chosen model, the proper sequencing of skills is particularly important for training group counsellors.   Training of group counsellors ranges from multi-modal plans that span over years, to a single course or nothing at all.  However, sequencing at least the didactic and experiential components using a theoretical basis is recommended to meet the personal and professional development of the trainee (Stockton & Toth, 1996; Barlow, 2004).  Experiential training is now reviewed. 2.3 Experiential learning and group counsellor training Of the four methods for training group counsellors, the experiential component is considered vital, but is faced with ethical and feasibility barriers (Barlow, 2004; Merta & Sisson, 1991; Shumaker et al., 2011).  These authors highlight a dilemma between protecting students from harmful dual relationships and ensuring that group counsellors meet regulatory standards, where the latter is argued to be much improved through the use of experiential group training.  This section will briefly explore the history of experiential group training,  38  review its theoretical underpinnings, and highlight ethical and feasibility barriers that such training must overcome.   Feiner (1998) reviewed the history of experiential group training for mental health professionals, stating that it began in the 1940s in the National Training Laboratories.  By the 1970s, such training was criticised for lack of coherent theory and uniformity, and remained outside of mainstream training programs.  However, in the 1980s, attention shifted to more efficient delivery of mental health services that emphasized experiential training.  However, previously-identified ethical concerns had yet to be addressed, causing such training to be avoided (Barlow, 2004).  The 1990s brought new interest in providing experiential training, but with more attention to providing a theoretical learning framework that integrated didactic learning, and to ensuring that ethical boundaries were not crossed.   Founding experiential group training on a theoretical footing not only brings clarity to interventions and evaluations, but avoids the pitfall of eclectic “play it by ear” designs (Kormanski, 1991, p. 215).  Contributions from pioneers in experiential learning, including Dewey’s theory of experience, Piaget’s developmental theory, and Lewin’s field theory, have coalesced into a holistic, constructivist, developmental model of human learning that underpins many experiential training programs (Kolb 1984; Kolb & Kolb, 2005).  These authors clarified experiential learning as more than a set of techniques to provide an experience to learners, and proposed a model explaining how optimum learning is achieved through a balance between four dialectic modes of experience.  They proposed that experiential learning begins with concrete experience, which is observed and reflected upon, followed by formulating new concepts and plans, and then testing them in new situations.  Called a feedback process, experiential learning theory bears striking resemblance to  39  Bandura’s (1978a; 1978b) reciprocal process model of learning, particularly where the influence of experiential action causes new self-referent thoughts to lead to new behaviours.  This provides additional theoretical support for SCT as a learning theory for experiential training.  A social cognitive model of counsellor training emerges from this process, in which the training environment and trainee self-efficacy jointly influence learning and performance (Larson & Daniels, 1998).   Barnes (2004) differentiated two approaches to applying SCT to experiential counsellor training for individual psychotherapy.  The first approach attempts to enhance self-efficacy as a way to impact performance.  This approach assumes that CSE predicts counselling performance.  Trainees were assessed for their level of self-efficacy and training interventions were aimed at enhancing these beliefs in order to improve actual counselling performance.  Interventions included supportive feedback, live modelling, and methods to reduce anxiety.  The author cautioned that while there are studies supporting this assumption, there have also been mixed findings in research.  She concluded that this approach may work better for trainees who have very low CSE or high anxiety that interferes with performance.  In the second approach, the contexts from which self-efficacy beliefs are formed in supervision were examined.  Here, supervisors collaborated with trainees to understand how CSE beliefs may not be lining up with actual skill levels.  Interventions for this approach included accurate and consistent self-reflection, combined with formative feedback and evaluation of performance.  However, this approach required high self-scrutiny of the trainee.  The author suggested this approach for trainees who have over- or under-estimated CSE beliefs, and who can tolerate high scrutiny and constructive feedback.  In summary, experiential group training has roots in experiential learning theories of Kolb and Dewey that  40  share theoretical congruence to new theoretical formations drawing from social cognitive theory.  Armed with theory, we now turn to pedagogy.   Several methods to experientially teach group counselling have been proposed, including observing expert models, leading a group, personal participation in group therapy, personal psychotherapy, and supervision (Merta, et al., 1993; Tschuschke & Greene, 2002; Yalom & Leszcz, 2005).  Published experiential-based training programs utilize a mix of these components, which can greatly vary what is meant by experiential group training in research studies.  From a social cognitive perspective, a key distinction involves whether an experiential training program utilizes the mechanisms theorized to increase CSE, particularly in combination with mastery and modelling.  The next section will categorize such programs based on this criteria with emphasis on the most influential method for increasing CSE, namely whether trainees actually lead a group, thus actively enacting leader behaviours (Bandura, 1982).  First, barriers to experiential group training are reviewed. Experiential training is not without controversy.  Ethical risks have been identified, including dual relationships, informed consent, and privacy concerns among students and faculty (Merta & Sisson, 1991; Shumaker et al., 2011).  Dual relationships involve two sets of risks.  The first arises because experiential groups may coerce the disclosure of personal information that students would have preferred not to share with colleagues.  As demonstrated above, effective group therapy elicits self-disclosure as a norm, and students may not be able to help themselves from saying or showing too much among their peers.  In addition, the limits of confidentiality in groups, particularly non-therapeutic training classes, mean that personal information about students may be shared among peers or with instructors.  The second dual relationship can occur if instructors are present or made aware  41  of disclosures by students that may impact their evaluations of them.  Very recently, St. Pierre (2014) found that 58% of experiential group training programs are faculty-led.  While there is a gate-keeping function for instructors to protect the public from potentially unsuitable individuals, students in an experiential group may experience a surfacing of thoughts or feelings that would not normally arise in professional practice (Feiner, 1998).  Studies of student’s perspectives of such groups reveals that ethical boundaries can be inadvertently crossed (Davenport, 2004).   Some solutions to these ethical problems have been proposed that range from no experiential component at all, to an instructor-led group that endorses high self-disclosure (Merta, & Sisson, 1991).  Variants within this range include leaderless groups, peer-led groups, feedback or no-feedback groups, and instructor observed groups (Merta et al., 1993).  Shumaker et al. (2011) provided four safeguards for students that employ informed consent.  These include making the experiential component voluntary, providing informed consent prior to their enrollment in the course, explaining the kinds of self-disclosure that would be expected and appropriate in the group, and informing students that participation is not evaluated for grading purposes.   Finally, a non-ethical constraint to experiential groups is one of finance.  Romano (1998) stated that many programs only offer a single group course, and that counselling programs have many competing demands for student time and faculty resources.  He suggested peer-led experiential groups as a possible solution to this problem.   Counsellor educators largely agree that these ethical risks are worth taking, although they should be minimized with safeguards in order for students to reap the benefits of experiential group training (Kottler, 2004; Shumaker et al., 2011).  In addition, studies of  42  students’ perspectives suggest that experiential approaches do facilitate personal and professional growth and development, despite risks of interpersonal boundary violations (Anderson & Price; 2001; Ieva, Ohrt, Swank & Young, 2009; Kiweewa et al., 2013).   2.3.1 The impact of experiential group training on counsellor development. Barlow (2004) stated that two-thirds of training programs employ experiential teaching, but that the usefulness of such training has yet to be definitively determined.  One confound to examining experiential training is found in St. Pierre’s (2014) survey of experiential groups.  In it, eighty-five percent of students reported receiving vicarious learning as a group member, but that nearly half of students denied receiving direct group leading experience.  Therefore, greater clarity must be applied when evaluating and grouping experiential training programs.  Although the research on self-efficacy and experiential group training is scarce, only studies that examine how experiential learning impacts self-efficacy, or that place trainees in the position of group leader, are reviewed.   2.3.2 Self-efficacy from experiential training. Urbani, Smith, Maddux, Smaby, Torres-Rivera and Crews (2002) employed a repeated-measures comparison study to examine the impact of a skills training program on individual counsellor self-efficacy.  Treatment and comparison groups were both enrolled in 45-hour courses that covered counselling theories (group in the first study and individual in the second).  However, the treatment group also participated in a 36-hour skills-training program.  Groups were pre and post tested with a self-report measure of individual counsellor self-efficacy (COSE; Larson et al., 1992) and a video-recorded six-minute mock interview in which students demonstrated as many counselling skills as they could and trained raters coded the quality of the behavioural skills.  As expected, the authors found that  43  the experiential program was associated with significantly higher skill levels and higher self-efficacy.  They also discovered that the treatment group rated their skills lower than independent raters, while the opposite perspectives were found before the training, in that students rated their skills more effective than raters.  They concluded that experiential training has an impact on counsellor self-efficacy as well as on performance in an individual counselling context. Strengths of this study included the use of a comparison group, and behavioural measures of performance.  In addition, the study employed a training model that was seated in the theoretical underpinnings of self-efficacy.  However, this study has several statistical and design limitations.  First, the groups were not randomly assigned, and no mention was made as to why students would choose a course offering a skills model versus one with didactic-only instruction.  Second, the treatment and comparison groups were greatly unbalanced, with 53 treatment participants versus 9 controls.  Statistical comparisons used the combined sample size for calculation of degrees of freedom, calling into question the validity of the results.  Third, the self-efficacy measure was pretested, yet only post-test results were provided.  This raises the question of whether the scores of the treatment group were significantly higher than the comparison group before the study began (perhaps explaining the non-equivalence between groups raised in the first limitation).  This also meant that the authors did not report whether self-efficacy changed in treatment or control groups over time. Next, Daniels and Larson (2001) used a repeated measures randomized control experiment to examine how performance feedback impacted individual CSE and anxiety levels.  They examined two hypotheses.  First, that positive feedback would result in  44  increased self-efficacy over time, while negative feedback would do the reverse.  Second, they hypothesized that positive feedback would lower anxiety levels and negative feedback would do the reverse.  Forty-five counselling students conducted a 10-minute mock interview with a trained actor, and were told they were being evaluated by the researcher through a two-way mirror.  Following the interview, the participants rated their own performances and completed a state anxiety measure.  They were randomly assigned to either receive positive or negative feedback on their performance.  Each group was then retested for self-perception of performance, self-efficacy and state anxiety.   The authors conducted multiple ANOVAs and found that all hypotheses were met.  They concluded that the intention of supervisory feedback is a significant determinant of counsellor self-efficacy and anxiety levels.  This meant that experiential groups that employ positive feedback reduced the anxiety levels of trainees.  According to SCT, this should also predict increased performance in situations that cause high anxiety.  Strengths of this study include the clarity of the research questions and design used to answer them.  The experimental design also allowed results to be more generalized to other counsellor trainee populations.  A weakness of the study was the contrived positive and negative feedback being more extreme than was likely to be experienced. Third, Hoyt et al., (2003) conducted a structural equation model design to determine relationships between leader self-efficacy, collective efficacy, and group performance in an organizational context.  Students in two laboratory studies participated in a hiring task and were tested for the three variables.  Leader self-efficacy and collective efficacy were measured by self-report, while performance was determined by the number of hiring evaluations completed by each group.  Results indicated that leader self-efficacy did not  45  predict performance, but that this relationship was mediated by collective efficacy.  In other words, high leader efficacy predicted collective efficacy of the group, which in turn predicted higher performance.  This study provided an indirect link between self-efficacy and performance, but in a management setting that was financially incented.  However, this result bears some similar results to Kivlighan and Tarrant’s (2001) finding that group leader intentions indirectly relate to group outcome via the group climate their intentions fostered.  The construct of collective self-efficacy bears resemblance to that of group cohesion, and may be an avenue for future exploration of the mediating impact of cohesion on outcome or productivity. Finally, Ohrt et al. (2013) conducted a quasi-experimental study to compare the effects of counsellor trainees participating in the two most common types of experiential groups (personal growth groups and psychoeducation groups).  They hypothesized that such experiential groups would impact group leader development of self-efficacy, as well as empathy, cohesion, catharsis, and insight.  The authors compared the two groups using pre/post measures of each construct, including the Group Leader Self-Efficacy Instrument (GLSI; Page et al., 2001).  The authors posited that vicarious learning would impact group leader self-efficacy, and hypothesized a difference between the two group experiences.   They first pre/post tested the dependent variables using self-report measures and conducted a MANCOVA to determine the effect of the group conditions while controlling for pretest scores, and found significant between-group differences for empathy and self-efficacy, with a large effect size.  Second, a discriminant analysis was conducted to determine the predictability of group condition to the dependent variables, and revealed that the predictors did differentiate between the two types of group with catharsis and insight  46  being the best predictors of personal growth groups.  Third, they conducted repeated measures ANOVA for pre/post effects of each dependent variable.  They found no significant within-group changes for empathy.  However, they did find significant within-group changes for self-efficacy. The authors concluded that such experiential groups may not be enough to cultivate empathy in trainees, but that they both had a large effect on trainee self-efficacy.  These findings support the suggestion that experiential learning has a similar impact to CSE in group training, as was previously found for individual counselling (Larson et al., 1992).  They also concluded that the vicarious learning may have been a significant factor due to students observing a successfully led group without adverse consequences.  Finally, they stated that since the GCSE increased in both group conditions, the mechanism of change for GCSE may be viewing leader functions firsthand, instead of the group format or how one interacts as a member. This study bears resemblance to the current study in that a similar pre-post design was employed, the GLSI was measured, and two groups were compared.  Strengths of this study include the use of comparison groups, and the use of group measures to address their research questions.  There are several weaknesses to this study.  First, there was not a control group, and comparison groups were not equivalent in design or even of similar sample size.  The personal growth groups (n = 27) had 80 minutes session length, lasted 10 weeks, and each contained 6-8 students in their first semester.  The psychoeducation groups (n = 47) were 60 minute sessions, lasting 12 weeks, and each contained 10-12 students in their second semester.  Although students were randomly assigned to one of four small groups within each condition, the matching of students was based primarily on availability to attend in the  47  day or night, further confounding the randomization process.  Second, only self-report measures were used, limiting the findings to statements of student’s perceptions of their abilities, and not to performance, which the Shechtman and Toren (2009) study, above, found to be measurably different.  Finally, the authors neglected to consider the mutual dependencies problem of group research (Kivlighan et al., 2000).  This means that the individual self-reports may have been impacted by group-level phenomena.  An example is that the goal of these groups is to help counsellor trainees improve self-efficacy.  Being in a group of people who are succeeding may have a salutary effect on counsellor trainees’ self-efficacy that may not be present when they run the group themselves.  In conclusion, this study did demonstrate how experiential training can impact GCSE, but failed to do so in a generalizable way. 2.3.3 Experiential training where students actually lead. Bruce-Sanford (1998) proposed a simulation experiential leading model that was integrated into a group counselling course.  The design provided students the opportunity to lead a simulated role-play of a group of their choosing for 30-45 minutes, followed by 30-45 minutes of process debriefing of their leading.  The author did not conduct a formal evaluation of the program, but anecdotally noted elements of cohesion between peers, and the ability for the design to foster creativity and flexibility in the group. Romano and Sullivan (2000) tested a simulated group counselling (SGC) model that provided graduate students with an experiential group experience while minimizing dual relationships and resource costs.  Counsellor trainees enrolled in a didactic group counselling course were placed by the instructor into peer-led groups that met weekly to experience group leading together.  The students co-led in pairs, received feedback and discussion  48  following each leading session, and rotated their roles every two weeks.  The instructor assigned a senior student to observe the group, to facilitate the feedback and discussion session, and to complete a measure on group and facilitator competencies.  The authors hypothesized that the model would approximate a well-known four-stage theory of group development.  They developed a group observer form that examined group process, member and leader behaviours.  This measure was subjected to factor analysis, revealing three factors: group cohesion, here-and-now focus, and conflict.  Taken together, the factors describe stages of group development.  For example, growing cohesion and here-and-now focus, combined with high conflict, is qualitatively indicative of the transition stage of group development (Corey, 2012). Results validated the peer-led model as a realistic training experience for group counsellors by approximating group stage characteristics and member behaviors.  Cohesion grew through all stages, while here and now focus peaked and levelled off in the second stage.  Conflict peaked in the second stage and then dropped off.  Some limitations of the findings included the fact that the groups lasted eight sessions, and students led only two sessions each.  This meant that the approximation of the four-stage group development model may have been confounded because each stage was represented by a different co-leading pair of students.  However, this study demonstrated the viability of experiential groups to provide an authentic training experience while minimizing ethical concerns. Fall and Levitov (2002) followed Romano and Sullivan’s (2000) SGC model but employed actors to play the roles previously taken by counsellor trainees.  The authors argued that this alleviated confusion students had in the various roles they took and reduced dual relationships present when student peers draw from their own experiences in a role.   49  Fifty-minute simulated sessions were conducted with other students observing fish-bowl style, through a one-way glass.  Sessions were video recorded for the student leaders to review, and sessions were debriefed in the following didactic class.  At the end of the semester, actors were invited to join the didactic class to provide feedback to the student leaders.  Results from the program were presented anecdotally by the authors.  Aside from a lack of a research focus, one criticism of this design is that counsellor trainees were denied the opportunity to vicariously experience their colleagues work with them in a role, and lost the experience of being a group client. Next, Toth and Stockton (1996) conducted a quasi-experimental examination of the impact of experiential training on counsellor trainee skill level.  The authors hypothesized that a skill-based experiential training program would increase student’s skills level, both against a placebo control and over time.  They randomly assigned 48 students into experimental and control groups.  The treatment group received a one-time, three-hour group consisting of didactic, observational, role-playing, and supervision experiences.  The control group received three hours of standardized didactic instruction and observed a video of group leading.  The authors pre and post video-recorded all participants each leading a 5 minute small-group interaction that were coded for skill interventions by trained and independent raters.  Leader statements were qualified as being in the here-and-now, linking members, or directed at the expression of feelings.  They used a 2 (group) x 2 (time) split-plot, factorial, repeated measures design and found that both hypotheses were met with significant results at the p < .01 level.  They concluded that such training does positively influence counsellor trainees’ ability to use group skills, and suggested that the observation of fellow students leading may invoke vicarious learning and improve performance.  This study provides  50  generalizable findings in support of the impact of experiential training on skill development, and compliments the efforts of the current study to identify changes to self-efficacy. Strengths of this study include the use of a randomized control group, and the use of behavioural ratings versus self-report.  Weaknesses in the study include that the treatment group size was twice the size of the control group, that effect sizes were not reported, and that the treatment was a single three-hour group.  The authors stated the biggest limitation was that the frequency of intervention use did not reflect the quality of intervention.  The authors mentioned the concept of self-efficacy in the discussion section, providing a rationale to experimentally examine this construct.  An improvement to this study would have been to not differentiate the learning of the control group from that of the treatment group, thereby making the treatment group’s experiential components the only difference between them, thus ensuring that the control group does meet the criteria for a placebo. Finally, Smaby, Maddux, Torres-Rivera, & Zimmick, (1999) conducted a study similar to the Urbani et al. (2002) study, but applied to group counselling.  The authors employed a repeated-measures comparison study to examine the impact of a skills training program on ratings of group counsellor skills (but not self-efficacy in this study).  The treatment and comparison groups were enrolled in 45-hour courses that covered counselling theories (group in the first study and individual in the second).  However, the treatment course contained a 36-hour skills training program adapted to group work, while the comparison group only received didactic instruction.  Groups were pre and post tested using a video-recorded six-minute group counselling session in which students demonstrated as many counselling skills as they could and trained raters coded the quality of the behavioural skills.  Similar to Urbani et al. (2002), the experiential program was associated with higher  51  group skill levels over time and over the comparison group.  Strengths and weaknesses in the study were also similar to the previous study. 2.4 Gap in the literature This review found that the most important and ubiquitous curative factor in group counselling is the development and maintenance of cohesion.  Temporally, cohesion does not grow spontaneously, but requires empirically-supported group-based skills and interventions, delivered competently and confidently by trained group leaders.  Such training is argued to require experiential learning that integrates skill acquisition with self-efficacious delivery of group interventions.  However, no study to date has experimentally explored whether an experiential training program that includes mastery and modelling components raises group counsellor self-efficacy and performance.  This gap is an important one to address, as many statements about the value of experiential training on group counsellor development and performance rest upon this assumption.  The next chapter introduces the Experiam program, including theoretical underpinnings, a detailed manual of the program, and a preliminary qualitative focus group case study.     52  Chapter 3: Experiam: A Peer-Based Experiential Training Program for Group Leader Development This article introduces and describes the experiential program called Experiam.  First, rationale and foundational thinking supporting Experiam are provided, and an integrated model of mastery (IMM) that integrates Experiam into a didactic course is proposed.  Next, step-by-step instructions of the program are outlined.  Finally, results are presented from a preliminary qualitative group case study that inform the current program design. 3.1 Rationale for Experiam: Peer-based simulated practice Experiam is a derived Latin word that broadly translates to “I test/practice/try”.  The mission of Experiam is to create a realistic, cohesive, and practical learning environment, where counsellor trainees can experientially practice recently-acquired group counselling skills.  Challenges that counsellor trainees face in group work include pre-existing fear of groups due to prior negative experiences, low self-efficacy arising from a lack of applied knowledge for problems unique to group work, and anxiety from the requirements of the leader role (Billow, 2001; Markus & Abernethy, 2001; Yalom, 1966).  A peer-led, experiential simulated group with limited faculty supervision was believed to address these issues.  The program was designed to promote safety and risk-taking as well as addressing ethical risks inherent in experiential training programs, including dual relationships, privacy concerns, and informed consent (Anderson & Price, 2001; Davenport, 2004; Goodrich, 2008).  Practical limiting factors were addressed, included financial and workload constraints to the trainees, curriculum limitations, and limits to faculty member supervision (Romano, 1998).  The program design was founded on Bandura’s (1978a, 1982) social cognitive theory (SCT) and applied using a model of peer group supervision for counsellors-in-training  53  (Westwood, 1989).  Founding the program on theory bridges the gap between domains of research and application and enables them to reciprocally inform one another (Kormanski, 1991).  Current thinking about group program development espouses an evolving process that seeks out and incorporates input from participants, leaders, faculty supervisors, and informal and formal studies (Stockton, 2010).   3.1.1 Foundational thinking and mechanisms of change There are two theorized mechanisms of change from SCT that are employed in Experiam.  First is the four-part process by which self-efficacy is enhanced, namely mastery, modelling, social persuasion, and affective arousal (Bandura, 1989).  Participants in Experiam experientially practice new skills, they vicariously learn from modelling by their peers, they receive in vivo feedback on their performances, and they are part of a cohesive group container that regulates strong emotion.  Experiential learning models have proven success at raising self-efficacy in individual counselling and in participant-observer group contexts (Daniels & Larson, 2001; Ohrt, et al., 2013).  