Development and validation of the Group Supervision Impact Scale

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content DEVELOPMENT AND VALIDATION OF THE GROUP SUPERVISION IMPACT
Copyright 2003
SCALE
by
Michele A. Getzelman
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment o f the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
EDUCATION (COUNSELING PSYCHOLOGY)
August 2003
Michele A. Getzelman
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3133275
Copyright 2003 by
Getzelman, Michele A.
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3133275
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This dissertation, written by
____________ Michele A. Getzelman_________
under the direction o f her dissertation committee, and
approved by all its members, has been presented to and
accepted by the Director o f Graduate and Professional
Programs, in partial fulfillment o f the requirements for the
degree of
TOR OF PHILOSOPHY
Director
Date December 17. 2001
Dissert
I
Chair
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dedication
To Bill Wilson and Doctor Bob Smith
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements
This project could not have been completed without the support and encouragement
so freely given by those that believed in me long before the first words were ever written.
My friends and family made it possible for me to keep at it, “one day at a time.” I
couldn’t have done it without you!
To my committee members, Dr. Rodney Goodyear, Dr. Dick Clark, and Dr. Bob
Turrill - Thank you for your insights and for taking the time to help me with this project.
To my advisor, Dr. Goodyear, thank you for your guidance, attention to my work, and
encouragement throughout my years at USC. To Dr. Clark, thank you for your
investment in teaching, sharing your ideas, and constant willingness to help. To Dr.
Turrill, thank you for your warm enthusiasm, contagious passion to make a difference,
and for your countless contributions to my learning and development.
To Peggy Albrecht, whose business card reads “An Investment In Human Dignity.”
Words will never be enough to thank you for investing in me. Your love and constant
support have enabled me to persevere through so many o f life’s obstacles. The
completion o f this project is one o f many endeavors I have had the courage to take on as
a result of your faith in me, and support o f my education. To John O’Connor, thank you
for setting such a fine example, believing in me, and supporting my education.
To Frank S. Leach, your love and friendship have carried me through - Thank you for
being there for me and helping me to remember the important things in life. Your
courage and strength have rubbed off on me at all the right times. I never would have
made it without your humor, laughter, encouragement and support.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To Valerie Gibson, for always believing that I could handle more than I thought I
could, and telling me so. Your friendship has lifted me up more times than I can count.
You are amazing and I am so lucky to have you in my life. To Janet “Little Feather” my
friend, thank you for your constant support and friendship. Dr. Edward Kohler, thank
you for showing me the way, making me laugh, and for sharing your brilliance. To Dr.
Jayson Mystkowski, what can I say? You are one o f a kind and I count myself blessed to
have you as a friend.
To Dr. Julia K. Getzelman, Richard A. Getzelman, and Kelly M. Getzelman - among
other things...you have dealt with my absence at special occasions due to school
commitments and listened to countless hours o f my ramblings. Thank you for being
proud of me, showing up for me, and for supporting me. I am so proud of each o f you
and your unique accomplishments, and grateful that you have shared in mine.
To Mom, Dad, Babu, Damian... I miss you!
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Dedication---------- 1
Acknowledgement---------------------------------------------------------------------------------------ii
List o f Tables-------------------- ■ ----------------------------------------------------- 'v iii
Abstract---------------------------------------------------------------------------------------------------- is
Chapter I: Introduction and Review o f the Literature----------------------------------------------1
Introduction---------------- 1
Overall Purpose---------------------------------------------------------------------------------------6
Literature Review-------------------------------------------- ■ --------------------------- ■ ---------------- 7
Defining Group Supervision--------------------------------------------------- 7
Why Investigate Group Supervision?----------------------------------------------------------- 10
Benefits o f Group Supervision------------------------------------------------------------------- 1 1
Disadvantages o f Group Supervision------------------------------------------------------------1 4
Helpful & Hindering Phenomena In Group Supervision ---------------------------------- 17
Helpful Phenomena ------------------------------- 1 8
Hindering Phenomena ------------------------------------------------------------------ 19
Assessment o f Supervision Outcome --------------------------------------------------------- 24
Predictors o f Group Supervision Outcome ---------------------------------------------------27
Trainee Satisfaction in Supervision-------------------------- 27
Supervisory Working Alliance -------------------------------------------------------- 28
Group Cohesion---------------------------------------------------------------------------30
Supervisee Self-Efficacy ----------------------------------------------------------32
Supervisor Leadership Roles------------------------------ —-----------35
Discriminant Validity: Theoretical Orientation--------------------------------------------- 37
Summary and Research Questions------------------------------------------------------------- 38
Chapter II: Methodology------------------------------------------------------------------------------- 40
Overview------------------------------------ 40
Phase Two ----------------------------------------------------------------------------------------------- 40
Item Development--------------------------------------------------------------------------------- 40
Expert Raters---------------------------------------------------------------------------------------- 4 1
Trainee Raters-------------------------------------------------------------------------------------- 42
Scale Item Revisions------------------------------------------------------------------------------- 42
Phase Three ---------------------------------------------------------------------------------------------43
Participants------------------------------------------------------------------------------------------ 43
Participants’ Supervisors--------------------------------------------------------------------------46
Measures ■ -------- 47
Group Supervision Impact Scale-------------------------------------------------------47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Supervisee Satisfaction Questionnaire --------------------------------------------- 47
Supervisory Working Alliance Inventory ------------------------------------------ 48
Group Climate Questionnaire----------------------------------------------------------- 48
Group Supervision Scale --------------------------------------------------------------- 49
Supervisee Self-Efficacy Scale ------------------------------------------------------- 50
Demographic Questionnaire ----------------------------------------------------------- 50
Procedures —----------* -----------------------------------------------------------------------------------5 1
Chapter III: Results-------------------------------------------------------------------------------------5 3
Overview — --------- > ---------------------------------------------------------- 5 3
Factor Analyses ----------------------------------------------------------------------------------------- 54
Adequacy o f Extraction------------------------------------------- 55
Retention o f Factors Criteria--------------------------------------------------------------------- 5 5
Adequacy o f Rotation----------------------------------------------------------------------------- 56
Orthogonal and Oblique Rotation o f Factors-------------------------------------------------57
Interpretation o f Item Loadings on Factors---------------------------- 58
Group Supervision Impact Scale Item Factor Structure------------------------------------ 6 1
Descriptive Statistics ----------------------------------------------------------------------------------62
Reliability------------------------------------------------------------------------------------------- 62
Preliminary Analyses------------------------ 65
Convergent Validity ------------------------------------------------------------------------------ 65
Multicollinearity and Singularity Analyses --------------------------------------------------67
Construct Validity--------------------------------------------------------------------------------------68
Group Supervision Impact Scale and Satisiaction ----------------------------------------- 68
Group Supervision Impact Scale and Supervisory Working Alliance -----------------69
Group Supervision Impact Scale and Supervisor R oles----------------------------------- 70
Group Supervision Impact Scale and Group Cohesion------------------------------------ 7 1
Group Supervision Impact Scale and Supervisee Self-Efficacy ------------------------ 72
Discriminant Validity -------------------------------------------------------------------------- 73
Summary o f Results --------------------------- 7 5
Factor Analyses ---------------------------------------------------------------------------------- 75
Reliability--------------------------■ -----------------------------------------------------------------75
Construct Validity--------------------------------------------------------------------------------- 75
Discriminant Validity----------------------------------------------------------------------------- 76
Chapter IV: Discussion------------------------------------------------------------------------ 77
Overview-------------------------------------------------------------------------------------------------- 77
Psychometric Properties o f the Group Supervision Impact Scale------------------------------77
Description o f Group Supervision Impact Scales ------------- 78
Group Supervisor Impact Scale -----------------------------------------------------------------79
Group Environment Impact Scale----------------------------------- 80
Peer Supervisee Impact Scale -------------------------------------------------------------------8 1
Convergent Validity-------------------------------------------------------------------------------------8 1
Overview------------------------- 8 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vii
Group Supervision Impact Scale and Satisfaction------------------------------------------85
Group Supervision Impact Scale and Working Alliance— -------------------------------85
Group Supervision Impact Scale and Group Cohesion-------------------- 86
Group Supervision Impact Scale and Supervisor Roles------------------------------------ 87
Group Supervision Impact Scale and Self-Efficacy---------------------------------------- 88
Limitations------------------------- 8 8
Implications for Theory, Research, and Practice-------------------------------------------------- 9 4 -
Theoretical Implications--------------------------------------------------------------------------- 95
Research Implications--------------------- 97
Practical Implications------------------------------- 99
Conclusion-----------------------------------------------------------------------------------------------100
References---------------------------------------------------------------------- 1 0 1
Appendices---------------------------------------------------------------------------------------------- 11 3
Appendix A: Supervisee Self-Efficacy Scale------------------ 113
Appendix B: Group Supervision Impact Scale---------------------------------------------- 114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 1------------------------------------------------------------------------------------------------------44
Ethnicity and Gender o f Participants and Their Supervisors.
Table 2------------------------------------------------------------------------------------------------------5 7
Component Correlation Matrix.
Table 3------------------------------------------------------------------------------------------------------59
Group Supervision Impact Scale Item Factor Loadings, Item-Scale Correlations,
Eigenvalues, and Variance Explained.
Table 4----------------------------------------------------------------------------------------------------- ©
Inter-Item Correlations for Group Supervisor Impact Scale.
Table 5----------------------------------------------------------------------------------------------------- 64
Inter-Item Correlations for Group Environment Impact Scale.
Table 6----------------------------------------------------------------------------------------------------- 64
Inter-Item Correlations for Peer Supervisee Impact Scale.
Table 7----------------------------------------------------------------------------------------------------- 6 7
Intercorrelations, Means, and Standard Deviations for Group Supervision Impact Scale
Total and Three Factors, Supervisee Satisfaction Questionnaire, Supervisee Self-Efficacy
Scale, Group Supervision Scale, Group Climate Questionnaire, and Supervisee Working
Alliance Inventory.
Table 8---------------------------------------------------------------------------------- 7 4
Theoretical Orientation: Means and Standard Deviations for Participants’ and
Supervisors’ Theoretical Orientation on Group Supervision Impact Scale Total and
Factor Scores.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The purpose of this study was to develop the Group Supervision Impact Scale (GSIS).
The GSIS measures the positive and negative impact o f three main aspects o f Group
Supervision: the Supervisor, the Group Environment, and the Peer Supervisee.
Participants (n=222) consisted o f a national sample o f psychology graduate students
mainly from clinical and counseling psychology programs who had experience in group
supervision. Factor analysis supported three factors: The Group Supervisor Impact Scale
(e.g, provision o f structure and group focus, constructive utilization o f group dynamics,
effective feedback, openness, humor, flexibility, competency, and validation), The Group
Environment Impact Scale (e.g., unresolved between-member conflicts, competition,
support, attraction to group, anxiety), and the Peer Supervisee Impact Scale (e.g.,
feedback, vicarious learning, generating ideas, treatment planning). Validity o f the GSIS
is supported by theoretically consistent relationships between the three scales and (a)
supervisee satisfaction, (b) stronger working alliance, (c) group cohesion, (d) supervisor
leadership roles, and (e) supervisee self-efficacy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Chapter I
INTRODUCTION
Psychotherapy training is complex and multifaceted. Supervision, a fundamental
component o f that training, facilitates the development o f a trainee’s clinical skills
and professional identity (Ducheny, Alletzhauser, CrandelL, & Schneider, 1997;
Nelson & Friedlander, 2001). As Holloway and Neufeldt (1995) noted, supervision
helps to maintain “the standards o f the profession” (p. 207) and, in turn, the efficacy
o f treatment.
At some point in the process o f graduate education in psychology an individual
almost certainly will have experience with group supervision. Most university
training programs and internships utilize a group supervision format (e.g., Bernard &
Goodyear, 1998; Riva & Cornish, 1995). Though it is clear that supervision is an
essential aspect o f training, the level o f empirical research does not appear to match
the extent to which this training modality is used (Gray, Ladany, Walker, & Ancis,
2001; Prieto, 1996; Worthen & McNeill, 1996). Numerous authors (e.g.,
Counselman & Gumpert, 1993; Holloway & Johnston, 1985; Savickas, Marquart, &
Supinski, 1986) have noted the limited amount o f studies that “empirically describe
group supervision practices” (Savickas et al., 1986; p. 17) and maintain that “it is
imperative that authors and researchers o f supervision begin a systematic examination
o f the group supervision process” (Holloway & Johnston, 1985; p. 338).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
In what was likely the first substantial review o f group supervision literature,
Holloway and Johnston (1985) emphasized the surprising lack o f research, especially
in contrast to the prevalence o f its use. More recently, Riva and Cornish (1995), in a
national survey o f psychology predoctoral internship programs found that a majority
(65%) o f the sites investigated utilized group supervision in their training programs.
Echoing other researchers (e.g., Holloway & Johnston, 1985; Prieto, 1996; Wilber,
Roberts-Wilber, Hart, Morris, & Betz, 1994) Riva and Cornish (1995) state that
“although widely practiced, group supervision is not well understood” (p. 523).
Prieto (1996), in a later review o f the group supervision literature further emphasized
the need for additional research to better understand the group supervision process
and asserts that “the continuing lack o f empirical research investigating the group
supervision o f trainees learning individual psychotherapy is indefensible” (p. 305). It
appears that a gap in the supervision literature still exists nearly 20 years after
Holloway and Johnston’s expressed concern that without more in depth examination
“group supervision will remain a weak link in our training programs, widely practiced
and poorly justified” (p. 339).
Authors (e.g., Holloway & Johnston, 1985; Prieto, 1996; Wilbur et al., 1994) have
been encouraging further research to support group supervision as a training format.
As previously mentioned, the widespread use o f group supervision and its link with
trainee development and client outcomes (Holloway & Neufeldt, 1995; Riva &
Cornish, 1995; Stoltenberg, McNeill, & Crethar, 1994) provides a solid impetus for
further investigation o f group supervision. Research is warranted in order to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
demonstrate actual empirical evidence for the plethora o f “professed benefits”
provided by group supervision (Wilber, Roberts-Wilber, Hart, Morris, & Betz, 1994;
p. 263) as well as events that can potentially thwart group process, trainee
development, and ultimately client outcomes (Patton & Kivlighan, 1997).
Counselman and Gumbert (1993), in an investigation o f the use of groups in
supervising individual and couples psychotherapy, noted that there are “various
advantages and some potential disadvantages o f group supervision”(p. 32). Several
other researchers (e.g., Gray, Ladanay, Walker, & Ancis, 2001; Nelson &
Friedlander, 2001; Patton & Kivlighan, 1997; Savickas et al., 1986; Worthen &
McNeill, 1996) also have discussed the presence o f both helpful and
counterproductive or hindering events that occur in supervision. Factors cited in the
discussions o f the benefits o f group supervision range from cost effectiveness to
vicarious learning (Bernard & Goodyear, 1998).
Researchers (e.g., Gray et al., 2001; Nelson & Friedlander, 2001; Veach, 2001)
also have documented counterproductive events, such as supervisors who are
perceived as being critical (Allen, Szollos, & Williams, 1986) or personality conflicts
between group members (Counselman & Gumpert, 1993). Yet research to identify
specific helpful and hindering events that occur during group supervision has been
minimal. Veach (2001) noted that much o f the focus o f events in supervision has
been with individual supervision, and states that several questions remain regarding
the potentially unique factors that might be identified in taking a closer look at group
supervision.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
Several studies (Ellis, 2001: Gray et al., 2001; Nelson & Friedlander, 2001;
Nelson, Gray, Friedlander, Ladany, & Walker, 2001; Ramos-Sanchez, Esnil,
Goodwin, Riggs, Touster, Wright, Ratanasiripong, & Rodolfa, 2002) have discussed
“supervisor behaviors and supervision components that contribute to trainee’s reports
o f satisfaction and dissatisfaction” (Gray et al., 2001, p. 371) but the focus has been
primarily on individual supervision. Veach (2001) notes that one study (Magnuson,
Wilcoxon, and Norem, 2000) did included trainees participating in both individual
and group supervision in their examination o f poor supervisory behaviors. Though
Magnuson et al. (2000) investigated supervisory behaviors experienced by trainees in
both individual and group supervision the aim o f the study was not specifically group
supervision processes and the sample size was limited (11 counselors).
In a recent study, however, Goodyear, Enyedy, Arcinue, Puri, Carter, and
Getzelman (in press) investigated group supervision in order to develop a
classification ofhelpful phenomena that occur during group supervision. In a
complementary study Enyedy, Arcinue, Puri, Carter, Goodyear, and Getzelman
(2003) examined hindering phenomena occurring during group supervision. These
two studies are the first to classify empirically the types of helpfiil and hindering
phenomena that occur in the process o f group supervision. Both studies found
sources o f group supervision impact (either helpful or hindering) included the
supervisor, peers in the group, as well as the overall safety and group climate.
Evaluation o f these group supervision events such as the impact o f the supervisor and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
peers in the group, as well as the emotional or affective tone o f the group would
contribute to filling the gap in the group supervision literature.
Researchers and clinicians alike could utilize an instrument that evaluates group
supervision processes related to the impact o f the supervisor, interactions o f group
members, and the overall safety and group climate. Such a measure could be used to
investigate: (1) individual difference variables in group supervision, (2)
developmental theories (e.g., developmental level o f trainee or developmental stage
o f the group), (3) group functioning and its impact upon supervisees, and (4)
differential impacts o f various group formats (e.g., experiential, focus on case
presentations, process focus, or focus on countertransference). Such a measure also
could help supervisors develop and tailor specific interventions in group supervision.
There are no doubt a plethora o f additional research and applied endeavors in which a
measure o f group supervision impact might be employed.
As this brief literature review has indicated, research to identify and clarify group
supervision processes is indeed at a beginning stage. There is a need to develop valid
means o f measuring these events that occur during group supervision. In creating a
more precise and in depth understanding o f the group supervision process, the need
for supervision specific instruments is brought to the surface. Bernard and Goodyear
(1998) state that “almost all instruments that have been used to assess supervision
process and outcome variables are ones that were originally developed for other
purposes” (p. 262). Following, they discuss two main problems with adapting
instruments for use in supervision research: (1) potential unknown changes to the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
psychometric properties o f the instrument, and (2) the use o f these instruments “can
perpetuate the use o f roles and metaphors from interventions other than supervision
(e.g., therapy or teaching)”(p.262). Prieto (1996) also noted that researchers should
avoid “ the practice o f using psychotherapy-based instruments in group supervision”
and highlighted the fact that “investigators need to develop and use supervision-
specific assessment instruments in future research” (pp. 304-305).
Ellis and Ladany (1997) stated that “one o f the most pernicious problems
confronting supervision researchers is the dearth o f psychometrically sound measures
specific to a clinical supervision context” (p.488). As previously had been noted by
Bernard and Goodyear (1998) and others (e.g., Ellis and Ladany, 1997; Lambert and
Ogles, 1997), Stoltenberg, McNeill, and Crethar (1994) conclude that “measures
intended for use in counseling may not tap important constructs present in the
supervision context” (p.426). Thus, it is clear that the development o f supervision
instruments is indeed a necessary step in helping to bridge the gap in our
understanding o f group supervision processes.
Purpose o f the Study
In response to the near-absence o f supervision-specific instruments, and the need
for research that investigates “group supervision of trainees learning individual
psychotherapy” (Prieto, 1996, p. 295), the purpose o f this study was to develop and
provide validation evidence for the Group Supervision Impact Scale (GSIS). The
GSIS is a self-report group supervision evaluation scale measuring the following
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
essential components o f group supervision: (1) supervisor impact, (2) group impact,
and (3) supervisee impact.
Initial items for the scales were generated in two preliminary studies (i.e., Enyedy
et al., 2003; Goodyear et a l, in press) in which both helpful and hindering phenomena
during group supervision were identified. Items from several o f the cluster-
analytically derived categories from these studies were used as the basis for the scale
described in this study. As well, this study describes the results o f factor analyses and
evidence o f reliability to demonstrate statistical support for the three scales o f the
GSIS (i.e., supervisor impact, group impact, and peer impact). Additional evidence
o f GSIS validity is described in terms o f relationships between its scales and (a)
trainee satisfaction with supervision, (b) the supervisory working alliance, (c) group
cohesion, (d) supervisor leadership roles, and (e) self-efficacy ratings. Discriminant
validity o f the GSIS will be determined by examining the relationships between the
scales o f the GSIS and (a) participants’ theoretical orientation, and (b) supervisor’s
theoretical orientation.
Literature Review
Defining Group Supervision
Several authors have offered definitions o f group supervision (Altfeld & Bernard,
1997; Bernard & Goodyear, 1998; Goodyear & Nelson, 1997; Holloway & Johnston,
1985; Prieto, 1996). These range from broad to more detailed descriptions, and
describe several types o f group supervision. Holloway and Johnston (1985) note that
“group supervision has referred to the supervision o f counselors in training in a group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
format,” but that “within this broad area...there are several specific applications o f
group supervision”(p.333). The applications Holloway and Johnston (1985)
described include:
(a) group supervision o f trainees in a practicum setting who are learning
individual counseling skills.
(b) group supervision o f trainees learning pre-practicum interviewing skills.
(c) leaderless groups in which trainees provide peer supervision in a group
format.
(d) group supervision o f trainees in a practicum setting focusing on learning
group factilitation skills (p. 333).
Though the particular application, background o f the group members (e.g., social
work, psychiatry, psychology trainee), and goals o f the group affect the definition,
Holloway and Johnston (1985) defined group supervision as “supervision in which
supervisors oversee a trainee’s professional development in a group of peers”(p. 333).