Interpersonal feedback adds to vicarious learning and social persuasion mechanisms, but requires careful structuring and monitoring (Toth & Erwin, 1998). The second mechanism of change is the link between fear and managing potentially aversive events (Bandura et al., 1982).  In Experiam, students are brought face-to-face with realistic yet aversive simulated group events.  They enact skills to facilitate the group and observe others model similar enactments in a structured and cohesive group environment.  Exposure to feared group incidents is therefore experientially paired with mechanisms known to increase self-efficacy.  This exposure is controlled three ways.  It is mediated within-session by structured norms and and cohesion in the peer group, between-session by having  54  early sessions deal with less aversive group incidents than later sessions, and within-person by ensuring the student leader retains control over scenario intensity and duration.  These two mechanisms of change guided the planning, implementation, and evaluation of the Experiam program. 3.1.2 Application of foundational thinking: The Integrated Model of Mastery (IMM) Experiam is delivered through a model of peer-based group supervision.  Westwood (1989) provided guidelines for the creation of an optimal learning environment for group supervision that increases cohesion, promotes trainee risk-taking, and decreases defensiveness.  Peer-based learning models have ethical, theoretical, and practical advantages.  Ethically, multiple relationships are an often-cited criticism of experiential group training programs (Feiner, 1998; Merta & Sisson, 1991).  In response, Romano (1998) argued that peer-based experiential group programs reduce faculty-student dual relationships that emerge when instructors either lead or observe students directly.  Furthermore, when students perform simulated client roles instead of acting as themselves in these groups, student-to-student dual-relationship concerns are further reduced (Fall & Levitov, 2002).   Peer-based groups are a practical option for academic programs with heavy curriculum requirements.  Counselling programs fall into this category because they commonly employ a scientist-practitioner training model to meet professional standards of public trust.  Therefore, additional training must be creatively applied.  Peer-based groups are common on university campuses, easy for students to organize and conduct, and proximal for faculty to supervise and research.  Universities also have a mandate for teaching and research which provides a strong rationale to support extra-curricular training.  55  Peer supervision groups are found to reduce anxiety, increase trust and group productivity, increase skill development and self-confidence, and improve participants’ understanding of the supervisory process (Borders, 1991; Romano & Sullivan, 2000; Starling & Baker, 2000).  In a recent qualitative exploration of experiential group training, counsellor trainees reported that peer-led groups enhanced self-awareness as counsellors, increased confidence in leading groups, and stated that the absence of an instructor fostered a climate of trust and safety (Ieva et al., 2009).  Yalom and Leszcz (2005) suggested that experiential peer-groups led by a competent group leader could be the single most valuable experience students receive in their curricula.  They cited benefits such as peer support, student exposure to a range of group therapy phenomena, and the opportunity for students to take a supervisory role when giving feedback to each other.  From a SCT perspective, self-efficacy is increased when people believe they have some control over those who have power, a concept called proxy control (Bandura, 1982).  Clearly, specific and important learning takes place when peers can effectively work together in groups – a type of learning that does not necessarily occur in the presence of an expert.   Finally, social cognitive mechanisms of change are experientially-applied by use of the psychodrama technique of role training in supervision groups (Blatner, 2000; Wilkins, 1995).  This technique enhances experiential rehearsal and refinement of a role (in this case specific group leading skills) in a realistic setting (the feared group incident).  Role training enables participants to enact client roles in service to the leading participant’s chosen scenario, while Westwood’s (1989) theory of group supervision provides rules of containment to safeguard against excess spontaneity that is inherent in standard psychodrama practice.  One safeguard was the establishment and modelling of group norms to create safety  56  and cohesion, such as enabling any participant, including the leading participant, to time out the scenario and receive support and debriefing.   A model for group counselling mastery (IMM) is proposed that integrates classroom learning with Experiam.  The model contains two axes, shown in figure 3-1.The vertical axis defines who holds power in the environment, ranging from peer-based to the presence of an expert (i.e. instructor/supervisor).  The horizontal axis defines methods of learning, and ranges between didactic instruction and experiential practice.  The axes define four learning environments.    Figure 3-1: Integrated model of mastery (IMM) Note: The vertical axis represents who holds power in the environment (expert versus peer) and the horizontal axis represents how learning takes place (instruction versus practice). Four learning environments are defined, where optimal learning occurs in the upper right and lower left environments. The most common environment is the lower left quadrant, in which an expert is didactically teaching new knowledge.  This environment is often a classroom setting, where Instruction PracticeExpertPeerAnxietyResistanceClassroomExperiam 57  the students listen and interact primarily with the expert.  When the expert is absent in a didactic teaching environment, resistance is argued to arise between peers, shown in the upper left quadrant.  In other words, peers resist being taught new knowledge by each other, suggesting that new knowledge transfer lies in the domain of the expert.  In the lower right quadrant, the presence of an expert while students attempt to practice new skills is argued to foster a climate of evaluation or anxiety, thus leading to lower self-efficacy and performance.  Finally, the upper right quadrant defines peers practicing together without an expert present.  This is the quadrant where Experiam takes place.  The arrows illustrate that optimum learning occurs when students move from knowledge transfer in the didactic/expert quadrant, to peer-based practice in the Experiam quadrant and back again.  According to this model, resistance in an Experiam group may reflect a climate of instruction rather than practice, while anxiety may indicate a perception of members taking an evaluative role with one another.  Once such group processes are identified, corrective steps can be taken to foster an optimal peer-learning environment. The model is enacted when students move iteratively between the didactic/expert environment, such as the classroom, and the practice/peers environment (i.e., Experiam).  This embraces the Kleinian “scissor” metaphor to student development, where one blade is experience and the other is knowledge (Barlow, 2004, p. 116).  For example, students learn and discuss skills in the classroom, then apply the skills in an environment that fosters risk taking and peer-based feedback.  Practice often brings new questions that can be brought back to the classroom for discussion and refinement from the expert.  Students can then return to practice with more finely tuned understanding of the skills to practice again, and  58  skill embodiment or mastery arises from this process, shown as an upward progression in figure 3-2.  Figure 3-2: Integrated model of master (IMM) Note:  Adding the third dimension to the IMM that describes how iterating between didactic classroom and Experiam environments increases embodiment or mastery of skills.  The IMM guided the integration of the Experiam lab into the didactic group counselling course.  A manual for Experiam and its integration into a classroom course is now presented. 3.2 A manual for Experiam This manual describes how to run Experiam sessions that are embedded in a course on group leading.  It begins with an overview of the program and glossary of terms, then describes the design, procedures and roles, and ends with steps to integrating the program into a didactic course. ExpertInstruction PracticePeerEmbodiment/Mastery 59  3.2.1 Overview of Experiam and glossary.  In one sentence, Experiam is a strength-focused peer group that allows participants to lead a video-recorded group through a pre-planned scenario for 10-15 minutes, followed by guided feedback from the group.  A distinction must be made between the levels of group process occurring, as Experiam creates a group-within-a-group.  The session defines the overarching group of counsellor trainees meeting to practice group leading skills.  The scenario defines a subordinate group in which realistic group process scenarios can be enacted and facilitated. Session:  A three-hour group session where up to six students and 1 or 2 lab leaders meet to practice video-recorded leading.   Theme:  A theme is set for each session and defines the kind of group in which the leading turns will take place.  This enables members to project themselves into a real group and take roles that would support such a group.  The theme also allows the leading participants to place their scenarios in a realistic context.  Common themes include anxiety groups, relationship groups, emotional regulation groups (e.g. anger, sadness).  The theme should be agreed to by the group to ensure all members can identify with the group, and to ensure no members find the theme too personally triggering to be a part of (e.g. a grief group for a student who recently experienced a personal loss would not be appropriate). Lab leader:  A peer to the participants who is not involved in the evaluation of the course, who has taken the Experiam program and who has taken  60  additional lab leader training.  The lab leader leads the overall session, sets the theme, facilitates scenario length and feedback, helps to refine expectations of roles and can intervene in the scenario if a time-out occurs or if the scenario becomes destabilized or harmful to participants.  Lab leaders otherwise act as participants during each scenario. Leading participant:  The participant leading each scenario.  They determine the intensity and spontaneity of the roles to be enacted, and define the type of feedback that other participants can provide.  Role participant:  Other participants take roles in service of the scenario.  A description of roles can be found in Benne and Sheats (1948) and Yalom and Leszcz (2005). Leading turn:  A leading turn consists of two parts: the scenario and the feedback, and takes in total 20-25 minutes.  All participants take a leading turn in each Experiam session. Scenario:  During each session, each participant leads a ten-minute simulated group scenario.  There are four specific scenarios that the course instructor will provide skills and modelling to facilitate, described below. Time out: A time out is called by the leading participant if she/he feels stuck or overwhelmed by the scenario.  It may also be called by the lab leader if she/he determines the scenario is becoming threatening to members of the group.    61  Feedback:  Feedback is provided for ten minutes, following the scenario.  Participants provide concrete feedback on specific verbal or nonverbal behaviours of the leading participant.  Interpersonal feedback is facilitated by the lab leader, using structured guidelines (Jung, Howard, Emory, & Pino, 1972, p. 112-113; Morran, Stockton, Cline, & Teed, 1998).  This ensures members can provide and receive constructive criticism that maximizes learning and minimizes defensiveness.  Note that feedback is not intended for discussion or analysis of the scenario, but to provide the leading participant with specific and concrete observations. Role residue A term to describe negative impacts a role participant may experience, following an enacted role.  Role residue is actively processed by the lab leader to ensure participants are not emotionally injured in a role, or that group cohesion is not reduced by participant reactions to a role. Up and out:  The lab leader debriefs the entire session for 5 to 10 minutes to provide a space for members to discharge any negative roles they took (i.e. role residue), and to process emotions arising from the session.  3.2.2 Design overview.  Experiam comprises six weekly 3-hour sessions that run concurrently with a didactic group counselling course.  Each Experiam lab can accommodate up to six students, meaning that for a class of 18 students, three Experiam labs would be conducted.  The limit of six participants was chosen for two reasons.  First, the optimal size for a process-oriented group is six to eight participants.  Fewer results in the group becoming too small to generate small  62  group dynamics.  Groups exceeding eight members have been shown to have decreased levels of interpersonal interaction (Yalom & Leszcz, 2005).  Second, all participants take a leading turn every session.  As leading turns take approximately 30 minutes to accomplish, six participants require three hours.  Labs exceeding three hours in duration may result in students becoming too fatigued to practice and learn effectively.  In addition, many graduate courses in universities have a three hour limit, making this a time period that students are accustomed to.  Having students lead every session accomplishes three goals.  It establishes a systematic desensitization to the fear of group leading, it creates a momentum of learning where skills are regularly practiced, and it fosters high cohesion and universality, as participants collectively attempt scenarios.  The lab runs for six sessions to enable students to practice four seminal scenarios and have two additional weeks to re-attempt a scenario or to try a new scenario.  The first four weeks have pre-set leading scenarios, described below.  The final two weeks are open for trainees to choose the scenario.  The six sessions provide 18 clock hours of student group leading and participation, which nearly doubles CACREP minimum standards for group counsellor training (CACREP, 2009). 3.2.3 Session procedure  Each session, the lab leader begins with a short check-in and sets the theme for the session.  Leading turns have the following structure: (1) The leading participant describes the stage and climate of the group (e.g. second session of a mandated group), the intensity of roles to be taken, and any specific feedback they are seeking; (2) They start the video recorder and begin the scenario, at 10 minutes, the lab leader signals time, either by a small gesture or alarm on a timer, and the leading participant takes up to 5 more minutes to  63  conclude the scenario; (4) Immediately following the scenario, the lab leader engages the group in 10 minutes of feedback.  This involves checking the leading participant’s emotional/cognitive/somatic state, then engaging each member for objective, concrete feedback, while limiting meaning making or scenario discussion; (5) If time permits, participants can take a short break while the next leading participant prepares his/her scenario.  (6) After the final scenario, the lab leader will lead a short up and out, involving debriefing the group members as themselves.  The goal is to prepare the members to leave the group without feeling like they are opened up to process or strong emotions and to integrate learning from the session.  If participants experience role residue, they are encouraged to express any feelings related to the role or to clear the air between members who had conflicting roles.   3.2.4 Roles  Enacting roles serves several functions.  They enable participants to fully engage the behaviours required for the scenario, and can be titrated if the leading participant wishes for more or less intense role enactment.  They can also protect role participants from perceptions that behaviours enacted in service of the scenario are reflective of their personal identities.  The last point is ethically responsible, as it reduces role residue.  All participants choose a pseudonym for the session, which distances them from the roles they undertake, and allows them to de-role at the end of the session.  For example, in discussing feedback involving a role, the group will refer to the pseudonym and not the role participant’s real name.  Finally, participants not enacting a specific role are asked to respond as they would in a real group.    64  3.2.5 Scenarios There are four scenarios that trainees must attempt in order to gain a range of leading skills, and are attempted in a sequence that is progressively more challenging and that is integrated with the skills taught by the course instructor.  The same scenario is attempted by all trainees in a session to minimize setup time for each leading scenario and to allow vicarious learning of others leading a similar group experience.   The first scenario is initiating a cold group, using basic listening skills such as empathy, paraphrasing and summarizing, and group leading skills including linking, consensus building, and information giving.  In this scenario, participants do not take roles.  The second scenario involves working with a challenging role, such as a withdrawer or aggressor, and requires skills such as blocking and norm setting for the aggressor, and perception checking, reframing to goals, and consensus building for the withdrawer.  The withdrawer, which Yalom and Leszcz (2005, p. 297) term the “silent client”, is of particular importance in group counselling.  Smokowski et al. (2001) found that withdrawn or timid clients tend to be most vulnerable to damaging group experiences and are most likely to suffer long term psychological distress from such experiences.  They advocated that leaders refine skills to ensure groups are properly structured and that withdrawn clients are carefully integrated into the group before they are perceived as aloof or uncaring by members. The third scenario involves facilitating feedback between two members who may be disagreeing or violating group norms by criticising one another.  This scenario requires skills including perception checking, empathy, paraphrasing, and open questioning.  Two helpful guides are in Jung et al. (1972) and Morran et al. (1998).  The fourth scenario involves the leader being challenged by a group member.  This scenario is very challenging to trainee  65  self-efficacy, and requires skills including empathy, clarifying, reframing to goals, and consensus building.  A helpful guide is provided in Donigian and Malnati (1987), and specific skills must first be introduced by the course instructor.  In such high conflict scenarios, remaining non-defensive is a critical intrapersonal skills that leaders must demonstrate (Stockton et al., 2004). Although the four scenarios are required, the program is fundamentally person-centred, and allows trainees to alter their scenarios according to their personal and professional development goals.  The trainee can choose the intensity and reactivity of members in scenarios that provide an optimum level of challenge.  For example, the challenge-the-leader scenario can be as intense as a member yelling at the leader about an unchangeable personal characteristic, or as benign as a member making a single, calm, dissatisfied comment about a concrete, changeable behaviour of the leader.  In this way, the trainee can attempt a feared scenario with proxy control over scenario intensity (Bandura et al., 1982).  According to these authors, placing scenario control in the hands of leading participants should increase self-efficacy by permitting them to prevent, attenuate, or terminate aversive events.  If the leading participant can control of the degree to which the event is fearsome, they will not become preoccupied with personal deficiencies or catastrophize potential difficulties in accomplishing the task.  The final two scenarios are open to the trainees to choose, allowing for reattempting of scenarios. 3.2.6 Lab leaders  Lab leaders are peers to participants.  While they are not involved in the evaluation process of the course, they are leaders of the Experiam session and are vested with authority to help the group achieve its goals.  They have taken a group counselling course and have  66  been trained in delivering Experiam, according to this manual.  They do not need to be highly experienced in group counselling, but must have a clear idea of their task and process roles. Task roles include instructing the group about how the session and scenarios will proceed, such as by establishing the session theme, and keeping time limits for leading turns.  Lab leaders also limit scenario length and complexity, according to time and group cohesion limits.  For example, a leading participant may wish to attempt a scenario that is too emotionally charged for the group, and so the lab leader will work to reduce the intensity of the scenario, such that all participants feel safe.  Finally, lab leaders direct role participants to stay within a window of tolerance for the leading participant (Ogden et al., 2006).  This involves ensuring the role is balanced between extremes of safety and realism.  Regarding safety, leading participants are working with new skills, and may not effectively facilitate role participants’ needs.  In a real group, not attending to one incident in a group can often spontaneously spawn more intense incidents.  While this can be therapeutically indicated in group counselling, the goal of Experiam is the practice of skills specific to a scenario in a strategic fashion.  Therefore, role participants are asked to enact the role only until the leading participant has done something to shift them out of it.  Regarding realism, role participants are directed to be as realistic as possible in the role and to not bail out the lead.  The empathic nature of many counsellor trainees may cause role participants to soften the role if the leading participant seems to be struggling.  This may limit the realism for the lead, or provide them with a false sense of accomplishment.   There are four process roles for lab leaders.  First, lab leaders create safety and cohesion in the group.  As peers, this is most effectively accomplished when they work collaboratively with their peers, are consistent to the design of the lab, and do not try to teach  67  their peers.  Second, lab leaders must enforce norms and feedback guidelines, as participants may inadvertently provide injurious feedback, or engage the group in scenario discussion.  Lab leaders can model feedback guidelines and raise awareness that leading participants may be vulnerable to critical feedback immediately following a scenario.  Toth and Erwin (1998) provided guidelines for teaching and modelling effective feedback in skills-based curricula. Third, the lab leaders take over group leadership in the case of a time out.  However, lab leaders must resist timing out a scenario at the first sign of the leading participant struggling with the scenario, as this can deny the leading participant the right to time out when they feel they must.  The lab leader should only time out the scenario if it is becoming unsafe for participants, or the scenario is spontaneously evolving beyond what was planned.  In both cases, the lab leader will neither instruct nor discuss the scenario.  Instead, they will engage the group to provide one piece of concrete feedback concerning what they would want as members of the scenario.  The leading participant can choose their next skill or intervention from these pieces of feedback and will resume the scenario.  If they feel too overwhelmed to continue with the scenario, the lab leader will provide them with a simple statement with which they can close the scenario, such as perhaps we can take a short break to cool down.  In this way, the leading participant (and not the lab leader) has successfully concluded the scenario, even if it was not facilitated exactly according to plan.   Finally, lab leaders keep the group in an action-oriented climate.  This requires that lab leaders take many characteristics of the director in psychodrama (Corey, 2012; Gladding, 2012).  The lab leader keeps the action moving and limits the lead, the group, and themselves, from engaging in discussion, providing rationales for ideas, or storytelling.  The anxiety of the group task can often cause group members to engage in this form of  68  conversation, and not only shifts the group from an experiential mode, but consumes valuable time not available in the design.   3.2.7 Integration with the course Two criteria should be established to ensure students in the course maximize each Experiam session.  First, the group counselling course should teach skills to be practiced prior to the Experiam sessions.  Group counselling courses can focus on many pedagogical themes, including theory, design, dynamics, roles, pathology, theoretical orientation, and so on.  However, skills must be taught prior to the Experiam sessions, else students may not navigate the scenario using skills demonstrated to increase cohesion.  In addition, skills relevant to the upcoming session scenario should be taught and modelled by the instructor to ensure students are adequately prepared to attempt a similar scenario.  For example, prior to the session involving the challenge-the-leader scenario, the instructor could explain the theory and practice of the scenario, and can demonstrate how to facilitate such a scenario in class. Second, the instructor should invite feedback and discussion following each session to ensure students can complement their practice with additional expert instruction.  Video recordings are vital to this process and can be used in assignments, as classroom case presentations, or in one-on-one instructor feedback.  Skills that were not understood can be refined or modelled by the instructor, which prepares students for subsequent sessions.   A preliminary qualitative group case study explored what was working and what could be improved in delivering the program, and is presented next.  69  3.3 A preliminary study of Experiam A preliminary group case study was conducted with a single group of five first-year masters students in counselling psychology (3 female, 2 male), co-facilitated by the two co-authors of the program, both of whom were also masters students (herein termed lab leaders).  The study received ethical approval from the institutional review board of the university.  Participants and lab leaders had completed the required graduate course in group counselling, and were known peers to each other.  The Experiam program was conducted for nine weeks in 2010 at a large APA-accredited, western-Canadian University under the supervision of a faculty member who was also a group counselling expert with over 25 years of experience.  Over the nine weeks, participants attempted a range of scenarios, ranked in complexity from easier to more challenging.  The faculty member was asked to join the seventh session, as the challenge-the-leaders scenario exceeded the peer group’s collective ability to facilitate. Following the Experiam program, lab leaders facilitated a videotaped, two-hour, semi-structured focus group interview to explore participants’ perceptions of the research question: What elements of peer-led, experiential group leading impacted (helped or hindered) counsellor trainee confidence and competence?  The interview questions explored elements of the group design, including situations that arose during the workshop.  For example, they were asked how the peer group impacted thoughts, feelings, or behaviors while group leading.  The interview was transcribed and results were analysed using enhanced critical incident technique (E-CIT; Butterfield, Borgen, Maglio, & Amundson, 2009).  E-CIT enables researchers in counselling psychology to collect observations from people about their own or others’ behavior and to divide these observations as either helpful  70  or hindering (Woolsey, 1986).  A third category, called wish list items, captures participants’ statements describing what could have been done to impact their experience.  Butterfield et al., (2009) stated that E-CIT is useful for exploring little-understood events, incidents, or factors that promote or detract from an experience or activity.  Statements relating to the research question were identified and later sorted into helpful or hindering categories, according to common themes that emerged and that had 100% agreement from participants.  The analysis compared the number incidents in each category and not the number of individuals reporting them.  This was because the focus group may have caused one participant response to influence the response rate of others.  The analysis was subjected to credibility checks defined by Butterfield et al. (2009), satisfying 8 out of 9 checks.  The unsatisfied check required saturation of categories across individual interviews, which was not achievable due to the focus group interview employed.   3.3.1 Results One hundred and eighty-eight participant statements were identified and organized into 161 critical incidents (CI) and 27 wish list (WL) items.  The CIs comprised 114 helpful incidents that were reported to increase participants’ confidence and competence as a group leader, and 47 hindering incidents that were reported to decrease confidence and competence.  CIs clustered into seven categories, sorted by highest frequency of statements.  These categories emerged inductively by grouping similar incidents within a frame of reference that sought to identify impacts on participants’ confidence and competence (see table 3-1).          71          Helpful Incidents (N = 114)  Hindering Incidents  (N = 47)  Wish list Items  (N = 27) Category n (%)   n (%)   n (%) Involvement of the expert 26 (23)   7 (15)   10 (37) Working as peers 28 (25)  3 (6)  4 (15) Leading scenarios 16 (14)  14 (30)  2 (7) Post-leading learning 20 (18)  0 (0)  6 (22) Role playing and observing others 14 (12)  5 (11)  1 (4) Group logistics and design 1 (1)  15 (32)  4 (15) Lab leader role 9 (8)   3 (6)   0 (0) Note:    Categories are sorted by highest frequency of combined statements. Table 3-1: Critical incident and wish list categories (N = 188) 3.3.2 Involvement of the expert.  This category contained the most statements, and involed the impact of the faculty member’s attendance in the seventh session. This suggested that the expert’s presence was valuable, yet was not without cost.  