Bernard and Goodyear (1998) provide the following more detailed definition of
group supervision:
group supervision is the regular meeting o f a group o f supervisees with a
designated supervisor, for the purpose o f furthering their understanding o f
themselves as clinicians, o f the clients with whom they work, and/or of service
delivery in general, and who are aided in this endeavor by their interaction with
each other in the context o f group process (p. 111).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
Wilbur et al. (1994) noted as well that “supervising trainees in groups also
provides opportunities for peer review, peer feedback, and personal insight into
interpersonal behaviors relevant to a counselor’s professional role” (p. 262).
Essentially, group supervision consists o f three or more supervisees working together
and facilitated by a supervisor for the purpose o f managing client care and developing
clinical expertise, as well as potential heightened self awareness. The processes
within the group as well as the content o f the clinical material presented are
productive areas o f focus for group supervision.
One difficulty in drawing conclusions ffom the group supervision literature is that
many types o f supervision groups are often lumped together under the heading of
group supervision. There are common elements in group supervision processes that
apply to groups o f novice trainees learning interviewing skills, group supervision o f
those learning group therapy skills, peer group supervision, group supervision o f
supervision, and group supervision o f trainees providing individual psychotherapy.
However, researchers (Altfeld, 1999; Holloway and Johnston, 1985; Prieto, 1996)
have noted that more didactic groups, those focusing on supervision o f group therapy,
and peer supervision groups differ enough that they should be examined separately.
In their reviews o f the group supervision literature both Prieto (1996) and Holloway
and Johnston (1985) offer evidence in support o f the “conceptual separation of
supervision group types” (Prieto, 1996, p.296).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
In both o f these reviews the grounding assumption was that the recipients o f
group supervision were primarily predoctoral trainees learning individual
psychotherapy. That assumption guided this study as well
Why Investigate Group Supervision?
Bernard and Goodyear (1998) state that “virtually all university training programs
employ group supervision at one point or another during their clinical courses” (p.
111). Yet, as previously mentioned, research to inform its practice remains scant.
This is an important ommission, for as Counselman & Gumbert (1993) noted “group
supervision is not individual supervision in the presence o f a few other people”(p.
30). Moreover, Ellis and Ladany (1997) concluded from their extensive integrated
review o f the supervision literature “ that the supervision process may be more
complex than previously postulated” (p. 463). This makes the near absence o f group
supervision research even more problematic.
There are a plethora o f areas to investigate that would fiirther our understanding
o f group supervision and contribute to both research and applied interests. So one
might ask, what specifically do we already know and what do we need to know about
group supervision? Kruger et al. (1988 b) believed that due to the “paucity of
research on group supervision, several o f the supposed advantages o f group
supervision have not been investigated” (p. 331), Holloway and Johnston (1985),
Riva and Cornish (1995), Prieto (1996) and others (e.g. Newman & Lovell, 1993;
Wilbur et al., 1994) have noted the potentially unique advantages o f supervision in a
group and highlight the need for empirical evidence o f helpful events in group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
supervision. Bernard and Goodyear (1998) extensively discuss the potential
advantages and limitations o f supervising in groups. Here I will summarize their
observations, though for a more in depth discussion I will refer you to their chapter
(Le., Bernard & Goodyear, 1998).
Whereas research on group supervision has been scant, a number o f authors have
offered their observations about benefits and disadvantages o f group supervision.
These will be summarized in the next two subsections.
Benefits and Disadvantages o f Group Supervision
Benefits. Bernard and Goodyear (1998) discuss the following 10 advantages of
group supervision:
1. Economies o f time, money, and expertise.
2. Minimized supervisee dependence.
3. Opportunities for vicarious learning.
4. Supervisee exposure to a broader range o f clients.
5. Feedback for the supervisee: greater quantity and diversity.
6. Feedback for the supervisee: greater quality.
7. A more comprehensive picture o f the supervisee.
8. Facilitated risk taking.
9. Greater opportunity to use action techniques.
10. Mirroring the supervisees’ intervention (specific to the supervision o f group
therapistis) (pp. 112-114).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
Many others (e.g., Counselman & Gumbert, 1993; Holloway & Johnston, 1985;
Kruger, Chemiss, Maher, & Leichtman, 1988; Newman & Lovell, 1993; Prieto, 1996;
Riva & Cornish, 1995; Sansbury, 1982; Wilbur, Roberts-Wilbur, Hart, Morris, &
Betz, 1994) have discussed these and other advantages o f group supervision.
Holloway and Johnston (1985), for example, state that group supervision is “an
economical use o f supervisory time” as well as providing “an opportunity for peer
review, peer feedback, and personal insight into interpersonal behavior relevant to the
professional role” (p. 332). Holloway and Johnston (1985) also noted that group
supervision can provide “a supportive atmosphere o f peers in which fledgling
counselors can realize their own feelings o f inadequacy as a common effect in the
training process” (p. 337).
Prieto (1996) added that in addition to the aforementioned benefits, supervision in
a group helps to normalize trainee anxiety surrounding clinical work, and can provide
peer support during the training process. Riva and Cornish (1995), Wilbur et al.
(1994), Kruger et al. (1988), and Newman and Lovell (1993) all highlighted these
advantages. In addition, Holloway and Johnston (1985) note that group supervision
has an “intuitive appeal” due to the opportunity to combine “the discussion o f case
material, group dynamics, and didactic material”(p. 337).
Hillerbrand (1989), in a discussion o f cognitive differences between experts and
novices and the implications for group supervision, noted some distinct advantages
for novice counselors. He observed that social interaction is a key factor in the
development o f cognitive skills. Group supervision provides modeling, feedback,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
and a focus on self-monitoring, and the interaction o f several novices (i.e., trainee)
with an expert (i.e., supervisor) provides a unique benefit due to the trainee’s
verbalization o f cognitive processes (Hillerbrand, 1989). The advantages o f group
supervision for novices include explicit cognitive processes, greater opportunity for
cognitive rehearsal, increased motivation to learn, and increased perceptions o f self-
efficacy (Hillerbrand, p. 295).
Hillerbrand (1989) contended that novices are more likely to utilize language that
is readily understandable by other novices; in addition, novices may key into the
nonverbal cues o f other novices more effortlessly. Supervision in a group with
trainees at various levels may produce increased motivation as a result o f watching
counselors at a similar training level and being able to model and provide feedback to
those less skilled (Hillerbrand). Finally, Hillerbrand noted that in observing a peer
review a specific skill, the novice is more likely to gain proficiency and experience an
increase in self-efficacy than if the trainee had merely observed an expert demonstrate
the skill. The application o f cognitive psychology constructs to group supervision
demonstrates the ways in which the benefits o f this modality are especially useful for
novice counselors.
Counselman and Gumbert (1993) discussed therapist-related benefits o f group
supervision that are more specific to therapists that are already established in the
professional community. In addition to such previously noted benefits as cost
effectiveness, parallel process, and vicarious learning, Counselman and Gumbert
(1993) asserted that group supervision provides these more experienced professionals
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
with regular collegial contact and “a place not only to continue to learn, but also to
let off steam, share one’s anxieties and insecurities, laugh, cry, and generally be with
other professionals who care about each other and share concerns about the work” (p.
27). Counselman and Gumbert contended that group supervision is a beneficial way
for therapists to help reduce burnout and isolation. Essentially, they conclude that
group supervision continues to offer benefits long after official graduate level training
has been completed.
It is clear, therefore, that there are widely acknowledged advantages o f group
supervision. Therefore, it would be unwise to refrain from the use o f group
supervision until there is more empirical evidence for the processes and outcomes.
Given the benefits suggested, further investigation o f these advantages and those
helpful events within group supervision that potentially produce them provide an
important research agenda.
Disadvantages o f group supervision. Bernard and Goodyear (1998) also discuss
several disadvantages o f group supervision:
1. The group format may not allow individuals to get what they need due to
(a) supervisees with heavy caseloads requiring more time, (b) heterogeneous
groups with respect to skill level may leave the more skilled members feeling
that they are not getting what they need, (c) the learning available to each group
member may be too diffuse to be worthwhile, and (d) an overpowering group
member might rob others o f their instructional needs, or the structure itself
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
might fit the majority o f the members, but offer virtually nothing to a distinct
minority o f the members.
2. Confidentiality concerns: in group supervision there is heightened concern
regarding (a) the clients who are the focus o f clinical attention, and (b) the
supervisees in the group.
3. The group format is not isomorphic with that o f individual counseling.
The group might be less likely to mirror some o f the individual processes that
occur in that counseling format.
4. Certain group phenomena can impede learning: Between-member
competition and scapegoating are among the phenomena that, if unchecked, can
interfere with learning and in some cases even result in deleterious effects to
one or more o f the supervisees.
5. The group may focus too much time on issues not o f particular relevance
to or interest for the other group members, (pp. 114-115).
Bernard and Goodyear noted as well that supervisees with heavy caseloads may
take more time in the group or require extra supervision, and Counselman and
Gumbert concur that a potential limitation is that some members might not have the
opportunity to present on a weekly basis. Bernard and Goodyear discuss group
phenomena that might impede learning. Altfeld (1999) and others (e.g., Counselman
& Gumbert, 1993; Ellis & Douce, 1994; Savickas et al., 1986) state that “too often
group members’ critical responses to the presenting therapist’s work elicit reactions
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
o f shame, embarrassment, and anxiety that do not provide an atmosphere for optimal
learning” (p. 238).
Counselman and Gumbert also noted that it is not only the potential between-
member problems or personality conflicts that are problematic, but it is the time as
well that is taken to resolve these issues in the group. Though one of the main
professed benefits o f group supervision has been the cost-effectiveness and time
efficiency, the group process may also create time obstacles.
Bernard and Goodyear discussed increased confidentiality concerns due to group
supervision. Other aspects o f working in a group also can become problematic. For
example, Altfeld (1999) states that it can become overwhelming for a supervisee to
receive a large amount of feedback from the group members. Counselman and
Gumbert discuss the process o f receiving feedback and having “more players” in the
mix than in individual supervision, which can become confusing for the supervisee
(p. 29). Counselman and Gumbert also note that the integration o f new members into
an already established and “cozy” group can become problematic (p. 29). This
process is similar to circumstances in group therapy with the addition o f new
members and can be handled in a similar fashion.
Finally, differences in terms o f theoretical orientation might become problematic.
Altfeld (1999) states that “a moderately experienced clinician can formulate dynamics
reasonably quickly...” which makes “the occasion ripe for disagreement as members
advance their own closely held viewpoints” (p. 238). Thus, conflicts regarding
theoretical orientation and treatment plans can become problematic in a group.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
Essentially, we are aware o f what some o f the advantages and limitations o f group
supervision are, but we lack an adequate means o f measuring these elements. A
group supervision evaluation scale would provide a useful tool in furthering our
understanding o f group supervision phenomena by clarifying and providing empirical
evidence o f the professed benefits and limitations. In particular, a scale measuring
the impact o f three main forces (e.g., supervisors, peers, and group climate) in group
supervision would enable researchers to clarify salient events for trainees. In
addition, through additional research utilizing a group supervision impact scale we
can begin to develop a greater understanding o f the implications that these events and
processes might have upon trainees’ development and clinical outcomes.
Helpful and Hindering Phenomena in Group Supervision
The benefits and disadvantages reviewed in the previous section are from the
supervisors’ perspective. At least as important is the supervisee’s perspective about
what is helpful or hindering to their learning. Helpful phenomena that occur during
group supervision are those that lead to desirable or positive consequences. For our
purposes “helpful phenomena” were defined as supervisees perceptions o f
phenomena that they discerned to be helpful. Hindering phenomena occurring during
group supervision are those that lead to negative or undesirable consequences. For
our purposes “hindering phenomena” were defined as those phenomena during group
supervision that supervisees believed had a negative effect on their functioning.
Only two studies (i.e., Enyedy et al., 2003; Goodyear et a l, in press) have
investigated helpful and hindering phenomena that occur specifically during group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
supervision. Though various authors have made observations about the advantages
and potential negative aspects o f group supervision, these have not been validated
empirically. Moreover, these have been offered from the perspective o f the
supervisor. Supervisees might have very different perceptions about what is helpful
or not. The Enyedy et al. (2003) and Goodyear et al. (in press) studies were designed
to assess those perceptions.
Helpful phenomena. Goodyear et a l (in press) identified specific helpful group
supervision incidents which they found could be mapped on two dimensions
(supervisor versus supervisee impact; and, acquiring self versus objective
knowledge). They also found the following six helpful phenomena categories: (1)
supervisor impact, (2) specific instruction, (3) self-understanding, (4) validation of
experience, (5) support and safety, and (6) peer impact. This study provides an
empirically derived classification o f helpful events in group supervision.
Others have investigated helpful, or “good,” supervision as it occurs in a one-on-
one context. Worthen and McNeill (1996), for example, identified general factors
indicative o f positive supervision experiences and several themes regarding the
characteristics that contribute to good supervision events. Many o f these support both
the findings o f Goodyear et a l (in press) and previously cited conceptual discussions
o f the advantages o f supervision. For example, Worthen and McNeill found that the
central factor in good supervision experiences involved the quality o f the supervisory
relationship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
Supervisee satisfaction is another important measure o f supervisees’ perceptions.
Factors that have been identified as enhancing satisfaction include: supervisor self
disclosure (Stoltenberg, 1981), providing feedback regarding both strengths and
weaknesses (Savickas et al., 1986), providing support, encouragement, and challenge
(Bernard & Goodyear, 1998; Gray et al., 2001), and providing competent instruction
(Nelson & Friedlander, 2001). Worthen and McNeill (1996) found that good
supervision was associated with supervisee perceptions o f encouragement to explore
and experiment, validation, a non-judgmental stance, openness, and an attitude of
empathy.
Goodyear et al. (in press) zoomed in even closer by identifying specific events in
group supervision that contribute to these good supervision experiences. For
example, following are some o f the specific helpful events that they identified:
collegial feedback, openness o f supervisor, specific feedback from supervisor on
audiotapes, conducing a case presentation, the opportunity to explore
difficulties/differences, and being able to ask peers about what did and did not work.
The specific events that have been identified not only provide support for the benefits
of group supervision, but they create opportunities for application in terms of
additional research and use in supervision groups to enhance positive experiences and
outcomes.
Hindering phenomena. It is just as important to know what supervisees perceive
as hindering their learning. Chicca-Enyedy et al. (2003) identified five types of
hindering phenomena: between-member problems; supervisor problems; supervisee
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
anxiety and other negative affects; logistical constraints; and poor group time
management. In terms o f between-member problems they identified two subclusters
consisting of negative supervisee behaviors (e.g., complaints regarding the behavior
o f peers in the group), and personal reactions to negative behaviors (e.g., the
consequence o f the negative behavior). These findings are consistent with several o f
the potential limitations or disadvantages previously discussed (e.g., competition in
the group, conflicts between supervisees, one member dominating the group,
difficulty with critical statements).
In the Enyedy et al. (2003) study, the largest cluster o f hindering phenomena
concerned supervisor problems. It consisted o f supervisor behaviors or
characteristics that were perceived as hindering learning. Some examples include:
personal style o f supervisor (e.g., supervisor taking group time to talk about
knowledge or experience that was not related to or helpful with cases, supervisor
going on tangents, or supervisor being rigid), perceived lack o f experience and
reduced clinical focus (e.g., supervisor had insufficient expertise, supervisor lacked a
theoretical focus, and supervisors confusion regarding confidentiality issues) and
miscellaneous problems with the supervisor (e.g., supervision being too laid back,
and the supervisor highlighting differences rather than discussing similarities).
Again, in terms o f previous literature, these findings provide empirical evidence for
the negative and hindering events in group supervision that have only been discussed
conceptually in the past.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
Enyedy et al. (2003) also identified negative emotional experiences o f the
supervisee such as fear, anxiety, and alienation that supervisees experience as
hindering events. Following are a few o f the specific experiences identified: being
the only (male, African-American, Latina, etc.), fear o f negative evaluation from
supervisor, feeling pressured to self-disclose, and wanting more positive feedback.
They also found that supervisees experienced time limitations (e.g., not having
enough time to discuss cases or ask questions) and logistical constraints (e.g., room
size, scheduling time o f supervision) as hindering events.
It is instructive to examine hindering or negative supervisory phenomena that
occur in a group context. Worthen and McNeill (1996) reported what they considered
a “surprising aspect” (p. 331) which was the experience of “less than fulfilling”
supervisory relationships and disappointment in supervision even though these were
not phenomena they had targeted. Only four other studies (Gray, Ladany, Walker, &
Ancis, 2001; Magnuson, Wilcoxon, & Norem, 2000; Nelson & Friedlander, 2001;
Ramos-Sanchez, Esnil, Goodwin, Riggs, Touster, Wright, Ratanasiripong, & Rodolfa,
2002) were identified in which negative or counterproductive events in supervision
were investigated (though only Magnuson et al. included participants experiences in
group as well as individual supervision).
Gray, Ladany, Walker, and Ancis (2001) investigated counterproductive events in
individual supervision. They identified the following four categories:
1. supervisor dismissed trainee’ thoughts and feelings or was indifferent
(e.g., conflict regarding case conceptualization, inappropriate supervisor self-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
disclosure, supervisor was not prepared for supervision, lack o f validation and
support).
2. supervisor denied trainee request (e.g., denied trainee’s request for more
supportive feedback, supervisor denied trainee’s request to shut the door during
supervision due to concerns about confidentiality).
3. supervisor misunderstood trainee (e.g., supervisor misunderstood trainee’s
desire to discuss negative transference instead suggesting that the trainee
change fields).
4. supervisor directed trainee to be different with clients (e.g., to change
theoretical approach without an explanation).
Gray et al. found that a negative interaction between supervisor and trainee was
likely to occur following the initial counterproductive event. They identified the
following negative interactions:
1. trainee trying to be agreeable/trying not to be defensive.
2. supervisor not listening/not responding to trainee.
3. supervisor disputed/challenged trainee.
4. supervisor pushed own agenda.
5. supervision work stilted, (p. 375).
In addition to identifying specific hindering events, Gray et al. found that these
counterproductive interactions weakened the supervisory relationship, changed the
manner in which supervisees approached their supervisors, negatively affected the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
trainees work with clients, and that most did not disclose the negative experience with
their supervisors.
In a qualitative study o f conflictual supervisory relationships Nelson and
Friedlander (2001) identified several negative events and interactions such as feeling
a lack o f support, power struggles, disagreements about what should take place in
supervision, theoretical disagreement, and differing world views (p. 388), In
discussing the findings o f Gray et al. (2001) and Nelson and Friedlander (2001), Ellis
(2001) noted that in both studies the following aspects had been identified as negative
supervision events:
1. lack o f supervisor empathy and support
2. the supervisee felt unsafe and withdrew from the relationship
3. the supervisee developed self-doubts, blamed him- or herself and
experienced lessened self-efficacy as a professional
4. supervisors were unaware o f the problem
5. the conflict was never disclosed to the supervisor (for about half o f the
supervisees)
6. the supervisees directly addressed the issue with the supervisor (for about
half o f the supervisees, half o f them were able to resolve the issue with the
supervisor)
7. the conflict was never resolved
8. the conflict fostered supervisee strength and professional development (p.
403).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
These salient findings are also consistent with data reported by Ramos-Sanchez et
al. (2002) and Magnuson et al. (2000) who reported negative supervision experiences
categorized by difficulties regarding interpersonal relationship and style, conflicts
regarding tasks in supervision and theoretical orientation, a lack o f validation and
support, and intolerance regarding differences. Just a handful o f studies have
investigated hindering events in supervision, but the common themes and descriptions
o f negative consequences for the supervisees demonstrate the importance o f
continuing to investigate and measure these experiences.
Assessment o f Supervision Outcome
So far this review has considered supervision processes. The ultimate goal,
however, is to maximize positive outcomes. Yet the criteria for supervision outcome
and the measurement o f them has not always been easy. Several researchers (e.g.,
Crethar, 1994; Goodyear & Guzzardo, 2000; Lambert & Ogles, 1997; McNeill, &
Efstation, Patton, & Kardash, 1990; Stoltenberg, Patton & Kivlighan, 1997) have
discussed the efficacy o f supervision, particularly the difficulty o f measuring its
impact on the supervisee and, ultimately, clients (Holloway & Neufeldt, 1995). At
first glance the most obvious test of supervision outcome appears to be client change
or treatment effectiveness. In fact, in an extensive review addressing the outcome o f
supervision as related to supervisees and clients, Ellis and Ladany (1997) noted that
“the impact o f clinical supervision on client outcome is considered by many to be the
acid test o f the efficacy o f supervision”(p. 485). Yet it has proven to be an elusive
goal. Holloway and Neufeldt (1995) highlight the fact that though “client outcome is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
a critical measure o f treatment efficacy within the context of psychotherapy.. .it is an
unfeasible standard as the only test o f competence, because o f the numerous
difficulties in controlling client and other external factors” (pp. 207-208).
Goodyear and Guzzardo (2000) and Ellis and Ladany (1997) noted the difficulty
o f drawing definitive conclusions regarding the effect of supervision on client
outcome. This point is based upon the limited number o f studies as well as what Ellis
and Ladany (1997) characterized as the poor methodological rigor o f extant studies.
Goodyear and Guzzardo stated that the process o f conducting efficacy studies in
supervision has been complicated by the nature o f supervision models, the lack o f
manuals or specific supervision protocols, and the inability to assign a supervisee to a
no-treatment control group (e.g., a novice clinician treating clients unsupervised).
Holloway and Neufeldt (1995) added that the difficulty o f investigating the
“effectiveness o f supervision in relation to client outcome is even further complicated
by the number o f intermediary variables between supervisor interventions and client
change”(p. 208). As a consequence, supervision outcome is often measured by
various supervisee characteristics, perspectives, and behaviors.
Holloway and Neufeldt (1995) identified the following sources of supervision
outcome:
1. Trainee’s acquisition in supervision o f attitudes, beliefs, and skills that
have been identified as relevant to treatment efficacy.
2. Trainee’s performance in the therapist role as related to supervision
factors.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
3. Interactional process events in therapy as related to supervision process
events.
4. Client change as related to supervision, (p. 208).
Within this framework more specific supervision outcome variables often include the
following: supervisee attitudes, specific trainee behaviors (Goodyear & Guzzardo,
2000); supervisee satisfaction, supervisees’ perceptions o f the interpersonal climate of
group or individual supervision meetings, supervisees’ perceptions of their
development o f counseling knowledge, supervisees’ perceptions o f their development
o f counseling skill (Kruger, Chemiss, Maher, & Leichtman, 1988); the acquisition o f
specific techniques, therapeutic style, development o f therapeutic strategies, and
development o f theoretical orientation (Guest & Beutler, 1988).
Essentially, though direct links between supervision and client outcomes have not
yet been established, supervision outcome has been assessed in terms o f the
acquisition o f specific skills and behaviors, as well as trainee attitudes and
perspectives. Goodyear and Guzzardo (2000) noted that though there is not an
established empirical link between supervision and client outcome, it is possible to
infer that therapeutic outcome is influenced by supervision (p. 87). For example, they
point out that we do know that the working alliance in supervision can impact the
therapeutic working alliance (Patton & Kivlighan, 1997), and the therapeutic working
alliance is associated with client outcomes. Thus, it appears that client outcomes are
indirectly affected by supervision. Further clarification o f supervision processes and
outcomes will provide additional evidence o f this link.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Predictors o f Group Supervision Outcome
A number o f factors predict group supervision outcomes. In the sections that
follow, some o f the more important o f these are reviewed. In each case, the particular
factor was used in validating the GSIS.
Trainee satisfaction in group supervision. Supervisee satisfaction is often
employed as an outcome variable. As previously discussed, researchers (e.g.,
Bernard & Goodyear, 1998; Enyedy et a l, 2003; Goodyear et al., in press; Gray et al.,
2001; Nelson & Friedlander, 2001; Savickas et a l, 1986; Stoltenberg, 1981; Worthen
& McNeill, 1996) have identified specific characteristics o f supervisors and
supervision experiences associated with satisfaction in supervision (e.g., amount and
type o f feedback, validation, competent instruction, level o f focus). Ellis and Ladany
(1997) highlight the fact “that aspects or components o f the supervisory relationship
are related to supervisee outcome (e.g., supervisee skill and satisfaction)”(p. 462). In
addition, supervision in a group creates the potential for supervisee satisfaction to be
affected by both the supervisor as well as other trainees in the group (e.g., between-
member competition, support and validation).
Ramos-Sanchez et a l (2002) found that participants who reported having
experienced negative supervision events demonstrated significantly lower levels of
satisfaction than did participants who did not report those experiences. Ramos-
Sachez et al. and others have identified trainee satisfaction as a valid outcome
variable and research has linked both helpful and hindering events with high and low
levels o f satisfaction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
A research question o f this investigation involved demonstrating evidence of
construct validity for the GSIS based upon significant correlations with perceptions o f
satisfaction. There seems to be a general consensus that both supervisors and
supervisees influence outcomes o f group supervision. Thus, evidence o f validity for
the group supervisor impact and peer supervisee impact scales will be evident in
terms o f a positive relationship with trainee perceptions o f satisfaction. The
Supervisee Satisfaction Questionnaire (SSQ) (Ladany, Hill, Corbett, & Nutt, 1996)
was used to assess this variable.
Supervisory working alliance. Numerous authors (e.g., Bernard & Goodyear,
1988; Ellis & Ladany, 1998; Goodyear & Guzzardo, 2000; Gray et al., 2001;
Lehrman-Waterman & Ladany, 2001; Nelson & Friedlander, 2001) have documented
the importance o f the supervisory relationship. Many indicate that supervisees
identified it as “the most critical element in supervision”(Ramos-Sanchez et al.,
2002)(p.l97). In fact, in their extensive review o f the supervision literature, Ellis and
Ladany (1998) found that “the quality o f the supervisory relationship is paramount to
successful supervision” (p. 496). Gray et al. (2001) noted that the “supervisory
relationship appears to be a salient component o f trainees’ experience o f good or
positive supervision” as well as focus o f intense negative emotion in the event of
counterproductive supervision events (p. 372). The supervisory relationship often has
been measured and investigated in terms o f the supervisory working alliance
(Efstation, Patton, & Kardash, 1990; Goodyear & Guzzardo, 2000). According to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Bordin (1983) the supervisory working alliance consists o f the following three
factors:
1. mutual agreements and understandings regarding the goals sought in the
change process
2. the tasks o f each o f the partners (i.e., supervisor and supervisee)
3. the bonds between the partners necessary to sustain the enterprise (p. 35).
In discussing the emotional bond between the supervisor and supervisee, Bordin
stated that the bond resembles something in between that o f “teacher to class
members and therapist to patient” (p. 37). He further stated that this applies to group
supervision as well, and that the “individualized performance character o f the process
turns us toward the bonds between a player and coach” (p. 37). As such, all factors o f
the working alliance might be positively or negatively affected by specific events or
disagreements generating from interactions between the supervisor, supervisee, and
other trainees in the group supervision setting.
Researchers have demonstrated a relationship between the supervisory working
alliance and both positive and negative outcomes for the supervisee. In addition,
research has shown that specific supervision events affect the strength o f the working
alliance. For example, Gray et aL (2001) and Ramos-Sanchez et a l (2002) found that
counterproductive events in supervision negatively influenced the supervisory
alliance and the overall experience o f the supervisee. Others (e.g., Worthen &
McNeill, 1996) have noted the positive relationship between good or helpful
supervision events and a strong supervisory working alliance. In addition, Patton and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
Kivlighan (1997) found a relationship between “the strength o f the working alliance
in both supervision and counseling,” (p. 111), further highlighting the importance o f
this aspect o f supervision
A research question o f this investigation involved demonstrating evidence of
construct validity for the GSIS via perceptions o f the supervisory working alliance.
We believe that results will demonstrate a significant positive relationship between
the Group Supervisor Impact Scale and perceptions o f the supervisory working
alliance. More specifically, strong ratings o f the supervisory working alliance will be
related to high scores on the Group Supervisor Impact Scale indicating that positive
supervisor impact is a predictor o f a strong working alliance. The Rapport Scale from
the trainee version o f the Supervisory Working Alliance Inventory (SWAI-T)
(Efstation, Patton, & Kardash, 1990) was used to assess this variable.
Group cohesion. Research regarding group cohesion has spanned several decades
(e.g., Dion, 2000; Evans & Dion, 1991; Festinger, Schachter, & Back, 1950) and
many consider this construct to be “one o f the most important properties, if not the
most important property, o f groups” (Dion, 2000, p. 7). Yalom (1995) employed the
following sweeping definition o f cohesiveness:
Cohesiveness is broadly defined as the resultant of all the forces acting on all
the members to remain in the group, or, more simply, the attractiveness o f a
group for its members. It refers to the condition o f members feeling warmth
and comfort in the group, feeling they belong, valuing the group and feeling, in
turn, that they are valued and unconditionally accepted and supported by other
members (p. 48).
Dion (2000), in a review o f psychological research on cohesion, noted that the
terms “ cohesion” or “cohesiveness” have been used to describe the processes that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
keep members o f a small group “together and united to varying degrees” (p. 7). In
addition, Dion added that the “groups’ net affective tone provides an additional or
more global affective element to the cohesion construct” (p. 7). More specifically, in
terms o f cohesiveness in group supervision the “net affective tone” involves empathy,
support, trust, attraction to and engagement in the group (Bernard & Goodyear, 1998;
Dion, 2000; Kruger et al., 1988).
Bernard and Goodyear (1998) pointed out that the development o f group cohesion
and trust allows group members to experience the peer support in the group and
ultimately to become more willing to expose their mistakes and weaknesses to the
group. This process o f increased vulnerability that is mediated by higher levels o f
cohesiveness often has the effect o f group members placing a greater value on the
group (Bernard & Goodyear, 1998). As previously discussed (see benefits and
disadvantages o f group supervision), several components o f the group such as
between-member competition and conflict have the potential to disrupt the group
process. It is likely that higher levels o f support, safety, and cohesiveness enable the
group to effectively manage competition or conflicts when they do arise. For
example, Kruger et al. (1988) found that those supervision groups evidencing the
highest interpersonal climate ratings (e.g. greater cohesiveness) also demonstrated
more frequent discussions o f “counselor problems” (p. 615). Though they reported
this relationship, they also noted that it was not possible to determine whether the
discussion o f difficulties improved the interpersonal climate, or whether the high
levels o f cohesiveness facilitated the discussion. Though the causal direction was not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
clear, the relationship was established. Thus, cohesion within groups appears to
facilitate a more open discussion among group members.
The cohesiveness o f a group enables group members to utilize the group
environment more effectively. Factors such as safety, support, and open discussion
are key to successful group functioning and have been found to represent supervisee
perspectives o f helpful events in group supervision and positive outcomes (Goodyear
et al., in press). This relationship between cohesiveness and effectively dealing with
negative group events appears to represent an environment in which it is safe to
discuss difficulties and receive mutual support.
A research question o f this investigation involved demonstrating evidence o f
construct validity for the GSIS by providing evidence o f a positive relationship
between cohesiveness and the safety/group climate scale. More specifically, we
believed that more cohesive groups would produce stronger perceptions o f group
safety and a supportive group climate. We used the Group Climate Questionnaire-
Short Form (GCQ-S) (Mackenzie, 1981) to assess this variable.
Self-efficacv. Based on Bandura’s theory (1977), self-efficacy is known as the
relationship between an individual’s perceived ability to perform a task within a
specific area, and the actual performance o f that task. Bernard and Goodyear (1998)
describe self-efficacy in supervision as “a person’s certainty that he or she can
execute a particular behavior successfully” (p. 253). Self-efficacy beliefs have been
linked with the behavioral choices an individual engages in, the effort expended on a
given task, tenacity and perseverance in the midst o f impediments, and behavioral
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
outcomes (Bandura 1982; Bernard & Goodyear, 1998; Heppner, Multon, Gysbers,
Ellis, & Zook, 1998). Essentially, the attributions people make regarding their
perceived ability to perform a task mediates their ability to effectively achieve
carrying out that task (i.e, skill acquisition, engaging in the group supervision process,
learning and executing counseling interventions). Investigations o f self-efficacy have
been an important aspect o f research in counseling, training, and skill acquisition for
quite some time (e.g., Bandura, 1982,1993; Bruning, Schraw, & Ronning, 1999;
Ladany, Ellis, & Friedlander, 1999; Wester & Vogel, 2002). Self-efficacy, given its
function in assisting individuals in the engagement o f learning and the willingness to
persist in a specific domain, is relevant to consider in terms o f evaluating supervision
and outcome.
According to Bandura’s (1977) model, four main forces affect self-efficacy
beliefs: (1) performing a given behavior, (2) vicarious learning experiences in which
an individual can witness another perform the specific behavior, (3) receiving
persuasive feedback that one can actually perform the behavior, and (4) arousal and
physiological responses (e.g., anxiety). Ladany et aL (1999) observe that these main
sources o f self-efficacy expectations are the very factors that should be present in
supervision. In group supervision, for example, a supervisee experiences vicarious
learning, receives feedback from the group supervisor and peer supervisees regarding
their ability to perform the required tasks, and potentially experiences support and/or
other emotional arousal (e.g., anxiety) in the course o f supervision. Thus, group
supervision provides an appropriate environment to either help or hinder self-efficacy
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
beliefs. To the extent that an individual’s group supervision experience provides
opportunities for positive feedback (verbal persuasion), vicarious learning,
performance o f skills, and support (versus anxiety), that supervision experience
would be linked with stronger self-efficacy beliefs (Ladany et al., 1999).
Research (Efstation et al., 1990; Ladany et al., 1999; Lehrman-Waterman &
Ladany, 2001; Williams et a l, 1997) has linked self-efficacy with satisfaction in
supervision and the supervisory working alliance. Overall, greater self-efficacy is
related to the experience o f positive or helpful supervision elements. Efstation et aL
(1990) linked self-efficacy with a strong supervisory working alliance, whereas
Ladany et al., (1999) did not find the working alliance to predict changes in trainee
self-efficacy. Ladany et al. suggested that a lack o f precise measurement in terms of
study variables and the use o f different assessment scales might have contributed to
their varied findings. Leherman-Waterman and Ladany (2001) demonstrated that
effective evaluation practices in supervision were associated with greater trainee self-
efficacy, indicating the contribution o f the supervisor in promoting self-efficacy.
Just as verbal persuasion, vicarious learning, and support can enhance self-
efficacy, negative supervision experiences can interfere with trainee’s judgements
about their ability to perform. Low self-efficacy can hinder the process o f developing
skills during supervision and training as a result o f a lesser degree o f willingness to
take risks, avoidance, or not persisting with tasks (Larson & Daniels, 1998; Wester &
VogeL, 2002). Thus, stronger self-efficacy beliefs in group supervision are more
likely to be present when the conditions are helpful versus hindering.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
A research question o f this investigation involved demonstrating evidence o f
construct validity for the GSIS by providing evidence o f a positive relationship
between self-efficacy and both the group supervisor impact and peer supervisee
impact scales. More specifically, we believed that higher scores on the Group
Supervisor Impact and Peer Supervisee Impact scales would be related to stronger
supervisee self-efficacy ratings. We predicted that self-efficacy ratings would be
correlated with both the supervisor and fellow supervisees given the overlap in
contributions to group supervision (e.g., providing feedback, modeling, support,
vicarious learning via group members or role playing with supervisor). The
Supervisee Self-Efficacy Scale (SSE) was developed for this study to assess this
variable.
Supervisor leadership roles. Arcinue (2002) recently developed and provided
preliminary reliability and validity evidence for a scale designed to measure
supervisees’ perceptions o f the group supervisor. The Group Supervision Scale
(GSS; Arcinue, 2002) consists o f the following three scales: (1) Group Safety
(representing supervisor’s attention to group safety), (2) Skill Development and Case
Conceptualization (assessing supervisor’s role in developing counseling intervention
skills and case conceptualization), and (3) Group Management (reflecting
supervisor’s attention to time management, group structure, and focus). Arcinue
demonstrated that group supervisor leadership roles, as measured by the GSS, are
linked with group cohesion, satisfaction with supervision, and client focus in the
supervisory working alliance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
Prietro (1996) and others (e.g., Holloway & Johnston, 1985; Wilber, Roberts-
Wilber, Hart, Morris, & Betz, 1994) have recommend against the use o f measures
developed for other purposes (e.g. psychotherapy evaluation) as dependent measures
in supervision research. Given the lack o f supervision specific measures, much o f the
literature has employed measures originally designed for evaluation o f psychotherapy
outcome or process issues. Thus, though the GSS is a new measure with initial
reliability and validity data, its particular strength is that it was developed specifically
for the group supervision domain. Following, the GSS assesses the role o f the group
supervisor and would provide a comparison measure for the Group Supervisor Impact
scale o f the Group Supervision Impact Scale.
A research question o f this investigat ion involved demonstrating evidence of
construct validity for the GSIS by providing evidence o f a positive relationship
between group supervisor leadership roles (as measured by the GSS) and the Group
Supervisor Impact Scale (GSIS). More specifically, we believed that high scores on
the Group Supervisor Impact scale (indicating helpful supervisor impact) would be
related to high scores on the GSS. Whereas the GSS (Arcinue, 2002) assesses just the
role o f the group supervisor in group supervision, the Group Supervision Impact
Scale extends Arcinue’s research by assessing the three main forces o f influence with
group supervision (i.e., the supervisor, the group climate, and the group members).
We predict that the GSS will be related to both the Group Supervisor Impact
scale, as well as the Peer Supervisee Impact scale, given that there is overlap in terms
o f the behaviors engaged in by both the supervisor and group members. For example,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
both the supervisor and fellow supervisees provide feedback, generate ideas regarding
treatment planning, and contribute to the overall group atmosphere. Though the GSS
is expected to be related to both the Group Supervisor Impact and Peer Supervisee
Impact scales, given that the GSS was developed to assess the specific role o f the
group supervisor we expect that the Group Supervisor Impact scale will be the best
predictor o f supervisor roles (as measured on the GSS).
Discriminant Validity
In the previous section, the case was made for variables that would help establish
concurrent validity. But it also is important that a scale not correlate significantly
with variables with which is has no logical relationship. This is discriminant validity.
Altfeld (1999) suggested that differences in terms o f theoretical orientation
between the supervisor and supervisees, or between peer supervisees, might become
problematic rather than providing an opportunity for group members to consider
differing viewpoints. An area o f discussion has been the potential conflicts that might
arise when group members assert strongly held theoretical beliefs and treatment
considerations that differ from other group members or the supervisor. Given the
potential contribution o f theoretical disputes to the overall impact in group
supervision, the theoretical orientations o f the group supervisor and participants were
examined to assess the discriminant validity o f the Group Supervision Impact Scale.
Though it is likely that there will be some similarities in terms o f theoretical
orientation, the potential relationship between supervisor and supervisee’s theoretical
orientation or potential conflicts regarding differences should not account for the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
variance on participants’ GSIS scale scores. In fact, in an extensive review o f the
supervision literature, Ellis and Ladany (1997) observed that supervisees were able to
identify their most and least effective supervisors regardless o f their theoretical
orientation or that o f their supervisors. Thus, any potential similarities or differences
in regard to theoretical orientation, though a potential source o f conflict or discussion,
should not be significantly related to participants’ perceptions o f the overall helpful or
hindering nature o f their group supervision experience.
Summary and Research Questions
This dissertation reports an instrument development study. The scale was
developed to measure three essential components o f group supervision: (1) supervisor
impact, (2) peer impact, and (3) the overall safety and group climate. We have the
following research questions related to providing evidence o f reliability and validity
for the Group Supervision Impact Scale.
1. Will the three scales intended to measure (1) supervisor impact, (2) group
member impact, and (3) group safety/climate be supported by factor analysis?
2. Will the scales demonstrate satisfactory internal consistency reliability?
3. Will the scales demonstrate satisfactory evidence o f construct validity?
Specifically, will construct validity will be evidenced by statistically significant
relationships between:
(a) the Group Supervisor Impact Scale and the Supervisee Satisfaction
Questionnaire (SSQ), the rapport scale o f the Supervisory Working Alliance
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
Inventory (SWAI), the Group Supervision Scale (GSS), and the Supervisee Self-
Efficacy Scale (SSE);
(b) the Peer Impact Scale and the Supervisee Satisfaction Questionnaire
(SSQ), and the Supervisee Self-Efficacy Scale (SSE);
(c) and, the Group Environment Impact Scale and the short form o f the Group
Climate Questionnaire (GCQ-S).
4. Will the scales on the GSIS demonstrate discriminant validity? Discriminant
validity will be evidenced via relationships between the scales o f the GSIS and: (a)
supervisor’s theoretical orientation, and (c) participant’s theoretical orientation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
Chapter II
METHOD
The development and validation o f the Group Supervision Impact Scale
(GSIS) encompassed three phases. The initial phase involved two preliminary studies
in which helpful and hindering phenomena in group supervision were empirically
identified. These results were utilized to generate initial items for the scale. The
second phase involved reviewing the initial pool o f items and selecting those clusters
that were generally complementary in each study and both relevant and appropriate
for the scale. Expert raters and graduate level psychology students then reviewed the
scale items to identify any obvious omissions and improve the clarity o f wording.
The third phase entailed administering the scale to a national sample o f graduate
students to demonstrate evidence o f reliability and validity.
This chapter describes each o f the steps. Phase one is described in the articles by
Enyedy et al. (2003) and Goodyear et al. (in press). Therefore, only phases 2 and 3
are described here.
Phase Two
Item development. The initial pool o f items were generated in two preliminary
studies o f helpful and hindering group supervision phenomena o f which Getzelman
and Goodyear were co-authors (i.e., Enyedy et al., 2003; Goodyear et al, in press).
The initial pool o f items were then reviewed by Getzelman and Goodyear for
inclusion in the scale based upon (a) previous findings regarding group supervision
events, (b) suitability for the scale, (c) to ensure a thorough representation o f group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
supervision events, and (d) likelihood o f occurring in any supervision group. The
clusters o f items that had a complementary version in both preliminary studies were
selected and all items from the other clusters were discarded. That is, both studies
found evidence for the impact (positive or negative) o f the supervisor, o f peers, and o f
group climate.