Participants identified helpful elements including the expert guiding them through the scenario, the expert providing a structured systematic process to address the scenario, and watching the expert demonstrate the scenario.  Hindering elements included the expert’s presence decreasing spontaneity and causing performance anxiety.  The wish list items in this category included wanting the expert to come earlier in the training program to allow more subsequent sessions for peer practice. 3.3.3 Working as peers.   This category captured elements of a working group of equals, including participant and lab leader collaboration in leading scenarios, taking roles, and providing feedback and  72  discussion.  Participants described several helpful elements: they felt a sense of safety and mutual respect working in a familiar and cohesive group of colleagues; they mutually discovered solutions to leading scenarios; they had the freedom in the peer group to make mistakes or try strategies that differed from the instructor; and they did not feel the pressure to conform or perform.  They further described the peer group as a way of learning that was very different from the classroom, specifically because there was no authority figure dictating a single way to lead groups.  They stated that the peer group allowed them to discover their own leading styles.  A hindering element was the unequal power differential created by their peers occupying lab leader roles.  The wish list items in this category focused on rotating the lab leader role among all participants, which participants felt could augment the peer-led nature of the experience and decrease the power differential between lab leaders and participants. 3.3.4 Leading scenarios.  This category contained all statements participants made about leading the 10-minute scenario.  Helpful elements included experiencing realistic roles played by other participants, success in leading complex scenarios, and increased confidence from the progression of weekly scenario complexity.  Hindering elements included feeling overwhelming responsibility as the leader to fix challenges that arose in the group, finding that the theory and micro skills interfered with spontaneous leading, feeling that harder scenarios were not optional, and choosing scenarios that were either too hard or too easy.  The wish list item involved participants having more freedom to discover their own leadership styles.  73  3.3.5 Post-leading learning.  In this category the impact of post-leading activities was described, such as interpersonal feedback, scenario discussion, video recordings, and self-reflections.  Participants reported specific behaviors they noted from the leading participant and the impacts these behaviors had on them.  Scenario discussion was a process where the group collectively reflected on components of the scenario itself, such as the leader’s interactions, and how the scenario progressed.  Participants reported no hindering incidents, and helpful elements included learning leading behaviors that others considered effective, and receiving concrete, objective feedback with which they could later compare to their video clips.  Wish list items included wanting more critical feedback from others, and having in-group time for personal journaling to augment the video clips as a historical record of group-leading experiences. 3.3.6 Role simulation and observing others.  This category described what impacted participants as they took roles, observed others in roles, and observed leading participants facilitate roles.  Helpful elements included gaining a felt-sense for different roles that clients present in groups, observing how leading interventions impacted them in their roles, and observing how peers facilitated members in conflict.  They also found it helpful to discuss the limits of roles before starting the scenario, which was particularly important when participants were asked to play stigmatizing roles, such as a group member with racial prejudices.  Hindering elements related to the emotional consequences that participants continued to experience after they took on roles, defined here as role residue.  For example, participants who agreed to take on roles where they verbally attacked each other found themselves in a persistent negative state after the session ended.   74  The wish list item in this category suggested that more effective de-roling might have enabled the participants to take greater risks in the group.   3.3.7 Group design.  This category captured statements about the group’s duration, time, and interruptions to schedule or absence of participants.  Hindering elements and wish list items focused on the importance of consistency in the planning, timing, and composition of the group.  3.3.8 Lab leader role.  This category was the least discussed and contained no wish list items.  Participants conceptualized the purpose of the lab leader role as one that established a cohesive working group, fostered risk-taking, and provided logistic duties, such as keeping time and bringing video cameras.  Helpful incidents identified included lab leaders fulfilling a necessary role for the effective functioning of the group and confronting participants who violated group norms.  Hindering incidents focused on how the power differential inherent in the lab leader role diminished safety and creativity in the peer group.  3.4 Discussion This case study sought to answer the question: What elements of peer-led, experiential group leading impacted (helped or hindered) counsellor trainee confidence and competence? The results provided answers to this question and extended existing research in the literature.  Three novel contributions are discussed. First, a window of optimal scenario challenge was suggested by between-session and within-session findings.  Between sessions, participants stated that the gradient of progressively-challenging scenarios positively impacted their confidence and competence.  Having choice along a gradient of scenario complexity enabled participants to maximize self- 75  learning, and substantiates Bandura’s (1982) experiments that demonstrated individuals’ increasing self-efficacy through mastery of ever-more threatening activities.  Within session, scenarios that were either too hard or too easy negatively impacted confidence and learning.  This finding supported a fine-tuning of scenario complexity by the leading participant to establish Bandura’s “enhanced range of perceived self-efficacy” (p. 126) in which people successfully execute tasks.  Going beyond this range required the presence of the expert.  The instruction and modelling of a highly threatening situation enabled most, if not all, participants to go beyond their comfort zone, leading to a sense of success.  This exemplifies Bandura’s (1982) definition of self-efficacy as a process of succeeding at a task that was previously thought to be just outside one’s abilities.  Participants reported feeling confident and motivated to re-attempt the scenario as peers the following week. Second, the tension between peer and expert models of learning exemplified the learning environments suggested by the IMM.  The expert was introduced to help participants navigate how to successfully facilitate a group member who challenged them as leaders.  Corey (2012) explained that leaders may endure such “attacks” (p. 88) expressed toward their personhoods, rather than toward the role they have undertaken as the group leader, which may account for the strong aversion participants experienced in anticipation of the scenario.  Participant reaction was mixed to the presence of the expert.  On one hand, they stated that they gained skills and confidence from watching and being guided by the expert.  On the other hand, the expert’s presence increased fear of evaluation and decreased spontaneity in the group.  A difficult balance was struck that enabled participants to face the feared scenario (increasing efficacy), while minimizing performance anxiety (protecting efficacy).  A trade-off between expert-based and peer-based learning supports how iterating  76  between IMM learning environments may enhance mastery of group leading, and warrants further definition and research.  Third, the peer group itself provided evidence supporting its value to experiential learning, as well as a caution.  The high value placed on working as peers and post-leading learning, particularly feedback, supported studies showing that peer-led feedback increased self-awareness, and that students desired effectively-delivered critical feedback (Borders, 1991; Starling & Baker; 2000; Toth & Erwin, 1998).  Feedback from group members further contributed to participants’ self-perceptions of success in group leading, which Bandura (1982) stated was a greater influence on self-efficacy than the actual leading performance on its own.  Evidence for the role residue phenomenon was found in Romano and Sullivan’s (2000) study, in which students reported a “spill over” (p. 374) of their real self into the character role.  While the authors did not consider this effect to be a major concern, Fall and Levitov (2002) extended their work by using trained actors to minimize what they termed “role slippage”, or a contamination between the student’s role and their personal lives.  Participants in the current study felt that greater attention to de-roling would have encouraged them to take riskier roles and to ask others to take such roles, thereby safely enhancing scenario realism.  Since they identified experiencing realistic roles as a helpful element to their confidence and competence, future studies using peers in roles has merit, if role residue can be minimized.  3.4.1 Practical implications of the study The findings from this study suggest that peer-led group leading sessions can make significant contributions to group leader self-efficacy beyond course-based work.   Three  77  practical implications for counsellor trainees, supervisors, and counsellor educators are discussed.   First, the lab leaders contended with dual roles of being leaders and peers to participants, and the group struggled with being led by those seen as equals.  Most participants agreed that the dual roles were a necessary inconvenience in order to preserve an effective peer-only group, but some struggled with a perceived power differential.  For example, when lab leaders enforced group feedback norms, some participants felt an uncomfortable power differential arise in the peer group, while others felt protected by the lab leader role.  Participants reported that the power differential was made more tolerable when lab leaders made self-disclosures, participated in leading scenarios, and received critical feedback.  However, a leader role was necessary to prevent the peer group’s inclination to become a rudderless “cocktail hour” (Yalom & Leszcz, 2005, p. 449).  This suggests that the role of the lab leader in peer-led groups requires a subtle balance of sharing control and power, which may be established early as a group norm.  Three ways to maintain this balance include having lab leaders co-lead to promote a peer environment and enable lab leader modelling, restricting lab leader instructing in favour of consensus taking, and ensuring lab leaders maintain consistency in their behaviours and enforcement of norms.  Participants also reported wishing for more critical feedback from each other in the post-leading learning, but lab leaders must carefully weigh this after-the-fact wish against the need for respect and safety in the group to promote risk-taking.  In other words, participants may have perceived critical feedback as more constructive when they knew they had already succeeded in the scenario, versus when they were actually leading the scenario.  78  Second, peer groups can flounder when members’ collective wisdom and expertise reaches a plateau.  In such instances, peer groups can benefit from an injection of new skills and experience from an outside expert.  On one hand, peer-based learning environments that attempt challenging scenarios without existing theoretical or evidence-based instruction may  result in spontaneous and creative discoveries.  However, Yalom and Leszcz (2005) warned that the absence of supervision, particularly where the probability for errors in leading are increased, would likely have led to a reinforcement of ineffective leadership by participants.  This supports the IMM hypothesis that iterating between expert-didactic and peer-experiential environments can foster mastery in group leading.  Timing and rationale for such iterations are key determinants of whether participants will experience the presence of an expert as helpful or hindering to their confidence and competence, and provides rationale to integrate the lab into a course (Feiner, 1998).   Finally, the realism of the leading experience was very important to participant self-efficacy, and peer groups must foster high cohesion and safety in order for participants to adopt realistic roles in service to one another.  Role residue from enacting a challenging role cannot be underestimated by leaders of peer groups.  Brenner (1991) proposed a process-play method that offers the “student’s mask”, where participants respond differentially to each other based on privately-chosen but superficial characteristics, such as the colour of one’s clothing (p. 147).  Methods for reducing the residue in this study included making time to debrief and support participants who took these roles, limiting the number of challenging roles that one participant enacted, and ending the group with guided meditation or progressive muscle relaxation exercises that enabled participants to leave their roles behind them.    79  3.4.2 Limitations and future research There are some limitations to this study.  There was not an alternative group to compare participant perceptions, which limits the findings of the study to this group.  In particular, the E-CIT method has not previously been applied to a focus group, which meant that one credibility check could not be tested.  Future studies could interview individuals independently from the group, or several focus groups could be interviewed to ensure all credibility checks can be met.  Next, the interview data may have been influenced by the group-based interview and the interviewers also being the facilitators of the group.  Future studies could compare student perceptions between peer-led and instructor-led experiential groups.  Another avenue of research could quantitatively explore differences in perceptions of leader efficacy between peer-led experiential groups and instructor-led groups.  Future studies could examine the mechanisms that impact self-efficacy, such as vicarious learning, or how a highly-cohesive group regulates affective control. In conclusion, peer-led experiential group training uniquely impacted the confidence and competence of counsellor trainees, and is an innovative augmentation to group counsellor training.  Peer-led groups that enable students to trail blaze their own leadership style, while maintaining some expert supervision, facilitates and promotes a balance between the competing demands in group counsellor training.    80  Chapter 4: Evaluating Experiential Training on Group Counsellor Development This study evaluated the impact of a peer-led experiential training program on the development of counsellor-trainee competence and confidence as group leaders.  The six-session peer-led group simulation training program, named Experiam, provided counsellor trainees with practice applying group leading skills to address a range of group process scenarios.  A randomized control study was conducted on thirty participants in two graduate-level group-counselling courses, where half the students in each course took the program and half were wait-listed.  Participants were measured for self-efficacy, anxiety, and performance.  Group leading self-efficacy was reported to increase significantly and to large effect, while anxiety remained unchanged.  Observed leading skills were employed significantly more frequently, but with a non-significant decrease in overall skill clarity, as the complexity of the program increased.  Implications for researchers, educators, and group practitioners are discussed. 4.1 Introduction For counselling students to develop into effective psychotherapy group leaders, they require training to perform a complex array of skills with both confidence and competence.  Clinicians, researchers and professional associations agree that training group counsellors requires the integration of didactic learning and experiential practice (AGPA, 2003; ASGW, 2000; Barlow, 2012; CACREP, 2009; Feiner, 1998).  However, many counsellor-training programs do not provide an experiential component, or limit experiential training to individual counselling skills (Barlow, 2004; Robison et al., 1996).  While some group counselling skills overlap with those of individual counselling, groups involve multiple levels  81  of communication that are subject to group dynamics, and counsellor trainees may experience these interactions as complex, rapid, unpredictable and uncontrollable (Barlow, 2004, 2012; Page et al., 2001).  Furthermore, ethical concerns have been identified in experiential group training, such as multiple relationships.  These concerns may partially explain the caution counselling educators have toward this recommended training method (Kottler, 2004; Merta & Sisson, 1991; Shumaker et al., 2011).   The low prevalence of use of experiential training in group counselling may limit trainee confidence, foster anxiety, and result in trainees attempting to apply their individual counselling skills to group counselling contexts, which are argued to be inappropriate for the development of therapeutic factors in groups, such as interpersonal learning and cohesion (Barlow, 2004; Hoyt et al., 2003; Page et al., 2001; Yalom & Leszcz, 2005).  Finally, even when experiential group training is offered in counselling programs, only about half of students actually lead a group (St. Pierre, 2014).  Therefore, experiential learning, while thought to be a critical training method, suffers from at least four barriers that may contribute to its underemployment in counsellor training, namely misapplication of individual skills, lack of a theoretical basis, ethical and resource barriers, and definition.   Frameworks for experiential group training have been proposed, including observing expert models, supervised leading, and personal participation in group therapy (Merta et al., 1993; Tschuschke & Greene, 2002; Yalom & Leszcz, 2005).  Experiential-based training programs have been developed that focus on skills development through the use of simulated group leading (Fall & Levitov, 2002; Romano & Sullivan, 2000; Toth & Erwin, 1998; Toth, Stockton, & Erwin, 1998).    82  Group leadership has been described using terms such as skills, micro skills, behaviours, and interventions.  The term intervention can be confused with a collection of skills delivered over a period of sessions, such as a manualized treatment intervention for anxiety.  Similar confusion can arise from the term behaviour, which could include non-verbal gestures, such as a frown, and which may lack empirical linkages to therapeutic outcomes.  Therefore, the term skill will be used in this manuscript to describe leader actions that are discrete, measurable, and that have empirical evidence as markers of group leadership performance. An experiential group-training program, named Experiam, was evaluated at a large, APA-accredited, North American University.  Experiam was the mandatory lab component of the group counselling course.  The goal of Experiam was to develop counsellor trainee confidence and competence in group leading while addressing many of the barriers of experiential group training.  First, the program is integrated into a course in group theory and scenarios are seated in group dynamics.  Second, the program employs a facilitated peer group that reduces dual relationships and is cost effective (Romano, 1998).  Third, the program includes multiple forms of experiential learning, including direct practice, observational learning, and feedback.  Fourth, the Experiam design is seated in social cognitive theory (SCT), which addresses the gap between theory and practice (Kormanski, 1991).   The primary mechanism of change in Experiam arises from factors believed to impact self-efficacy.  Bandura (1982) stated that self-efficacy is increased through four experiential sources, namely mastery, modeling, social persuasion, and affective arousal.  He defined mastery as the successful enactment of behaviours toward an outcome, and modelling as the  83  vicarious experiencing of others’ enactive attainment.  He stated that mastery experiences impact self-efficacy more than modelling, but that a combination of the two was most impactful.  Larson and Daniels (1998) applied SCT to counsellor training and defined counsellor self-efficacy as counsellors’ beliefs or judgments about their capability to effectively counsel a client in the near future.  This definition can be extended to group contexts by changing the focus from the individual client to the group, called group counsellor self-efficacy (GCSE; Page et al., 2003).  Within SCT, GCSE is the primary causal determinant of effective group leading performance, and hypotheses regarding its relation to anxiety and subsequent performance can be tested (Bandura, 1989, 2001).  Extant studies of experiential group training can be categorized into studies of groups where trainees actually lead the group, and studies of the impact of experiential learning on self-efficacy, both in group and individual contexts.  Regarding the former, Toth and Stockton (1996) compared a skills-based program integrating experiential and didactic learning to a didactic-only control, and found that group skills were used significantly more frequent in the integrated condition.  Several studies of students’ perspectives of experiential training suggested that such approaches facilitate personal and professional growth and development (Anderson & Price; 2001; Ieva et al., 2009; Kiweewa et al., 2013).  Regarding the impact of experiential training on self-efficacy, evaluations have demonstrated increased counsellor self-efficacy in domains of: individual counsellor training programs (Daniels & Ivey, 2007; Larson & Daniels, 1998; Kuntze, Molen, & Born; 2009); leading management groups (Hoyt et al., 2003); and for counsellor trainees as participants in groups (Ohrt et al., 2013).  Larson and Daniels (1998) stated that individual practicum experiences have increased counsellor self-efficacy by over one standard deviation.  However, no study to date  84  has evaluated the impact of experiential group leading on GCSE, despite a call for such evidence (Page et al., 2003).  In this manuscript, the term self-efficacy will be used to refer to GCSE. The purpose of this study was to address this gap in the literature by evaluating the impact of the experiential group-training program, Experiam, on counsellor-trainee leadership development.  Two research questions were proposed.  The first research question was: does participation in Experiam increase counsellor trainee confidence in group leading? Four directional hypotheses were proposed using an experimental design.  Hypotheses 1 and 2 were that Experiam would improve participant self-efficacy over controls and over time, as measured by the Group Leader Self-efficacy Instrument (GLSI; Page et al, 2001).  Hypotheses 3 and 4 were that participation in Experiam would decrease counsellor trainee anxiety over controls and over time, as measured by the State-Trait Anxiety Inventory (STAI; Spielberger, 1983).  The second research question was: does participation in Experiam increase group counsellor-trainee performance?  Using a quasi-experimental design, two directional hypotheses were examined using the Group Psychotherapy Intervention Rating Scale (GPIRS; Burlingame, 2013; Chapman et al., 2010): (1) that participation in Experiam will be associated with increased group counselling skill demonstration; and (2) the clarity of acquired skills would increase with participation in Experiam.  4.2 Method 4.2.1 Participants Participants were 30 graduate students ranging in age from 25 to 47 years (M = 29.97, SD = 4.64), who participated in one of two class sections of a group counselling course at a  85  large North American university.  The two class sections were taught by different faculty members who were group counselling experts, each with over 25 years of experience.  The first section comprised seventeen participants, and the second comprised thirteen.  Overall, nineteen participants identified as female (63%), eleven identified as male (37%).  Two-thirds identified as White/Caucasian (n = 20), and one-third was split between East Asian (n = 4), South Asian/Middle Eastern (n = 3), and participants identifying as bi-racial (n = 3).  Ninety percent of participants were in a master’s degree program for counselling (n = 27), of which all but one were in the first year of their program.  The remaining three participants were in either social work or psychology graduate programs.  See table 4-1 for description of participants’ prior group leading experience.            n (%) Hours: M (SD) Prior experience as group participant? Yes 7 (23) 28.7 (33.12) No 23 (77) 0  Prior group leading experience? < 6 Months 26 (87)   < 2 years 4 (13)   Prior supervised group leading? Yes 4 (13) 12.5 (8.39) No 26 (87)  0 Table 4-1: Participant group experience Participants in each course were randomly divided into treatment and waitlist control conditions.  During the first half of the course, or the experimental period, the treatment condition attended classroom lectures and Experiam sessions and the control condition only attended classroom lectures.  In the remaining half of the course, called the follow-up period, the waitlisted control condition attended Experiam sessions and the treatment condition only attended the classroom lectures.  During Experiam sessions in each period, participants were placed into working groups of size 4 to 6.  In total, four treatment groups and three waitlist  86  groups were created.  Experiam sessions occurred outside of class time, and were facilitated by one or two senior graduate students who took the group course and lab previously, and who took additional ten-weeks (30 hours) of training as lab leaders.  There were six lab leaders who provided facilitation for the labs in this study. 4.2.2 Materials A private meeting room with a video camera was designated for the labs.  The course instructor instructed the theory and skills for the group process scenarios to be practiced in Experiam sessions.  Readings were derived from articles in group counselling literature, and are summarized in five categories with suggested references: guidelines for giving feedback between participants (Morran et al., 1998); challenging or problematic roles that group leaders encounter (Benne & Sheats, 1948; Yalom & Leszcz, 2005, p. 391-427); group climates that help or hinder group process (Gibb, 1961; Kivlighan & Tarrant, 2001); group dynamics that destabilize new group therapists (Billow, 2001; Markus & Abernethy, 2001; Schimmel & Jacobs, 2011; Yalom, 1966); and group facilitator interventions, such as empathic statements, linking, strength confrontation and blocking (Corey, 2012; Gladding, 2012; Kline, 2003).  4.2.3 Measures From a social cognitive theory perspective, confidence and competence in group leading can be argued to centrally involve counsellor self-efficacy, management of anxiety, and task performance, and so group-based measures of these three constructs were administered (Larson & Daniels, 1998; Page et al., 2001).   Self-efficacy.  The thirty-six item GLSI (Page et al., 2001) measured counsellor trainee’s self-estimates in delivering skills in group leadership, process, and diversity.  This  87  self-report instrument was developed from the Counsellor Self-Estimate Inventory (COSE; Larson et al., 1992), but adapted to group work.  It assessed participants’ perceived self-efficacy for leading groups and answers are arranged in a Likert-type scale ranging from 1 (strongly disagree) to 6 (strongly agree).  Factor analysis was conducted using 204 counsellor trainees from 16 American universities, and confirmed a one-factor solution, called group leader self-efficacy, that accounted for 38% of total variance, with high internal consistency (Cronbach’s α = .95).  The 2-week test-retest reliability of the measure was calculated to be .72 (DeLucia-Waack & Bridbord, 2004).  Discriminant validity was obtained through non-significant correlations between the GLSI and the S-Anxiety scale of the STAI and the neuroticism, extroversion, and openness to experience subscales of the NEO five factor personality inventory.  The GLSI was selected for this study because it was the only peer-reviewed and validated instrument in the literature that specifically measured self-efficacy for leading counselling groups. Anxiety (STAI). The State-Trait Anxiety Inventory consists of two 20-item subscales using a 4-point Likert-type scale anchored from 1 to 4 (thus ranging from 20 to 80 per subscale), where higher scores represent higher levels of anxiety (Spielberger, 1983).  The trait subscale is a measure of anxiety levels that are stable over time, while the state subscale assesses the person’s current state and context.  This scale has been normed on men and women over three age ranges, and across working adults, college students, high school students, and military recruits.  Daniels and Larson (2001) stated that self-efficacy is more congruent to state versus trait anxiety, and so the state measure was selected.  The transient nature of state anxiety makes the test/retest less meaningful, while Cronbach’s alpha was found to be over .9 for male and female college students.  Construct validity was established  88  for college students by comparing state subscale levels from a control of a regular class, and finding higher scores during examination conditions and lower scores after relaxation training.  Whereas the STAI is not related to intelligence or aptitude, high levels of state anxiety have been associated with a wide array of conditions that induce stress, including exam periods for college students. Performance (GPIRS). To assess counsellor performance, video-recorded group leading was coded using the GPIRS (Chapman et al., 2010).  This observer-rated scale is used to evaluate the frequency and clarity of group leader skills that enhance outcomes in group psychotherapy (Chapman, 2010).  The earliest published scale was a thesis from the University of Amsterdam and contained 26-items, each defining a group leading skill, rated on a 4-point scale ranging from 0 (poor) to 3 (excellent), and not rated if not observed (Chausovsky, Trijsburg, Snijders & Spiering, 2005).  