The items in each o f the retained clusters were then reviewed. Duplicate items
were discarded as were those that were clearly inappropriate for the scale. Items were
further reviewed, rewording some items for clarity and to ensure parallel form. At
this stage there were 24 items: 11 for the Group Supervisor Impact Scale, 6 for the
group environment impact scale, and 7 for the peer supervisee impact scale
Expert raters. Five expert raters (1 Asian American female, 1 Asian American
male, 2 White females, and 1 White male) were recruited to review the scale and
provide evidence o f content validity. These raters were considered to be experts
based upon their extensive experience providing group and individual supervision
(more than 10 years), their status as licensed psychologists, and by recommendations
from training directors. The raters were provided with a brief review o f the literature
detailing events in group supervision (a summary o f the literature review), a copy o f
the GSIS scale, and a feedback form. For each scale item, raters were asked to
provide feedback regarding: (a) the extent to which the item reflected events and/or
processes in group supervision, (b) whether there were any obvious omissions or
items that clearly did not fit, and (c) the clarity and ease o f instructions and overall
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
format. Expert raters also were asked to review the supervisee self-efficacy scale and
provide feedback regarding the content validity.
Trainee raters. Five advanced (completed 4 or more years o f graduate school)
graduate students in clinical and counseling psychology programs were recruited to
review the scale items and provide additional evidence o f content validity. These
raters were selected based upon their experience as a supervisee in group supervision
(at least 2 different supervision groups each), completion o f at least 4 semesters of
practicum/field placement, and completion o f graduate coursework. These raters
consisted o f 3 clinical psychology graduate students (1 Asian-American female, 2
White females, and 1 White male) and 1 counseling psychology graduate student
(White female). The student raters were provided with a brief review o f the literature
detailing events in group supervision (a summary o f the literature review), a copy of
the GSIS scale, and a feedback form.
For each scale item, raters were asked to provide feedback regarding: (a) the
extent to which the item reflected events and/or processes in group supervision, (b)
whether there were any obvious omissions or items that clearly did not fit, and (c) the
clarity and ease o f instructions and overall format. Trainee raters also were asked to
review the supervisee self-efficacy scale and provide feedback regarding the content
validity.
Scale item revision. Feedback from the expert and graduate student raters was
incorporated. Items were modified for clarity and additions were made following
rater recommendations. The final scale consisted o f 27 items: (a) supervisor impact
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
(14 items), (b) peer impact (4 items), (c) group safety/climate (9 items). Three items
were added to the Group Supervisor Impact Scale to reflect the supervisor’s use o f
group dynamics, management o f group time, and maintenance o f group focus. Three
items were added to the Group Environment Impact Scale to reflect group members’
attraction to the group, whether or not group members liked one another, and whether
or not the group was a safe place to ask questions. Three items from the Peer
Supervisee Impact Scale that dealt with supervisees role in management o f group
time, maintaining focus within the group, and group structure were removed. Trainee
raters believed that these items reflected the role o f the group supervisor greater than
that o f peer supervisees. The supervisee self-efficacy scale (SSE) was also reviewed
and revised with the final version consisting o f 4 items. Only clarity of wording
changes were made to the SSE.
Phase Three
Participants. Participants consisted o f a national sample o f 222 trainees from
psychology graduate programs who indicated that they were either currently involved
in group supervision or had been within the past six months. The participants were
84.3% female and 15.7% male, ranging in age from 22 to 58 (M-29.35, SD=6.16).
Participants were .9% (n=2) Native American/Alaskan Native, 5.8% (n=T3)
Asian/Pacific Islander, 2.7% (n=6) Black/African American, 3.6% (n=8)
Hispanic/Latino, 1.3% (n=3) International Student, 4.5% (n=T0) Multi-Ethnic/Racial,
78.9% (n=176) White/Non-Hispanic, and 2.2% (n=5) indicating “other” ethnicity.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
The demographic composition o f participants in this study is representative of
demographic characteristics o f national samples o f students enrolled in accredited
graduate psychology programs (Norcross, Sayette, Mayne, Karg, & Turkson, 1998).
Norcross, Hanych, and Terranova (1996) noted that entering graduate students in
1992 consisted o f 67% to 70% women and 17% ethnic minority backgrounds.
Norcross et al. (1998) reported that two thirds o f entering students were women and
one fifth were minorities. In addition, Bernal, Sirolli, Weisser, Ruiz, Chamberlain,
and Knight (1999), based upon raw data from APA’s Office o f Program Consultation
and Accreditation, reported that 17.4% o f students in 1994-1995 were o f ethnic
minority background. As shown in Table 1, participants in this sample were 18%
minority and 84% female (slightly above the reported range o f 67 to 70%). Thus, the
demographic characteristics o f this sample are fairly representative o f students
enrolled in graduate programs in psychology.
Table 1
Ethnicity and Gender o f Participants and Their Supervisor’s_____________________
Participants Supervisors
________________________ % _________ n______________________ % ________ n
Ethnicity:
Native American/
Alaskan Native
.9 2 .4
Asian/
Pacific Islander
5.8 13 4
Black/African
American
2.7 6 4.9
Hispanic/
Latino
3.6 8 2.7
Multi-Ethnic/
Racial
4.5 10 .9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Table 1 (continued)
Ethnicity and Gender o f Participants and Their Supervisor’s
Participants Supervisors
% _________ n______________________ % ________ n
White/ Non-
Hispanic
78.9 176 84.3 188
International 1.3 3
— —
Other
Gender:
2.2 5 1.8 4
Male 15.7 35 49.3 110
Female 84.3 188 50.2 112
Note: Participants were 78.9% WMte/Non-Hispanic and 18.8% ethnic minority
background; Participants’ Supervisors were 84.3% White/Non-Hispanic and 12.9%
ethnic minority background.
Participants were enrolled in APA accredited graduate programs and pursuing
Ph.D. (90.1%), Psy.D. (4.9%), or terminal MA (4%) degrees enrolled in clinical
(61%), counseling (34.5%), or combined professional-scientific (4%) programs.
Participants were currently (or within the past six months) working at a field
practicum/field placement (85.2%) or pre-doctoral internship (14.8%) which utilized
a group supervision format. The type o f agency participants reported having received
group supervision was mainly in university-based counseling centers (56.1%), with
15.7% in community agencies, 9.9% in hospital/inpatient settings, and 18.4% other.
Participants’ reported theoretical orientations were 42% eclectic, 17.9% cognitive,
10.8% psychodynamic/neo-Freudian, 9.4% behavioral therapy/learning, 4%
humanistic, 4% systems, 2.7% existential, 2.7% Sullivanian/interpersonal, 1.3%
Rogerian/person-centered, .4% gestalt, .4% Adlerian, .4% Jungian, and 2.7% other.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
The smallest number o f respondents (10.3%) were in their first year o f graduate
school, 25.6% were 2n d year students, 17% were 3r d year students, 18.4% were 4t h
year students, and 27.4% had completed more than four years. The average number
o f practicum hours accrued up to the date o f participation was 863.75 (SD=934) with
a range o f 20 to 6000. The average number o f completed semesters of
practicum/field placement was 5.25 (SD=3.26) ranging from 0 to 20. The size o f the
group in which respondents were currently involved or last participated ranged from 2
to 15 members with an average o f 4.79 (SD=2.13) members. The participants’
supervision groups that were the basis o f their responses had been meeting an average
o f27.25 weeks (SD=21.73). Participants represented 52 different psychology
training programs from 32 different states.
Participants’ supervisors. Participants reported that their supervisors were 50.2%
female and 49.3% male (.4% missing data), with an average age o f 43.79 (SD=8.87)
ranging from 27 to 68. As shown in Table 1, supervisors were .4% (n=l) Native
American/Alaskan Native, 4% (n=9) Asian/Pacific Islander, 4.9% (n=Tl)
Black/African American, 2.7% (n=6) Hispanic/Latino, .9% (n=2) Multi-
Ethnic/Racial, 84.3% (n=188) White/Non-Hispanic, and 1.8% (n=4) indicating
“other” ethnicity. Participants reported that their supervisors theoretical orientations
were 20.6% eclectic, 16.6% psychodynamic/neo-Freudian, 15.7% behavioral
therapy/learning, 15.2% cognitive, 7.6% systems, 4.5% Sullivanian/interpersonal,
3.6% Rogerian/person-centered, 3.1% humanistic, 1.8% psychoanalytic/Freudian,
1.8% existential, 1.3% gestalt, .4% Adlerian, 5.6% other, and 1.8% not reported.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Supervisors’ approximate number o f years experience providing group supervision
averaged 10.57 years (SD=8.0) ranging from 0 to 35.
Measures
Group supervision impact scale (GSIS). The GSIS, the focus o f this study, was a
self-report measure developed by the investigators to assess the impact o f the
supervisor, peers, and the overall group climate in group supervision. The Group
Supervision Impact Scale consists o f 27 items rated on a 7-point Likert-type scale
(e.g., 1= never. 7 -alwavs). Participants were asked to indicate the extent to which the
items (see Appendix B for items and instructions) reflected their current or most
recent group supervision experience.
Supervisee satisfaction questionnaire fSSOL The SSQ (Ladany, Hill, Corbett, &
Nutt, 1996) is an eight-item self-report measure that was adapted from the Client
Satisfaction Questionnaire (Larson, Attkisson, Hargreaves, & Nguyen, 1979).
Ladany et al. modified the questionnaire by replacing the terms “counseling” and
“services” with the term “supervision.” Supervisees rate their satisfaction with
various aspects o f supervision on a 4-point scale ranging from low (1) to high (4).
Scores range from eight to 32, with higher scores indicating greater satisfaction.
Example items include “How would you rate the quality o f the supervision you have
received?” and “Did you get the kind o f supervision you wanted?” Ladany et al.
reported that the SSQ was related to supervisee non-disclosures that involved
negative reactions to the supervisor. Previous studies have found the internal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
consistency (alpha) o f the SSQ to be .96 (Ladany et al., 1996; Gray et a l, 2001). The
Cronbach alpha coefficient was .91 for this sample.
Rapport scale supervisory working alliance inventory-trainee form (SWAI-T).
The SWAI (Efstation, Patton, & Kardash, 1990) is a measure o f the relationship in
counselor supervision and consisting o f a supervisor version and a trainee version.
The trainee version is a 19-item scale in which participants indicate on a 7-point
Likert type scale ranging from (1) almost never to (7) almost always their agreement
with statements regarding the supervision working relationship. The scale is scored
based upon a mean rating o f the items with higher scores suggesting a stronger
working alliance. The trainee version measures two factors, (1) rapport (representing
the trainees perception of support from the supervisor), and (2) client focus (reflecting
the emphasis the supervisors placed on promoting the trainees’ understanding o f the
client)(Efstation et al., 1990).
The SWAI demonstrates adequate reliability and evidence o f convergent and
divergent validity were established by examining its relation to selected scales from
the Supervisory Styles Inventory (Friedlander & Ward, 1984), the Personal Reactions
Scale-Revised (Holloway & Wampold, 1983), and complementarity in supervision
relationships (Chen & Bernstein, 2000). For the purposes o f this study only the 12
items o f the rapport scale were used. The Rapport Scale’s Cronbach alpha coefficient
for this sample was .91.
Engagement scale group climate questionnaire-short form (GCO-S). The GCQ-S
(MacKenzie, 1981) is a 12-item instrument that measures group members’
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
perceptions o f the group atmosphere and it has been described as a session impact
measure (Brossart, Patton, & Wood, 1998). The items are scored on a 7-point Likert
type scale ranging from not at all to always/definitely. Factor analysis demonstrated
the following three scales: (1) Engagement (positive working environment), (2)
Avoidance (members avoiding personal responsibility for group work), and (3)
Conflict (a negative atmosphere characterized by anger and mistrust). One item is
used to measure anxiety.
Only the engagement scale was used in the current study. Evidence o f construct
validity has been demonstrated in studies finding a relationship between successful
groups and high scores on the engagement scale (MacKenzie, Dies, Coche, Rutan, &
Stone, 1987), and evidence that the engagement scale is related to cohesion and desire
to attend the group. Alpha coefficients have been reported as ranging from .80 to .94
(Brossart et al., 1998; Kivlighan & Goldfine, 1991). The Cronbach alpha coefficient
for this sample was .83.
Group supervision scale fGSS). The GSS (Arcinue, 2002) is a 16-item self-report
instrument developed to measure the supervisee’s perceptions o f the group
supervisor. The items are scored on a 6-point likert type scale ranging from strongly
disagree to strongly agree. Factor analysis provided evidence for the following three
scales: (1) Group Safety (representing supervisor’s attention to group safety), (2) Skill
Development and Case Conceptualization (assessing supervisor’s role in developing
counseling intervention skills and case conceptualization), and (3) Group
Management (reflecting supervisor’s attention to time management, group structure,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
and focus). Evidence o f construct validity has been demonstrated via relationships
with group cohesion, satisfaction with supervision, and client focus in the supervisory
working alliance (Arcinue, 2002). Arcinue (2002) reported alpha coefficients o f .96
for the full-scale GSS, .91 for the Group Safety scale, .94 for the Skill Development
and Case Conceptualization scale, and .93 for the Group Management scale. The
Cronbach alpha coefficient was .95 for the total GSS score for this sample.
Supervisee self-efficacv IS SET The SSE was developed by the researchers to
assess self-efficacy perceptions related to participants’ experience in group
supervision. The scale consists o f 4 items scored on a 7-point likert type scale
ranging from never to always. Participants are given the following instructions for
rating the items: “As compared to how you were prior to participating in this group,
to what extent are you insert items.” The four scale items are: (1) better at
understanding your client’s process issues, (2) better at conceptualizing client
problems, (3) better at understanding your own countertransference, and (4) more
confident in your role as a therapist (see Appendix A). Item-scale correlations for
this scale ranged from .64 to .76. The Cronbach alpha coefficient was .86 for this
sample.
Demographic information. A self-report questionnaire was designed to gather
information regarding participants’: (1) age, (2) gender, (3) ethnicity, (4) degree
program and type, (5) type o f clinical placement, (6) type o f agency, (7) year in
graduate program, (8) theoretical orientation, (9) approximate number o f completed
practicum hours, (10) number o f semesters o f completed practicum/field placement,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
(11) number supervisees in supervision group, and (12) number o f weeks the
supervision group has met. Participants were also asked to report their supervisors’:
(1) ethnicity, (2) approximate age, (3) gender, (4) theoretical orientation, and (5)
approximate number o f years experience providing group supervision. Participants
had the option o f reporting the name o f their university and/or the state in which they
attend graduate school.
Procedures
This study was conducted online by creating a website on which the study details,
IRB consent information, and the study questionnaire were posted. Training directors
from 192 psychology graduate training and internship programs were contacted by
email. They all were members o f one o f the following: the Association o f Counseling
Center Training Agencies [ACCTA]; the Council o f Counseling Psychology Training
Programs [CCPTP]; or the Council ofUniversity Directors o f Clinical Psychology
[CUDCP]. They were provided with (1) a description o f the study and its potential
contributions, (2) eligibility requirements for participation (currently receiving group
supervision or within the past six months), (3) IRB information, and (4) a hyperlink to
the study website. They then were asked to forward the email to their students for
participation. Participants’ responses were processed through a secure server and
were sent anonymously directly to the investigator.
Given this process, it is not possible to determine the number o f students that
were initially contacted for participation. However, participants were asked to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
indicate the name o f the university in which they were enrolled and/or the state in
which they were attending graduate school and based upon this information it appears
that the sample was quite diverse. In order provide confidentiality participants were
given the option o f whether or not to respond to the items regarding name and state o f
graduate school. Of the 222, 171 provided the name o f their university, identifying
52 different universities. Two hundred and nine responded to the item regarding the
state in which they attended graduate school, identifying 32 different states. Thus, it
appears that the study information was widely distributed.
Students choosing to participate were directed to the study website which
contained detailed information regarding the study, eligibility requirements (currently
involved in group supervision or involved within the past six months), IRB
information and informed consent, the study measures, and contact information for
the investigators. Participants were informed that completion o f the online survey
constituted consent to participate in this research study.
The online presentational style o f the study’s measures resembled that o f a
traditional paper and pencil measure. Participants were able to type responses in text
boxes for each item. Upon completion o f the survey, participants were able to submit
their responses by checking the “submit” box at which time the completed survey was
routed to a secure server and anonymously sent via email to the investigator.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
Chapter HI
RESULTS
This chapter describes the statistical procedures utilized in the analyses and the
obtained results. Included in this report are the results o f the factor analyses
conducted to determine the underlying fector model that best fit the Group
Supervision Impact Scale (GSIS) items. These results are followed by a description
o f the final scales (Supervisor Impact, Group Environment Impact, and Peer
Supervisee Impact) and corresponding scale items.
The factor analyses results are followed by descriptive statistics, GSIS reliability
results, and preliminary analyses describing the relationships between the Group
Supervision Impact Scale items and the measures employed for convergent validity
(i.e., Supervisee Satisfaction Questionnaire -SSQ, Supervisory Working Alliance
Inventory -SW AI, Supervisee Self-Efficacy Scale -SSE , Group Supervision Scale -
GSS, and the Group Climate Questionnaire short form -GCQ-S). Next, results
demonstrating the divergent validity o f the Group Supervision Impact Scale (GSIS)
are presented.
Finally, regression analyses that assessed for the effects o f the three GSIS scales
(Group Supervisor Impact, Group Environment Impact, Peer Supervisee Impact) on
the criterion variables (Le., SSQ, SWAI, SSE, GSS, and GCQ-S) are presented.
Regression analyses were conducted to demonstrate evidence o f construct validity
corresponding to the primary research questions. This section ends with a summary
of the results.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
Factor Analyses
Participants (n=222) ratings of the 27 GSIS items were subjected to principal
components factor analysis to determine the probable number and composite o f
factors (Tabachnick & Fidell, 1996). Researchers (e.g., Bryant & Yarnold, 1998;
Tabachnick & Fidell, 1996) have recommended principal components factor analysis
as a first step in exploratory factor analysis (EFA) in order to extract maximum
variance from the data set and bring to light the composition o f the factors.
Pedhazur and Schmelkin (1991) suggested that when employing principal
components factor analysis the factorability o f the data should first be determined and
the first two factors should account for a large percentage (approximately 50%) o f the
total variance. The following results demonstrate that these guidelines have been
met.
On the basis o f Bartlett’s test o f sphericity (Chi-square=3429.64, p<.000), the
correlation matrix is appropriate for factor analysis. However, Bartlett’s test of
sphericity is used as a general test for determining rejection o f the null hypothesis that
the matrix is an identity matrix and is often highly sensitive and dependent upon
sample size (Pedhazur & Schmelkin, 1991). Tabachnick and Fidell (1996)
recommend employing additional tests o f factorability especially in sample sizes with
more than five cases per variable. Therefore, given the subject to variable ratio in this
study (STV=8.2/1) the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy
was used with suggested values o f .6 and above indicating appropriateness for factor
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
analysis (Tabachnick & Fidell, 1996). The KMO was .90, demonstrating the
factorability for this data.
Adequacy o f extraction. To assess the adequacy o f extraction and the number o f
factors, two widely used criteria for determining the appropriate number o f
eigenvectors (factors) were employed. First, Kaiser’s stopping rule extracts only
eigenvectors with values o f at least 1 (Bryant & Yamold, 1998). Tabachnick and
Fidell (1996) suggest that this “criterion may either over- or underestimate the
number o f factors in the data set” (p. 672) and interpretability o f the factors should be
considered. A second criterion is Catell’s scree test which plots eigenvalues against
factors in descending order and the slope is examined for the point at which there is a
change (Tabachnick & Fidell, 1996). The factors retained are found along the initial
slope on which a line might be drawn through the points, those factors occurring after
a noticeable change in the slope are discarded and not considered in further analyses
(Tabachnick & Fidell, 1996).
Retention o f factors. The initial principal components EFA produced 5 factors
with eigenvalues greater than 1, accounting for approximately 66.48% o f the total
variance. Cattell’s scree test, though, indicated a clear retention o f 3 main factors, and
a potential fourth factor. Researchers have highlighted the inexact nature o f the scree
test, noting that it requires a judgement regarding the point at which the discontinuity
in eigenvalues occurs (Bryant & Yamold, 1998; Tabachnick & Fidell, 1996).
Tabachnick and Fidell (1996) suggest repeating the factor analyses specifying a
different number o f factors each time, and examining the scree plot and residual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
matrix. Solutions were requested for 5 and consecutively fewer factors to determine
the most appropriate modeL Factorists also recommend that the utility, meaning,
interpretability and comprehensibility o f the factors remain a key criterion in the
extraction o f eigenvectors. The 3 factor model, accounting for approximately 57% o f
the variance, produced the most interpretable solution.
Adequacy o f rotation. Tabachnick and Fiddell (1996) recommend conducting
both orthogonal and oblique rotations in order to produce the best model and identify
items that might be excluded from subsequent analyses. The decision regarding
orthogonal versus oblique rotation is determined as a result o f correlations between
factors. Tabachnick and Fiddell (1996) recommend that when observed factor
correlations are greater than .32 there is enough overlap in variance among factors
(10% or more) to justify oblique rotation. Once the decision regarding rotation type
is made, the adequacy o f rotation can be assessed both by examining the pattern o f
correlations and via the criterion o f simple structure.
The criterion of an ideal simple structure is characterized by a rotation solution in
which each indicator (item) has a high loading on only one factor, and each factor has
high loadings for only a portion o f the total indicators (Pedhazur & Schmelkin, 1991).
Bryant and Yamold (1998) add that simple structure is demonstrated when (a) each
indicator has at least one loading close to zero on at least one eigenvector, (b) there
are at least as many indicators with loadings close to zero on each eigenvector as
there are number o f eigenvectors (e.g., a 3 factor model indicates 3 indicators should
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
load with values close to zero on each factor), and (c) for each pair of eigenvectors
there should be several indicators that load on only one eigenvector.
Orthogonal and oblique rotation o f factors. In order to produce a simple structure
and increase interpretability, principal factors extraction o f 3 factors with varimax
rotation (orthogonal) was requested. The varimax rotation solution was interpretable.
However, given the recommendations noted above and considering the perspective o f
construct validation Pedhazur and Schmelkin (1991) suggest conducting both
orthogonal and oblique rotations. In addition, given the potential for intercorrelations
among aspects o f the 3 GSIS factors, requesting oblique rotations was warranted.
Principal factors extraction o f 3 factors with direct oblimin rotation (oblique) was
requested. The direct oblimin rotation solution was interpretable with clearly defined
factors. Examination o f the pattern matrix reveals a clear distinction between
indicators, and as shown in Table 2 the component correlation matrix indicates that
the correlation between factor 1 and factor 2 is .40. Though the other factor
correlations are modest and below the .32 criterion, the theoretical conception that
separate aspects o f impact in group supervision are correlated, combined with the
results noted above (factor 1 and factor 2, r_= .40) indicated that oblique rotation was
appropriate. Therefore, the 3 factor direct oblimin solution was interpreted.
Table 2 Component Correlation Matrix
Factors:________ 1 ___________ 2____________3
1
2 .40
3 ______________ .26_________ .19__________ ~
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Interpretation o f item loadings on factors. General guidelines for interpretation
o f factors indicate that there is a great deal o f variability in what constitutes
appropriate cutoff scores for item loadings on specific factors, ranging from .30 to
.40. Bryant and Yamold (1998) suggest that item loadings o f .30 are considered
appropriate in consideration o f interpreting the meaning o f eigenvectors. Pedhazur
and Schmelkin (1991) suggest that researchers tend to consider loadings meaningful
at .30 or .40. Tabachnick and Fiddell (1996) recommend that only those items with
loadings o f .32 and above be interpreted, noting that determining the cutoff value is a
matter o f preference. Stevens (1986) suggested that a variable should share at least
15% o f its variance with a given factor, which indicates retaining items with loadings
o f .40 or greater for interpretation. Based upon these recommendations, the cutoff
value for interpreting items for the three GSIS factors was set at .40.
Table 3 presents the item factor loadings, eigenvalues, variance explained by each
factor, and item-scale correlations. As shown in Table 3, three items did not have
factor loadings reaching the cutoff (items 9,10, and 26) and as recommended by
researchers (Stevens, 1986; Tabachnick and Fidell, 1996) were not considered in
further analyses. The following deleted items were originally from the supervisor
impact (items 9 and 10) and group impact (item 26) scales.
Item 9: Did your supervisor allow non-supervision issues to intrude into supervision?
Item 10: Was your supervisor late for group sessions?
Item 26: Did you feel pressure to self-disclose more than you were comfortable
doing?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
Though the loading cutoff was set at .40, most o f the factor loadings that reached
the cutoff were considerably higher and considered excellent to very good. A
majority (13) o f the items demonstrated loadings above .71 (excellent with 50%
overlapping variance), 4 items were above .63 (very good with 40% overlapping
variance), 3 items were above .55 (good with 30% overlapping variance), and 4 items
had loadings between .51 and .55 (fair- with at least 26% overlapping variance and
considered meaningful for interpretation).
Table 3
Item
E l
F2 F3 Item
Scale Correlation
1. Did your supervisor demonstrate
openness
.77 .17 -.02 .75
2. Did your supervisor demonstrate
a sense of humor
.69 -.04 .11 .60
3. Did your supervisor seem flexible .76 .02 .02 .64
4. Was your supervisor competent .85 -.25 -.09 .46
5. Did your supervisor provide
you with useful feedback
.84 -.14 .14 .65
6. Did your supervisor give you
validation
.82 -.06 .12 .68
7. Did your supervisor help you to
better understand your clients
.86 -.09 .15 .71
8. Did your supervisor structure
group time effectively
.68 .08 .07 .63
9. Did your supervisor allow non
supervision issues to intrude
into supervision
.31 .36 -.23 .39
10. Was your supervisor late
for group sessions
.26 .20 .09 .39
11. Was your supervisor’s
feedback unfairly negative
.51 .35 -.18 .57
12. Was your supervisor reliable .69 .01 .08 .60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
Table 3 (Continued)
GSIS Item Factor Loadings. Ttem-Scale Correlations. Eigenvalues, and Variance Explained
FI F2 F3 Item
Item__________________________________________________________________ Scale Correlation
13. Did your supervisor
constructively use group dynamics
.58 .12 .28 .70
14. Did you receive candid
feedback from other group members
.03 .07 .74 .40
15. Did you receive useful
treatment planning from other
group members
.06 .12 .84 .53
16. Did your fellow group
members help generate ideas
through group discussion
.18 .09 .79 .58
17. Were you able to benefit from
mistakes others in the group
reported having made
.29 .08 .57 .57
18. Was there between-member
competition within the group
-.11 .87 -.09 .44
19. Were there between-member
conflicts that did not get resolved
-.04 .84 -.06 .51
20. Did some members dominate
the group time
-.05 .68 .15 .50
21. Was the group a safe place
for you to ask questions
.53 .38 .13 .76
22. Did you like the other
people in the group
-.05 .74 .15 .53
23. Did group members seem
mutually supportive
.04 .75 .19 .64
24. Did you feel anxious when
you presented your work
to the group
-.09 .51 .09 .38
25. Did you feel part o f the group .13 .59 .27 .65
26. Did you feel pressure to
self-disclose more than you were
comfortable doing
.31 .39 -.21 .43
27. Did your supervisor effectively
resolve conflicts in the group
.51 .39 .03 .70
Eigenvalue 10.419 2.918 2.064
% of Variance 38.587 10.809 7.645
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Group Supervision Impact Scale Item Factor Structure
Factor 1 o f the GSIS consisted o f 13 items (items 1-8, 11-13, 21, & 27)
accounting for approximately 38.58% o f the variance. As shown in Table 2,
descriptors loading high on this factor (.51 to .86) reflected the supervisors’
contribution (e.g., supervisor openness, competency, effectiveness) to the impact o f
group supervision, therefore this factor was designated Group Supervisor Impact
(GSI). Examples o f items from this scale include: “Did your supervisor
constructively use group dynamics,” “Was your supervisor competent,” and “Did
your supervisor effectively resolve conflicts in the group?”
Factor 2 o f the GSIS consisted o f 7 items (items 18-20, & 22-25) accounting for
approximately 10.80% o f the variance. Descriptors loading high on this factor (.51 to
.87) reflected the impact o f the overall feel o f the group in group supervision (e.g.,
support, conflict, safety) and was designated Group Environment Impact (GET).
Examples o f items from this scale include: “ Did you feel anxious when you presented
your work to the group,” “Were there between-member conflicts that did not get
resolved,” and “Did group members seem mutually supportive?”
Factor 3 o f the GSIS consisted o f 4 items (items 14-17) accounting for
approximately 7.64% o f the variance. Descriptors loading high on this factor (.57 to
.84) reflected the impact o f other supervisees within the group (e.g., vicarious
learning, receiving feedback, generating ideas) and was designated Peer Supervisee
Impact (PSI). Examples o f items from this scale include: “Were you able to benefit
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
from mistakes others in the group reported having made,” and “ Did you receive
useful treatment planning from other group members?”
Descriptive Statistics
Descriptive statistics for the GSIS and other measures are reported in Table 4.
The average item scores for the three scales o f the GSIS were 5.4 (M=71.1,
SD=T4.69) for the Group Supervisor Impact Scale, 5.6 (M=39.5, SD=7.66) for the
Group Environment Impact Scale, and 5.1 (M=20.5, SD-4.96) for the Peer
Supervisee Impact Scale. The average item score for the total GSIS was 5.4
(MM 30.7 (SD=23.66). Overall, participants’ scores reflect positive/helpful group
supervision experiences. Participants’ ratings suggest they experienced the impact o f
the supervisor (5.4), the group environment (5.6), and peer supervisees (5.1)
relatively equally.
The average scores for the convergent validity measures were as follows:
Supervision Satisfaction Questionnaire (M=25.37, SD -5.35). Group Climate
Questionnaire-S (M=24.60, SD-6.22T Group Supervision Scale (M=76.05,
SDM6.11), Supervisory Working Alliance Inventory (M=65.50, SDM5.34), and an
average score o f 20.02 (SD=4.98) for the Supervisee Self-Efficacy Scale.
Reliability. Cronbach’s alpha was used in order to estimate the internal
consistency reliability o f the GSIS. Alpha coefficients for this sample were .93 for
the Group Supervisor Impact Scale (GSIS-GSI), .85 for the Peer Supervisee Impact
Scale (GSIS-PSI), and .85 for the Group Environment Impact Scale (GSIS-GSI). The
total scale Cronbach alpha coefficient was .93 for this sample. As shown in Table 3,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
item-scale correlations ranged from .56 to .82 for Group Supervisor Impact Scale, .40
to .74 for Group Environment Impact Scale, and from .57 to .80 for the Peer
Supervisee Impact Scale. The overall item-scale correlations for the GSIS ranged
from ,37 to 78.
As shown in Table 4, the inter-item correlations for the Group Supervisor Impact
Scale ranged from .32 to .79.
Table 4
Inter-item Correlations for Group Supervisor Impact Scale
Item# 1 2 3 4 5 6 7 8 11 12 13 21 27
1 —
2 .66
. . .
3 .73 .63
. . .
4 .44 .55 .50
. . .
5 .62 .49 .55 .67 —
6 .76 .60 .67 .48 .71 —
7 .67 .58 .63 .67 .79 .74 —
8 .52 .43 .44 .48 .59 .50 .56
. . .
11 .56 .33 .44 .32 .37 .46 .45 .37 —
12 .52 .39 .48 .42 .53 .51 .53 .51 .39 —
13 .56 .46 .53 .38 .53 .54 .59 .56 .39 .56 —
21 .68 .56 .54 .40 .53 .59 .62 .46 .56 .38 .55 —
27 .60 .42 .53 .39 .47 .47 .47 .49 .44 .47 .53 .60 —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
As shown in Table 5, the inter-item correlations for the Group Environment
Impact Scale ranged from .29 to .77.
Table 5
Inter-item Correlations for Group Environment Impact Scale
Item# 18 19 20 22 23 24 25
18
—
19 .72
—
20 .48 .47
—
22 .55 .55 .40
—
23 .58 .59 .47 .77
—
24 .34 .33 .34 .29 .29
—
25 .42 .40 .46 .56 .63 .27 —
As shown in Table 6, the inter-item correlations for the Peer Supervisee Impact
Scale ranged from .37 to .76.
Table 6
Inter-item Correlations for Peer Supervisee Impact Scale
Item# 14 15 16 17
14
—
15 .62
—
16 .52 .76
—
17 .37 .58 .64 —
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
Preliminary Analyses
The potential confounding influence o f demographic variables was tested via a
series o f multiple regression analyses. The demographic variables were listed as the
independent variables, and scores on the GSIS total and scales, SSQ, SSE, GSS,
GCQ-S (engagement scale), and SWAI (rapport scale) served as the dependent
variables (convergent validity measures). The results indicate that none o f the
demographic variables (age, gender, ethnicity, degree program and type, type of
clinical placement, type o f agency, year in graduate program, theoretical orientation,
approximate number o f completed practicum hours, number o f semesters of
completed practicum or field placement, and number supervisees in supervision
group) were significantly related to participants scores on the GSIS or the convergent
validity measures (Supervisee Satisfaction Questionnaire, Supervisee Self-Efficacy
Scale, Group Supervision Scale, Group Climate Questionnaire- engagement scale,
and the rapport scale o f the Supervisory Working Alliance Inventory). Results,
respectively, were: GSIS Total Score, F(12,190)= .386, p=.96, x2= .02; GSIS Factor 1
(GS1), F(12,190)= .216, p=.99, r2= .01; GSIS Factor 2 (GEI), F(12,188)= 1.531,
p=.12, r2= .08; GSIS Factor 3 (PSI), F(12,189)= .748, p=.70, r2= .04; SSQ,
F(12,190)= .20, p= .99, r2= .01; SSE, F(12,188)= .58, p=85, r2= .03; GSS,
F(12,185)= .18, p= .99, r2= .01; GCQ-S, F(12,184)= .57, p=.85, r2 - .03; and SWAI,
F(12,186)= .41, £= 95, r2= .02.
Preliminary analyses o f relationships between the group supervision impact scale
and measures employed for convergent validity. As shown in Table 7, there were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
strong bivariate correlations between the total scores on the GSIS and the 5 measures
used to estimate convergent validity (Supervisee Satisfaction Questionnaire,
Supervisee Self-Efficacy Scale, Group Supervision Scale, Group Climate
Questionnaire- engagement scale, and the rapport scale o f the Supervisory Working
Alliance Inventory). The results demonstrate significant positive relationships
between the total score on the GSIS and total scores on the SSQ (r=.68), p<.01, SSE
(r= 63), p<.01, GSS (r=.83), p<.01, GCQ-S (engagement scale) (r= 70), p<.01, and
SWAI (rapport scale) (r=.79), p<.01. In addition, the results show that the three
scales o f the GSIS are significantly correlated with the measures o f convergent
validity as hypothesized.
The Group Supervisor Impact (GSI) Scale demonstrated significant positive
relationships with the SSQ (r=.80), p<.01, SWAI (rapport scale) (r=.86), p<.01, SSE
(r=.68), p<.01, and GSS (r=.90), g<.01. The GSI scale was also significantly
correlated with the GCQ-S (engagement scale) (r=.54), g<.01.
The Group Environment Impact (GEI) Scale demonstrated a significant positive
relationship with the GCQ-S (engagement scale) (r=.61), g<.01. The results also
showed that the GEI scale demonstrated low to moderate positive significant
relationships with the other measures.
The Peer Supervisee Impact (PSI) Scale demonstrated positive significant
relationships with the GCQ-S (engagement scale) (r=.60), p<.01, and the SSQ (r=.35),
gc.Ol. The PSI scale was also significantly related to the other criterion measures
(see Table 7).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Table 7
Intercorrelations, means, and standard deviations for GSIS total and 3 factors, SSQ,
SSE, GSS, GCQ-S, and SWAI.
Variable 1 2 3 4 5 6 7 8 9 M SD
1.GSIS Total — 130.7 23.66
2.GSIS-GSI .91 -- 71.1 14.69
3.GSIS-GEI .73 .45 — 39.5 7.66
4.GSIS-PSI .70 .49 .41 -- 20.5 4.96
5.SSQ .68 .80 .30 .35 — 25.37 5.35
6.SSE .63 .68 .27 .42 .68 ~ 20.02 4.98
7. GSS .83 .90 .41 .50 .81 .71 — 76.05 16.11
8.GCQ-S .70 .54 .61 .60 .37 .48 .51 — 24.60 6.22
9.SWAI .79 .86 .43 .40 .69 .61 .80 .49 - 65.50 15.34
p<.01 for all correlations. Note: 1. GSIS total, 2. Group Supervisor Impact Scale, 3.
Group Environment Impact Scale, 4. Peer Supervisee Impact Scale, 5. SSQ
(Supervisee Satisfaction Questionnaire), 6. SSE (Supervisee Self-Efficacy Scale), 7.
GSS (Group Supervision Scale), 8. GCQ-S (Engagement Scale o f the Group Climate
Questionnaire short form), and 9. SWAI (Rapport Scale o f the Supervisory Working
Alliance Inventory).
Multicollinearitv and singularity analyses. Multicollinearity and singularity are
problems that may arise due to variables that are potentially too highly correlated.
The term multicollinearity has been used both to describe the extent to which
predictor variables are correlated, and to indicate whether or not intercorrelations
have reached a critical level (Licht, 1998). Singularity occurs when a variable is
accounted for by a combination o f other variables, making it redundant (Tabachnick
& Fidell, 1996).
Given the high intercorrelations among variables (see Table 7) collinearity
diagnostics were requested. To evaluate collinearity diagnostics, Tabachnick and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
Fidell (1996) recommend the criteria o f the condition index >30 and no more than 2
variance proportions >.50 for a given root number, and tolerance statistics (tolerance
= 1 - SMC) that are not too low (.01) (lower tolerance statistics indicate greater
multicollinearity). Reported tolerance statistics were .67 (GCQ-S), .42 (SSE),
.29 (SSQ), .33 (SWAI), and .21 (GSS). Results of collinearity diagnostics indicate
that the condition index, variance proportions, and tolerance statistics are within
appropriate limits for all measures.
Construct Validity
A series o f multiple regression analyses were employed to test the primary
hypotheses o f this study. Multivariate multiple regression was used for all main
analyses in order to control for the potential intercorrelations among the predictor and
criterion variables (Isaac & Michael, 1997). To determine the unique contributions of
each factor, initial analyses involved simultaneous regression o f the 3 GSIS Scales
(predictor variables) and the five criterion variables (Supervisee Satisfaction
Questionnaire, Supervisee Self-Efficacy Scale, Supervisory Working Alliance
Inventory, Group Climate Questionnaire, and Group Supervision Scale). Stepwise
regression analyses were then conducted to determine the best predictor o f each
criterion variable. Results o f these analyses are discussed in the following sections.
GSIS and satisfaction. A simultaneous multiple regression analysis was used to
test the hypothesized relationship between the three factors o f the GSIS (i.e.,
supervisor impact, group environment impact, and peer supervisee impact) and
supervisee satisfaction with supervision (SSQ). The predictor variables were ratings
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
on the three GSIS scales and the criterion variable was the Supervisee Satisfaction
Questionnaire rating. The overall proportion o f variance in the satisfaction accounted
for by the three GSIS scales was significant, F(3,216)= 134.207, p<.01, accounting
for 65% o f the known variance. The Group Supervisor Impact Scale was
significantly related to satisfaction with supervision (B= .847, t= 17.248, p<.01),
whereas the Group Environment Impact (& = -.084, t= -1.796, p=.074) and Peer
Supervisee Impact Scales (S= -.012, t= -.240, p-.81) were not significantly related.
The stepwise regression analysis first considered the Group Supervisor Impact
Scale, entered as the only variable in the prediction equation. The overall proportion
o f variance explained by supervisor impact was significant, F(l,218)=395.850, p<.01,
accounting for 64% o f the known variance in supervisee satisfaction as measured on
the SSQ. The Group Environment Impact and Peer Supervisee Impact Scales were
excluded variables. Therefore, the Group Supervisor Impact Scale was considered
the best indicator o f supervisee satisfaction.
GSIS and rapport scale o f supervisory working alliance. A simultaneous multiple
regression analysis was conducted to determine the unique contributions o f the three
GSIS scales (predictor variables) to scores on the Rapport Scale o f the Supervisee
Working Alliance Inventory (criterion variable). The overall proportion o f variance
in the SWAI Rapport Scale explained by the 3 GSIS scales was significant, F(3,212)=
205.309, p<.01, accounting for 74% o f the known variance. The Group Supervisor
Impact Scale was significantly related to the supervisory working alliance, (B= .85, t -
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
20.162, jK.01), whereas the Group Environment Impact Scale (R= .064, t= 1.594,
p=.112) and the Peer Supervisee Impact Scale (B- -.048, t= -1.161, p=247) were not
significantly related.
The stepwise regression analysis first considered the Group Supervisor Impact
Scale, entered as the only variable in the prediction equation. The overall proportion
o f variance explained by Group Supervisor Impact Scale was significant,
F(1,214)=609.383, p<.01, accounting for 74% o f the variance in the rapport aspect o f
the supervisory working alliance. The other two GSIS Scales were excluded
variables. Therefore, the Group Supervisor Impact Scale was considered the best
indicator o f supervisor-supervisee rapport as measured by the SWAI.
GSIS and supervisor roles. Simultaneous and stepwise regression analyses were
conducted to test the hypothesized relationship between the three factors o f the GSIS
and supervisor roles in supervision (GSS). The predictor variables were scores on the
three GSIS Scales, and the criterion variable was the GSS score. In the simultaneous
regression the overall proportion o f variance on the GSS explained by the 3 GSIS
scales was significant, F (3,211)= 310.125, p_<.01, accounting for 81% o f the known
variance. In terms o f the unique contributions o f the GSIS scales, B=.871 was
significant (t= -.156, p= 876) for the Group Supervisor Impact Scale, 6= -.015 was
not significant (t= -.446, p=.656) for the Group Environment Impact Scale, and B=
.073 was significant (t= 2.070, p=.04) for the Peer Supervisee Impact Scale.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
The stepwise regression analysis first considered the Group Supervisor Impact
Scale, entered as the only variable in the prediction equatioa The overall proportion
of variance explained by supervisor impact was significant, F(l,213)=916.242, p<.01,
accounting for 81.1% o f the variance in supervisor roles as measured on the GSS.
Entering the other two scales as predictors, 81.5% o f the variance in supervisor roles
was accounted for (a very slight increase o f .004%). Therefore, the Group Supervisor
Impact Scale was considered the best predictor o f supervisor roles as measured by the
GSS.
GSIS and group cohesion (engagement). Simultaneous and stepwise regression
analyses were conducted to test the hypothesized relationship between the three GSIS
Scales and the engagement scale o f the GCQ-S. The predictor variables were scores
on the three GSIS Scales, and the criterion variable was the GCQ-S score. In the
simultaneous regression the overall proportion o f variance on the GCQ-S explained
by the 3 GSIS scales was significant, F(3,210)= 87.356, p<.01, accounting for 55% o f
the known variance. In terms o f the unique contributions o f the GSIS scales, 6=. 193
was significant (t= 3.453, p<.01) for the Group Supervisor Impact Scale, fi= .385 was
significant (t= 7.221, g<.01) for the Group Environment Impact Scale, and 6= .350
was significant (t= 6.394, jK.01) for the Peer Supervisee Impact Scale.