Items were grouped into three broad categories and ten subscales, and were scored both for frequency and clarity.  In 2010, Christopher Chapman and faculty at Brigham Young University published works further developing and testing the GPIRS (Chapman 2010; Chapman et al., 2010).  The authors increased the number of items to forty-eight, derived from empirical and clinical evidence of leader behaviours that increase group cohesion, herein referred to as leader skills  (Burlingame et al., 2002).  They further categorized items into three domains: (a) Group Structuring, comprising skills that define group goals, norms, and boundaries, (b) Verbal Interactions, comprising skills that model and facilitate self-disclosure and feedback between members, and (c) Emotional Climate, comprising skills that help members express feelings in a constructive manner, and to provide appropriate balance of support and challenge.   89  Raters scored leader behaviours using a 4-point Likert-type scale of skill clarity and frequency, ranging from 1 (poor) to 4 (excellent), and 0 was coded if the skill was not observed.  The number of times each skill occurred was also recorded, providing a measure of skill frequency.  The authors stated that the internal consistency of the scale is high, with alpha = .93.  They tested concurrent validity by comparing the scale to measures of similar and dissimilar scales.  Concurrent validity was shown via correlations to an individual psychotherapy analogue, called the Comprehensive Psychotherapy Intervention Rating Scale (CPIRS; Trijsburg, Frederiks, Gorlee, Klouwer, Hollander & Duivenvoorden, 2002).  They further determined that higher rated leader skills on the GPIRS were significantly positively correlated to lower levels of hostility between group members, as measured by the Hill Interaction Matrix (HIM), considered to be the gold standard of group climate scales.  Recent unpublished research and development of the GPIRS has resulted in its reduction to 36 items and division into two 18-item subscales, one containing Basic skills and one containing Advanced skills (Burlingame, 2013; G. Burlingame & K. Bruer, personal communication, August 2013).   These researchers further reduced the scale anchors to three: intervention did not occur (0), ambiguous – intervention occurred but clarity could be improved (1), or intervention was performed with clarity (2).  Items were not coded if those skills were not observed.  Finally, videos were coded in two-minute intervals.  The coding scheme is provided in appendix B, and the revised scale is provided in appendix C. In this study, the GPIRS coding scheme was further clarified to consider the absence of skills needed to foster or maintain cohesion in the group.  Coding for the absence of leader actions has empirical support in studies demonstrating that the formation of cohesion is fostered by leaders who provide structure to groups in early sessions, and that passivity in  90  leaders can result in harm to group members (Kivlighan et al., 2000; Smokowski 2001).1  Therefore, in an emotionally-charged group climate or dynamic, cohesion may be impacted by what leaders fail to do.  For example, a conflictual group climate has been shown to reduce therapeutic outcomes in group therapy (Kivlighan & Tarrant, 2001).  Therefore, a group leader not addressing member behaviours to decrease interpersonal conflict would be assigned 0 for the item: Stopped attacking and judgmental situations and expressions between members.  In contrast, a scenario with no interpersonal conflict would likely have no ratings for this item.  Therefore, for this study, the above-mentioned anchor definitions were employed, with the (0) rating defined as intervention did not occur and was therapeutically indicated.  As in prior incarnations of the GPIRS, items were not rated if they both did not occur and were not deemed necessary.   In reviewing behavioural measures of group process variables, Chapman et al. (2010) noted that many are focused on member interactions and not leader skills, or that leader skills are not, themselves, theoretically or empirically derived.  The GPIRS was created specifically to measure empirically-tested leader skills that increase cohesion in groups, while requiring minimal rater training and inference. Therefore the GPIRS most directly addresses the research question of performance in group leading.   4.2.4 Procedure Once ethical approval was granted from the university, students in two different sections of the course who agreed to take part in the study were randomly divided into treatment and waitlist control conditions.  The treatment condition took the program in the                                                  1 For a theoretical argument of the impacts of absentee group leadership, the reader is directed to the first chapter of Bion`s (1961) theory of group development.  91  first half of the course (called the experimental period) while the control condition waited until the latter half of the course (called the follow-up period).  Self-report measures were administered at three time periods for the experimental period (pre, mid, post), with two ancillary administrations during the follow-up period.  State anxiety measures were also gathered at the beginning of the first, third, and fifth sessions of all Experiam sessions, while video recordings were collected on the first and sixth leading session.  See appendix D for an illustration of measure administration during each period. 4.2.5 Course procedure. The course instructor introduced course curriculum relevant to Experiam sessions by teaching skills to facilitate four group process scenarios.  The scenarios were: initiating a first session of a group; working with a challenging role such as a withdrawer or aggressor; facilitating interpersonal feedback between participants; and facilitating a challenge to the leader by a group member.  These scenarios were chosen based on the faculty member’s expertise and current literature on group incidents, and were arranged to be progressively more challenging to the leader (Donigian & Malnati, 1987; Tyson, Perussem, Whitledge & Kelly, 2004).  Video recordings were required for course assignments and feedback from the course instructor. 4.2.6 Experiam procedure. Experiam draws from other experiential group training programs in the literature, including role-play simulation of group process events called scenarios, practicing specific group-leading skills, and enabling participants to provide and receive feedback from each  92  other (for a review of experiential programs, see Ieva et al., 2009)2.  The program design is seated in social cognitive theory and simulates empirically-derived group scenarios in a structured, peer-led environment.  Experiam functions as a group-within-a-group, in that simulated group experiences are conducted within a structured peer-led group.  In other words, the simulated leading is designed to closely approximate real-world group processes, but these simulations are conducted as a larger group, called the peer group.  Unlike the structured scenarios that limit spontaneity to encourage skill development, the peer group can elicit spontaneous, here-and-now group processes, especially if interpersonal conflict arises between trainees.  Therefore group leaders, called lab leaders, were necessary for the Experiam session itself.  These were senior graduate counselling students who facilitated the Experiam sessions by setting and enforcing group norms, assisting in structuring the leading scenarios, and facilitating group consensus and feedback.  The structured peer group invoked their collective learning from the course instructor to address scenario challenges and supported one another in attempting new skills.  The Experiam lab consisted of six sessions, of which the first four were matched to the four scenarios presented by the course instructor, described above.  The final two sessions were left open for participants to reattempt a scenario after having received further course instruction or instructor feedback.   Each session, group members took turns as leading participants in a video-recorded 30-minute leading turn that consisted of ten minutes of leading a simulated scenario, followed by fifteen minutes for feedback and discussion by the group.  Sessions were set at three hours, and all participants (maximum six) were required to lead each session.                                                    2 For the Experiam manual, please contact the first author.  93  The Experiam design employed elements of the psychodrama technique of role training within a framework of group supervision for counsellors-in-training (Blatner, 2000; Westwood, 1989).  Role training enabled participants to enact client roles in service to the leading participant’s chosen scenario, while the framework for group supervision provided rules to safeguard against excess spontaneity that is inherent in standard psychodrama practice.  One safeguard was the establishment and modeling of group norms by lab leaders to create safety and cohesion, such as enabling any participant to time out the scenario and receive support and debriefing.   Sessions began with lab leaders establishing a theme for the simulated group, providing a realistic purpose for the group that continued through all scenarios in the session.  Each leading participant prepared for his/her simulated leading experience, or scenario, by explaining what skills he/she planned to use, and participants negotiated roles they were willing to take in service of the scenario. The lab leader facilitated a discussion around the limits of the role and the scenario, to collaboratively strike a balance between a realistic experience and a safe learning environment for all participants.  The leading participant also primed the group to observe specific verbal or nonverbal behaviors he/she wanted feedback from, following the scenario.  For example, the leading participant might ask peers to observe his/her body posture throughout the scenario and report how it impacted them in their roles.  The role-playing members then enacted the scenario with the leading participant practicing the specific skills.  After ten minutes, the scenario was concluded, and the group provided the leading participant with fifteen minutes of feedback.  Leading participants retained a video recording of their scenario for course assignments and instructor feedback.  94  4.2.7 Data analysis. Self-report measures were entered by two independent raters, and video recordings were coded by two independent raters, who coded pilot video recordings and calibrated to achieve 100% agreement in use of the coding scheme (see appendix B).  The digital video recordings permitted the GPIRS to be coded in two-minute time intervals, and so codes were cross-validated between items and across time.  Best practices for inter-observer agreement in small-sample behavioural analysis requires that the primary investigator code 100% of videos, with an independent rater coding 20% to 33% of videos at several points during the coding process (Gast, 2010; Maione, & Mirenda, 2006; McComas, Wacker, Cooper, Asmus, Richman & Stoner, 1996).  Videos from the study were randomly sorted, then coded in a three-stage validation process: after the primary author coded one-third of videos, the second rater coded a random selection of 25% of those videos and inter-observer agreement was calculated.  Percentage agreement was high with a mean frequency agreement of 95%.  To account for agreement due to chance, a kappa statistic was calculated according to Fleiss (1981), and the inter-raters achieved over 80% agreement across all three stages (k = .83), which Landis and Koch (1977) stated is indicative of excellent agreement.  Correlation between mean clarity scores was moderate, at 64.2%, indicating that while inter-raters agreed on interventions demonstrated, they did not as strongly agree on the scores given for clarity of intervention.  All data were analyzed using the software package SPSS version 22.   4.3 Results Hypotheses were based on the three dependent variables collected in the study and were tested at the 95% confidence level.  Effect sizes are calculated for between-subject comparisons using Cohen’s d with pooled standard deviations, while within-subject effect  95  sizes were downward-adjusted by the correlations within the groups, as suggested by Dunlap, Cortina, Vaslow and Burke (1996).  As found in common practice, the eta-squared statistic, 2, was reported with the ANOVA results.  4.3.1 Confidence variables: Self-efficacy and anxiety. Regarding the first research question, two-way, mixed model ANOVAs were conducted with self-efficacy and anxiety as the dependent variables.  The independent variables included one between-subjects variable, condition, with two levels (treatment, control), and one within-subject variable, time, with three levels (pre, mid, post).  Assumptions of homogeneity of variance, normality, and sphericity were all met. See table 4.2 for descriptive statistics and figure 4-1 for a plot of the group means.                     Level of Time   Level of Condition   Pre (Week 1)   Mid (Week 5)   Post (Week 8) Measure n M (SD)   M (SD)   M (SD) GLSI Treatment 17 3.96 (0.75)   4.06 (0.48)   4.62 (0.40)   Control 13 4.06 (1.06)   3.92 (0.80)   3.90 (0.67) STATE Treatment 17 35.35 (6.37)   34.76 (6.32)   35.35 (10.95)   Control 13 34.76 (6.82)   34.77 (11.81)   38.31 (13.80) Table 4-2: Means and standard deviations for dependent variables: Group Leader Self-Efficacy (GLSI) and State Anxiety (STATE) The first hypothesis was that Experiam would improve participant self-efficacy over controls, as measured by the GLSI.  There was a significant interaction between condition and time (F(2,56) = 7.65, p = .001, MSE = .18, 2 = .22; see figure 4-1).  Therefore, group leader self-efficacy significantly changed over time, depending on the condition.  Simple main effects contrasts with Bonferroni adjustment were performed at each time period, and revealed that self-efficacy between treatment and control conditions did not significantly change over the pre and mid time periods (Pre: t(28) = -.36, p = .72, d = .12; Mid: t(28) = .58,  96  p = .56, d = .20).  However, in the post time period, self-efficacy did significantly increase in the treatment condition over controls (t(28) = 3.71, p = .012, d = 1.27).  The effect size of the between-condition difference at post-test was large, using a pooled standard deviation (Cohen, 1992).  Because the mean of the treatment condition was higher than controls at post-test (see table 4.2), hypothesis 1 is confirmed, meaning that participation in Experiam caused significantly higher self-efficacy over classroom instruction alone.  Figure 4-1:  Mean Group leader self-efficacy (GLSI) for treatment and control groups over three measurement periods during the experimental period.  Note: Error bars represent 95% confidence intervals of the standard errors. The second hypothesis was that self-efficacy would increase over time for the treatment condition.  A one-way ANOVA was conducted on the treatment group, and a significant effect was found (F(2,32) = 11.67; p < .001, MSE = .19, 2 = .42), and a trend contrast resulted in a significant quadratic trend (F(1,16) = 4.88, p = .04, 2 = .23).  Pair-wise comparisons with Bonferroni adjustment of the treatment group revealed a significant difference between Mid and Post means (t(16) = 5.08, p < .001, r = .46, d = 1.28), and Pre 1 2 3 4 5 6 Pretest,  Week 1 Midtest,  Week 5 Posttest,  Week 8 Group Leader Self-Efficacy Experimental period Treatment (n=17) Control (n=13)  97  and Post means (t(16) = 3.88, p = .004, r = .38, d = 1.05), but not between Pre and Mid means (t(16) = .64, p = 1.00, d = .32, r = .53, d = .15).  As shown in table 4-2, the means for the treatment group increased over the Mid and Post time periods, therefore hypothesis two is supported and we may conclude that participants in Experiam experienced a significant increase in self-efficacy over time.  A one-way ANOVA of the control group was conducted to compare the impact of time on GCSE in pre, mid, and post test adminstrations, and no significant change was found (F(2,24) = .72, p = .50, 2 = .06).  As shown in table    4-2, control group means decreased over time, indicating that self-efficacy was static-to-declining for participants in the classroom-only condition. The third and fourth hypotheses stated that participation in Experiam would decrease counsellor trainee anxiety over controls, and over time, respectively, as measured by the STAI.  Anxiety was measured both in  class, and in the first, third, and fifth sessions of the lab.  A one-way ANOVA was conducted on trait anxiety and results indicated non-significant change between time periods (F(2,32) = .01, p = .99, 2 = .001), thus demonstrating the relative stability of respondent trait anxiety over the short term. A mixed model ANOVA, similar to the above analysis, was conducted where the dependent variable was state anxiety in both environments.  Results indicated no interaction or main effects between group and time for either variable (F(2,56) = .72, p = .49, 2 = .03).  Therefore, these hypotheses were disconfirmed, indicating that Experiam did not reduce anxiety over controls, nor over time.  However, the raw state anxiety scores were in a normal range for college students, according to the STAI manual (Spielberger, 1983).    98  4.3.2 Competence variable: Performance. To answer the second research question, a quasi-experimental, pretest-posttest design was employed to examine performance in group leading, as measured by the GPIRS.  All participants (N = 30) provided ten-minute leading scenarios of first and sixth sessions of Experiam, however two participant videos did not follow the protocol for the program and were deemed unacceptable for analysis. The remaining fifty-six participant videos were coded by observable utterances of the leading participants and categorized by session.  An utterance was defined as beginning when the leading participant spoke and ended when another participant spoke, or the scenario ended. As shown in table 4-3, there were 684 utterances across all videos (M = 12.21, SD = 2.96). These were coded into 4,352 rated items using the 36-item GPIRS, therefore a mean of 6.4 items were rated, per utterance.     Utterances N Utterances per video  M (SD) Rated items N Ratings per utterance  M (SD) Session 1 325 11.61 (3.53) 1917 5.90 (1.37) Session 6 359 12.82 (2.16) 2436 6.79 (1.79) Total 684 12.21 (2.96) 4353 6.36 (1.66) Table 4-3: Utterances and rated items in the GPIRS (N=28) As shown in table 4-4, pair-wise comparisons between the treatment and waitlist control groups and between course offerings determined that groups were not statistically different in frequency or clarity of skill demonstration, with the exception of the treatment group having significantly lower session 1 skill frequency than the waitlist group (t(26) = 5.25, p < .001).  Therefore, the groups can be grouped together and compared across Experiam sessions.   99       Treatment  (n = 17) M (SD) Control   (n = 11) M (SD) t(26) (p)   Course 1  (n = 15) M (SD) Course 2  (n = 13) M (SD) t(26) (p) Skill clarity  Session 1 1.83 (0.1) 1.87 (0.2) -0.72 (0.48)   1.85 (0.2) 1.85 (0.1) 0.01 (0.99) Session 6 1.73 (0.2) 1.83 (0.2) -1.20 (0.24)   1.71 (0.2) 1.83 (0.2) -1.42 (0.17)                   Skill frequency Session 1 56.29 (14.1) 87.27 (17.0) 5.25 (0.00)*   71.47 (23.5) 65.00 (19.1) -0.79 (0.44) Session 6 89.35 (18.8) 83.36 (15.7) -0.87 (0.39)   89.00 (16.5) 84.69 (19.3) -0.64 (0.53) Note: * p < .05           Table 4-4: Descriptive statistics for performance data (GPIRS) (N=28)  100  The first hypothesis was that participation in Experiam would be associated with increased skill demonstration.  The data were analysed between basic and advanced subscales, and across the three domains of the GPIRS (Chapman et al., 2010). As shown in table 4-5, there were significant increases in skill usage in two of three basic and advanced skill domains.  Structural skills in both domains remained unchanged or decreased  in session 6. Paired t-tests were conducted to compare the frequency of total skill demonstration across two levels, session 1 and session 6.  There was a significant overall increase in frequency of session 6 skills (M = 87.0, SD = 17.6) over session 1 skills (M = 68.5, SD = 21.5); t(27) = 3.39, p = .002, r = -.08, d = .94).  In addition, participants in session 6 demonstrated six skills not observed in session 1, in domains of Verbal Interaction and Emotional Climate; five of the skills were in the Advanced  skills subscale.  Therefore hypothesis 1 was supported, and we can conclude that participation in Experiam is associated with higher frequency and diversity of skill demonstration.      Observed items Frequency       Session 1 N Session 6 N Session 1 M (SD) Session 6 M (SD) t (27) (p) Basic Skills Structure 4 4 13.3 (4.1) 13.5 (2.7) 0.2 (0.83) Verbal 7 7 6.4 (3.8) 10.6 (5.8) 3.3 (0.00)* Climate 5 6 43 (14.1) 50.4 (9.5) 2.3 (0.03)*               Advanced Skills Structure 2 2 0.8 (1.2) 0.1 (0.3) -3.0 (0.01)* Verbal 4 7 3.1 (2.8) 6.6 (5.1) 2.9 (0.01)* Climate 4 5 1.9 (2.4) 5.8 (3.3) 4.0 (0.00)*   Total 26 31 68.5 (21.5) 87.0 (17.6) 3.4 (.002)* Note:  * p < .05.  Table 4-5: Frequency of skills across two subscales and three domains of performance (GPIRS) The second hypothesis was that clarity of skill demonstration would improve with participation in Experiam over time. A paired t-test was conducted to compare the scores for total skill clarity across two levels, session 1 and session 6.   There was a non-significant   101  decrease in clarity of session 6 skills (M = 1.77,  SD = .23) over session 1 skills (M = 1.85, SD = .14); t(27) = 1.52,  p = .14, r = -.09, d = .43).  Therefore, the evidence does not support hypothesis 2, that participation in Experiam would improve leader skill clarity.  However, two pieces of evidence refute the conclusion that Experiam decreases skill clarity. First, the mean session 1 clarity score was nearly the maximum ratable score and the distribution of scores was negatively skewed, which suggests that the initial session ratings may have suffered from a ceiling effect and that clarity skills could not statistically increase in the 6th session.  Second, thirteen of the twenty-eight participants did improve in overall skill clarity, with seven of the eight lowest session 1 clarity scores (M = 1.68) improving in session 6, although not significantly so (M = 1.84); (t(7) = 2.03, p = .08, r = .-.12, d = 1.07).  In other words, those who struggled most to lead the initial group scenario demonstrated higher performance in the more complex final scenario.  While not statistically significant, this evidence suggests that Experiam may improve participant skill clarity where improvement is possible.   4.3.3 Ancillary analysis  Four ancillary analyses were conducted, one on the GLSI data, and three on the GPIRS data.  First, self-efficacy data was collected for two follow-up time periods, to capture ongoing effects of Experiam on treatment and waitlist groups.  A 2x2, mixed model ANOVA  was conducted with self-efficacy as the dependent variable.  The independent variables included one between-groups variable, group, with two levels (treatment, waitlist), and one within-subject variable, time, with two levels (follow-up, final).  Assumptions of homogeneity of variance and normality were met.  See table 4-6 for descriptive statistics.    102       Level of Time     Follow-up 1  Follow-up 2 Level of Group n M (SD)  M (SD) Treatment 17 4.65 (0.41)  4.71 (0.42) Waitlist 13 4.10 (0.64)  4.76 (0.47) Note:  The waitlist group comprises the control group participants who took Experiam during the follow-up period. The treatment group had classroom-only instruction during this period. Table 4-6: Group Leader Self-Efficacy (GLSI) for follow-up period There was a significant interaction between groups and time (F(2,56) = 24.09, MSE = 1.34, p < .001, 2 = .046).  This indicates that group leader self-efficacy significantly changed over time, depending on the group.  Simple main effects contrasts with Bonferroni adjustment revealed that self-efficacy for the waitlist group significantly increased between follow-up administrations (t(12) = 5.44; p < .001, r = .73, d = 1.11).  Self-efficacy for the treatment group did not significantly change between these periods (t(16) = 1.02, p = .33, r = .85, d = .13), nor was there a significant difference in final levels of self-efficacy between treatment and waitlist groups (t(28) = -.31, p = .76, d = .11).  Therefore, a significant increase in self-efficacy was associated with the waitlist group taking both Experiam and classroom instruction, and not with the treatment group having only classroom instruction.  This finding, along with the results of the experiment, suggests that the self-efficacy of the waitlist group caught up to the treatment group, once they participated in Experiam, and reveals a possible ceiling to self-efficacy gains, since the treatment group did not gain additional self-efficacy following Experiam.  However, this inference is associative and not generalizable due to the absence of an independent control group. The second ancillary analysis was an item analysis of acquired skills, using the GPIRS.  Four criteria were applied to the data. First, the mean frequency each skill was coded per  103  scenario, resulting in four skills that were prevalent in nearly every leader utterance: Provide structure that facilitates member interaction—note, not member-member interaction. (e.g. feedback exercises) (94.4%), Maintained an active engagement with the group and its work (99.6%), Used nonjudgmental language with members (97.1%), and Modelled expressions of open and genuine warmth to all members throughout session (99.5%).  These items may be considered a broader class of skills that could be rated with every leader utterance, and were therefore excluded from the analysis of item frequency.  The resulting thirty-two items may be considered specific skills and were observed more frequently in basic over advanced subscales.  However, the ratio of basic to advanced specific skill use increased from session 1 (3:1) to session 6 (2:1).  This provides further evidence that Experiam scenarios provide opportunity for the practice of more advanced group leading skills.  Some skills were rarely demonstrated, so the second criterion was to include skills with a mean frequency exceeding 1.0 across all participants, and in at least one session. Some skills were only present in one session and not the other, so the third criterion was to exclude items unless present in both sessions.  Finally, the coding of several items revealed possible ceiling effects, as the distribution of scores were negatively skewed.  Therefore, the fourth criterion was to exclude items with a mean clarity above 1.8 out of a possible 2.0 in both sessions.  These criteria resulted in eight items that all fell in the basic skill subscale and within verbal and climate skill domains (see figure 4-2).   Two extensions of earlier results were drawn from this analysis.  Regarding the finding of higher skill frequency in session six, two skills with the largest increase in session 6 frequency were Maintaining functional control of group and Leader was not defensive (when leader made a mistake or an intervention failed or when confronted by a member.  These skills would be  104  expected in more conflictual group environments and may reflect appropriate leader reactions.  in six of eight items.  but lower clarity in overall skill demonstration Regarding the finding of lower skill clarity in session six, participants demonstrated improved clarity in two skills pertained to facilitating emotional expression.  Therefore, participants appeared to struggle with enacting skills related to group conflict, but did attempt more skills and with higher clarity when encouraging emotional expression among group members.  0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00Present–centered to something explicitMaintain functional controlFacilitate member-member interactionEncourage self-disclosureNot defensive when failed or confrontedEn ouraged emotional engagement between membersRespond with Empathy Assisted mbrs in describing emotionsVerbal DomainClimate DomainFrequency of intervention occurrence (GPIRS)GPIRS SkillSession 1Session 6 105   Figure 4-2: Frequency and clarity of selected GPIRS items between first and sixth sessions (N = 28).    Note that coding scheme descriptors are displayed, as shown in appendix B.  The full description of items is provided in the revised scale in appendix C.  The third ancillary analysis was to examine the qualification made to the GPIRS coding scheme.  As mentioned in the measure description, skills not observed and not expected were not given a rating, whereas a rating of 0 was given when a skill was absent, but deemed necessary to have occurred.  In all videos (N=56), a code of 0 was assigned 137 times, or 3.15% of all ratings (N=4353).  Half of the GPIRS items were rated as absent at least once (n = 18).  Items were rated 0 more frequently in session 6 (n = 98) than session 1 (n = 39), with the most frequently absent skills in the last session being: Recognized and responded to the meaning of groups members’ comments (n = 24), Encouraged self disclosure relevant to the current group agenda without “forcing it” (n = 14), and Modelled expressions of open and genuine warmth to all members throughout session (n = 12).  Therefore, participants in this session were less likely to respond with empathy or warmth, and may fail to encourage member disclosures by responding too often or lecturing.  These skills were demonstrated frequently in both sessions, thus providing 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00Present–centered to something explicitMaintain functional controlFacilitate member-member interactionEncourage self-disclosureNot defensive when failed or confrontedEncouraged emotional engagement between membersRespond with Empathy Assisted mbrs in describing emotionsVerbal DomainClimate DomainClarity of intervention rating (GPIRS)GPIRS SkillSession 1Session 6 106  evidence that empathy, warmth, and encouraging member disclosures may be skills most frequently dropped during more conflictual Experiam scenarios. The fourth ancillary analysis was an examination of the GPIRS as a time-series.  Utterances were coded in 2-minute intervals (see appendix B for coding scheme).  This provided the opportunity to analyse the data as a 5-point time series (see figure 4-3).  A 2x5 repeated measures ANOVA was conducted with skill clarity as the dependent variable.  The independent variables included one between-groups variable, session, with two levels (session 1, session 6), and one within-subject variable, time interval, with five levels (0 – 2 minutes, 2 – 4 minutes, 4 – 6 minutes, 6 – 8 minutes, 8 – 10 minutes).  Assumptions of homogeneity of variance, normality, and sphericity were only partially met, and so Huynh-Feldt statistics are reported.  There was not a significant interaction effect (F(3.41,197.7) = 1.72, p = .16, 2 = .03).    Figure 4-3: Skill clarity (GPIRS) between session 1 and session 6, coded in five 2-minute time intervals.   Note: Mean (SD) provided at each interval.  1.84 1.85 1.84 1.85 1.851.861.761.711.751.780.00 51.01.52.00-2min 2-4min 4-6min 6-8min 8-10minSkill Clarity (GPIRS)Time intervalSession 1 (n=28)Session 6 (n=28)(.21) (.24) (.17) (.18) (.19)(.19)(.34)(.35)(.31) (.27) 107  One-way repeated measures ANOVAs were conducted on each session.  Skill clarity was not found to significantly change in session 1 (F(2.6,71.1) = .03, p = .99, 2 = .001), or session 6 (F(2.8,75.0) = 1.84, p = .15, 2 = .064).  Session 1 scores for each time interval were above the 95th percentile for rating, and variance was lower than session 6 in all time intervals except the 0 – 2 minute interval.  Therefore, session 1 scenarios did not demonstrate variation in skill clarity, and may have been subject to ceiling effects.  A within-subject trend contrast of session 6 scores revealed a significant quadratic trend (F(1,27) = 7.31, p = .01, 2 = .213).  Therefore, the final session exhibited a u-shape where skill clarity dropped from beginning to the middle of the scenario, and then made a recovery towards the scenario’s end. 4.4 Discussion Two research questions were addressed by this study.  First: does participation in Experiam increase counsellor trainee confidence in group leading? Four hypotheses were tested; specifically that participation in Experiam would improve self-efficacy over controls and over time and would decrease anxiety over controls and over time.  Second: does participation in Experiam increase group counsellor-trainee performance? Two hypotheses were proposed; specifically that participation in Experiam would be associated with increased frequency and clarity of group skill demonstration.  The study employed self-report and behavioural observation measures that were empirically-validated and congruent to the theoretical framework of the Experiam design, thus providing psychometric validity to the evaluation of experiential group counsellor training (Romano & Sullivan, 2000; Toth & Erwin, 1998; Toth & Stockton, 1996). Experiam was found to significantly increase group leader self-efficacy and to a large effect, but to not have a significant impact on anxiety.  Ancillary analyses were conducted that  108  revealed significantly increased self-efficacy for the waitlisted group after they participated in Experiam.  In addition, frequency of existing and new skill usage increased between first and final practice sessions, while clarity of skill demonstration decreased but not significantly so.  An item analysis was performed, highlighting that in session 6, participant leaders more frequently attempted to maintain control of the group and of themselves, by staying non-defensive to members.   There are several novel contributions of this study.  The primary aim was to demonstrate that an experiential training program increased counsellor trainee confidence and competence in group leading.  The significant gains in self-efficacy for the treatment condition over controls replicates, in a group context, similar findings for individual counsellor-training programs (Daniels & Larson, 2001; Urbani et al., 2002).  Participant self-efficacy increased by a large effect size at the same time a more diverse set of advanced skills were applied to more complex critical group incidents.  This evidence is supported by Bandura’s (1982) experiments demonstrating increasing self-efficacy through mastery of ever more threatening activities, and provides an argument that Experiam effectively increased counsellor-trainee confidence in group leading.  The constancy of state anxiety in a normal range revealed that participants were not overwhelmed by anxiety while performing a range of new skills.  The experimental design further permits the causal inference that Experiam raises self-efficacy to be generalized to a population of counsellor trainees. A related finding was that the classroom-only group did not improve in self-efficacy, and in fact reported a small decline.  After taking Experiam, this group did report significantly higher self-efficacy and to large effect, while the treatment group did not report any further significant increases in self-efficacy during the same period.  These findings support the SCT proposition  109  that self-efficacious beliefs are not associated with observational learning alone, such as would be presented in a classroom setting, but that didactic learning must also be accompanied by experiential learning (Bandura, 1982, Barlow, 2012; Larson & Daniels, 1998).  The assumption that experiential learning is a critical ingredient in the development of group leading self-efficacy, above didactic instruction alone, is therefore supported.   Several novel findings arose from the analysis of the performance variable in the second research question.  Participants demonstrated a diversity of group leading skills from every domain and subscale of the GPIRS.  They demonstrated significantly more skills in the last session, particularly advanced skills that promoted verbal interactions between members and that fostered a climate of emotional expression.  Given that later sessions involved more complex group scenarios, the presence of a more diverse set of skills suggests that Experiam was effective in preparing students to perform skills known to build cohesion in multiple group process scenarios.  This replicates Toth and Stockton’s (1996) experimental design that demonstrated a skills-based training program integrating experiential and didactic learning significantly increased the frequency of group skills used, over classroom-only controls.  More broadly, this finding addresses the pedagogical call to educate counsellor trainees in skills seated in group dynamics, rather than applying individual counselling skills in a group setting (Barlow, 2012).   The decline in skill clarity over time, while non-significant, was unexpected and suggests that the first session was subject to ceiling effects.  The first session entailed initiating a group for the first time, and ceiling effects can be explained by group dynamics.  Yalom and Leszcz (2005, p. 310) stated “the first group therapy session is invariably a success”.  They described high anxiety on the part of members, and stated that leader interventions involve establishing the relevance of the group to members (i.e. goal setting), and to provide structure that helps members  110  find a sense of belonging in the group.  They stated that members look to the leader to provide a sense of universality, and that a silent leader may amplify anxiety or aggression in members.  Therefore, any active leader interactions that encourage member disclosure may suffice, including active listening and other individual counselling skills that trainees may already possess.  The flat-to-decreased skill clarity over time may have reflected the increasing complexity of leading scenarios that require group leading skills new to the trainees.  Decreased skill clarity may be a pedagogical marker in that trainees were practicing new skills, rather than relying on individual counselling skills, or eclectic reactions.  Examining skills that were absent but therapeutically indicated revealed skills that students struggled to enact.  In this study, warmth, empathy, and encouraging member disclosures arose as skills most in need of development in the final Experiam scenario.  Therefore, leading performance may have been reduced due to delivering the expected group skill with poor clarity, which could be addressed by additional practice, or from ignorance or avoidance of an indicated skill, which may require classroom instruction or supervision.   Finally, analyzing only those items that were demonstrated in both sessions, and that were not skewed to the ceiling of the score scale, resulted in eight basic skills.  Several items related to establishing structure and control, while also remaining non-defensive and empathic to conflictual members.  High structure has been demonstrated to increase cohesion in groups, particularly in early sessions (Kivlighan et al., 2000).  This suggests that trainees were attempting skills known to increase cohesion when conflict may have threatened the cohesion and safety of the group.  In particular, clarity suffered as trainees maintained control of the group and compartmentalized their internal anxiety, but trainees demonstrated higher clarity in encouraging emotional expression among members.   111  In summary, trainees performed with lower clarity, but higher frequency, which describes a sustained effort in the face of challenge: a hallmark of self-efficacy (Bandura, 2001).  However, skill clarity did increase, albeit non-significantly, for those participants who had the lowest session 1 scores, suggesting that Experiam may be particularly helpful for students who struggle with basic group leading skills.  Extrapolating these results, the Experiam program may be associated with trainees perceiving critical group events as a surmountable challenge, and thus were engaged in Bandura’s (2001) triadic reciprocal causation for human agency: they derived motivation from the group (environment) to proactively plan their scenario (cognitive and affective factors) and then successfully enact the plan (behaviour).  These three determinants interacted to increase self-efficacious beliefs and motivation to master more complex skills. 4.4.1 Limitations and implications for research and pedagogy There were some limitations to this study.  First, the sample size was capped by the availability of participants within a single academic year.  This may have limited the power to detect statistically significant effects, thus resulting in findings that were due to sampling error.  Future studies could expand the sample size over 40, which Hsu (2003) stated would improve the chance of equivalence between comparison groups despite possible nuisance variables.  In particular, the frequency of skills in session 1 was significantly higher for the waitlist group, which suggests that classroom participation may have confounded that group’s first-session skill knowledge (see table 4-4). A related limitation arose because the anxiety measure was taken during class time and at the beginning of three Experiam sessions.  The former measurement period produced anxiety relevant only to the classroom.  The latter took place at the beginning of the session, and before the scenarios began, raising the question whether anxiety was higher immediately prior to  112  leading.  Therefore, measuring anxiety just prior to each leading participant beginning his or her scenario may have better captured anxiety in group leading.  Future studies could augment self-report measures of anxiety with physiological measures, such as a galvanic skin response device.  Second, creating an average score from individual group member responses suffers from methodological problems in group research, including nested data and mutually dependent data (Coco et al., 2015; Kivlighan et al., 2000). Nested data is defined as how members within a group do not constitute independent units of observation. Called the “unit of analysis problem”, Pollack (1998) defined it as choosing one level of analysis (i.e. the group) and forcing all variables of interest to that level (e.g. by averaging across individuals), so that statistical analysis may be performed on variable relationships.  Kivlighan et al. (2000) stated that treating all members as individuals and ignoring the group may violate the independence assumption of hypothesis testing, which can lead to confounded statistical results. The second methodological problem lies in mutual dependency of responses. This occurs when individual member or leader personalities or climates have an impact on group members. Kivlighan et al. (2000) named effects that account for this, such as group norming, perceiver and target effects, and dyadic effects. These problems are theoretically reduced in this study because the participants were not measured as a group, but as individual group leaders.  However, the influence of the group-as-a-whole could be examined in future studies by employing multiple levels of analysis (Coco et al., 2015). Third, a limitation of the GPIRS was that items emphasized skills that increased cohesion in groups, but did not accommodate behaviours that decrease cohesion (Burlingame et al., 2002).  Cohesion may be negatively affected by the use of helpful leading skills delivered without a foundation in group dynamics (Barlow, 2012).  For example, leaders may clearly demonstrate  113  skills from the GPIRS, but may persistently insert responses after any and all member statements.  This interaction style is illustrated by Yalom and Leszcz (2005, p. 121) as a “hub and spoke”, where members may only interact through the leader.  The authors stated that this denies members the ability to interact freely with each other.  The GPIRS lacks items that take into account such leader behaviours, and may show effective leadership where it was not in fact present.  The qualification of the GPIRS 0-code in this study presents one avenue for future research into the identification of unhelpful or absent leader behaviours.  Literature on ineffective group leading is scarce, so future studies could examine leader behaviours that reduce cohesion in groups and a subscale of the GPIRS could be proposed to highlight such behaviours.  Another avenue of research could utilize the time series analysis of the GPIRS, along with 0-codes, to more accurately identify group processes and leader behaviours that reduce cohesion.  They may also serve as pedagogical screening tools to help trainees specify learning moments in their group leading, and gauge their progress in group leading. Fourth, session 1 clarity scores were near the ceiling for the measure, suggesting that this session may not be an accurate baseline of comparison for group leading performance.  One explanation for the high initial quality scores is that Experiam operates as a group within a group in that simulated group leading occurs within a non-simulated peer-based group that is, itself, subject to norms and stages of development.  Therefore, the initial Experiam session would be characterized by members seeking inclusion needs and the group requiring a basis of cohesion in order to proceed with the task of simulated leading (Schutz, 1958, Yalom & Leszcz, 2005).  Therefore, the performance of participants (leading and role) in the first session may have been confounded by a lack of risk taking behaviours, as group members sought to establish interpersonal trust with each other.  With higher cohesion comes increased risk-taking, and so  114  subsequent sessions may build on successful prior performances.  The Experiam design has ethical merit by erring on the side of lower initial risk-taking to ensure students are not psychologically harmed (Kottler, 2004; Shumaker et al., 2011).  For those participants who did struggle in the first session, skill quality was shown to increase, and so a more accurate baseline may be dependent on individual trainee performances over time.  Future evaluations could compare several sessions, and over a longer time period, to better identify variance due to the stage of the group.   Finally, randomized control trials of groups have been criticized for making comparisons between the treatment of choice and no treatment at all (Greene, 2012).  This study was no exception, and so results must be limited to Experiam versus classroom-only instruction, and cannot claim Experiam as the program of choice in experiential learning.  An interesting future study could compare the developmental advantages between experiential programs, including Experiam. As a pedagogical tool, Experiam was demonstrated to cause a large increase in self-efficacy in group counsellor trainees, compared to students enrolled in a classroom-only condition.  This provides strong evidence that experiential training programs, such as Experiam, add something unique that classroom settings alone do not necessarily provide, and supports the integration of these methods, while adding a much-needed theoretical framework for experiential learning and evaluation (Brabender, 2010; Kormanski, 1991; Toth & Stockton, 1996).  The peer-led design of Experiam not only addresses ethical concerns by reducing faculty-student and student-student multiple relationship issues, but may consume fewer resources than faculty-led labs, and provides senior counselling lab leaders with much-needed group experience (Romano,  115  1998).  From theoretical, ethical, financial, and empirical perspectives, the bang of experiential training certainly seems to be worth the buck. In conclusion, the experiential group training program, Experiam, significantly impacted self-efficacy in counsellor trainees over classroom teaching.  This not only supports Experiam as an efficacious group counsellor training program, but also provides a strong rationale for the importance of integrating experiential training in the preparation of group counsellors.   116  Chapter 5: General Discussion 5.1 Summary of findings  The goal of this dissertation was to define and evaluate the Experiam program.  First, a preliminary, qualitative, focus-group study of five participants and two peer leaders was conducted to address the question: what elements of a peer-led, experiential group leading program impacted (helped or hindered) counsellor trainee confidence and competence? The E-CIT method (Butterfield et al., 2009) was applied to the transcribed focus group interview and seven categories were identified including the availability of an expert, leading scenarios, sharing feedback as a peer group, taking roles and vicarious observation.  These results helped evolve the program, and led to its integration into a didactic group counselling course.  Second, a randomized control study of Experiam was conducted to answer two research questions.  First, does Experiam increase counsellor trainee confidence in group leading? Hypotheses that Experiam would increase group leader self-efficacy over time and over controls were confirmed, while hypotheses that Experiam would decrease counsellor trainee anxiety over time and over controls were not confirmed.  The second research question was: Does an experiential training program increase counsellor-trainee performance? Experiam was shown to significantly increase frequency of group leading skills, but clarity of skill delivery was not significantly increased over time.   This chapter comprises novel contributions from the evaluations, followed by implications for research, practice, and pedagogy.  Limitations and future directions for research and the Experiam program are then discussed.    117  5.2 Novel contributions  Several novel contributions were made by these studies; four are noteworthy.  The first and primary finding concerns the relationship between Experiam and self-efficacy in group leading.  A significant increase in self-efficacy occurred for the treatment condition that did not occur for controls.  Random selection of the contrasted conditions and likelihood of their pretreatment equivalence to each other means that Experiam likely caused higher self-efficacy in trainees over classroom instruction alone (Hsu, 2003).  During the follow-up period, the treatment condition was in the classroom, and did not exhibit significant changes in self-efficacy.  This demonstrates that acquired self-efficacy did not diminish over time for these students, and that classroom-only participation was not significantly associated with further changes in self-efficacy, although an interaction between the two cannot be ruled out.  Finally, when the waitlist group participated in Experiam, participants reported significantly higher self-efficacy.   Therefore, the experimental result provides a more generalizable and causal argument that Experiam raised self-efficacy in counsellor trainees.  Quasi-experimental results add associative strength to the argument, although the absence of another control group makes this finding less robust.  Finally, the increase in self-efficacy reported following Experiam was over one standard deviation.  Examining the statistical significance and effect size results within the context of related research reveals that similar findings have been found in studies of individual counsellor training programs (Daniels & Ivey, 2007; Larson & Daniels, 1998; Kuntze et al., 2009), for leading management groups (Hoyt et al., 2003), and for counsellor trainees as participants in groups (Ohrt et al., 2013).  Therefore, the effect size can be considered large, based on Cohen’s (1992) guidelines, but must be tempered by similar findings in other studies.   118  This finding addresses a gap in the literature identified by Page et al. (2001) by demonstrating the impact of experiential group leader training on counsellor trainee self-efficacy. Evidence from the preliminary case study provides non-experimental support to these findings.  Participants reported increased confidence and competence in group leading from leading scenarios and in observing others lead and take roles.  Together, these findings support what Bandura (1989) stated was the most impactful combination of the four causal mechanisms for the promotion of self-efficacy, namely the successful enactment of behaviours toward an outcome, and the vicarious experiencing of others’ enactive attainment.  Not only does this evidence provide a justification for SCT as the theoretical framework for the Experiam design, it also supports experiential training as an effective method for group counsellor development. The second contribution concerns the relationship between experiential training and fear.  A distinction between Experiam and other training programs is that counsellor trainees are purposefully, yet thoughtfully, exposed to feared situations that may arise in professional work.  Failure in group psychotherapy has been attributed to therapist fear (Harpaz, 1994).  Fear arousal is explained by SCT as perceived inefficacy in coping with potentially aversive events (Bandura, 1982).  Fear is therefore reduced by the belief that one can prevent or lessen the severity of the event, meaning that higher self-efficacy diminishes fear and avoidance of previously feared events.  The hypothesis that Experiam would reduce anxiety in trainees was not confirmed, as measured anxiety did not change over time or between groups, and levels were consistently within a normal range for college students.  While design issues described in the limitations section below, may help explain the absence of an effect, it is noteworthy that anxiety did not increase despite increases in scenario complexity and emotional intensity.    119  Scenarios believed to challenge novice counsellors were chosen based on the literature and the practical experience of faculty members specialized in group psychotherapy (Billow, 2001; Donigian & Malnati, 1987; Markus & Abernathy, 2001; Shechtman & Toren, 2009; Yalom, 1966).  Evidence from these sources informed an anecdotal ordering of scenarios from least to most challenging.  This enabled participants to adopt attainable sub-goals that cumulatively contributed to larger goal attainment, which Bandura (1982) referred to as proximal self-motivation.  For example, in succeeding in running the first scenario, the cold group, participants gained a sense of efficacy to working with a challenging role in the next scenario.  Bandura (1982) theorized that proximal self-motivation developed self-efficacy and sustained motivation in working towards a goal.   According to SCT, individuals are thought to achieve some control over a feared event in two ways: behaviourally, by taking actions, and cognitively by believing the threat is manageable.  These are considered distinct forms of fear control (Bandura, 1982).  Experiam brings these two forms of control into close contact by sandwiching cognitive processing on either side of an enacted behavioural intervention.  Before each scenario, leading participants discussed the scenario, the roles, and their plan of action to facilitate the scenario.  After the scenario, they were given the opportunity to speak about their experience of the scenario.  The proximity of cognitive processing to behavioural action may align self-efficacious thoughts with a recently attained enacted behaviour.  Participants in the preliminary study reported that post-leading learning provided them with immediate feedback to their enacted behaviours, which in turn increased confidence.  The results of the second study provided associative evidence that trainees exposed to ever-more threatening group leading events did not perceive higher anxiety.  120  The third contribution was that participants demonstrated more advanced and diverse skills over time.  Evaluations of skills-based group training programs (where trainees actually lead groups), have shown increases in the frequency of skill usage and the development of therapeutic group climates (Romano & Sullivan, 2000; Smaby et al., 1999; Stockton & Toth, 1996).  Therefore, this finding supports existing literature.  Examining specific skills revealed that in the more complex scenario, participants invoked skills related to group control and encouraging emotional expression, while remaining non-defensive towards members.  In emotionally-charged group processes, groups can become harmful to members, and leaders can inadvertently take part in causing harm.  In such moments, effective leadership is never more needed, as leaders who cannot strike a balance between over-passivity and over-control, or who permit damaging group processes to continue, can allow group members to be psychologically harmed (Smokowski, 2001).  Experiam therefore enabled trainees to practice skills vital to facilitating critical group incidents, maintaining cohesiveness, and reducing damaging group processes.  An examination of those skills that were not demonstrated, but therapeutically-indicated, revealed moments where participants possibly failed to lead effectively.  In session six, students more frequently failed to demonstrate warmth and empathy, or did not encourage member self-disclosure during a challenging scenario.  Using this coding scheme, students and faculty can be provided feedback about what skills need further instruction or clarification in the classroom.  The Integrated Model of Mastery (IMM) in chapter three proposed that students iterate from classroom instruction to practice sessions, and back again to the classroom for further instruction and refinement of skills.  A measure of skills-not-demonstrated provides a specific source of instruction.  What makes this finding invaluable is that students may not know what they do not  121  know, or may subconsciously avoid enacting certain group skills, and can return to the classroom without understanding what they are failing to do as leaders (Harpaz, 1994).  Counsellor educators may similarly struggle to justify skills to refine in the classroom, particularly if they are not present during the practice sessions.  This refinement of the GPIRS coding scheme therefore has pedagogical merit within the IMM.   The fourth contribution involved how skill clarity evolved across sessions. Although unexpected, this non-significant result revealed insights into Experiam.  Within-session analysis of the behavioural measure was examined over two-minute intervals, and trends in skill clarity emerged.  Although not statistically significant in this study, such trend analysis may reveal trainee group leading performance in reaction to a critical group process event.  Such time series analysis can illustrate how Experiam scenarios challenged participants to try new skills that may not be delivered with high performance.  Also, the u-shape of the final session was reminiscent of research showing that good individual therapy outcomes are associated with a curvilinear-shaped alliance development over time, usually U or V-shaped.  Kivlighan and Shaughnessy (2000) tested a brief treatment, where good outcomes correlated to the alliance beginning strongly, with high optimism, followed by a subsequent period of frustration and lower alliance, and concluding with a stronger alliance, based more in reality for the client.  They concluded that, based on Bordin’s (1979) concept of “tear and repair”, the patient’s pathology necessitates a period of low alliance, which improves with successful therapeutic interventions, increasing both the alliance and the therapeutic outcomes (p. 368).  The salutary properties of a curvilinear alliance relationship may have extension to group cohesion.   Before moving to implications for research and practice, the bridge between scientist and practitioner worlds requires discussion.  The gap between research designs demonstrating  122  efficacy, and practitioner designs demonstrating effectiveness, has been widely discussed in fields of health care, including counselling psychology (Fuhriman & Burlingame, 2001).  Efficacy is defined as a conclusion that observed benefits from a treatment are due to the effects of the treatment and not due to chance or confounding factors (Nathan Stuart & Dolan, 2000).  Chambless and Hollon (1998) argued that efficacy is most clearly demonstrated by the randomized control trial.  They stated that replication of a treatment effect in at least two independent trials is required to make the claim that the treatment is efficacious.  Therefore, until this study has been replicated independently, findings can only be labelled as “possibly efficacious” (p. 8).  The authors cautioned that efficacy statements must be tempered by asking whom the treatment benefits, as RCT criteria can sometimes exclude the very population the treatment seeks to help.  This caution provides a rationale for effectiveness in research, which seeks to demonstrate that a treatment can be shown to work in clinical practice.  These authors further stated that efficacious treatments should be augmented by other forms of research inquiry that includes non-experimental methods.  Such efforts provide alternative sources of causal agency, a term describing whether an observed change can be attributed to the treatment.  Effectiveness, or clinical utility, is related not only to external validity, but to ease of dissemination (feasibility) and cost utility.  Effectiveness requires answers beyond the statistical differences between test means, yet the authors argue that the answers they provide are no less important to the value of a treatment. The studies presented here demonstrate Experiam to be possibly efficacious and effective in developing confidence and competence in group leading.  Beyond the findings of the RCT, all students were eligible to participate in the lab, including 10% not in the counselling field.  While this may seem a confounding inclusion criterion, in fact it provides evidence of effectiveness to a  123  wider population.  The eligibility for students in related health disciplines to take the group counselling course was based on satisfying the same course requirements that any first year counselling student would have to have to take the course.  Therefore, the generalizability of the results can be argued to apply to students in health-related fields, primarily counselling, who meet requirements for group counselling instruction.  In addition, the peer group was shown in the qualitative study to have feasibility and cost utility implications.  The former was demonstrated by the extracurricular nature of the peer group and the desire for students to work as peers.  The latter was demonstrated by the low resource cost of the lab on faculty, program curricula, and students.  Finally, the focus on the peer group as a mode of learning demonstrates that Experiam can adapt to changing student needs, and can integrate into existing didactic group counselling curricula, thus providing evidence of ecological validity in Experiam.  The following two sections examine implications for scientist and practitioner paradigms. 5.3 Implications for research  The Experiam design was founded on SCT, which addresses a criticism that group training often overlooks theory in favour of practical experience (Kormanski, 1991).  The mechanisms of change in SCT are well explained and researched, which provides opportunities to extend research in experiential training, and offers many empirically-tested measures of underlying constructs, such as mechanisms related to self-efficacy.  Bandura’s (1982) four mechanisms that impact self-efficacy can be mapped into separable elements of the Experiam design.  For example, enactive attainments may be represented by the leading scenario, vicarious experiences are represented by observing others lead the same scenario and by taking roles, verbal persuasion is represented by support and feedback from the group, and regulated physiological arousal is represented by the cohesion in the group-as-a-whole.  Therefore, the  124  significant increase in self-efficacy associated with the program may be attributed to one or more of these causal mechanisms.  Future studies could examine which of these mechanisms contribute most to self-efficacy.  For example, Bandura (1982) theorized that enactive mastery was the most influential mechanism on self-efficacy, but that combining enactive attainment with modelling was most effective.  A future study could compare Experiam groups that separate modelling elements from leading elements to determine the relative contributions of each mechanism. Next, counselling psychology research has historically employed video recording (Young, Valach & Domene, 2005).  The video recordings in Experiam serve not only as a pedagogical and training tool, but as rich evidence of manifest behaviour that can be examined jointly by researchers and participants.  For example, Kagan, Schauble, Resnikoff,  Danish and Krathwohl’s (1969) interpersonal process recall (IPR) method involves video-recording participants in an active process, then later replaying the video for each participant, with a trained interviewer.  Larson, Flesaker and Stege (2008) stated that video replay enhances participants’ ability to recall and interpret feelings, thoughts, and behaviours that arise during interactions, and Hartson and Kruce (1973) demonstrated IPR to be effective in group settings. Video recordings can also serve as a journal of experiential learning across a counsellor’s training would provide an invaluable research narrative of counsellor experience and of the effectiveness of experiential training methods. Third, the relationship between Experiam and anxiety was not strongly demonstrated in the primary study, yet SCT has much to say about the link between fear and self-efficacy.  Because Experiam was designed to expose trainees to feared group events, future studies could use multiple sources of information to examine the mediating role of fear on the self- 125  efficacy/performance relationship.  The experiential nature of Experiam may limit participants’ ability to recall moments where anxiety or fear change, thus invalidating self-report measures, and behavioural observation measures are limited to the perceptions of an outside rater.  Using physiological measures of fear or anxiety may provide novel insights.  Galvanic skin response methods have become more precise and tools are more accessible to researchers (Bach & Friston, 2013).  These tools measure a variety of psychological constructs, such as stress, by capturing changes in skin conductance in areas of the body that have been shown to react to stimuli (i.e., hands and feet).  This tool avoids social desirability bias and may pinpoint processes that precede and follow impactful group leader behaviour.  Data is also recorded as a time series that can be synchronized to behavioural observations and IPR participant responses to provide precedents and antecedents to significant group events.  Together, these independent sources of information can be examined in relative synchrony and provide precise and robust information during moments where fear gives way to efficacious behaviours in group leading. Fourth, the university peer group represents an under-examined area of literature (Topping, 2005).  Qualitative evidence of this study and others suggests that a different kind of learning takes place when peers provide feedback to each other (Ieva at al., 2009; Ohrt et al., 2013; Toth & Erwin, 1998).  The SCT concept of collective efficacy may be applied here, and is defined as the sense that people can solve problems and improve their lives through concerted effort (Bandura, 1982).  The design of Experiam required that all members of the group participate every session, which may foster a sense of universality among peers, but also to the belief that participants have control over the design itself.  Peers in the preliminary study indicated a wish to have more control over the group and such input has contributed to the evolution of the design.  This echoes Orr and Hulse-Killacky (2006), who stated that when  126  members of a group have a direct stake in constructing the group’s knowledge, they become active resources in the creation of a new kind of “mutually-constructed knowledge” (p. 195).  Future studies could examine the impact of collective efficacy on the individual efficacy and performance of Experiam participants.  Such examination would provide a counselling contrast to the work of Hoyt et al. (2003), who examined these variables in management leadership contexts.  Fifth, the ancillary analysis of chapter four provided a new approach to coding the GPIRS that enhanced its use as a process research tool.  Scenarios were divided into timed intervals and coded within each interval, in this case, five 2-minute intervals within a 10-minute scenario.  This not only increased the amount of data available, and added more time intervals to the coding, but it enabled inter-raters to more closely calibrate.  Future studies may employ this technique to gain greater specificity to group leader interventions.  For example, using the GPIRS as a time series allows it to be used in growth curve analysis methods of group research (Kivlighan et al., 2000).  This method uses recent developments in statistical analysis, such as hierarchical linear modelling (HLM) to explain time series changes in groups and can separate group from individual variances.  Although previous studies of the GPIRS examined it based on quantity and clarity of intervention, how such analyses are translated to practical implications is scarce (Chapman et al., 2010; Chausovsky et al., 2005).  This study provides an improvement on the method of coding the observational data.  Sixth, the items of the GPIRS described those skills that were theoretically and empirically found to increase cohesion.  However, this meant that leader interventions and behaviours that decreased cohesion were either not coded or could only be coded as an unclear delivery of a helpful skill.  This presented a limitation of the GPIRS as an evaluation tool for  127  cohesion development in groups, as researchers are faced with the conundrum of how to code for the absence of a needed skill or the presence of an unhelpful skill.  For example, leaders may enact skills that do not harness group dynamics, such as responding to each member statement.  This may creates a hub and spoke style of interaction, where the leader is the centre of communication and members are not free to interact with one another (for a diagram, see Yalom & Leszcz, 2005, p. 121).  The GPIRS can only be coded based on what the leader does, and so many helpful skills may be coded that mask an unhelpful interaction style.  The 0-code can provide some insight into absent leader skills , but requires greater clinical interpretation by the coder, which can reduce coding validity.  Future studies could examine what leaders do that decrease cohesion and a subscale of the GPIRS could be developed that capture these dynamics. Seventh, findings from the preliminary study revealed that participants experienced tension with the lab leader.  Bandura (1982) exemplified his concept of proxy control by stating that the presence of a highly confident individual can lower the self-efficacy of others nearby.  Lab leaders must share control with peers to maximize the perceived control of participants while maintaining structure and safety.  This balance could be explored by measuring the control style of various Experiam lab leaders and correlating this to group member self-efficacy.  For example, highly anxious participants may benefit from lab leaders who are more structured, while more confident participants may learn more in sessions led by hands-off leaders.   Finally, the integrated model of mastery (IMM) proposed in chapter 3 contains many areas for future research.  This model theorized that high confidence and competence, or  mastery arose from a purposeful interplay between peer/experiential and expert/didactic learning environments.  Because Experiam clearly delineates between peer work and the presence of the expert, future studies could examine the potency with which iterating between each domain  128  causes increases in constructs related to skill mastery, such as self-efficacy.  If substantiated, the model may provide a useful tool for evaluating other training programs that integrate didactic-experiential learning environments. 5.4 Implications for practice and pedagogy  Experiam is a training tool that was designed for graduate-level training of counselling students.  Six practical implications relevant to pedagogy are discussed.  First, although Larson and Daniels (1998) articulated a social cognitive model of counsellor training, the authors limited the impact of supervision to the role of instructors and expert supervisors.  Experiam furthers this theoretical framework by integrating peer supervision.  Models of peer learning evolved from developmental and cognitive models including Piaget and Vygotsky, and affective models that involve trusting peer relationships (Topping, 2005).  This author situated peer learning in literature for child and adolescent development and primary school instruction, but stated that it has begun to take hold in university education contexts.  The IMM proposed in chapter three is a pedagogical framework for integrating Experiam into applied course work at the university level.  Optimal learning is theorized to occur when peers learn didactically from an expert instructor, and practice without the instructor present.  The two axes defined the four learning environments.  The IMM is fully realized when students iterate between instruction/expert and practice/peer environments and can bring forward new learning from one environment into the other.  An analogy to this model is found in attachment theory.  Wallin (2007) invoked the Kleinian term refueling to describe how children iterate between exploring the unknown, then returning to parental security, and then exploring again, as if refuelled by their secure base.  This idea is not meant to infantilize students or parentify instructors.  A rationale for the iteration between practice (exploring) and didactic instruction/supervision (secure base) is that group  129  counsellor trainees experience groups as complex, rapid, unpredictable and uncontrollable (Page, et al., 2001).  While the Experiam program sets out guidelines for the peer environment to function, the IMM also requires instructor buy-in, leadership, and supervision (Westwood, 1989). Second, the GCSE instrument may be a valuable tool in the screening of participants for experiential training groups. Instructors of group counselling have the opportunity to divide their students into groups according to group composition guidelines. Yalom and Leszcz (2005) stated that cohesion is the primary guideline for group composition, and that members be grouped homogeneously for ego strength and heterogeneously for conflict areas. They stated that this would maximize cohesion while also enabling members to learn from each other.  Self-efficacy may serve as a measure of ego strength in learning new tasks, and SCT provides evidence of how to use self-efficacy measures in a group context. Bandura (1982) stated that observing success in those perceived to have similar competency can increase one’s self-efficacy, and that observing more efficacious peers may actually reduce perceived self-efficacy. For experiential group training programs that target self-efficacy, this measure may allow educators to screen student perceptions of self-efficacy and form groups based on similar levels of self-efficacy.  While evidence on the efficacy of group composition remains anecdotal, the GCSE as a composition screener may provide educators with a low-cost way to improve student confidence in experiential settings.   Third, the prevalence of experiential training in counselling programs cannot be increased just by calling on curriculum planners to do more with the same resources.  Instead, smarter program designs must be developed to meet training needs in an ethically and financially feasible way.  Experiam presents such a program by utilizing a structured peer group.  The peer group  130  addresses what Romano (1998) called the “most problematic” (p. 120) ethical problem in experiential training, namely dual relationships between student and faculty members that can arise in such environments.  Experiam allows counsellor trainees to practice newly-acquired group leading skills without fear that they are being evaluated for such practice.  Another ethical concern involves student-student dual relationships, in that experiential activities can result in students revealing personal material that would not otherwise be disclosed among peers (Fall & Levitov, 2002).  This concern is mitigated in Experiam by the establishment of roles that allow participants to enact group member behaviours without such actions reflecting on them personally.  Finally, the peer group can provide a safe and realistic learning environment without an instructor present, by the use of structured feedback (Westwood, 1989).  Effective feedback groups, such as outlined by Toth and Erwin (1998) are easily taught to counsellors by employing skills they have already acquired.  In addition, effective peer groups may be more easily established by counselling students as they require skills that dovetail with counselling training.  In addition to being financially unburdensome, once students understand how to create and participate in effective peer groups, they could be applied in other areas of counsellor training requiring experiential learning.  One such area is multicultural counselling, where experiential training has already been shown to be effective for counsellor trainees (Arthur & Achenbach, 2002).  Fourth, the experiential peer group is not only financially and ethically sound, it promotes student development.  Counsellor trainees in experiential training groups reported twelve growth factors that included therapeutic group factors such as universality and cohesion (Kiweewa et al., 2013).  When in a peer supervision group, counsellors reported clearer goals, increased confidence, and a discovery of the value of interacting with peers (Starling & Baker, 2000).   131  Counsellor trainees receive feedback from instructors, supervisors, and their own clients, yet peer group feedback integrates universality into interpersonal learning.  In other words, peers may hear something more genuine in feedback to one another because they share similar feelings and concerns (Borders, 1991; Christensen & Kline, 2001).  Certainly peers in the preliminary study of Experiam reported valuing the feedback they received and wished for more critical feedback from each other.  Peers compete with each other in academic settings, and may observe one another in ways that instructors and supervisors may not; this may lead to more incisive insights and feedback.  Therefore, a commitment to peer-based learning can provide ongoing personal and professional benefits to students after they complete their education. Fifth, role residue was a finding of the preliminary study that warrants further discussion.  This term has been described in literature as a spilling or leaking of experiencing between the role and the role participant’s real life (Goodrich, 2008; Romano & Sullivan, 2000).  Role residue raises an ethical dilemma between the need for counsellor trainees to be protected from emotionally charged or even injurious experiencing, and the need for such trainees to competently clients who present with such stories.  A solution is presented by first acknowledging that role residue exists and can interfere with risk-taking and learning of counsellor trainees.  Next, roles should be assigned with the spirit of Carl Rogers’ (1961) “as if” quality to empathic understanding (p. 284).  In other words, roles allow us to sense the private world of clients who present them, but we must not lose the as if quality of the role.  The Experiam design embraces this distinction by having participants use pseudonyms in scenarios and allotting time for de-roleing participants.  While not all roles may require these safeguards, this two-stage process enables participants to take greater risks in the roles they enact and in the roles they ask others to enact.  Realism of the scenario is therefore increased, while maintaining  132  safety within the participant and between participants.  As successive Experiam scenarios call for more realistic roles, and as cohesion and safety between participants grows, the issue of role residue must be attended to.  Future research could examine the growth in role realism as Experiam groups become bonded and attempt more threatening scenarios.  New methods for attenuating role residue could be sought, and studies could seek where the balance between realism and safety can be best set. Sixth, the choice of scenarios that prepare group counsellors warrants close scrutiny and further discussion.  As stated above, scenarios were chosen based on what prominent writers in group counselling saw in their students, supervisees, and presumably themselves.  An alternate perspective is to examine what group processes cause harm to clients in group counselling.  Smokowski et al. (2001) delimited three categories of damaging group experiences, including leader attributes and actions, attributes of the group and attributes of the casualties themselves.  The latter two findings provide areas of growth for Experiam scenarios and for group counsellor training in general.  For example, a damaging group attribute that the authors found was a lack of tolerance for diverse points of view.  The Experiam scenario of facilitating feedback could be altered to incorporate an intolerant group consensus and to practice skills to ensure the offended participant has been heard by the group.  In this way, scenarios may become more finely tuned to empirically-supported needs of group clients who may otherwise become group casualties.  The attributes of casualties that contribute to their injury from groups included shyness, social inhibition, reluctance to participate, and being slow to warm up to the group.  These attributes are surprisingly similar to the role of the withdrawer or “silent client” (Yalom & Leszcz, 2005, p. 397 – 400).  This signals the withdrawer as a critical role for counsellor trainees to understand and to gain experience working them into the group.     133  The seventh implication for practice and pedagogy is the use of video recordings.  Video recordings of a group leading session may be the next best thing to having the supervisor physically present in the group, and may be superior in that the group leader is not distracted by a figure of authority in the group.  An additional benefit of video is that student and supervisor can re-watch and analyse significant moments together.  Additionally, the advent of smart phone technology makes high-definition video recording accessible and inexpensive.  Such recording tools introduce ethical issues to experiential training, as confidentiality and privacy are changing in the Internet era.  For example, current smart phones offer free and automatic upload of videos to backup servers that do not necessarily guarantee privacy.  It seems inevitable that counselling programs will have to address the ethical issue of the growingly-recordable world.  This effort may have a benefit for programs that also harness video recordings in training and supervision. Finally, nearly fifty years ago Yalom (1966) stated that admitting errors or shortcomings were among the hardest things new group therapists face.  This seems as true today as then.  Admitting real or perceived errors may not only bruise the counsellor trainee’s fledgling esteem and self-efficacy, but it may expose the trainee to feeling shamed at the group level.  A literature review failed to discover what group psychotherapists and counsellors can do to apologise to their groups for mistakes.  However, a roadmap to accepting appropriate blame for a rupture to the therapeutic alliance in individual therapy may provide informative to group contexts (Safran & Muran; 1996; Safran, Muran, Samstag & Stevens; 2001).  These authors defined an alliance rupture as an impairment or fluctuation in the quality of the alliance between therapist and client.  They proposed that therapists first allow clients to express negative feelings about the rupture and then to empathically respond to these expressions openly and non-defensively.  They then suggest that therapists explore the rupture experience with the client so that the appropriate level  134  of responsibility can be accepted by therapist and client alike.  Rupture and repair is analogous to the curvilinear shape of trainee performance, and may hold clues to cohesion formation in groups.  In Experiam, the challenge-the-leader scenario could be adapted to include admission of errors or shortcomings to a group.  This scenario would likely be perceived as highly threatening to counsellor trainees, and should be reserved until later sessions.  In addition, instructor feedback may be required to help trainees differentiate appropriate levels of responsibility to the individual versus the group and to use group skills to facilitate ruptures in group cohesion.  This scenario could reduce trainee fear of such a group process, and demonstrates how classic group issues can be brought forward and enacted using the Experiam design.  It may also enable leaders to better identify and respond to cohesion ruptures with a view to repair and better group outcomes. 5.5 Limitations and delimitations  Limitations are the potential weaknesses of a study that are beyond the control of the researcher.  A primary limitation of both studies was the small sample size.  The quantitative study was particularly vulnerable, as studies with sample sizes less than 40 are argued to have a higher probability that nuisance variables may make the randomized comparison groups unequal (Hsu, 2003).  As only 30 students were accepted per year into the counselling program where the study was conducted, the sample size limitation was due to time constraint.   Second, the use of self-report measures may have resulted in socially desirable responding.  Because the behavioural measure did not examine a similar construct, correlational analysis between these measures was not meaningful for validity purposes.  Future evaluations could behaviourally-analyse all participants leading at the same moment during specified intervals to provide more equivalent comparisons.  Third, the waitlist group responses may have  135  been biased by perceptions that the treatment group was given more time to complete a course assignment that was based on the video recordings of the lab.  Participants waiting for the lab to start may have reported higher anxiety and lower self-efficacy due to worry over their grade in the course.  In addition, the waitlist group had attended classes prior to starting Experiam, which may have introduced non-equivalence between these groups.  Future studies could examine other designs, such as a non-didactic condition that only took Experiam.  Fourth, some Experiam sessions did not take place in the same setting, due to room availability.  This not only may have introduced a nuisance variable between groups, it may have impacted the behavioural coding, as inter-raters may react differently to leaders in different settings.  Future studies should secure a consistent lab setting to reduce this potential confound.  Finally, Experiam sessions 1 and 6 were chosen for comparison to capture the cumulative effects of the program.  However, the sixth session may not be comparable to the first session in at least two ways.  First, the scenarios in session 6 were chosen by the participant to permit practice of those skills most important to them, while the session 1 scenario was set as the starting of a new group.  Second, the stage of development or cohesion of the group-as-a-whole may have changed across sessions.  This may have resulted in a higher proportion of the variance of change in self-efficacy being attributable to the group level over the individual level.  Future students could examine this difference with higher-order linear modelling, or by randomly mixing the Experiam groups each session.  One clear drawback to mixing groups, however, is the reduction in the risk-taking that may arise in less cohesive groups.  Delimitations are choices made by the researcher that define the boundaries of the study and that deserve acknowledgement.  In the primary study, the choice of a quantitative method using a control group was chosen as the best method to isolate the impact of the training lab on  136  self-efficacy.  The preliminary qualitative study pointed to self-efficacy as a likely construct that was impacted by the experiential lab, which provided a natural rationale for the quantitative method.  A self-report measure of anxiety was chosen over physiological measures, such as a galvanic skin response device, due to the existence of empirically-validated measures within the domain of counselling education, and validating such devices was deemed to be beyond the scope of the research questions.  Next, the choice to only test self-efficacy and anxiety using two measures was made to allow for ease of administration.  Additional measures may have provided greater validation of the construct sought, however a pilot study of the experiment demonstrated that multiple measures were found by participants to be onerous and yielded highly-correlated results.  Therefore, a single measure for each construct was chosen based on its fidelity to the construct and its published psychometric base.  Finally, the population was chosen to be graduate students at a single university.  Extending the research question to include other universities would have required cross-university collaboration regarding the didactic and experiential group course.  This undertaking was beyond the ability of the researcher to facilitate.   5.5.1 Limitations to Experiam Wicker’s (1998) concept of providing counter arguments to strongly-held ideas has merit for the Experiam program as a whole.  Therefore, limitations or criticisms of the Experiam program are provided with the hope that future ideas may improve the program.  First, the peer-based element of the program has a weakness if the peer leaders do not follow the non-expert role.  This arose in the preliminary study and in some video recorded segments of the second study.  However, peer leaders are vested with authority by design, which makes them susceptible to the expert role.  This may mean that the peer group cannot function as effectively, based on the peer leader, and an expert-run training group may be preferred for some counsellor trainees.   137  Alternatively, Experiam could be run with only student participants who rotate the lab leading role.  Second, the program is entirely focused on short vignette-style leading turns.  Leading for such short time periods denies the leading participant the understanding of larger group processes, such as the developmental stage of the group, meaning that counsellor trainees may require additional training for this learning to occur.  Romano’s (1998) simulated group training program achieves this by having only one or two trainees lead for a longer duration.  This further demonstrates that a balance be struck between student experiencing and time/financial constraints.   Third, counsellor trainees may not receive realistic enough training with peers enacting roles to prepare them for clients as themselves.  The success of Experiam relies on the ability for counsellor trainees to translate rehearsed practice into real world scenarios.  If they do not learn experientially or struggle to find realism in the scenarios, then the theoretical mechanisms of change will not apply to their learning, and they would be better served co-leading real groups with a competent supervisor.  This is a converse to Hoyt et al. (2003), who noted that their self-efficacy model may not generalize to in-tact groups with a history.  Fourth, the link between self-efficacy and performance was not established in this study, and remains a theoretical link in SCT.  Other studies have been stymied in demonstrating this link (Barnes, 2004; Hoyt et al., 2003).  The Experiam design relies upon self-efficacy as the underpinning construct to counsellor development, making performance improvements still a theoretical assumption.  Should performance require something additional or different from self-efficacy, then the Experiam design would need to be adapted to this new assumption.  138  5.6 Conclusion The Experiam design provides high-impact, experiential group leader training that addresses ethical and resource concerns.  The primary study experimentally tested the impact of group leader training on group counsellor self-efficacy.  The theorized mechanisms of change of social cognitive theory were at work in the design of Experiam, as group counselling self-efficacy significantly increased while anxiety remained at normal levels and the complexity of scenario was reflected in higher frequency and diversity of skill usage.  The spirit of Experiam was to provide a safe and structured way for peers to master group counselling skills and embrace the personal qualities of successful group leaders.  In closing, Yalom and Leszcz (2005) stated that group therapy must be constantly tended to avoid being choked by old weeds of criticism, such as superficiality, dangerousness, and a second-rate to individual therapy.  Experiential training programs like Experiam must rise above these criticisms by applying evidence-based interventions within theoretically-sound underpinnings, while also harnessing the creative and formative power that experiential work provides.   139  Bibliography American Group Psychotherapy Association (AGPA). (2003). Certification requirements. Retrieved 6 June 2015 from http://agpa.org/cgp-certification/how-to-apply. Anderson, R. D., & Price, G. E. (2001). Experiential groups in counselor education: Student attitudes and instructor participation. Counselor Education and Supervision, 41(2), 111-119.  doi:10.1002/j.1556-6978.2001.tb01275.x Arthur, N., & Achenbach, K. (2002). Developing multicultural counseling competencies through experiential learning. Counselor Education and Supervision, 42(1), 2-14. doi:10.1002/j.1556-6978.2002.tb01299.x  Association for Specialists in Group Work (ASGW). (2000). Professional standards for the training of group workers. Alexandria, VA: Author. Retrieved 6 June 2015  from http://www.asgw.org/PDF/training_standards.pdf. Bach, D. R., & Friston, K. J. (2013). Model‐based analysis of skin conductance responses: Towards causal models in psychophysiology. Psychophysiology, 50(1), 15-22. doi:10.1111/j.1469-8986.2012.01483.x Bandura, A. (1978a). Self-efficacy: Toward a unifying theory of behavioral change. Advances in Behavior Research and Therapy, 1(4), 139-161. doi:10.1016/0146-6402(78)90002-4 Bandura, A. (1978b). The self-system in reciprocal determinism. American Psychologist, 33(4), 344-358. doi:10.1037/0003-066x.33.4.344 Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37(2), 122-147.  doi:10.1037/0003-066X.37.2.122  140  Bandura, A., Reese, L., & Adams, N. E. (1982). Microanalysis of action and fear arousal as a function of differential levels of perceived self-efficacy. Journal of Personality and Social Psychology, 43(1), 5-21. doi:10.1037//0022-3514.43.1.5 Bandura, A. (1984). Recycling misconceptions of perceived self-efficacy. Cognitive Therapy and Research, 8(3), 231-255. doi:10.1007/bf01172995 Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44(9), 1175-1184. doi:10.1037//0003-066x.44.9.1175 Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52(1), 1-26. doi: 10.1146/annurev.psych.52.1.1 Barlow, S. H. (2004). A strategic three-year plan to teach beginning, intermediate, and advanced group skills. Journal for Specialists in Group Work, 29(1), 113-126. doi:10.1080/01933920490275600 Barlow, S. H. (2012). An application of the competency model to group-specialty practice. Professional Psychology: Research and Practice, 43(5), 442-451. doi:10.1037/a0029090 Barlow, S. H., Burlingame, G. M., & Fuhriman, A. (2000). Therapeutic applications of groups: From Pratt's “thought control classes” to modern group psychotherapy. Group Dynamics: Theory, Research, and Practice, 4(1), 115-134. doi:10.1037/1089-2699.4.1.115 Barlow, S. H., Fuhriman, A., & Burlingame, G. M. (2004). The history of group counseling and psychotherapy. In J. L. DeLucia-Waack, D. A. Gerrity, C. R. Kalodner, & M. T. Riva (Eds.). Handbook of group counseling and psychotherapy (pp. 3-21). Thousand Oaks, CA: Sage Publications Inc. doi:10.4135/9781452229683.n1  141  Barnes, K. L. (2004). Applying self‐efficacy theory to counselor training and supervision: A comparison of two approaches. Counselor Education and Supervision, 44(1), 56-69. doi:10.1002/j.1556-6978.2004.tb01860.x Bedi, R. P., Haverkamp, B. E., Beatch, R., Cave, D. G., Domene, J. F., Harris, G. E., & Mikhail, A. M. (2011). Counselling psychology in a Canadian context: Definition and description. Canadian Psychology/Psychologie Canadienne, 52(2), 128-138. doi:10.1037/a0023186 Benne, K. D., & Sheats, P. (1948). Functional roles of group members. Journal of Social Issues, 4(2), 41-49. doi:10.1111/j.1540-4560.1948.tb01783.x Beutler, L. E., Machado, P. P., & Neufeldt, S. A. (1994). Therapist variables. In A. E. Bergin & S. L. Garfield (Eds.). Handbook of psychotherapy and behavior change (4th ed.). (pp. 229-269). New York, NY: John Wiley & Sons. Billow, R. M. (2001). The therapist's anxiety and resistance to group therapy. International Journal of Group Psychotherapy, 51(2), 225-242. doi:10.1521/ijgp.51.2.225.49856 Billow, R. M. (2010). Anarchy. Group Analysis, 43(1), 65-80. doi:10.1177/0533316409356464 Bion, W. R. (1961). Experiences in groups and other papers. New York, NY: Basic Books. doi:10.4324/9780203359075 Blatner, A. (2000). Foundations of psychodrama: History, theory, and practice. New York, NY: Springer.  Boote, D. N., & Beile, P. (2005). Scholars before researchers: On the centrality of the dissertation literature review in research preparation. Educational Researcher, 34(6), 3-15. doi:10.3102/0013189x034006003 Borders, L. D. (1991). A systematic approach to peer group supervision. Journal of Counseling & Development, 69(3), 248-252. doi:10.1002/j.1556-6676.1991.tb01497.x  142  Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice, 16(3), 252-260. doi:10.1037/h0085885 Brabender, V. (2010). The developmental path to expertise in group psychotherapy. Journal of Contemporary Psychotherapy, 40(3), 163-173. doi:10.1007/s10879-010-9142-4 Brenner, V. (1999). Process-play: A simulation procedure for group work training. Journal for Specialists in Group Work, 24(2), 145-151. doi:10.1080/01933929908411426 Bruce-Sanford, G. (1998). A simulation model for training in group process. International Journal of Group Psychotherapy, 48(3), 393-399. doi:10.1080/01933929908411426 Bugental, J. F. (1988). What is" failure" in psychotherapy? Psychotherapy: Theory, Research, Practice, Training, 25(4), 532-536. doi:10.1037/h0085378 Burlingame, G.M. (2013). Group Psychotherapy Intervention Rating Scale (GPIRS). Unpublished revised scale, Brigham Young University: Provo, UT. Burlingame, G. M., Fuhriman, A. J., & Johnson, J. E. (2001). Cohesion in group psychotherapy. Psychotherapy: Theory, Research, Practice, Training, 38(4), 373-379. doi:10.1037/0033-3204.38.4.373 Burlingame, G. M., Fuhriman, A. J., & Johnson, J. E. (2002). Cohesion in group psychotherapy. In J. C. Norcross (Ed.), Psychotherapy relationships that work: Therapist contributions and responsiveness to patients (pp. 71–88). New York, NY: Oxford University Press. Burnes, B. (2004). Kurt Lewin and the planned approach to change: A re‐appraisal. Journal of Management Studies, 41(6), 977-1002. doi:10.1111/j.1467-6486.2004.00463.x Butterfield, L. D., Borgen, W. A., Maglio, A. S. T., & Amundson, N. E. (2009). Using the enhanced critical incident technique in counseling psychology research. Canadian Journal of Counselling, 43(4), 265-282.   143  Chambless, D. L., & Hollon, S. D. (1998). Defining empirically supported therapies. Journal of Consulting and Clinical Psychology, 66(1), 7-18. doi:10.1037/0022-006x.66.1.7 Chapman, C. L. (2010). Clinical prediction in group psychotherapy. Doctoral dissertation. Provo, UT: Brigham Young University.  Chapman, C. L., Baker, E. L., Porter, G., Thayer, S. D., & Burlingame, G. M. (2010). Rating group therapist interventions: The validation of the Group Psychotherapy Intervention Rating Scale. Group Dynamics: Theory, Research, and Practice, 14(1), 15-31. doi:10.1037/a0016628 Chausovsky, F., Trijsburg, W. R., Snijders, H.J.A. & Spiering, M. (2005). The Group Psychotherapy Interventions Rating Scale: Research about the validity of an observation instrument for group psychotherapies. Thesis Department of Clinical Psychology. Amsterdam: University of Amsterdam.  Christensen, T. M., & Kline, W. B. (2001). The qualitative exploration of process-sensitive peer group supervision. The Journal for Specialists in Group Work, 26(1), 81-99. doi:10.1080/01933920108413779 Coco, G. L., Gullo, S., Prestano, C., & Burlingame, G. M. (2015). Current issues on group psychotherapy research: An overview. In O.C.G Gelo, A. Pritz & B. Rieken (Eds.). Psychotherapy Research (pp. 279-292). New York, NY: Springer-Verlag Wien. doi: 10.1007/978-3-7091-1382-0 Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159. doi:10.1037/0033-2909.112.1.155 Corey, G. (2012). Theory and practice of group counseling (8th ed.). Belmont, CA: Brooks/Cole.   144  Cota, A. A., Evans, C. R., Dion, K. L., Kilik, L., & Longman, R. S. (1995). The structure of group cohesion. Personality and Social Psychology Bulletin, 21(6), 572-580. doi:10.1177/0146167295216003 Counselman, E. F., & Weber, R. L. (2004). Organizing and maintaining peer supervision groups. International Journal of Group Psychotherapy, 54(2), 125-143. doi:10.1521/ijgp.54.2.125.40391 Council for the Accreditation of Counseling and Related Educational Programs (CACREP). (2009). 2009 Standards. Alexandria, VA: Author. Retrieved 6 June 2015 from http://www.cacrep.org/doc/2009%20Standards%20with%20cover.pdf Daniels, J. A., & Larson, L. M. (2001). The impact of performance feedback on counseling self‐efficacy and counselor anxiety. Counselor Education and Supervision, 41(2), 120-130. doi:10.1002/j.1556-6978.2001.tb01276.x Daniels, T., & Ivey, A. (2007). Microcounseling: Making skills training work in a multicultural world. Springfield, IL: Charles C Thomas Publisher. Davenport, D. (2004). Ethical issues in the teaching of group counseling. Journal for Specialists in Group Work, 29(1), 43-49. doi:10.1080/01933920490275376 DeLucia-Waack, J. L., & Bridbord, K. H. (2004). Measures of group process, dynamics, climate, leadership behaviors, and therapeutic factors. In J. L. DeLucia-Waack, D. A. Gerrity, C. R. Kalodner, & M. T. Riva (Eds.). Handbook of group counseling and psychotherapy (pp. 120-136). Thousand Oaks, CA: Sage Publications Inc. doi:10.4135/9781452229683.n9 Donigian, J., & Malnati, R. (1987). Critical incidents in group therapy. Belmont, CA:Wadsworth.   145  Dunlap, W. P., Cortina, J. M., Vaslow, J. B., & Burke, M. J. (1996). Meta-analysis of experiments with matched groups or repeated measures designs. Psychological Methods, 1(2), 170-177. doi:10.1037/1082-989x.1.2.170 Fall, K. A., & Levitov, J. E. (2002). Using actors in experiential group counseling leadership training. Journal for Specialists in Group Work, 27(2), 122-135. doi:10.1177/0193392202027002002 Feiner, S. E. (1998). Course design: An integration of didactic and experiential approaches to graduate training of group therapy. International Journal of Group Psychotherapy, 48(4), 439-460.  Fleiss, J. L., Levin, B., & Paik, M. C. (1981). The measurement of interrater agreement. Statistical Methods for Rates and Proportions, 2, 212-236. Fuhriman, A., & Burlingame, G. M. (2001). Group psychotherapy training and effectiveness. International Journal of Group Psychotherapy, 51(3), 399-416. doi:10.1521/ijgp.51.3.399.49889 Fuhriman, A., Drescher, S., Hanson, E., Henrie, R., & Rybicki, W. (1986). Refining the measurement of curativeness: An empirical approach. Small Group Research, 17(2), 186-201. doi:10.1177/104649648601700204 Furr, S. R., & Barret, B. (2000). Teaching group counseling skills: Problems and solutions. Counselor Education and Supervision, 40(2), 94-104. doi:10.1002/j.1556-6978.2000.tb01241.x Gardner, D. C., & Beatty, G. J. (1980). Dissertation proposal guidebook: How to prepare a research proposal and get it accepted. Springfield, IL: Charles C Thomas Pub.  146  Gast, D. L. (2010). Single subject research methodology in behavioral sciences. New York, NY: Routledge. doi:10.4324/9780203877937 Gibb, J. R. (1961). Defensive communication. Journal of Communication, 11(3), 141-148. doi:10.1111/j.1460-2466.1961.tb00344.x  Gladding, S. T. (2012). Groups: A counseling perspective. Upper Saddle River, NJ: Pearson Education Inc.  Goodrich, K. M. (2008). Dual relationships in group training. The Journal for Specialists in Group Work, 33(3), 221-235. doi:10.1080/01933920802204981 Greene, L. R. (2003). The state of group psychotherapy process research. International Journal of Group Psychotherapy, 53(1), 130-134. doi:10.1521/ijgp.53.1.129.42810 Greene, L. R. (2012). Group therapist as social scientist, with special reference to the psychodynamically oriented psychotherapist. American Psychologist, 67(6), 477-489. Grigg, G. (2006). Designs and discriminations for clinical group supervision in counselling psychology: An analysis. Canadian Journal of Counselling and Psychotherapy/Revue Canadienne de Counseling et de Psychothérapie, 40(2), 110-122. Harpaz, N. (1994). Failures in group psychotherapy: The therapist variable. International Journal of Group Psychotherapy. 44(1), 3-19. Hartson, D. J., & Kunce, J. T. (1973). Videotape replay and recall in group work. Journal of Counseling Psychology, 20(5), 437-441. doi:10.1037/h0035387 Hinshelwood, R. D. (2007). Bion and Foulkes: The group-as-a-whole. Group Analysis, 40(3), 344-356. doi:10.1177/0533316407076115  147  Hoyt, C. L., Murphy, S. E., Halverson, S. K., & Watson, C. B. (2003). Group leadership: Efficacy and effectiveness. Group Dynamics: Theory, Research, and Practice, 7(4), 259-274. doi:10.1037/1089-2699.7.4.259. Hsu, L. M. (2003). Random sampling, randomization, and equivalence of contrasted groups in psychotherapy outcome research. In A. E. Kazdin (Ed.) Methodological issues and strategies in clinical research 3rd ed. (pp.147-162). Washington, DC: American Psychological Association. doi:10.1037/10109-018 Ieva, K. P., Ohrt, J. H., Swank, J. M., & Young, T. (2009). The impact of experiential groups on master students. Journal for Specialists in Group Work, 34(4), 351-368. doi:10.1080/01933920903219078 Jung, C., Howard, R., Emory, R., & Pino, R. (1972). Interpersonal communications: Leader's manual. Portland, OR: Northwest Regional Educational Laboratory. Kagan, N., Schauble, P., Resnikoff, A., Danish, S. J., & Krathwohl, D. R. (1969). Interpersonal process recall. The Journal of Nervous and Mental Disease, 148(4), 365-374.  Kirk, R. E. (1995). Experimental design: Procedures for the behavioral sciences. Pacific Grove, CA: Brooks/Cole Publishing Company. Kivlighan, D. M., Coleman, M. N., & Anderson, D. C. (2000). Process, outcome, and methodology in group counseling research. In S. D. Brown & R. W. Lent (Eds.), Handbook of counseling psychology (3rd ed., pp. 767-796). New York, NY: Wiley. Kivlighan, D. M., & Tarrant, J. M. (2001). Does group climate mediate the group leadership-group member outcome relationship? A test of Yalom's hypotheses about leadership priorities. Group Dynamics, 5(3), 220-234. doi:10.1037//1089-2699.5.3.220  148  Kivlighan, D. M., & Shaughnessy, P. (2000). Patterns of working alliance development: A typology of client's working alliance ratings. Journal of Counseling Psychology, 47(3), 362-371. doi:10.1037/0022-0167.47.3.362 Kiweewa, J., Gilbride, D., Luke, M., & Seward, D. (2013). Endorsement of growth factors in experiential training groups. Journal for Specialists in Group Work, 38(1), 68-93. doi:10.1080/01933922.2012.745914 Kline, W. B. (2003). Interactive group counseling and therapy. Upper Saddle River, NJ: Pearson Education Inc.  Kline, W. B., Falbaum, D. F., Pope, V. T., Hargraves, G. A., & Hundley, S. F. (1997). The significance of the group experience for students in counselor education: A preliminary naturalistic inquiry. Journal for Specialists in Group Work, 22(3), 157-166. doi:10.1080/01933929708414377 Kolb, D. A. (1984). Experiential learning: Experience as the source of learning and development. Englewood Cliffs, NJ: Prentice-Hall. Kolb, A. Y., & Kolb, D. A. (2005). Learning styles and learning spaces: Enhancing experiential learning in higher education. Academy of Management Learning & Education, 4(2), 193-212. doi:10.5465/amle.2005.17268566 Kormanski, C. (1991). Using theory as a foundation for group training designs. Journal for Specialists in Group Work, 16(4), 215-222. doi:10.1080/01933929108416155 Kottler, J. A. (2004). Realities of teaching group counseling. Journal for Specialists in Group Work, 29(1), 51-53. doi:10.1080/01933920490275385  149  Kuntze, J., Molen, H. T., & Born, M. P. (2009). Increase in counselling communication skills after basic and advanced microskills training. British Journal of Educational Psychology, 79(1), 175-188. doi:10.1348/000709908X313758 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. doi:10.2307/2529310 Larson, D, Flesaker, K., & Stege, R. (2008). Qualitative interviewing using interpersonal process recall: Investigating interpersonal experiences during professional-client conversations. International Journal of Qualitative Methods, 7(1), 18-37.  Larson, L. M., & Daniels, J. A. (1998). Review of the counseling self-efficacy literature. The Counseling Psychologist, 26(2), 179-218. doi:10.1177/0011000098262001 Larson, L. M., Suzuki, L. A., Gillespie, K. N., Potenza, M. T., Bechtel, M. A., & Toulouse, A. L. (1992). Development and validation of the counseling self-estimate inventory. Journal of Counseling Psychology, 39(1), 105-120. doi:10.1037//0022-0167.39.1.105  Lee, F., & Bednar, R. L. (1977). Effects of group structure and risk-taking disposition on group behavior, attitudes, and atmosphere. Journal of Counseling Psychology, 24(3), 191-199. doi:10.1037/0022-0167.24.3.191 Lieberman, M. A. (1990). Understanding how groups work: A study of homogeneous peer group failures. International Journal of Group Psychotherapy, 40(1), 31-52. Lieberman, M. A., Yalom, I. D., & Miles, M. B. (1973). The impact of encounter groups on participants: Some preliminary findings. The Journal of Applied Behavioral Science, 8(1), 29-50. doi:10.1177/002188637200800102  150  Luft, J. & Ingham, H. (1955). The Johari window, a graphic model of interpersonal awareness. In Proceedings of the western training laboratory in group development. Los Angeles, CA: UCLA. Maione, L., & Mirenda, P. (2006). Effects of video modeling and video feedback on peer-directed social language skills of a child with autism. Journal of Positive Behavior Interventions, 8(2), 106-118. doi:10.1177/10983007060080020201 Markus, H. E. & Abernethy, A. D. (2001). Joining with resistance: Addressing reluctance to engage in group therapy training. International Journal of Group Psychotherapy, 51(2), 191-204. doi:10.1521/ijgp.51.2.191.49854 McComas, J. J., Wacker, D. P., Cooper, L. J., Asmus, J. M., Richman, D., & Stoner, B. (1996). Brief experimental analysis of stimulus prompts for accurate responding on academic tasks in an outpatient clinic. Journal of Applied Behavior Analysis, 29(3), 397-401. doi:10.1901/jaba.1996.29-397 Merta, R. J., & Sisson, J. A. (1991). The experiential group: An ethical and professional dilemma. Journal for Specialists in Group Work, 16(4), 236-245. doi:10.1080/01933929108416158 Merta, R. J., Wolfgang, L., & McNeil, K. (1993). Five models for using the experiential group in the preparation of group counselors. Journal for Specialists in Group Work, 18, 200–207. doi:10.1080/01933929308413755 Morran, D. K., Stockton, R., Cline, R. J., & Teed, C. (1998). Facilitating feedback exchange in groups: Leader interventions. Journal for Specialists in Group Work, 23(3), 257-268. doi:10.1080/01933929808411399  151  Nathan, P. E., Stuart, S. P., & Dolan, S. L. (2000). Research on psychotherapy efficacy and effectiveness: Between Scylla and Charybdis? Psychological Bulletin, 126(6), 964-981. doi:I0.1037//0033-2909.126.&.9S4 Norcross, J. C., & Wampold, B. E. (2011). Evidence-based therapy relationships: Research conclusions and clinical practices. Psychotherapy, 48(1), 423-430. doi:10.1093/acprof:oso/9780199737208.003.0021 Ogden, P., Pain, C., & Fisher, J. (2006). A sensorimotor approach to the treatment of trauma and dissociation. Psychiatric Clinics of North America, 29(1), 263-279. doi:10.1016/j.psc.2005.10.012 Ohrt, J. H., Ener, E., Porter, J., & Young, T. L. (2014). Group leader reflections on their training and experience: Implications for group counselor educators and supervisors. Journal for Specialists in Group Work, 39(2), 95-124. doi:10.1080/01933922.2014.883004 Ohrt, J. H., Robinson III, E. M., & Hagedorn, W. B. (2013). Group leader development: Effects of personal growth and psychoeducational groups. Journal for Specialists in Group Work, 38(1), 30-51. doi:10.1080/01933922.2012.732982 Orr, J. J., & Hulse-Killacky, D. (2006). Using voice, meaning, mutual construction of knowledge, and transfer of learning to apply an ecological perspective to group work training. Journal for Specialists in Group Work, 31(3), 189-200. doi:10.1080/01933920600777824 Osborn, C. J., Daninhirsch, C. L., & Page, B. J. (2003). Experiential training in group counseling: Humanistic processes in practice. The Journal of Humanistic Counseling, Education and Development, 42(1), 14-28. doi:10.1002/j.2164-490x.2003.tb00165.x  152  Page, B. J., Pietrzak, D. R., & Lewis, T. F. (2001). Development of the group leader self-efficacy instrument. Journal for Specialists in Group Work, 26(2), 168-184. doi:10.1080/01933920108415736 Pollack, B. N. (1998). Hierarchical linear modeling and the" unit of analysis" problem: A solution for analyzing responses of intact group members. Group Dynamics: Theory, Research, and Practice, 2(4), 299-312. doi:10.1037//1089-2699.2.4.299 Riva, M. T., Wachtel, M., & Lasky, G. B. (2004). Effective leadership in group counseling and psychotherapy: Research and practice. In J. L. DeLucia-Waack, D. A. Gerrity, C. R. Kalodner, & M. T. Riva (Eds.). Handbook of group counseling and psychotherapy (pp. 37-48). Thousand Oaks, CA: Sage Publications Inc. doi:10.4135/9781452229683.n3 Robison, F. F., Jones, E. N., & Berglund, K. E. (1996). Research on the preparation of group counselors. Journal for Specialists in Group Work, 21(3), 172-177. doi:10.1080/01933929608412248 Rogers, C. (1961). On becoming a person: A therapist's view of psychotherapy. Boston, MA : Houghton Mifflin. Romano, J. L. (1998). Simulated group counseling: An experiential training model for group work. Journal for Specialists in Group Work, 23(2), 119-132. doi:10.1080/01933929808411385 Romano, J. L., & Sullivan, B. A. (2000). Simulated group counseling for group work training: A four-year research study of group development. Journal for Specialists in Group Work, 25(4), 366-375. doi:10.1080/01933920008411680  153  Safran, J. D., & Muran, J. C. (1996).  The resolution of ruptures in the therapeutic alliance.  Journal of Consulting and Clinical Psychology, 64(3), 447-458. doi:10.1037//0022-006x.64.3.447 Safran, J. D., Muran, J. C., Samstag, L. W., & Stevens, C. (2001). Repairing alliance ruptures. Psychotherapy: Theory, Research, Practice, Training, 38(4), 406-412. doi:10.1037//0033-3204.38.4.406 Schimmel, C. J., & Jacobs, E. E. (2011). When leaders are challenged: Dealing with involuntary members in groups. The Journal for Specialists in Group Work, 36(2), 144-158. doi:10.1080/01933922.2011.562345 Schutz, W. C. (1958). FIRO: A three-dimensional theory of interpersonal behaviour. New York, NY: Holt, Rinehart and Winston, Inc. doi:10.2307/3708975 Shechtman, Z., & Toren, Z. (2009). The effect of leader behavior on processes and outcomes in group counseling. Group Dynamics: Theory, Research, and Practice, 13(3), 218-233. doi:10.1037/a0015718 Shumaker, D., Ortiz, C., & Brenninkmeyer, L. (2011). Revisiting experiential group training in counselor education: A survey of master's-level programs. Journal for Specialists in Group Work, 36(2), 111-128. doi:10.1080/01933922.2011.562742. Smaby, M. H., Maddux, C. D., Torres-Rivera, E., & Zimmick, R. (1999). A study of the effects of a skills-based versus a conventional group counseling training program. Journal for Specialists in Group Work, 24(2), 152-163. doi:10.1080/01933929908411427 Smokowski, P. R., Rose, S. D., & Bacallao, M. L. (2001). Damaging experiences in therapeutic groups: How vulnerable consumers become group casualties. Small Group Research, 32(2), 223-251. doi:10.1177/104649640103200205  154  Spielberger, C. D. (1983). State-trait anxiety inventory. Palo Alto, CA: Mind Garden Inc. doi:10.1037/t06496-000 Starling, P. V., & Baker, S. B. (2000). Structured peer group practicum supervision: Supervisees' perceptions of supervision theory. Counselor Education and Supervision, 39(3), 162-176. doi:10.1002/j.1556-6978.2000.tb01229.x Stockton, R. (2010). The art and science of group counseling. Journal for Specialists in Group Work, 35(4), 324-330. doi:10.1080/01933922.2010.515904 Stockton, R., Morran, K., & Krieger, K. M. (2004). An overview of current research and best practices for training beginning group leaders. In J. L. DeLucia-Waack, D. A. Gerrity, C. R. Kalodner, & M. T. Riva (Eds.). Handbook of group counseling and psychotherapy (pp. 65-75). Thousand Oaks, CA: Sage Publications Inc. doi:10.4135/9781452229683.n5 Stockton, R., & Toth, P. L. (1996). Teaching group counselors: Recommendations for maximizing preservice instruction. Journal for Specialists in Group Work, 21(4), 274-282. doi:10.1080/01933929608412259 St. Pierre, B. K. (2014). Student attitudes and instructor participation in experiential groups. The Journal for Specialists in Group Work, 39(3), 194-211. doi:10.1080/01933922.2014.919048 Topping, K. J. (2005). Trends in peer learning. Educational Psychology, 25(6), 631-645. doi:10.1080/01443410500345172 Toth, P. L., & Erwin, W. J. (1998). Applying skill‐based curriculum to teach feedback in groups: An evaluation study. Journal of Counseling & Development, 76(3), 294-301. doi:10.1002/j.1556-6676.1998.tb02545.x  155  Toth, P. L., & Stockton, R. (1996). A skill-based approach to teaching group counseling interventions. Journal for Specialists in Group Work, 21(2), 101-109. doi:10.1080/01933929608412238 Toth, P. L., Stockton, R., & Erwin, W. J. (1998). Application of a skill-based training model for group counselors. Journal for Specialists in Group Work, 23(1), 33-49. doi:10.1080/01933929808411380 Trijsburg, R. W., Frederiks, G. C. F. J., Gorlee, M., Klouwer, E., Hollander, A. D., & Duivenvoorden, H. J. (2002). Development of the Comprehensive Psychotherapeutic Interventions Rating Scale (CPIRS). Psychotherapy Research, 12(3), 287-317. doi:dx.doi.org/10.1093/ptr/12.3.287 Tschuschke, V., & Greene, L. R. (2002). Group therapists' training: What predicts learning? International Journal of Group Psychotherapy, 52(4), 463-482. doi:10.1521/ijgp.52.4.463.45522 Tyson, L. E., Perusse, R., Whitledge, J., & Kelly, V. A. (2004). Critical incidents in group counseling. Alexandria, VA: American Counseling Association. Urbani, S., Smith, M. R., Maddux, C. D., Smaby, M. H., Torres‐Rivera, E., & Crews, J. (2002). Skills‐based training and counseling self‐efficacy. Counselor Education and Supervision, 42(2), 92-106. doi:10.1002/j.1556-6978.2002.tb01802.x Wallin, D. J. (2007). Attachment in psychotherapy. New York, NY: Guilford Press.  Westwood, M. J. (1989). Group supervision for counsellors-in-training. Canadian Journal of Counselling, 23(4), 348-353. Wicker, A. W. (1998). Getting out of our conceptual ruts: Strategies for expanding conceptual frameworks. In A. E. Kazdin (Ed.) Methodological issues and strategies in clinical  156  research (pp. 39-58). Washington, DC: American Psychological Association. doi:10.1037/10109-016 Wilkins, P. (1995). A creative therapies model for the group supervision of counsellors. British Journal of Guidance and Counselling, 23(2), 245-257. doi:10.1080/03069889500760251 Woolsey, L. K. (1986). The critical incident technique: An innovative qualitative method of research. Canadian Journal of Counselling, 20(4), 242-254.  Yalom, I. D. (1966). Problems of neophyte group therapists. International Journal of Social Psychiatry, 12(1), 52-59. doi:10.1177/002076406601200107 Yalom, I. D., & Leszcz, M. (2005). The theory and practice of group psychotherapy. New York, NY: Basic Books. Young, R. A., Valach, L., & Domene, J. F. (2005). The action-project method in counseling psychology. Journal of Counseling Psychology, 52(2), 215-223. doi:10.1037/0022-0167.52.2.215   157  Appendices Appendix A  : Detailed methodology for primary study (chapter 4) This study employed a repeated-measures randomized control trial design, CR-2, where group counselling students from two classes (n = 30) were randomly divided into treatment and waitlist control conditions. While all participants attended classroom lectures, the treatment condition also attended an experiential training lab for 6 sessions. Both conditions were tested for self-reported measures of state/trait anxiety and counsellor self-efficacy at five points during the course. In addition, the performance of each student was measured by pre/post video-recording of group leading for 10 minutes of the treatment period. After the treatment period, the wait-list control condition attended the lab and was pre/post video-recorded for group leading. All video recordings were coded by independent raters. This study gained approval from the behavioural research ethics board (BREB) prior to commencing. This study addressed two research questions. First, does the experiential training program, Experiam, increase counsellor trainee confidence in group leading? Four directional hypotheses were tested using an experimental design.  (a) The first hypothesis was that Experiam will improve student GCSE over controls, as measured by the Group Leader Self-efficacy Instrument (GLSI; Page et al, 2001).  (b) The second hypothesis was that participation in Experiam will cause GCSE to increase over time, as measured by the Group Leader Self-efficacy Instrument (GLSI; Page et al, 2001). (c) The third hypothesis was that participation in Experiam will decrease counsellor trainee anxiety over controls, as measured by the State-Trait Anxiety Inventory (STAI; Spielberger, 1983).   158  (d) The fourth hypothesis was that participation in Experiam will decrease counsellor trainee anxiety over time, as measured by the State-Trait Anxiety Inventory (STAI; Spielberger, 1983).  The second research question asked: does participation in Experiam increase counsellor-trainee performance? Two directional hypotheses were tested using a quasi-experimental design. (e) The first hypothesis was that frequency of skill demonstration will increase with participation in Experiam over time, as measured by the Group Psychotherapy Intervention Rating Scale (GPIRS; Chapman et al., 2010).  (f) The second hypothesis was that clarity of skill demonstration will increase with participation in Experiam over time, as measured by the Group Psychotherapy Intervention Rating Scale (GPIRS; Chapman et al., 2010).   The research questions were seated in both experiential learning theory and social cognitive theory and tests the assumption that experiential learning increases self-efficacy, decreases anxiety, and increases performance; all hypotheses were directional (Bandura, 1978a; 1982; 2001).  This design was chosen to best capture the impact of the experiential lab on the development of group counselling leadership with minimal nuisance variables. The experimental design also allowed for results to be generalized to a larger population than the sample. Stockton et al. (2004) reviewed research on effective group leadership and discovered that the extent literature has made some progress from its anecdotal roots in supervision case studies, but that current research must employ experimental or quasi-experimental methods that maximize generalizability.   159  Experimental research validity. Kirk (1995) identified several threats to valid inference making in experimental research. These include threats to statistical, internal, and external validity. The randomized control group alleviates many of the validity threats of this study. Hsu (2003) examined equivalence of randomized groups in psychotherapy outcome research and concluded that sample sizes of at least 20 are likely equivalent, assuming a minimum of nuisance variables. Using his calculated table and the sample size of 30, the probability of non-equivalence between the conditions in this study is less than 10% with one nuisance variable. Therefore, a careful analysis of any possible nuisance variables between the conditions is required.  Nuisance variables are defined as undesired sources of variation that affect the dependent variable (Kirk, 1995). Examples include variance in presentation of instructions to different subjects, measurement error, environmental factors, and personal characteristics such as gender. To control for these variables, he suggested several strategies. First, any hidden nuisance variable was held constant for all subjects by randomly assign subjects to treatment and control conditions. Second, Experiam groups were led by different senior leaders, reducing leader characteristic bias. Third, the participants took one of two courses that are taught by different senior faculty members, thus reducing bias of a single instructor.  Statistical validity. Low statistical power may lead to inflated type II errors, due to an inadequate sample size, uncontrolled sources of variation, or inefficient test statistics. Kirk (1995) suggested that power be at least .8 as a minimum acceptable probability of correctly rejecting a false null hypothesis. A within-group pilot (n = 21) of three self-efficacy measures was conducted and a  160  power analysis for effect size > 1 revealed that a sample size of n > 20 was sufficient to minimize the chance of Type II error at the p = .01 level.  Threats to internal validity. Kirk (1995) described several threats to internal validity that may impact this study. First, maturation bias may be caused because the treatment groups are split between two courses offered in different semesters. This was addressed by comparing levels of both treatment groups at the beginning of each course, thus capturing if participants in the second semester course have maturation bias. However, retesting bias may occur, due to increased repetition of measures, and so each treatment and control group was limited to three measures each. Second, compensatory rivalry may arise between treatment and control groups because both groups were interacting in the classroom. This was mitigated by the fact that treatment group participation was normed to confidentiality within the groups. Third, experimenter bias may arise due to the behavioural raters knowing the purpose of the experiment. This was reduced by blinding the raters to the groups and time periods.  Finally, Kivlighan et al. (2000) identified methodological problems in group research, such as nested data and mutually dependent data. Nested data which describes how members within a group do not constitute independent units of observation. Called the “unit of analysis problem”, Pollack (1998) defined it as choosing one level of analysis (i.e. the group) and forcing all variables of interest to that level (e.g. by averaging across individuals), so that statistical analysis may be performed on variable relationships. Kivlighan et al. (2000) stated that treating all members as individuals and ignoring the group may violate the independence assumption of hypothesis testing, which can lead to confounded statistical results. The second methodological problem lies in mutual dependency of responses. This occurs when individual member or leader  161  personalities or climates have an impact on group members. Kivlighan et al. (2000) named effects that account for this, such as group norming, perceiver and target effects, and dyadic effects. These problems are theoretically reduced in this study because the participants were not measured as a group, but as individual group leaders. Threats to external validity. The CR-2 design controls for the interaction between testing and treatment, thus ensuring external validity (Gardner & Beatty, 1980). One threat to external validity warrants discussion. The selection of subjects into groups may interact with treatment, in that the participant availability for the groups that may thwart a simple randomization process. Other threats included experimenter expectancy effect, which is minimized by having the primary researcher not involved in the treatment. Demand characteristics were likely not a factor, since the experiment is congruent to the treatment. In other words, the participants were motivated to do well in the lab as a way to succeed in the course, but the success of each student was only limited by the very factors the experiment is measuring. However, this congruence may lead to subject-predisposition effects, in that the participants may have a vested interest in responding overly positively to the lab.  Participants The participants (n = 30) of this study comprised graduate students taking a first course in group counselling. All students had, at a minimum, an undergraduate course in basic listening skills as a prerequisite to the course. Demographic information was assessed of the students, including gender, age, previous education/occupation, months enrolled in current program, and hours of prior group leading experience.   162  Measures Self-efficacy.  The thirty-six item GLSI (Page et al., 2001) measured counsellor trainee’s self-efficacy in delivering a broad-base of group leader skills.  This self-report instrument was developed from the Counsellor Self-Estimate Inventory (COSE; Larson et al., 1992), but adapted to group work.  It assessed participants’ perceived self-efficacy for leading groups and answers are arranged in a Likert-type scale ranging from 1 (strongly disagree) to 6 (strongly agree).  Factor analysis was conducted using 204 counsellor trainees from 16 American universities, and confirmed a one-factor solution, called group leader self-efficacy, that accounted for 38% of total variance, with high internal consistency (Cronbach’s α = .95).  The 2-week test-retest reliability of the measure was calculated to be .72 (DeLucia-Waack & Bridbord, 2004).  Discriminant validity was obtained through non-significant correlations between the GLSI and the S-Anxiety scale of the STAI and the neuroticism, extroversion, and openness to experience subscales of the NEO five factor personality inventory.  The GLSI was selected for this study because it was the only peer-reviewed and validated instrument in the literature that specifically measured self-efficacy for leading counselling groups. Anxiety (STAI): The State-Trait Anxiety Inventory has two 20-item subscales that have a 4-point Likert-type scale anchored from 1 to 4 (thus ranging from 20 to 80 per subscale), where higher scores represent higher levels of anxiety (Spielberger, 1983).  The trait subscale measures anxiety levels that are stable over time, while the state subscale assesses the person’s current state and context.  This scale has been normed on men and women over three age ranges, and across working adults, college students, high school students, and military recruits.  Daniels and Larson (2001) stated that self-efficacy is more congruent to state versus trait anxiety, and so the state measure was selected.  The transient nature of state anxiety makes the test/retest less  163  meaningful, while Cronbach’s alpha was found to be over .9 for male and female college students.  Construct validity was established for college students by comparing state subscale levels from a control of a regular class, and finding higher scores during examination conditions and lower scores after relaxation training.  While the STAI is not related to intelligence or aptitude, high levels of state anxiety have been associated with a wide array of conditions that induce stress, including exam periods for college students. Performance (GPIRS): To assess counsellor performance, video-recorded group leading was coded using the GPIRS (Chapman et al., 2010).  This observer-rated scale is used to evaluate the frequency and clarity of group leader skills that enhance outcomes in group psychotherapy (Chapman, 2010).  The earliest published scale was a thesis from the University of Amsterdam and contained 26-items, each defining a group leading skill, rated on a 4-point scale ranging from 0 (poor) to 3 (excellent), and not rated if not observed (Chausovsky, Trijsburg, Snijders & Spiering, 2005).  Items were grouped into three broad categories and ten subscales, and were scored both for frequency and clarity.  In 2010, Christopher Chapman and faculty at Brigham Young University published works further developing and testing the GPIRS (Chapman 2010; Chapman et al., 2010).  The authors increased the number of items to forty-eight, derived from empirical and clinical evidence of leader behaviours that increase group cohesion, herein referred to as leader skills  (Burlingame et al., 2002).  They further categorized items into three domains: (a) Group Structuring, comprising skills defining group goals, norms, and boundaries, (b) Verbal Interactions, comprising skills modelling and facilitating self-disclosure and feedback between members, and (c) Emotional Climate, comprising skills that help members express feelings in a constructive manner, and to provide appropriate balance of support and challenge.   164  Raters scored leader behaviours using a 4-point Likert-type scale of skill clarity and frequency, ranging from 1 (poor) to 4 (excellent), and 0 was coded if the skill was not observed.  The number of times each skill occurred was also recorded, providing a measure of skill frequency.  The authors stated that the internal consistency of the scale is high, with alpha = .93.  They tested concurrent validity by comparing the scale to measures of similar and dissimilar scales.  Concurrent validity was shown via correlations to an individual psychotherapy analogue, called the Comprehensive Psychotherapy Intervention Rating Scale (CPIRS; Trijsburg, Frederiks, Gorlee, Klouwer, Hollander & Duivenvoorden, 2002).  They further determined that higher rated leader skills on the GPIRS were significantly positively correlated to lower levels of hostility between group members, as measured by the Hill Interaction Matrix (HIM), considered to be the gold standard of group climate scales.  Recent unpublished research and development of the GPIRS has resulted in its reduction to 36 items and division into two 18-item subscales, one containing Basic skills and one containing Advanced skills (Burlingame, 2013; G. Burlingame & K. Bruer, personal communication, August 2013).   These researchers further reduced the scale anchors to three: intervention did not occur (0), ambiguous – intervention occurred but clarity could be improved (1), or intervention was performed with clarity (2).  Items were coded N/A if those skills were not observed.  Finally, videos were coded in two-minute intervals.  The coding scheme is provided in appendix B, and the revised scale is provided in appendix C. In this study, the GPIRS coding scheme was further clarified to consider the absence of skills needed to foster or maintain cohesion in the group.  Coding for the absence of leader actions has empirical support in studies demonstrating that the formation of cohesion is fostered by leaders who provide structure to groups in early sessions, and that passivity in leaders can  165  result in harm to group members (Kivlighan et al., 2000; Smokowski 2001).3  Therefore, in an emotionally-charged group, cohesion may be impacted by what leaders fail to do.  For example, high group conflict has been shown to reduce therapeutic outcomes in group therapy (Kivlighan & Tarrant, 2001).  A group leader not addressing member behaviours that increase interpersonal conflict by limiting members from attacking one another would be assigned 0 for the item: Stopped attacking and judgmental situations and expressions between members.  In contrast, a scenario with no interpersonal conflict would likely have no ratings for this item.  Therefore, for this study, the most current anchor definitions were employed, with the (0) rating defined as intervention did not occur and was therapeutically indicated.  As in prior incarnations of the GPIRS, items were not rated if they both did not occur and were not deemed necessary.   In reviewing behavioural measures of group process variables, Chapman et al. (2010) noted that many focus on member interactions and not leader skills, or that leader skills are not, themselves, theoretically or empirically derived.  The GPIRS was created specifically to measure empirically-tested leader skills that increase cohesion of groups, while requiring minimal rater training and inference. Therefore the GPIRS most directly addresses the research question of performance in group leading.   Procedure  Two sections of the 13-week group counselling course were examined in this study, one in the fall term and one in the winter term. Two different faculty members taught the sections, each with 30+ years of group counselling experience. Students were made aware of the study and were given the choice to participate prior to the beginning of the first class. Participants were                                                  3 For a theoretical argument of the impacts of absentee group leadership, the reader is directed to the first chapter of Bion`s (1961) theory of group development.  166  randomly assigned into treatment and waitlist (control) conditions4. Hsu (2003) stated that for sample experiments of more than 30-40 subjects, simple random assignment will likely result in equivalence between treatment and control conditions, unless there are significant nuisance variables. Such variables were considered in section 3.1. Each treatment group was subsequently randomly assigned to groups of size 4-6, forming two treatment groups, because six is the maximum size for each experiential lab. All participants were administered measures of self-efficacy, state and trait anxiety prior to group assignment, at the first, fourth, seventh, tenth, and thirteenth weeks of the course. All participants attended the 3-hour weekly classroom course. As illustrated in appendix D, the treatment groups began the experiential lab in week 2, following the first three hours of classroom instruction. These first hours of classroom instruction provided students with basic group counselling micro-skills. Treatment groups met for six consecutive, weekly, 3-hour experiential lab sessions to practice micro-skills according to the Experiam manual (see chapter 3). At the beginning of the first, third, and sixth lab session, the state anxiety measure was administered to the treatment participants. Each participant led a group for 10 minutes during each session, and leading videos were archived for coding of behavioural group leading interventions. After the third experiential lab (week 4), all students in both group conditions completed the self-efficacy and state anxiety measures. Following the final experiential session (week 8 of the course), the self-efficacy measure was once again be administered to both treatment and control groups. The treatment groups met one final time in the 8th week of the course to recreate the first week scenario, and this was video-recorded to provide an equivalent                                                  4 Note that this is a randomization or random assignment model, in that the two groups are not independently drawn from a larger population (called a random sampling model), but together comprise all the available participants for the study (Hsu, 2003).  167  comparison of group leading scenario to the control group’s first week of the experiential lab. The control group then participated in the six-week experiential lab from weeks 8-13. As stated above, only the first session of the control group was behaviourally coded to test hypothesis 5. Data Collection Students who volunteered to participate in the study signed an informed consent and were assigned a number used to anonymously match their scores on the instruments. The students completed a demographic questionnaire, the GLSI, and the STAI. The hard copies of the instruments were stored in a locked filed cabinet in the researcher’s office. All self-report measures were entered by two independent raters to correct for data entry errors; the data was entered 99.5% error-free, and data entry errors were corrected.  Appendix D illustrates the data collection plan for the study. The pretest self-report measures were collected prior to the first class and took approximately 15 minutes to complete. During each administration stage, the self-report measures were taken prior to the beginning of the class.  Video recordings The video recordings comprised the first and final sessions of treatment and control groups. The recordings were collected for both offerings of the course before coding, then were randomized between the two raters. This minimized the effect of rater bias that could arise from knowing which session or group the recording came from. For a total sample of 30, there were 60 sessions to be coded (26 for the pre and post-test control and 34 for pre and post-test treatment). The second independent rater was a senior graduate student at the doctoral level, who participated in advanced group courses and facilitated the lab. The first independent rater was myself (SH), as I met those criteria and am not otherwise involved in the collection of data.  168  The video coding inter-observer process occurred as follows. The raters met for approximately 10 hours to train using the GPIRS with pilot videos. Both raters continued to code videos until percentage of agreement and inter-observer agreement (kappa) exceeded 75%, which are indicative of excellent agreement (Fleiss, 1981). Best practices for inter-observer agreement in small-sample behavioural analysis requires that the primary investigator code 100% of videos, with an independent rater coding 20% to 33% of videos at several points during the coding process (Gast, 2010; Maione, & Mirenda, 2006; McComas et al., 1996). A schedule of rater meetings was established where the independent rater coded 33.3% of sessions and these codings were compared to the primary rater on four occasions during the coding process using Cohen’s kappa.  Data Analysis The analysis of the first research question involved one independent variable with two levels (treatment and waitlist control conditions), and mean differences were analyzed on three dependent variables (self-efficacy, anxiety, and performance). All of the data was entered and analyzed in SPSS version 22.0 by the researcher.  The main analysis was a repeated-measures analysis of variance (ANOVA) of pre/mid/post-test impacts on treatment and control conditions. The research question stated that participation in an experiential group counselling lab will increase counsellor confidence and competence. For research question 1, hypotheses 1-4 were tested as follows:      169  1) The first hypothesis will be tested using a between-subjects analysis of the GCSE measure examining, a. Null hypothesis: H0 :  μ(GCSE)Treatment = μ(GCSE)Control 2) The second hypothesis was tested using a within- subjects analysis of the GCSE measure at the three time periods pre (t=0), mid (t=1), and post (t=2);  a. Null hypothesis: H0 :  μ(GCSE)t=0 = μ(GCSE)t=1 = μ(GCSE)t=2 3) The third hypothesis was tested using a between- subjects analysis of the STAI measure,  a. Null hypothesis: H0 :  μ(STAI)Treatment = μ(STAI)Control 4) The fourth hypothesis was tested using a within- subjects analysis of the STAI measure at the three time periods pre (t=0), mid (t=1), and post (t=2);  a. Null hypothesis: H0 :  μ(STAI)t=0 = μ(STAI)t=1 = μ(STAI)t=2 For research question 2, a quasi-experimental design was required, as there was no control group that did not take Experiam. Hypotheses 1 and 2 were tested as follows:  1) The first hypothesis was tested using a within- subjects analysis of frequency of GPIRS items at pre and post-test administrations; a. Null hypothesis: H0 : μ(GLSIfreq)t=0 = μ(GLSIfreq)t=1 2) The sixth hypothesis was tested using a within-groups analysis of clarity scores of the GPIRS at pre and post-test administrations; a. Null hypothesis: H0 :  μ(GLSIclarity)t=0 = μ(GLSIclarity)t=1 Several plausible rival hypotheses challenged the findings of this design. First, that completing the lab itself may bring about heightened perceptions of self-efficacy and that completing any extra lab work would bring about this effect.   170  Delimitations A delimitation of this design is that the experiential groups began after the first three hours of classroom instruction. This occurred to allow all students to participate in the 6 week lab while preserving the independence of the control condition and allowing the waitlisted students to also participate in Experiam. This was tested by comparing self-efficacy and anxiety levels at similar points in the lab between the treatment and control groups. Next, the performance variable was not pre-tested for the control group, limiting between-group comparisons only to the post-test. This was necessary design limitation, as attempting to gain pre-test control-group performance data would entail putting those participants into a treatment-style environment, thus biasing the control group. Finally, this study was conducted at a single university over a single calendar year.     171  Appendix B  : Coding scheme for GPIRS (adapted from Burlingame 2013)   GROUP PSYCHOTHERAPY INTERVENTION RATING SCALE (GPIRS)Rater:_______ VidID:______Group Structuring: Setting treatment expectation1 Set group agendas2 Describe rationaleA Discuss fears/concernsGroup Structuring: Establishing group procedures and roles3 Discuss Group rules4 Structure that facilitates mbr interactionB Discuss roles and responsibilityVerbal Interaction: Verbal style5 Present–centered to something explicit6 Model informational self-disclosure7 Model feeling self-disclosure8 Maintain functional controlC Model member-member behaviorD Facilitate member-member interactionE Disclosure achieves resolutionF Elicited mb-mbr feeling disclosureVerbal Interaction: Self disclosure9 Inviting probe/question (non forcing)10 Present-centered vs. story-telling11 Interrupted dominant member12 Share brief personal experienceVerbal Interaction: FeedbackGReframed injurious feedbackHRestated corrective feedback IUsed consensus for m-m or m-l  feedbackJBalance positive and corrective m-m feedbackK Encouraged +ve m-m feedbackL Apply mbr-mbr feedback to out-of-groupEmotional Climate: Leader contribution13 Not defensive when failed or confronted14 Maintain active engagement w/ group15 Nonjudgmental language with mbrs16 Modeled open/ genuine warmth M Adv. Empathy (concerns/motives)N Encouraged emotional engagemnt betw mbrsO Fostered supportive challengeEmotional Climate: Member contribution17 Respond w/ Empathy 18 Stopped attacking between membersP Assisted mbrs in describing emotionsQ Elicited mbr-mbr linking expressionsR No hostility to negative mbr behavior0-2min 2-4min 4-6min 6-8min 8+min0-2min 2-4min 4-6min 6-8min 8+min0-2min 2-4min 4-6min 6-8min 8+min 172  Appendix C  : Revised GPIRS Scale (Burlingame, 2013) Side 1—Basic Skills Intervention did not occur = 0 Ambiguous - occurred but clarity could be improved = 1 Intervention was performed with clarity = 2  Group Structuring  May be applied in any group session, at any time during a session, but more effective in early sessions and beginning of a group session. Setting treatment expectation □ 0 □ 1 □ 2 □ N/A 1.  Set group agendas (such as discussion topic or group activities)  □ 0 □ 1 □ 2 □ N/A 2.  Describe rationale underlying group session treatment/activity  Establishing group procedures and roles □ 0 □ 1 □ 2 □ N/A 3.  Discuss/remind Group rules (such as time, attendance, absences, tardiness, confidentiality, and participation; need to discuss at least one of these.  Just mentioning is not enough)  □ 0 □ 1 □ 2 □ N/A 4.  Provide structure that facilitates member interaction—note, not member-member interaction. (e.g. feedback exercises)   Verbal Interaction Verbal style □ 0 □ 1 □ 2 □ N/A 5.  Interact in a present – centered manner to something occurring explicitly in that group session  □ 0 □ 1 □ 2 □ N/A 6.  Model informational self-disclosure  □ 0 □ 1 □ 2 □ N/A 7.  Model feeling self-disclosure; has to have a feeling word (sad)  □ 0 □ 1 □ 2 □ N/A 8.  Maintain functional control of group  Self disclosure  □ 0 □ 1 □ 2 □ N/A 9.  Encouraged self disclosure relevant to the current group agenda without “forcing it”  □ 0 □ 1 □ 2 □ N/A 10.  Encouraged present – centered vs. story-telling disclosure (past or future)  □ 0 □ 1 □ 2 □ N/A 11.  Interrupted ill-timed or excessive member disclosure that dominates group talk  □ 0 □ 1 □ 2 □ N/A 12.  Leader shared brief personal experience (not as an observer but participant) relevant to group agenda (without being judgmental or overly-intellectual)   Emotional Climate Leader contribution □ 0 □ 1 □ 2 □ N/A 13.  Leader was not defensive (when leader made a mistake or an intervention failed or when confronted by a member)\  □ 0 □ 1 □ 2 □ N/A 14.  Maintained an active engagement with the group and its work  □ 0 □ 1 □ 2 □ N/A 15.  Used nonjudgmental language with members  □ 0 □ 1 □ 2 □ N/A 16.  Modeled expressions of open and genuine warmth to all members throughout session  Member contribution  □ 0 □ 1 □ 2 □ N/A 17.  Recognized and responded to the meaning of groups members’ comments  □ 0 □ 1 □ 2 □ N/A 18.  Stopped attacking and judgmental situations and expressions between members     173  Side 2—Advanced Skills Intervention did not occur = 0 Ambiguous - occurred but clarity could be improved = 1 Intervention was performed with clarity = 2  Group Structuring  May be applied in any group session, at any time during a session, but more effective in early sessions and beginning of a group session. Setting treatment expectation □ 0 □ 1 □ 2 □ N/A A.  Discuss fears/concerns about participating in group  Establishing group procedures and roles □ 0 □ 1 □ 2 □ N/A B.  Explicit discuss roles and responsibility  (Member and leader) beyond setting agenda (i.e., roles can relate to session agenda)  Verbal Interaction Verbal style □ 0 □ 1 □ 2 □ N/A C.  Model appropriate member-member behavior  □ 0 □ 1 □ 2 □ N/A D.  Facilitate member-member interaction; not member-group  Self disclosure  □ 0 □ 1 □ 2 □ N/A E.  Educate members understand that disclosed issues achieve more resolution than undisclosed issues  □ 0 □ 1 □ 2 □ N/A  F.  Elicited member – member feeling disclosure (versus informational disclosures  Feedback  □ 0 □ 1 □ 2 □ N/A G.  Reframed injurious feedback (interrupting, if necessary)  □ 0 □ 1 □ 2 □ N/A H.  Restated corrective feedback by member  □ 0 □ 1 □ 2 □ N/A I.  Used consensus to reinforce m-m or m-l  feedback that is present in the group  □ 0 □ 1 □ 2 □ N/A J.  Attended to balance between positive and corrective member-member feedback  □ 0 □ 1 □ 2 □ N/A K.  Encouraged positive member-member feedback  □ 0 □ 1 □ 2 □ N/A L.  Therapist helped members apply in-group member-member feedback to out-of-group situations  Emotional Climate Leader contribution  □ 0 □ 1 □ 2 □ N/A M.  Helped members recognize why they feel a certain way (identifying underlying concerns or motives)  □ 0 □ 1 □ 2 □ N/A N.  Encouraged active emotional engagement between group members  □ 0 □ 1 □ 2 □ N/A O.  Fostered a climate of supportive challenge  Member contribution  □ 0 □ 1 □ 2 □ N/A P.  Assisted members in describing their emotions (requires feeling word (sad, angry, joyful)  □ 0 □ 1 □ 2 □ N/A Q.  Elicited verbal expressions that connect group members to one another  □ 0 □ 1 □ 2 □ N/A R.  Refrained from conveying personal feelings of hostility and anger in response to negative member behavior     174  Appendix D  : Design for primary study (chapter 4)  Experimental Period      Follow-up Period   Class: Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Week 7 Week 8 Week 9 Week 10 Week 11 Week 12 Week 13Group BuildingWithdrawn MemberInterpers FeedbackLeader challengeParticipant choiceParticipant choiceGroup BuildingWithdrawn MemberInterpers FeedbackLeader challengeParticipant choiceParticipant choiceGCSE, STAI PreSTAI: Session 1STAI: Session 3GCSE, STAI MidSTAI: Session 5GCSE, STAI PostSTAI: Session 1STAI: Session 3GCSE, STAI Follow-upSTAI: Session 5GCSE, STAI FinalBehaviour rating(GPIRS)Behaviour rating(GPIRS)GCSE, STAI PreGCSE, STAI MidGCSE, STAI PostGCSE, STAI Follow-upGCSE, STAI FinalFollow-up PeriodTreatment GroupControl GroupBehaviour rating(GPIRS)Behaviour rating(GPIRS)Experimental PeriodHyp1,3Hyp1,3Hyp 5&6Hyp1,3Hyp 5&6Hyp 2, 4

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166291/manifest

Comment

Related Items