The stepwise regression analysis first considered the Group Environment Impact
Scale, entered as the only variable in the prediction equation. The overall proportion
o f variance explained by Group Environment Impact Scale was significant,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
F(1,212)=130.66Q, g<.01, accounting for 38.1% o f the variance in group cohesion as
measured on the GCQ-S. Entering the other two scales as predictors, 53% o f the
variance in group cohesion was accounted for (an increase o f 14.9%). The overall
proportion o f variance in group cohesion explained by all three GSIS scales was 55%
(an slight increase o f 2.5%). Therefore, the Group Environment Impact Scale was
considered the best indicator o f the engagement aspect o f group cohesion as measured
by the GCQ-S.
GSIS and supervisee self-efficacv. Simultaneous and stepwise regression
analyses were conducted to test the hypothesized relationship between the three GSIS
scales and supervisee self-efficacy ratings as measured by the SSE. The predictor
variables were scores on the three GSIS scales, and the criterion variable was the SSE
score. In the simultaneous regression the overall proportion o f variance on the SSE
explained by the 3 GSIS scales was significant, F(3,215)= 65.701, p<01, accounting
for 48% o f the known variance. In terms o f the unique contributions o f the GSIS
scales, £H65 was significant (t= 10.826, p<.01) for the Group Supervisor Impact
Scale, R= -.08 was not significant (t= -1.339, 182) for the Group Environment
Impact Scale, and S— .13 was significant (t= 2.197, £=.02) for the Peer Supervisee
Impact Scale.
The stepwise regression analysis first considered the Group Supervisor Impact
Scale, entered as the only variable in the prediction equation. The overall proportion
of variance explained by Group Supervisor Impact Scale was significant,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
F(l,217)=188.484, £<.01, accounting for 46% o f the variance in supervisee self-
efficacy as measured by the SSE. The other two scales were excluded variables.
Therefore, the Group Supervisor Impact Scale was considered the best indicator of
supervisee self-efficacy ratings.
GSIS and measures employed for divergent validity. Results demonstrate that
participants’ and supervisors’ theoretical orientations were mildly correlated (r=. 17,
p<.01). To provide evidence o f divergent validity, a series ofMANOVAs were
conducted to assess the relationships between the GSIS total score, and the 3 scales,
and both participants’ and supervisor’s theoretical orientation. Results indicate that
the GSIS total score is not significantly related to participants’ theoretical orientation
F( 12,219)= .965, p>.05, or supervisor’s theoretical orientation F(12,217)= 1.359,
p>.Q5, The Group Supervisor Impact Scale was not significantly related to
participants’ theoretical orientation F( 12,219)= 1.191, p>.05, or supervisor’s
theoretical orientation F( 12,217)= 1.105, g>.05. The Group Environment Impact
Scale was not significantly related to participants’ theoretical orientation F( 12,217)=
1.129, £>.05, or supervisor’s theoretical orientation F(12,215)= 1.225, p>.05. The
Peer Supervisee Impact Scale was not significantly related to participants’ theoretical
orientation F( 12,218)= .689, £>.05, or supervisor’s theoretical orientation F( 12,216)=
1.527, £>.05. Results demonstrate that the GSIS total score and the three scale scores
are not significantly related to either participants’ or their supervisors’ theoretical
orientation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
Table 8: Theoretical Orientation___________________________________________
Means and Standard Deviations for Participants’ and Supervisors’ Theoretical Orientation
on GSIS Total and Factor Scores.
GSIS Supervisor Impact Group Impact Peer Impact
Theory Total Score Scale Scale Scale
Participants: M SD M SD M SD M SD
Eclectic 129.9 24.9 70.0 15.1 39.9 7.1 20.5 5.1
Cognitive 135.5 19.9 75.1 10.7 40.4 7.1 20.9 5.4
Behavioral/Learning 119.3 28.9 63.7 18.1 36.0 10.4 19.6 4.9
Psychodynamic/ 133.3 16.7 74.3 11.5 37.7 7.7 21.2 3.8
Neo-Freudian
Rogerian/ 129.3 47.8 66.3 30.7 39.0 14.7 24.0 3.0
Person-Centered
Humanistic 136.2 15.2 73.2 10.7 42.1 5.4 20.9 4.4
Existential 124.3 22.2 71.3 17.6 34.7 6.1 18.3 3.7
Gestalt 161.0
-------------
88.0
------------
45.0
—
28.0
—
Systems 133.9 31.7 72.4 20.5 40.2 9.7 21.2 5.1
Sullivanian/ 130.5 16.3 67.1 12.7 43.1 5.1 20.2 2.1
Interpersonal
Adlerian 120.0
-------------
67.0
—
34.0
—
19.0
—
Jungian 162.0
-------------
91.0
—
48.0
—
23.0
—
Other 129.3 24.9 70.7 12.4 40.5 6.9 18.2 6.7
Supervisors:
Eclectic 130.7 24.9 71.7 12.5 39.8 5.6 20.1 5.1
Cognitive 134.4 19.9 71.9 15.4 40.4 8.2 22.1 4.5
Behavioral/Learning 119.8 28.9 64.4 17.6 36.4 10.1 19.0 5.6
Psychoanalytic/ 117.7 21.3 62.5 16.4 35.2 4.8 20.0 5.6
Freudian
Psychodynamic/ 130.9 29.6 72.4 15.8 39.7 7.9 20.6 4.5
Neo-Freudian
Rogerian/ 139.1 14.3 74.7 10.9 42.7 3.9 21.6 5.5
Person-Centered
Humanistic 134.8 22.1 74.7 10.3 39.0 6.9 21.1 6.7
Existential 115.0 23.4 65.0 18.0 35.5 3.4 14.5 3.3
Gestalt 141.7 2.5 78.3 1.5 44.7 3.5 18.7 4.2
Systems 137.5 20.0 74.1 14.8 41.5 5.6 21.8 4.4
Sullivanian/ 135.9 19.5 73.8 15.7 41.2 7.2 20.9 4.4
Interpersonal
Adlerian 120.0
-------------
73.0
—
30.0 — 17.0
—
Other 137.3 15.3 74.8 10.2 40.0 7.7 22.5 3.7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
Summary o f Results
The intent o f this study was to develop and provide evidence for the reliability
and validity o f the Group Supervision Impact Scale (GSIS). The GSIS was designed
to measure positive and negative group supervision experiences by evaluating three
essential components o f group supervision: (1) Group Supervisor Impact, (2) Group
Environment Impact, and (3) Peer Supervisee Impact. Following is a synopsis o f the
results associated with the primary research questions regarding the reliability and
validity o f the GSIS. As summarized below, all reliability and validity research
questions were supported by the results.
Factor analyses. The three GSIS scales were supported by factor analysis. The
final version consisted o f 24 items: The Group Supervisor Impact Scale (13 items),
the Group Environment Impact Scale (7 items), and the Peer Supervisee Impact Scale
(4 items) (see appendix B for complete group supervision impact scale).
Reliability. The results indicated that the full scale as well as the three GSIS
scales demonstrated robust reliability. Cronbach alpha coefficients and item-scale
correlations were significant, indicating the internal consistency reliability o f the
scale.
Construct validity. The results provided construct validity evidence for the three
GSIS Scales. As hypothesized, the Group Supervisor Impact (GSI) Scale was
significantly related to supervisees’ perceptions o f satisfaction with supervision, of
supervisor-supervisee rapport in the supervisory working alliance, supervisor
leadership roles, and self-efficacy ratings. Multivariate regression analyses indicated
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
that GSIS-GSI was the best predictor o f satisfaction, rapport in the supervisory
working alliance, supervisor leadership roles, and supervisee self-efficacy ratings.
The Group Environment Impact (GEI) Scale was significantly related to group
cohesion. Multivariate regression analyses indicated that the GSIS-GEI was the best
predictor o f group cohesion as measured by the group climate questionnaire.
The Peer Supervisee Impact (PSI) Scale was significantly related to satisfaction
with supervision and supervisee self-efficacy ratings. Multivariate regression
analyses indicated that the GSIS-PSI was a significant indicator o f self-efficacy
ratings, though the GSIS-GSI was found to be the best predictor for supervisee self-
efficacy ratings as measured by the SSE.
Divergent validity. As hypothesized, the results indicate that there were no
significant relationships between the Group Supervision Impact Total Scale and three
scales, and (a) participants’ theoretical orientation or (b) supervisors’ theoretical
orientation. Thus, divergent validity for the three GSIS scales was established.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
Chapter IV
DISCUSSION
The purpose o f this study was to develop and provide validation evidence for the
Group Supervision Impact Scale (GSIS). The GSIS is a self-report group supervision
evaluation scale measuring the following essential components o f group supervision:
(1) Group Supervisor Impact, (2) Group Environment Impact, and (3) Peer
Supervisee Impact (see appendix B). The GSIS makes a unique contribution and will
benefit investigations o f group supervision by allowing researchers to utilize a
measure specifically developed to assess the distinct components and experiences
occurring within group supervision.
This chapter will provide a discussion o f the results. First is a brief summary o f
the psychometric properties o f the GSIS, followed by a more detailed description o f
the GSIS scales. Third, a discussion o f the validity results is presented, indicating the
relationships between the GSIS and measures employed for convergent and divergent
validity. The validity discussion will begin with the overall findings, followed by a
review o f the predicted research questions. Fourth, are observations regarding the
limitations o f the current study. Finally, is a discussion o f the potential benefits o f the
Group Supervision Impact Scale and the implications for theory, research, and
practice.
Psychometric Properties o f the GSIS
In terms o f the psychometric properties o f the Group Supervision Impact Scale,
the results indicate strong preliminary evidence o f reliability and validity. The three
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
scales o f the GSIS were supported via exploratory factor analysis and accounted for
approximately 57% o f the variance. The three scales were distinct with strong item
loadings on the respective factors. Evidence o f the internal consistency reliability o f
the GSIS was high for the total scale (.93), as well as the three scales (.93 for the
Group Supervisor Impact Scale, .85 for the Group Environment Impact Scale, and .85
for the Peer Supervisee Impact Scale). Evidence o f validity was supported via
significant relationships with theoretically pertinent criterion variables. Thus, results
o f this study provide preliminary evidence o f the three-factor structure and reliability
o f the Group Supervision Impact Scale.
Description o f GSIS Scales
The GSIS scales measure three main areas o f impact within group supervision
(i.e., supervisor, group environment, and peer supervisee). Each scale assesses the
unique contributions made by the supervisor, the overall group climate, and group
members. An individual’s score on the GSIS reflects a supervisee’s perception o f
positive and negative group supervision experiences associated with the three
potential sources o f impact (i.e., supervisor, group climate, peer supervisees).
Overall, the scales combine to provide a measure of an individual group supervision
participant’s experience. Higher scores reflect positive and helpful group supervision
experiences, whereas lower scores reflect hindering or unhelpful group supervision
experiences. On average, participants’ scores were in the top quarter (<75% on each
scale) for all three scales, indicating that participants in this study experienced their
group supervision as helpful and positive.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Group supervisor impact scale. The Group Supervisor Impact Scale assesses
specific tasks o f the supervisor such as (a) providing structure, (b) maintaining
appropriate focus in the group, (c) constructively utilizing group dynamics, (d)
assisting in case conceptualization, and (e) providing effective feedback. In addition,
this scale assesses the impact o f specific supervisor characteristics associated with
supervisees’ perceptions o f positive and negative supervision experiences such as (a)
supervisor openness, (b) sense o f humor, (c) flexibility, and (d) providing validation.
The specific items o f the Group Supervisor Impact scale measure supervisees’
perceptions o f supervisor tasks and characteristics that are associated with positive
and negative supervision experiences (e.g., Enyedy et al., 2003; Goodyear et a i, in
press). Participants’ average scores (M - 71.1, SD= 14.69) on this scale were well
above the median split (45.5). Overall, participants rated their group supervisor
highly on this scale, indicating helpful or positive group supervision experiences
associated with the contributions o f the supervisor. Group supervisors may wish to
examine their individual characteristics or roles within the group in order to increase
the likelihood o f positive and helpful outcomes.
One item (i.e., Was the group a safe place for you to ask questions) on the Group
Supervisor Impact scale was predicted to correlate with the Group Environment
Impact scale. Though this item did load on both factors 1 (supervisor impact) and 2
(group environment impact), it’s strongest loading was on the first factor (supervisor
impact). Given the strong factor loading (.53) and item-scale correlation (.76), this
item was retained on the Group Supervisor Impact scale (factor 1). The wording o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
this item might have inferred that the group was a safe place to ask questions o f the
supervisor or was simply to vague, rather than tapping into the overall issue o f group
safety as it was originally intended. Though initially intended to asses an aspect o f
group environment, this item theoretically suits the Group Supervisor Impact scale as
well, demonstrating the importance o f an atmosphere that encourages trainees to ask
questions. In addition, it suggests that the supervisor’s role in creating an
environment in which trainees can feel comfortable asking questions is an important
consideration.
Group environment impact scale. The Group Environment Impact scale assesses
specific characteristics o f group milieu associated with an individual supervisee’s
perceptions o f potential positive and negative group supervision experiences. For
example, items from this scale measure a supervisee’s experience o f (a) unresolved
between-member conflicts, (b) between-member competition, (c) mutual support, (d)
attraction to group, and (e) potential anxiety experienced when presenting work. The
specific items o f the Group Environment Impact scale measure supervisee’s
perceptions o f the group environment that are associated with positive and negative
group supervision experiences. Participants’ average scores (M= 39.5, SD= 7.7) on
this scale were well above the median split (24.5) indicating positive group
supervision experiences associated with the contributions o f the overall group
environment. This finding suggests that the particular elements o f the group
environment measured by this scale are relevant considerations in the functioning o f a
supervision group. Supervisors facilitating a group might utilize this scale to assess
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
the group environment and identify specific areas in need o f intervention within the
group.
Peer supervisee impact scale. The Peer Supervisee Impact scale assesses the
supervisee’s perceptions o f experiences within the group associated with learning and
specific tasks. For example, this scale measures supervisees’ perceptions o f group
tasks and experiences provided by group members such as (a) providing feedback, (b)
vicarious learning, (c) generating ideas, and (d) treatment planning assistance. The
items from this scale measure a supervisee’s perceptions o f positive and negative
group supervision experiences that are linked with other supervision group members
contributions. Participants’ average scores (M= 20.5, SD= 4.9) on this scale were
well above the median split (14) indicating positive group supervision experiences
associated with the impact o f their peer supervisees.
Convergent Validity
Results o f this study demonstrate evidence o f validity for the Group Supervision
Impact Scale, As predicted, the GSIS total score and the three scales were related to
supervisee satisfaction (SSQ), the supervisory working alliance (SWAI), supervisee
self-efficacy (SSE), group supervisor leadership roles (GSS), and overall group
cohesiveness (GCQ-S). These findings indicate that the GSIS is an appropriate
measure for tapping into constructs that have previously been shown to be related to
group supervision outcome (e.g., Ellis & Ladany, 1998; Kruger et al., 1988; Worthen
& McNeill, 1996). In addition, the distinct scales were related to the criterion
variables as predicted, demonstrating the validity and practical utility o f the scales.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
The Group Supervisor Impact scale was related to satisfaction in supervision, the
supervisory working alliance, group supervisor leadership roles, and perceptions o f
supervisee self-efficacy. Previous research (e.g., Bernard & Goodyear, 1998; Dion,
2000; Goodyear & Guzzardo, 2000; Kruger et al., 1988) has demonstrated that the
perceived quality o f supervision is related to the aforementioned aspects o f
supervision outcome (i.e., satisfaction, working alliance, etc.). The specific group
supervisor tasks and characteristics outlined above (e.g. providing feedback,
providing assistance with case conceptualization, maintaining the focus o f the group,
being open and flexible, and having a sense o f humor), as measured by the Group
Supervisor Impact scale, contribute to supervisees’ positive group supervision
experiences. Researchers (e.g., Goodyear and Guzzardo, 2000; Gray et al., 2001;
Magnuson et a l, 2000; Nelson & Friedlander, 2001; Ramos-Sanchez et al., 2002)
have shown that the supervisor qualities assessed by this scale are essential
components o f group supervision. For example, researchers (e.g., Heppner et al.,
1998; Larson & Daniels, 1998; O’Brien et al., 1997) have discussed the importance o f
self-efficacy in the development o f counseling skills. The findings of this study,
which link specific group supervisor behaviors with supervisee self-efficacy ratings
extend previous research and provide a means o f assessing aspects o f group
supervision functioning that relate to positive and negative outcomes. Researchers
and clinicians alike could utilize this scale to assess the impact o f the group
supervisor and infer the supervisee’s perceptions of the quality and helpfulness o f the
supervision.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
The Group Environment Impact scale was related to perceptions of group
cohesion (GCQ-S). Researchers (e.g., Bernard & Goodyear, 1998; Dion, 2000;
Kruger et al., 1988) have discussed cohesiveness in group supervision in terms o f the
level o f support, trust, empathy, and engagement in the group. Groups evidencing
greater cohesion are more likely to enable its members to be vulnerable, to have open
discussions, and facilitate problem resolution. Further, this allows the group
members to receive mutual support and openly discuss relevant issues. As previously
noted, the Group Environment Impact scale assess key aspects o f the overall affective
tone and functioning o f the group by evaluating levels o f support, conflict, and
attraction to the group (e.g., unresolved between-member conflicts, between-member
competition, mutual support). Results of this study indicate that the Group
Environment Impact scale, which is designed to assess the contribution o f the overall
group atmosphere to group supervision outcome, could be a useful tool in measuring
the perceived helpfulness o f group supervision. The overall tone o f the group
environment can either help or hinder the progress made by the group as a whole and
its individual members. The Group Environment Impact scale might enable
clinicians in assessing potential areas to intervene in groups that are experiencing
difficulties, or provide researchers with a much need tool for measuring key aspects
o f group functioning.
The Peer Supervisee Impact scale was related to satisfaction in supervision and
supervisee self-efficacy ratings. Research regarding satisfaction in supervision has
demonstrated that it is a key outcome variable and is indicative o f perceived helpful
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
supervision experiences. Results o f this study establish that the contributions o f peer
supervisees in group supervision are linked with participants’ satisfaction ratings.
The current findings show that the specific peer supervisee behaviors (i.e., providing
feedback, vicarious learning, generating ideas, and treatment planning) assessed by
the Peer Supervisee Impact scale are important elements o f group supervision and
relevant to consider in the assessment o f group supervision functioning.
The specific behaviors assessed by the Peer Supervisee Impact scale also reflect
potential elements in the development o f self-efficacy (Lehrman-Waterman and
Ladany, 2001). Lehrman-Waterman and Ladany (2001) suggest that self-efficacy is
an instrumental component o f developing competence in trainees. In addition, the
demonstrated link between self-efficacy ratings and peer supervisee behaviors
suggests that the feedback and learning provided by the dynamic quality o f group
supervision contributes to an overall helpful group supervision experience.
Participants in this study who reported strong self-efficacy beliefs and high
satisfaction with their group supervision experience, scored on the high end o f the
Peer Supervisee Impact scale indicating the relationship between group members
behaviors and perceived outcomes (e.g. satisfaction, self-efficacy). These findings
extend previous research (e.g., Larson et al., 1992) and suggest that researchers and
clinicians should consider specific supervisee behaviors (i.e., evidenced in the peer
supervisee impact scale) when assessing group supervision functioning and outcome.
Following is a discussion o f the Group Supervision Impact Scale and its specific
associations with supervision satisfaction, supervisory working alliance, group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
cohesion, supervisor roles, and supervisee self-efficacy. Included is a discussion o f
the regression analyses, which indicates the unique contributions o f each o f the GSIS
scales. As evidenced by the findings presented in the previous chapter, and the
following discussion, all validity hypotheses were supported.
Group supervision impact scale and satisfaction. The results supported the first
validity hypothesis that the Group Supervisor Impact scale would be related to
satisfaction in supervision. As predicted, higher scores on the Group Supervisor
Impact scale were predictive o f greater supervisee satisfaction. In addition, the Group
Supervisor Impact scale was found to be the best predictor o f satisfaction, excluding
the other two scales (i.e., group environment impact and peer supervisee impact).
This finding supports previous studies investigating the link between supervision
outcome and satisfaction (e.g., Bernard & Goodyear, 1998; Chicca-Enyedy et a l,
2003; Goodyear et al., in press; Gray et al., 2001; Nelson & Friedlander, 2001;
Savickas et al., 1986; Stoltenberg, 1981; Worthen & McNeill, 1996) and extends
prior research by specifically linking behaviors o f the group supervisor. In addition,
the group supervision specific component adds to the small amount o f literature
utilizing measures designed to assess group supervision. Though previous research
has demonstrated the importance o f supervisor behaviors in producing satisfaction
with supervision, this study focuses more precisely on the specifics o f group
supervision.
Group supervision impact scale and working alliance. The results supported the
second validity hypothesis that the Group Supervisor Impact scale would be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
predictive o f a stronger supervisory working alliance. As hypothesized and supported
by previous research (e.g., Ladany et al., 1999; Lehrman-Waterman & Ladany, 2001),
higher ratings on the Group Supervisor Impact scale were predictive o f higher ratings
o f the working alliance. Ladany et al. (1999) report that a strong emotional bond
between supervisor and supervisee (an aspect o f the working alliance) was associated
with greater satisfaction, and Lehrman-Waterman and Ladany (2001) found that
effective evaluation practices were predictive o f a stronger working alliance. The
current findings extend prior research by illustrating the relationship between specific
helpful group supervisor behaviors and the rapport aspect o f the supervisory working
alliance in group supervision. Following, group supervisors might consider the
characteristics and behaviors outlined in this scale (e.g., providing structure and group
focus, utilizing group dynamics, effective feedback, openness, flexibility, and humor)
when assessing a means o f positively influencing the supervisory working alliance in
group supervision.
Group supervision impact scale and group cohesion. The third validity hypothesis
regarding the Group Environment Impact scale was also supported by the findings.
Namely, that the Group Environment Impact scale would be predictive o f group
cohesion. Though the initial simultaneous regression analysis indicated that all three
scales were related to group cohesion (with the group environment impact scale
demonstrating the strongest 6), the stepwise regression analysis supported the
expectation that the Group Environment Impact scale would be the best predictor o f
group cohesion. As predicted, positive and helpful group environments, as measured
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
by the items o f the Group Environment Impact scale, were related to more cohesive
groups. This finding is supported by previous research (e.g., Dies and Mackenzie,
1983; Evans & Jarvis, 1986; Mackenzie, 1981) that speaks to the importance of
cohesiveness in groups, and extends prior research by validating the importance of
specific aspects o f the group supervision environment. Researchers and clinicians
may utilize this scale to identify specific problem areas for intervention, thereby
increasing the likelihood o f positive group supervision outcomes and greater
cohesiveness within the group.
Group supervision impact scale and supervisor roles. The results supported the
validity hypothesis that the Group Supervisor Impact scale would be predictive o f
ratings o f supervisor leadership roles. As predicted, the Group Supervisor Impact
scale was the best predictor o f supervisor roles. Though not the best predictor, the
Peer Supervisee Impact scale contributed to a small proportion o f the variance in
group supervisor roles as measured by the GSS. Given that the GSS measures
specific group supervisor tasks such as provision o f feedback and case
conceptualization, there is some overlap with the peer supervisee behaviors measured
on the Peer Supervisee Impact scale o f the Group Supervision Impact Scale such as
feedback, generating ideas, and treatment planning assistance. Though the Group
Supervision Scale (Arcinue, 2002) and the Group Supervision Impact Scale (GSIS)
were highly correlated, the GSIS assesses two areas o f group supervision (Le., group
environment and group members) that Arcinue’s scale (GSS) does not account for.
Thus, the current study extends prior research in group supervision outcome and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
provides a measure o f supervisor impact that has been shown to predict supervisor
leadership roles in group supervision.
Group supervision impact scale and self-efficacy. The results supported the
validity hypothesis that both the Group Supervisor Impact and Peer Supervisee
Impact scales would be related to supervisee self-efficacy. The stepwise regression
analysis indicated that the contributions o f the supervisor served as the best predictor
o f self-efficacy ratings. This indicates that study participants who rated the impact of
the group supervisor positively, also reported strong self-efficacy ratings regarding
their role as a counselor. More specifically, participants believed that they were
better able to understand their clients’ process issues, conceptualize client problems,
understand their own countertransference, and felt more confident on their role as a
therapist. These findings are supported by previous research linking the impact o f the
supervisor with self-efficacy beliefs (e.g., Heppner et al., 1998; Ladany et al., 1999;
Larson et al., 1992: Lehrman-Waterman & Ladany, 2001; O’Brien et al., 1997). This
study extends prior research and makes a unique contribution by demonstrating the
relationship between specific behaviors that a group supervisor or group members
engage in that might contribute to and predict a trainee’s self-efficacy beliefs.
Limitations
One o f the main limitations o f this study was the unavoidable use o f measures that
were not originally developed to assess supervision-specific variables. As previously
discussed at length (Chapter 1) and in essence providing the impetus for this study,
reliance upon measures not designed to measure group supervision processes might
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
impact the results in unforeseen ways. In order to reduce the potential difficulties, the
current study utilized measures that have been employed by other researchers (e.g.,
Arcinue, 2002; Dion, 2000; Friedlander & Ward, 1984; Gray et a l, 2001; Heppner et
al., 1998; Kruger et al., 1988; Ladany et al., 1999; Lehrman-Waterman & Ladany,
2001; Ramos-Sanchez et al., 2002) assessing supervision-related variables.
In addition, the current study included the Group Supervision Scale (Arcinue,
2002) as a comparison for the Group Supervisor Impact scale. Inclusion of this
measure (GSS) was important given that it was developed specifically to assess the
role o f the supervisor in group supervision. A thorough review o f the literature did
not reveal any other measures that were developed for the purpose o f assessing group
supervision. Thus, though many o f the measures employed in this study have been
adapted by researchers in order to assess supervision variables, attempts were made to
minimize potential difficulties by: (1) selecting measures widely used in supervision
research, (2) selecting measures recommended by supervision researchers (e.g.,
Bernard & Goodyear, 1998), and (3) including the Group Supervision Scale (Arcinue,
2002) for comparison. Finally, future research investigating group supervision
variables will not have to rely so heavily upon measures designed for other purposes.
A second limitation that follows from the first concerns the high intercorrelations
among the measures. As noted in the results chapter, the collinearity diagnostics
were within tolerance limits indicating that the correlations among the measures had
not reached a critical level. However, researchers (e.g., Grimm and Yarnold, 1995;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
Tabachnick & Fidell, 1996) point out that any level o f correlation has the potential to
affect multiple regression analyses.
Grimm and Yarnold (1998) suggest that the greater the degree o f multicollinearity
the more likely that the partial regression coefficients will be unstable.
Multicollinearity can present difficulties in making theoretical interpretations o f the
data due to the potential for shared variance. A means o f assessing the potential
redundant variance is to compare the relative sizes o f the bivariate correlations and
the regression coefficients (Tabachnick & Fidell, 1996).
Examination o f the data suggests that the intercorrelations among measures did
not significantly hinder the ability to determine the best predictor for each regression
equation. In addition, Grimm and Yarnold (1998) suggest that even with variables
that are highly correlated, the decision regarding whether or not to include them
should be based on the theoretical considerations regarding the hypotheses o f the
study. Thus, though there were high intercorrelations among the measures the
strength o f the correlations does not appear to have presented significant difficulties
in this study. As previously discussed, luture investigations would be improved by
the use o f different measures which would remedy this particular limitation.
A third limitation o f this study stems from the ex post facto nature o f the design
that limits the interpretative nature o f the results (Isaac & Michael, 1997). Given the
inability to manipulate variables and randomly assign participants to conditions (e.g.,
negative versus positive group supervision), causal inferences cannot be made. For
example, the results suggest that specific group supervisor behaviors predict trainee
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
satisfaction, a stronger working alliance, and stronger self-efficacy beliefs. However,
it is possible that a strong supervisory working alliance and supervisee that
demonstrated satisfaction with supervision enabled the group supervisor to engage in
the behaviors suggestive o f positive impact. In a similar vein, results indicated that
the Group Environment Impact scale was predictive o f greater cohesiveness in the
group, but given the design o f the current study it can not be made clear whether
groups with existing cohesiveness produced more positive supervision group
environments.
A fourth limitation involves threats to external validity. Though results from the
demographic survey indicate that participants represented a very diverse group in
many ways (e.g., geographic region, type o f program, theoretical orientation, year in
graduate school), participants were mostly (84.3%) female, White (78.9%), and
working in university based counseling centers (56.1%). Thus, it is important to note
that the results can only be generalized to supervisees with similar demographic
characteristics.
A fifth limitation o f this study concerns the limited number o f items on the Peer
Supervisee Impact scale. Though the four items on this scale demonstrated strong
factor loadings and the alpha for this sample was high, the scale might be improved
and more comprehensive by the inclusion o f additional items. For example, items
addressing peer supervisee’s contributions and feedback regarding theoretical
considerations or specific counseling styles might strengthen this scale and increase
the overall ability o f the scale in addressing a greater degree o f group supervision
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
features. In addition, the scale might benefit from items assessing peer supervisees
contributions to facilitating group dynamic issues. Given the nature o f group
supervision, it is probable (and hopeful) that group members will comment on and
attempt to utilize group dynamics in their attempts to resolve group issues, in
providing feedback to other group members, or in attempts to understand client
dynamics. Though the issue o f group dynamics is addressed by the Group Supervisor
Impact scale (e.g., item stems: did your supervisor constructively use group dynamics
and did your supervisor effectively resolve conflicts in the group), it is likely that
group members also contribute in the area. Thus, a limitation o f the current scale
involves the limited number o f items on the Peer Supervisee Impact scale.
A sixth limitation o f this study involves the use o f individual participants rather
than intact supervision groups. This limits the ability to assess in depth the actual
utility o f the scale in that there is no information about specific within group
functioning. Though individual group members would still be likely to provide
varying information based upon individual differences, perceptions, and potential
scapegoating, the assessment o f intact supervision groups would provide a more
comprehensive view o f the GSIS validity. The inclusion o f additional items that
address participants overall assessment o f the helpful or hindering nature o f their
group supervision experience might help to account for some o f the information that
is lost as a result o f using individual participants. However, asking participants to
rate overall how helpful they perceived their group supervision experience to be
might have added additional threats to the validity by creating a halo effect. For
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
example, participants who reported positive or helpM group supervision might be
influenced in their subsequent ratings as a result o f formulating their initial
impressions.
A seventh limitation of this study concerns the lack o f GSIS items assessing the
component o f evaluation in group supervision. One o f the responsibilities o f a
supervisor, be it individual or group, is to evaluate the trainee. Much has been written
about evaluation processes in individual supervision (Bernard & Goodyear, 1998;
Holloway, 1997; Ellis & Ladany, 1997) ranging from specific criteria for evaluation
and qualitative measures o f assessing trainees to supervisor and trainee emotional
reactions to evaluation. In their integrative review o f the supervision literature, Ellis
and Ladany (1997) observe that supervisee evaluation is another area in great need o f
research attention (p. 484). Evaluation in group supervision in particular is an area
that will benefit from empirical research. Thus, a limitation o f the current study is the
inclusion o f evaluation impact in group supervision. However, given the complexity
o f the evaluation process, it is likely an area requiring an independent scale.
A final limitation o f this study involves the preliminary nature o f the findings.
Exploratory factor analysis allows researchers to develop theoretical models and
indicates the underlying factor model. The current findings need to be replicated via
confirmatory factor analysis in order to test the goodness o f fit for the model.
Replication and confirmation o f the three factor model o f the Group Supervision
Impact Scale is an appropriate next step in establishing the stability o f the model and
the validity o f the scale.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
The limitations o f the current study suggest specific areas to be addressed in
further research regarding this scale. As previously stated, the overall purpose was to
develop and provide preliminary reliability and validation evidence for the Group
Supervision Impact Scale. The results o f this study suggest that the preliminary
evidence for the validity and utility o f the GSIS has been demonstrated and further
development is warranted.
Implications for Theory, Research, and Practice
The development and validation o f the Group Supervision Impact Scale (GSIS)
makes a unique contribution to the existing literature by providing an instrument that
can be used in both research and applied settings. Supervision researchers have
repeatedly suggested that employing measures that were originally designed for other
purposes creates multiple problems for the utility and interpretation o f supervision
research. The Group Supervision Impact Scale (GSIS) is the first measure designed
to assess the three main influences in group supervision: the supervisor, the group
environment, and the group members. Development o f the GSIS provides an initial
step in the direction o f limiting the use o f psychotherapy measures in group
supervision research. Future research will not have to rely as heavily upon measures
originally designed for other purposes. In addition, the preliminary validation
evidence of the GSIS begins to fill the gap in terms o f a training format that has been
described as “widely practiced and poorly justified” (Holloway & Johnston, 1985, p.
339) due to the lack o f empirical evidence.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
Theoretical implications related to the development and validation o f the GSIS
include the identification o f specific behaviors in group supervision as well as the
relationships with the measures employed for convergent validity (i.e., satisfaction,
supervisory working alliance, supervisor leadership roles, group cohesion, and self-
efficacy). Though theorists have considered the role o f the supervisor, group climate,
and the interaction o f group members, there has been a considerable lack o f empirical
evidence to support the assumptions about positive group supervision processes and
the unique contributions o f different aspects o f group supervision. Results o f this
study have shown that not only are specific tasks such as providing feedback,
structure, and focus important areas o f focus in group supervision, but that
characteristics o f the supervisor such as openness and humor contribute to trainee
satisfaction, the strength o f the working alliance, and self-efficacy beliefs. In
addition, the contributions and impact o f the group members have been found to have
highly significant relationships with group cohesion, trainee satisfaction, and self-
efficacy beliefs. Theorists may consider approaches to group supervision that specify
ways in which supervisors and supervisees might blend relevant individual qualities
with the tasks o f the supervision.
In terms o f the contributions o f the group supervisor, the specific behaviors (e.g.,
feedback, structure, utilization o f group dynamics, conflict resolution) and supervisor
characteristics (e.g., openness, flexibility, humor) identified by the Group Supervisor
Impact scale have been found to be significantly related to and predictive o f trainee
satisfaction, the supervisory working alliance, the roles o f the group supervisor, and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
self-efficacy beliefs. Models o f group supervision interventions might consider and
integrate these behaviors and characteristics in explanations o f group supervision
process and outcome as well as suggesting specific approaches. For example,
Bernard and Goodyear (1998) identified the following five distinct supervisor tasks
and areas o f focus in group supervision: didactic presentations, case
conceptualization, supervisee individual development, group development, and
organization issues (p. 118). In their review o f the specific tasks and foci o f the group
supervisor, Bernard and Goodyear (1998) note that many models (e.g., Sansbury,
1982; Kruger et al., 1988; Wilber et al., 1994) indicate didactic components, case
conceptualization, interpersonal process focusing on individual and group
development, and problem resolution. The findings o f this study suggest that an
important area o f focus might also be the manner in which those supervision tasks are
carried out (e.g., via individual characteristics o f flexibility, humor, openness, support
and validation).
Thus, rather than accounting for specific characteristics under the heading of
individual differences, certain qualities and behaviors that have been shown to
contribute to the overall positive impact o f group supervision should be incorporated
into the discussion and approaches to facilitating and participating in group
supervision. In addition, the findings o f this study provide empirical evidence in
support o f the benefits o f group supervision and factors related to assessment o f
group supervision outcome.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
The GSIS provides the first empirically validated measure assessing the three
main forces o f impact in group supervision. A next step will be to replicate the
current findings and examine the psychometric properties o f the GSIS via
confirmatory factor analysis. Following, research employing more experimental
designs utilizing the GSIS might assist in identifying aspects o f causality related to
specific helpful and hindering phenomena in group supervision. For example, the
GSIS might be utilized to asses group supervision processes as related to the criterion
variables utilized in this study (e.g., self-efficacy, cohesion, supervisory working
alliance) as well as numerous others (e.g., client outcomes, specific skill acquisition,
evaluation practices within group supervision).
One potential area o f research utilizing the GSIS would be research examining
intact supervision groups. Assessing intact groups may enable researchers to identify
interactions among the supervisor, group environment, and group members. In
addition, researchers might examine whether or not the positive impact o f the
supervisor mediates negative impact in the group environment. Further, supervision
groups in which participants self-identify as having positive or negative experiences
might be compared to provide additional validation and discrimination evidence for
the GSIS as well as indicating the relationship o f these variables to specific outcomes.
Another area for further research involves the investigation o f individual
difference variables in group supervision. The GSIS might help to identify the
relationship that individual differences have with particular group processes and
outcomes. In addition to investigating individual differences, the GSIS might be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
employed to aid in the investigation o f differences related to group format (e.g.,
experiential, focus on case presentations, process focus), developmental theories (e.g.,
developmental level o f trainee, developmental stage o f group), or theoretical
differences related to positive and negative group supervision experiences.
For example, one question that might be addressed involves differences regarding
the developmental stage o f the group. Groups typically move through five stages
(Tuckman & Jensen, 1977): (1) forming (members develop initial comfort level with
one another), (2) storming (resolution o f issues related to power and competition), (3)
norming (identifying norms, expectations, responsibilities and consequences for
violating the established norms), (4) performing (highly productive stage), (5)
adjourning (processing ending o f the group). Bernard and Goodyear (1998) discern
that during the performing stage group cohesion is high allowing for members to
openly reveal their individual differences, and ideally there should be “no stars
(including the supervisor)” (p. 125). Research could investigate whether or not the
relative impact o f the three main forces in group supervision are stable at this stage
during the group supervision process. Following, the degree and importance o f
impact (as measured by the three GSIS scales) might be compared during different
stages o f the group to provide more empirical evidence in support o f specific
behaviors and tasks during each stage.
The GSIS might also be utilized in research investigating a plethora o f diversity
issues as related to group supervision and perceived outcome. For example,
researchers might investigate perceptions o f positive and negative group supervision
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
as related to multicultural issues, gender differences, sexual orientation issues, or age
differences. Essentially, the existence o f the GSIS provides a valid group supervision
measure to assist in the investigation o f the numerous aspects that contribute to
positive and negative group supervision experiences and the potential relevant
outcomes associated with supervision and counseling.
The practical utility o f the GSIS includes benefits for training, evaluation, and
intervention. For example, just as research investigating group supervision processes
and outcomes is in the beginning stages, so is the relative state o f empirical evidence
for models addressing the training o f group supervisors (Bernard & Goodyear, 1998).
The GSIS can be used as a tool to assess the supervisor trainee in selecting areas to
attend to as they learn to function as a group supervisor. Supervisors in training
might benefit from discussions regarding awareness o f the impact that specific
behaviors have upon the functioning o f the group and supervisee’s perspectives o f
helpful supervision experiences. The GSIS can be used to monitor growth as new
supervisors gain experience, as well as being employed as a tool for the new
supervisor to assess overall group functioning.
Supervisors facilitating a group might utilize the GSIS to assess the group
environment and identify specific areas in need o f intervention within the group. In
addition, the GSIS can be used to monitor changes in the group over time in order to
assess the effectiveness o f specific types o f feedback, case discussions, or problem
resolution (e.g., conflict, competition, anxiety, lack o f adequate feedback from group
members or supervisor). Thus, the GSIS can assist the supervisor in tailoring
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
interventions in group supervision as well as provide a means o f assessing the group
members perceptions.
Conclusion
In 1985 Holloway and Johnston conducted the first review o f the group
supervision literature and emphasized the lack o f research that focused on group
supervision despite its prevalent use in training sites. More than a decade later, Prieto
(1996) conducted a review o f the group supervision literature and further emphasized
that “the continuing lack o f empirical research investigating the group supervision o f
trainees learning individual psychotherapy is indefensible” (p.305). Due to the lack
o f research investigating group supervision, despite its prevalent use in training sites,
and the absence o f specific measures to assess group supervision process and
outcome, group supervision has been referred to as a “weak link” (Holloway &
Johnston, 1985). The development and validation o f the Group Supervision Impact
Scale takes a step in the direction o f bridging the gap between the practice o f group
supervision and empirical evidence in support o f its helpfulness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
REFERENCES
Altfeld, D. A. (1999). An experiential group model for psychotherapy
supervision. International Journal o f Group Psychotherapy. 49, 237 - 254.
Altfeld, D.A., & Bernard, H.S. (1997). An experiential group model for group
psychotherapy supervision, (pp. 381-399). In C.E. Watkins, Handbook o f
psychotherapy supervision. New York: Wiley.
Allen, G.J., Szollos, S.J., & Williams, B.E. (1986). Doctoral students’
comparative evaluations o f best and worst psychotherapy supervision. Professional
Psychology: Research and Practice. 17, 91-99.
Arcinue, F. (2002). Validation o f the group supervision scale. Dissertation
abstracts.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory o f behavior change.
Psychological Review. 84,191-215.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American
Psychologist, 37, 122-147.
Bandura, A. (1993). Perceived self-efficacy in cognitive development and
functioning. Educational Psychologist. 28, 117-148.
Bernal, M.E., Sirolli, A.A., Weisser, S.K., Ruiz, J.A., Chamberlain, V.J., &
Knight, P.G. (1999). Relevance o f multicultural training to students’ applications to
clinical psychology programs. Cultural Diversity and Ethnic Minority Psychology.
5(1), 43-55.
Bernard, J. M., & Goodyear, R. K. (1998). Fundamentals of clinical supervision.
(2n d edition). Boston: Allyn & Bacon.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
Bloch, S., Reibstein, J., Crouch, E., Holroyd, P. & Themen, J. (1979). A method
for the study o f therapeutic factors in group psychotherapy. British Journal of
Psychiatry, 134,257-263.
Borgen, F. H., & Barnett, D. C. (1987). Applying cluster analysis in counseling
psychology research. Journal o f Counseling Psychology, 34, 456 - 468.
Brossart, D.F., Patton, M.J., & Wood, P.K. (1998). Assessing group process: An
illustration using the tuckerized growth curves. Group Dynamics: Theory. Research,
and Practice. 2(1), 3-17.
Bruning, R.H., Schraw, G.J., & Ronning, R.R. (1999). Cognitive Psychology and
Instruction (3r d ed.). New Jersey: Prentice Hall, Inc.
Bryant, F.B., & Yarnold, P.R. (1998). Principal-components analysis and
exploratory and confirmatory factor analysis, (p. 99-136). In Grimm, L.G., &
Yarnold, P.R. (1998). Reading and Understanding Multivariate Statistics.
Washington, D.C.: American Psychological Association.
Cartwright, D., & Zander, A. (1968). Group dynamics: Research and Theory
(3rd ed.). New York: Harper & Row.
Chen, E.C., & Bernstein, B.L. (2000). Relations o f complementarity and
supervisory issues to supervisory working alliance. Journal o f Counseling
Psychology. 47(4), 485-497.
Christensen, T. M., & Kline, W. B. (2000). A qualitative investigation o f the
process o f group supervision with group counselors. Journal for Specialists in Group
Work. 25. 3 7 6 -3 9 3 .
Counselman, E. F., & Gumpert, P. (1993). Psychotherapy supervision in small
leader-led groups. Group. 17,25 - 32.
Crocker, L., & Algina, J. (1986). Introduction to Classical & Modem Test
Theory. New York: Holt, Rinehart and Winston, Inc..
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
Daughtry, D., & Kunkel, M. A. (1993). Experience o f depression in college
students: A concept map. Journal o f Counseling Psychology. 40, 3 1 6 - 323.
Dies, R.R., & MacKenzie, K.R. (1983). Advances in Group Psychotherapy:
Integrating research and practice. New York: International Universities Press, Inc.
Dion, K.L. (2000). Group cohesion: From “field o f forces” to multidimensional
construct. Group Dynamics: Theory. Research, and Practice. 4(1), 7-26.
Doxsee, D. J., Kivlighan, D. M., Jr. (1994). Hindering events in interpersonal
relations groups for counselor trainees. Journal o f Counseling and Development. 72.
6 2 1 -6 2 6 .
Ducheny, K., Alletzhauser, H.L., Crandell, D., & Schneider, T.R. (1997).
Graduate student professional development. Professional Psychology: Research and
Practice. 28(11. 87-91.
Efstation, J.F., Patton, M.J., & Kardash, C.M.(1990). Measuring the working
alliance in counselor supervision. Journal o f Counseling Psychology. 37(3), 322-329.
Ellis, M. V., & Douce, L. A. (1994). Group supervision o f novice clinical
supervisors: Eight recurring issues. Journal o f Counseling & Development. 72. 520 -
525.
Ellis, M.V., Krengel, M., & Beck, M. (2002). Testing self-focused attention
theory in clinical supervision: Effects on supervisee anxiety and performance. Journal
o f Counseling Psychology. 49(1),
Ellis, M.V., & Ladany, N. (1997). Inferences concerning supervisees and clients
in clinical supervision: An integrative review, (pp. 447-507). In C.E. Watkins,
Handbook o f Psychotherapy Supervision. New York: Wiley.
Ellis, M.V. (2001). Harmful supervision, a cause for alarm: Comment on Gray et
al. (2001) and Nelson & Friedlander (2001). Journal o f Counseling Psychology.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Elliott, R. & Wexler, M. M. (1994). Measuring the impact o f sessions in
process—experiential therapy o f depression: The Session Impacts Scale. Journal of
Counseling Psychology. 41,166-174.
Emerson, S. (1996). Creating a safe place for growth in supervision.
Contemporary Family Therapy: An International Journal. 18, 393 - 403.
Enyedy, K., Arcinue, F., Puri, N., Carter, J.W., Goodyear, R.K., & Getzelman,
M.A. (2003). Hindering phenomena in group supervision: Implications for practice.
Professional Psychology: Research and Practice. 34(3), 213-317.
Evans, N.J., & Jarvis, P.A. (1986). The group attitude scale: A measure of
attraction to group. Small Group Behavior. 17.203-216.
Evans, C.R., & Dion, K.L. (1991). Group cohesion and performance: A meta
analysis. Small Group Research. 22, 175-186.
Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in informal
groups: A study ofhuman factors in housing. Stanford: Stanford University Press.
Friedlander, M.L., & Ward, E.G. (1984). Development and validation of the
supervisory styles inventory. Journal o f Counseling Psychology. 4, 541-557.
Goodyear, R.K., Enyedy, K., Arcinue, F., Puri, N., Carter, J.W., & Getzelman,
M.A. (in press). Helpful phenomena in group supervision: A concept map.
Professional Psychology: Research and Practice.
Goodyear, R.K., & Guzzardo, C.R. (2000). Psychotherapy training and
supervision, (pp. XX) In S.Brown & R.Lent, Handbook o f counseling psychology.
(3r d ed.). New York: Wiley.
Goodyear, R. K., & Nelson, M. L. (1997). The major supervision formats, (pp.
328 - 344) In C. E. Watkins, Handbook o f Psychotherapy Supervision. New York:
Wiley.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
Gray, L.A., Ladany, N., Walker, J.A., Ancis, J.R. (2001). Psychotherapy trainees
experience o f counterproductive events in supervision. Journal o f Counseling
Psychology, 48(4),
Guest, P.D., & Beutler, L.E. (1988). Impact of psychotherapy supervision on
therapist orientation and values. Journal o f Consulting and Clinical Psychology.
56(5), 653-658.
Hair, J. F., & Black, W. C. (2000). Cluster analysis. In L. G. Grimm & P. R.
Yarnold (Eds.), Reading and Understanding More Multivariate Statistics (pp. 1 4 7 -
206). Washington, DC: American Psychological Association.
Hawkins, P., & Shohet, R (1989). Supervision in the helping professions. Milton
Keynes, U.K.: Open University Press.
Heppner, M.J., Multon, K.D., Gysbers, N.C., Ellis, C.A., & Zook, C.E. (1998).
The relationship o f trainee self-efficacy to the process and outcome o f career
counseling. Journal o f Counseling Psychology. 45(4), 393-402.
Hillerbrand, E. (1989). Cognitive differences between experts and novices:
Implications for group supervisioa Journal o f Counseling and Development 67,
293-296.
Holloway, E.L. (1997). Structures for the analysis and teaching of supervision. In
C.E. Watkins Jr. (Ed.), Handbook o f Psychotherapy Supervision (pp.249-276).
New York: Wiley.
Holloway, EX., & Neufeldt, S.A. (1995). Supervision: Its contributions to
treatment efficacy. Journal o f Consulting and Clinical Psychology. 63(21.207-213.
Holloway, E. L., & Johnston, R. (1985). Group supervision: Widely practiced but
poorly understood. Counselor Education and Supervisioa 24. 332 - 340.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
Holloway, E.L., & Wampold, B.E. (1983). Patterns o f verbal behaviors and
judgments o f Satisfaction in the supervision interview. Journal o f Counseling
Psychology. 30. 227-234.
Holmes, G. R., Stader, S. R., Swaim, K. F., Haigler, E. D., & Myers, D., Jr.
(1998). Adolescent group psychotherapy supervision in a group format: An emerging
model. Journal o f Child & Adolescent Group Therapy. 8. 1 9 7 - 206.
Hutt, C.H., Scott, J., & King, M. (1983). A phenomenological study o f
supervisees’ positive and negative experiences in supervision. Psychotherapy:
Theory. Research, and Practice. 20,118-123.
Isaac, S., & Michael, W.B. (1997). Handbook in Research and Evaluation (3r d
ed.). San Diego: Educational and Industrial Testing Services(EdITS).
Radushin, A. (1985). Supervision in social work 12nd ed.). New York: Columbia
University Press.
Kennard, B.D., Stewart, S.M., & Gluck, M.R. (1987). The supervision
relationship: Variables contributing to positive versus negative experiences.
Professional Psychology: Research and Practice. 18, 172-175.
Kivlighan, D. M., & Quigley, S. T. (1991). Dimensions used by experienced and
novice group therapists to conceptualize group process. Journal o f Counseling
Psychology. 3 8 .4 1 5 -4 2 3 .
Kivlighan, D. M., Jr., Multon, K. D., & Brossart, D. F. (1996). Helpful impacts
in group counseling: Development o f a multidimensional rating system. Journal of
Counseling Psychology. 43, 347
Kruger, L.J., Chemiss, C., Maher, C., & Leichtman, H. (1988). Group
supervision o f paraprofessional counselors. Professional Psychology: Research and
Practice. 19(6), 609-616. (a)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
Kruger, L., Chemiss, C., Maher, C., & Leichtman, H. (1988). A behavioral
observation system for group supervision. Counselor Education and Supervision. 27.
331-343. (b)
Kunkel, M. A., Chapa, B., Patterson, G., & Walling, D. D. (1995). The
experience o f giftedness: A concept map. Gifted Child Quarterly. 3 9 .126 - 134.
Ladany, N ., Ellis, M.V., & Friedlander, M.L. (1999). The supervisory working
alliance, trainee self-efficacy, and satisfaction. Journal o f Counseling and
Development. 77, 447-455.
Ladany, N., Lehrman-Waterman, D., Molinaro, M., & Wolgast, B. (1999).
Psychotherapy supervisor ethical practices: Adherence to guidelines, the supervisory
working alliance, and supervisee satisfaction. The Counseling Psychologist. 27,443-
475.
Ladany, N ., & Friedlander, M.L. (1995). The relationship between the
supervisory working alliance and supervisee role conflict and role ambiguity.
Counselor Education and Supervision. 34, 220-231.
Ladany, N., Hill, C.E., Corbett, M.M., & Nutt, E.A. (1996). Nature, extent, and
importance o f what psychotherapy trainees do not disclose to their supervisors.
Journal o f Counseling Psychology. 43(1), 10-24.
Lambert, M.J., & Ogles, B.M. (1997). The effectiveness o f psychotherapy
supervision, (pp. 421-446). In C.E. Watkins, Handbook o f psychotherapy
supervision. New York: Wiley.
Larsen, D.L., Attkisson, C.C., Hargreaves, W.A., & Nguyen, T.D. (1979).
Assessment o f client/patient satisfaction: Development o f a general scale. Evaluation
and Program Planning. 2. 197-207.
Larson, L.M., & Daniels, J.A. (1998). Review o f the counseling self-efficacy
literature. The Counseling Psychologist. 26, 179-218.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
Larson, L.M., Suzuki, L.A., Gillespie, K.N., Potenza, M.T., Bechtel, M.A., &
Toulouse, A.L. (1992). Development and validation o f the counseling self-estimate
inventory. Journal o f Counseling Psychology. 39(1), 105-120.
Lehrman-Waterman, D., & Ladany, N. (2001). Development and validation of
the evaluation process within supervision inventory. Journal o f Counseling
Psychology. 48(2). 168-177.
Licht, M.H. (1998). Multiple regression and correlation (p. 19-64). In Grimm,
L.G., & Yarnold, P.R. (1998). Reading and understanding multivariate statistics.
Washington, D.C., American Psychological Association.
Mackenzie, K.R. (1981). Measurement o f group climate. International Journal o f
Group Psychotherapy, 31(1), 287-295.
Mackenzie, K.R. (1983). The clinical application o f a group climate measure. (In
R.R. Dies & K.R. Mackenzie (Eds.), Advances in Group Psychotherapy: Integrating
Research and Practice), (pp. 159-170).
Magnuson, S., Wilcoxon, S.A., & Norem, K. (2000). A profile o f lousy
supervision: Experienced counselors perspectives. Counselor Education and
Supervision. 39. 189-202.
Mahrer, A. R. & Nadler, W. P. (1986). Good moments in psychotherapy: A
preliminary review, a list, and some promising research avenues. Journal o f
Consulting and Clinical Psychology. 5 4 .10-16.
Morran, D. K., Stockton, R., Cline, R. J., & Teed, C. (1998). Facilitating
feedback exchange in groups: Leader interventions. Journal for Specialists in Group
Work. 2 3 .2 5 7 -2 6 8 .
Nelson, M. L., & Friedlander, M. L. (2001). A close look at conflictual
supervisory relationships: The trainee's perspective. Journal o f Counseling
Psychology. 48, 384 - 395.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
Nelson, M.L., & Friedlander, M.L. (2001). A close look at conflictual supervisory
relationships: The trainees perspective. Journal o f Counseling Psychology. 48(4),
384-395.
Nelson, M.L., Gray, L.A., Friedlander, M.L., Ladany, N., & Walker, J.A. (2001).
Toward relationship-centered supervision: Reply to Veach (2001) & Ellis (2001).
Journal o f Counseling Psychology. 48(4),
Newman, J.A., & Lovell, M. (1993). Supervision: A description of a supervisory
group for group counselors. Counselor Education and Supervision. 33, 22-31.
Norusis, M. J. (1994). SPSS professional statistics 6.1. Chicago, IL: SPSS, Inc.
Norcross, J.C., HanychJ.M, & Terranova, R.D. (1996). Graduate study in
psychology. American Psychologist. 51(6), 631-643.
Norcross, J.C., Sayette, M.A., Mayne, T.J., Karg, R.S., & Turkson, M.A. (1998).
Selecting a doctoral program in professional psychology: Some comparisons among
PhD counseling, PhD clinical, and PsyD clinical psychology programs. Professional
Psychology: Research and Practice. 29(6), 609-614.
O’Brien, K.M., Heppner, M.J., Flores, L.Y., & Bikos, L.H. (1997). The career
counseling self-efficacy scale: Instrument development and training applications.
Journal o f Counseling Psychology. 44(1), 20-31.
Oik, M.E., & Friedlander, M.L. (1992). Trainees’ experiences o f role conflict and
role ambiguity in supervisory relationships. Journal o f Counseling Psychology. 39,
389-397.
Patton, M.J., & Kivlighan, D.M. (1997). Relevance o f the supervisory alliance to
the counseling alliance and to treatment adherence in counselor training. Journal o f
Counseling Psychology. 44(1). 108-115.
Paulson, B. L., Truscott, D., & Stuart, J. (1999). Clients' perceptions o f helpful
experiences in counseling. Journal o f Counseling Psychology. 46,317 - 324.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Pedhazur, E.J., & Pedhauzer-Schmelkin, L. (1991). Measurement. Design, and
Analysis: An Integrated Approach. New Jersey: Lawrence Erlbaum Associates.
Prieto, L. R. (1996). Group supervision: Still widely practiced but poorly
understood. Counselor Education and Supervision. 35,295 - 307.
Ramos-Sanchez, L., Esnil, E., Goodwin, A., Riggs, S., Touster, L., Wright, L.K.,
Ratanasiripong, P., & Rodolfa, E. (2002). Negative supervision events: Effects on
supervision satisfaction and supervisory alliance. Professional Psychology. Research
and Practice. 33(2),
Ray, D., & Altekruse, M. (2000). Effectiveness o f group supervision versus
combined group and individual supervision. Counselor Education and Supervision.
40, 19-30.
Riva, M. T., & Cornish, J. A. E. (1995). Group supervision practices at
psychology predoctoral internship programs: A national survey. Professional
Psychology: Research and Practice. 26, 523 - 525.
Sansbury, D.L. (1982). Developmental supervision from a skills perspective.
Counseling Psychologist. 10, 53-58.
Savickas, M. L., Marquart, C. D., & Supinski, C. R. (1986). Effective supervision
in groups. Counselor Education & Supervision. 26, 1 7 -2 5 .
Stenack, R.J., & Dye, H.A. (1982). Behavioral descriptions o f counseling
supervision roles. Counselor Education and Supervision. 21. 295-304.
Stevens, J. (1986). Applied multivariate statistics for the social sciences. New
jersey. Lawerence Erlbaum Associates, Inc.
Stoltenberg, C.D. (1981). Approaching supervision from a developmental
perpective: The counselor complexity model. Journal o f Counseling Psychology. 28,
59-65.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I l l
Stoltenberg, C.D., McNeill, B.W., & Crethar, H.C. (1994). Changes in
supervision as counselors and therapists gain experience: A review. Professional
Psychology: Research and Practice. 25(4), 416-449.
Strupp, H. E L , & Hadley, S. W. (1977). A tripartite model o f mental health and
therapeutic outcomes. American Psychologist. 32. 187 - 196.
Tabachnick, B.G., & Fidell, L.S. (1996). Using Multivariate Statistics. New York:
Harper Collins Publishers Inc.
Tebb, S., Manning, D. W., & Klaumann, T. K. (1996). A renaissance o f group
supervision in practicum. Clinical Supervisor. 14, 39 - 51.
Trochim, W. (1989). An introduction to concept mapping for planning and
evaluation. Evaluation and Program Planning. 12(1), 1 -16.
Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological
Bulletin. 63. 384 - 399.
Tuckman, B.W., & Jensen, M.A.C. (1977). Stages o f small group development
revisited. Group and Organizational Studies. 2. 419-427.
Veach, P. (2001). Conflict and counterproductivity in supervision- when
relationships are less than ideal: Comment on Nelson & Friedlander (2001) and Gray
et al. (2001). Journal o f Counseling Psychology. 48(4),
Ward, J. H. (1963). Hierarchical grouping to optimize objective function.
Journal o f the American Statistical Association. 58,236 - 244.
Walter, C. A., & Young, T. M. (1999). Combining individual and group
supervision in educating for the social work profession. The Clinical Supervisor. 18.
73 - 89.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
Westwood, M. J. (1989). Group supervision for counsellors-in-training. Canadian
Journal o f Counselling. 23. 348 - 353.
Wetchler, J. L., Piercy, F .P., & Sprenkle, D. H. (1989). Supervisors’ and
supervisees’ perceptions o f the effectiveness o f family therapy supervisory
techniques. American Journal o f Family Therapy. 17(1). 35 - 47.
Wilbur, M., Roberts-Wilbur, J., Hart, G., Morris, J., & Betz, R. (1994).
Structured group supervision (sgs): A pilot study. Counselor Education and
Supervision, 33. 262-279.
Williams, E.N., Judge, A.B., Hill, C.E., & Hoffman, M.A. (1997). Experiences of
novice therapists in prepracticum: Trainees’, clients’, and supervisors’ perceptions o f
therapists’ personal reactions and management strategies. Journal o f Counseling
Psychology. 44(4), 390-399.
Worthen, V., & McNeill, B.W. (1996). A phenomenological investigation o f
“good” supervision events. Journal o f Counseling Psychology. 43(1), 25-34.
Worthington, E.L. Jr., & Roehlke, H. (1979). Effective supervision as perceived
by beginning counselors in training. Journal o f Counseling Psychology. 26, 64-73.
Yalom, I.D. (1995). The theory and practice o f group psychotherapy (4t h ed.).
New York: Basic Books.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
Appendix A: Supervisee Self-Efficacy Scale
Supervisee Self-Efficacv Scale (SSE)
Please rate each item on a scale o f 1-7 where l=never and 7=always.
Compared to how you were prior to participating in this group, to what extent are
you:
1. Better at understanding your client’s process issues?
Never Always
1 2 3 4 5 6 7
2. Better at conceptualizing client problems?
Never Always
1 2 3 4 5 6 7
3. Better at understanding your own countertransference?
Never Always
1 2 3 4 5 6 7
4. More confident in your role as a therapist?
Never Always
1 2 3 4 5 6 7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
Appendix B: Group Supervision Impact Scale
Group Supervision Impact Scale
This scale is intended to get your perceptions o f your experience in your current or
most recent group supervision. Please circle the appropriate number immediately
beneath each item. Please use a scale where 1 - never and 7 = always___________
To what extent...
(Supervisor Impact)
1.did your supervisor demonstrate openness?
1 2 3 4 5 6 7
2.did your supervisor demonstrate a sense o f humor?
1 2 3 4 5 6 7
3 .did your supervisor seem flexible?
1 2 3 4 5 6 7
4.was your supervisor competent?
1 2 3 4 5 6 7
5 .did your supervisor provide you with useful feedback?
1 2 3 4 5 6 7
6.did your supervisor give you validation?
1 2 3 4 5 6 7
7.did your supervisor help you to better understand your clients?
1 2 3 4 5 6 7
8.did your supervisor structure group time effectively?
1 2 3 4 5 6 7
9.did your supervisor constructively use group dynamics?
1 2 3 4 5 6 7
10.did your supervisor effectively resolve conflicts in the group?
1 2 3 4 5 6 7
11.was the group a safe place for you to ask questions?
1 2 3 4 5 6 7
12.was your supervisor reliable?
1 2 3 4 5 6 7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
13.was your supervisor’s feedback unfairly negative?
1 2 3 4 5 6 7
(Peer Supervisee Impact)
14.did you receive candid feedback from other group members?
1 2 3 4 5 6 7
15.did you receive useful treatment planning help from other group members?
1 2 3 4 5 6 7
16.did your fellow group members help generate ideas through group discussion?
1 2 3 4 5 6 7
17.were you able to benefit from mistakes others in the group reported having made?
1 2 3 4 5 6 7
(Group Environment Impact)
18 .was there between-member competition within the group?
1 2 3 4 5 6 7
19.was there between-member conflicts that did not get resolved?
1 2 3 4 5 6 7
20.did some members dominate the group time?
1 2 3 4 5 6 7
21.did you like the other people in the group?
1 2 3 4 5 6 7
22.did group members seem mutually supportive?
1 2 3 4 5 6 7
23.did you feel anxious when you presented your work to the group?
1 2 3 4 5 6 7
24.did you feel a part o f the group?
1 2 3 4 5 6 7
Note: Items 13, 18,19,20, and 23 are reverse scored.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Asset Metadata

Creator Getzelman, Michele A. (author)

Core Title Development and validation of the Group Supervision Impact Scale

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Education (Counseling Psychology)

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,psychology, clinical,psychology, psychometrics

Language English

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-482933

Unique identifier UC11340078

Identifier 3133275.pdf (filename),usctheses-c16-482933 (legacy record id)

Legacy Identifier 3133275.pdf

Dmrecord 482933

Document Type Dissertation

Rights Getzelman, Michele A.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA